Genetically Modified Food and Global Welfare

FRONTIERS OF ECONOMICS AND GLOBALIZATION 10

Series Editors: HAMID BELADI University of Texas at San Antonio, USA E. KWAN CHOI Iowa State University, USA

ABOUT THE SERIES: FRONTIERS OF ECONOMICS AND GLOBALIZATION

This series is aimed at economists and financial economists worldwide and will provide an in-depth look at current global topics. Each volume in the series will focus on specialized topics for greater understanding of the chosen subject and provide a detailed discussion of emerging issues. The target audiences are professional researchers, graduate students, and policy makers. It will offer cutting-edge views on new horizons and deepen the understanding in these emerging topics. With contributions from leading researchers, each volume presents a fresh look at today’s current topics. This series will present primarily original works, and employ references appropriate to the topic being explored. Each volume will bring a set of highly concentrated chapters that will provide in-depth knowledge to a target audience, while the entire series will appeal to a wide audience by providing them with deeper knowledge on a broad set of emerging topics in the global economy. The Frontiers of Economics and Globalization series will publish on topics such as: – – – – – – – – – – – – – – – –

Frontiers Frontiers Frontiers Frontiers Frontiers Frontiers Frontiers Frontiers Frontiers Frontiers Frontiers Frontiers Frontiers Frontiers Frontiers Frontiers

of of of of of of of of of of of of of of of of

Trade Negotiations Derivative Pricing International Lending and Debt Problems Economics Integration Trade and Environment Foreign Exchange International Finance Growth of Open Economies Futures Pricing International Financial Markets Investment Banking Mergers and Acquisitions Government Policy and Regulations Multi-Sector Growth Models Intellectual Property Rights Fragmentations and Outsourcing Hamid Beladi E. Kwan Choi Series Editors

GENETICALLY MODIFIED FOOD AND GLOBAL WELFARE

FRONTIERS OF ECONOMICS AND GLOBALIZATION VOLUME 10

GENETICALLY MODIFIED FOOD AND GLOBAL WELFARE Edited by

Colin A. Carter Department of Agricultural and Resource Economics, University of California, USA

GianCarlo Moschini Department of Economics, Iowa State University, USA

Ian Sheldon Department of Agricultural, Environmental & Development Economics, Ohio State University, USA

United Kingdom – North America – Japan India – Malaysia – China

Emerald Group Publishing Limited Howard House, Wagon Lane, Bingley BD16 1WA, UK First edition 2011 Copyright r 2011 Emerald Group Publishing Limited Reprints and permission service Contact: [email protected] No part of this book may be reproduced, stored in a retrieval system, transmitted in any form or by any means electronic, mechanical, photocopying, recording or otherwise without either the prior written permission of the publisher or a licence permitting restricted copying issued in the UK by The Copyright Licensing Agency and in the USA by The Copyright Clearance Center. No responsibility is accepted for the accuracy of information contained in the text, illustrations or advertisements. The opinions expressed in these chapters are not necessarily those of the Editor or the publisher. British Library Cataloguing in Publication Data A catalogue record for this book is available from the British Library ISBN: 978-0-85724-757-5 ISSN: 1574-8715 (Series)

Emerald Group Publishing Limited, Howard House, Environmental Management System has been certified by ISOQAR to ISO 14001:2004 standards Awarded in recognition of Emerald’s production department’s adherence to quality systems and processes when preparing scholarly journals for print

ABOUT THE EDITORS

Colin Carter is professor of agricultural and resource economics at the University of California, Davis and the Director of the University of California’s Giannini Foundation of Agricultural Economics. Colin has published widely in the areas of international trade, agricultural policy, futures and commodity markets, the economics of China’s agriculture, and the economics of biotechnology adoption in agriculture. He was named Fellow of the American Agricultural Economics Association in 2000 in recognition of his many contributions to the field of agricultural economics. GianCarlo Moschini is professor of economics at Iowa State University and holder of the Pioneer Hi-Bred International Chair in Science and Technology Policy. He has published widely on modeling demand and production systems, decisions under risk and, more recently, the economics of agricultural biotechnology, intellectual property rights, and the economics of biofuels. He is a former editor of the American Journal of Agricultural Economics (1998–2000) and was named a Fellow of the Agricultural & Applied Economics Association (formerly the American Agricultural Economics Association) in 2003. Ian Sheldon is currently Andersons Professor of International Trade in the Department of Agricultural, Environmental, and Development Economics at Ohio State University. He has published widely in the areas of international trade and industrial organization, and recently completed a term as Chair of the International Agricultural Trade Research Consortium. He is a former Editor of the American Journal of Agricultural Economics, and is currently Featured Articles Editor for Applied Economic Perspectives and Policy.

About the volume Genetically modified (GM) (or transgenic) crops are produced using plant biotechnology to select desirable characteristics in plants and transfer genes from one organism to another. As a result, crops can survive under harsher conditions, costs are lowered, chemical application is reduced, and yields are improved. Scientists are introducing genes into plants that will give them resistance to herbicides, insects, disease, drought, and salt in the soil. The application of modern biotechnology to crop and food

viii

About the Editors

production is one of the most significant technological advances to impact modern agriculture. The future of GM technology holds further promises of continued benefits. But the potential of GM product innovations has been overshadowed by significant controversy over this technology. The regulatory activism that has accompanied the diffusion of GM technology has given rise to a complex situation that is replete with obstacles for current and future GM innovations. This is particularly true for the European Union (EU), which has implemented restrictive policies that undoubtedly constrain the current status and the future potential of biotechnology. The discourse on biotechnology applied to food and agriculture is at a crossroads due to rising food prices and concerns about adequate food supplies. Over the last decade, a large body of applied economics work has addressed the key questions surrounding the application of this technology to food production. It is now time to take stock of the results of these efforts, and consolidate the methodological, analytical, and empirical findings. The challenge is to strengthen the consensus of what economics has to offer in the analysis of the complex issues surrounding the ongoing development of GM products for the agricultural and food sector. The task is to provide a new perspective on the most pressing policy questions and to help foster an intellectual climate conducive to achieving meaningful progress and lasting solutions. That is the motivation for this volume. It brings together fresh insights from top agricultural economists in the areas of consumer attitudes, environmental impacts, policy and regulation, trade, investment, food security, and development.

LIST OF CONTRIBUTORS

Volker Beckmann

Law and Economics Faculty, Ernst-Moritz-ArndtUniversity Greifswald, Greifswald, Germany

Antoine Boue¨t

Markets, Trade and Institutions Division, International Food Policy Research Institute, Washington, DC, USA; Laboratoire d’Analyse et de Recherche en Economie et Finances Internationales, Universite´ Montesquieu Bordeaux IV, Bordeaux, France

Jean-Paul Chavas

Department of Agricultural and Applied Economics, University of Wisconsin-Madison, Madison, WI, USA

El Hadji Fall

UNDP, Dakar, Senegal

Elise Golan

Economic Research Service, US Department of Agriculture, Food Economics Division, Washington, DC, USA

Guillaume P. Grue`re

Environment and Production Technology Division, International Food Policy Research Institute, Washington, DC, USA

Robert W. Herdt

Department of Applied Economics and Management, Cornell University, Ithaca, NY, USA

Ruifa Hu

Center for Chinese Agricultural Policy, Chinese Academy of Sciences, Institute of Geographical Sciences and Natural Resources Research, Beijing, China

Jikun Huang

Center for Chinese Agricultural Policy, Chinese Academy of Sciences, Institute of Geographical Sciences and Natural Resources Research, Beijing, China

Wallace E. Huffman

Department of Economics, Iowa State University, Ames, IA, USA

Fred Kuchler

Economic Research Service, US Department of Agriculture, Washington, DC, USA

x

List of Contributors

Jayson L. Lusk

Department of Agricultural Economics, Oklahoma State University, Stillwater, OK, USA

Ira Matuschke

Food and Agriculture Organization of the United Nations, Rome, Italy

Simon Mevel

Formerly at the World Bank, Washington, DC, USA

Latha Nagarajan

Department of Agriculture, Food and Resource Economics, Rutgers University, New Brunswick, NJ, USA

Rebecca Nelson

Department of Plant Pathology and Plant-Microbe Biology, Cornell University, Ithaca, NY, USA

Carl E. Pray

Department of Agriculture, Food and Resource Economics, Rutgers University, New Brunswick, NJ, USA

Matin Qaim

Department of Agricultural Economics and Rural Development, Georg-August-University of Goettingen, Goettingen, Germany

Bharat Ramaswami

Indian Statistical Institute, New Delhi, India

Alan Randall

Agricultural and Resource Economics, The University of Sydney, Sydney, NSW, Australia; Department of Agricultural, Environmental, & Development Econimics, The Ohio State University, USA

Terri Raney

Food and Agriculture Organization of the United Nations, Rome, Italy

Sara Scatasta

Rural Development Theory and Policy, Universita¨t Hohenheim, Stuttgart, Germany

Steven E. Sexton

Department of Agricultural and Resource Economics, University of California, Berkeley, CA, USA

Guanming Shi

Department of Agricultural and Applied Economics, University of Wisconsin-Madison, Madison, WI, USA

Claudio Soregaroli

Alta Scuola in Economia Agro-Alimentare, Universita` Cattolica del Sacro Cuore, Cremona, Italy

List of Contributors

Kyle W. Stiegert

Department of Agricultural and Applied Economics, University of Wisconsin-Madison, WI, USA

Justus Wesseler

Technische Universita¨t Mu¨nchen, Center of Life and Food Sciences Weihenstephan, Technische Universita¨t Mu¨nchen, Freising, Germany

David Zilberman

Department of Agricultural and Resource Economics, University of California, Berkeley, CA, USA

xi

INTRODUCTION

The application of modern biotechnology to crop and food production is one of the most significant technological advances to impact modern agriculture. Barely a dozen years since their introduction, genetically modified (GM) crops are currently grown on more than 300 million acres worldwide. GM (or transgenic) crops are produced using plant biotechnology to select desirable characteristics in plants and transfer genes from one organism to another. As a result, crops can survive under harsher conditions, costs are lowered, and yields are improved. Scientists are introducing genes into plants that will give the plants resistance to herbicides, insects, disease, drought, and salt in the soil. Crop research in bioengineering is also aimed at improving the nutritional quality of food, such as providing healthier vegetable oils. Pharmaceutical and industrial crops (or ‘‘pharma’’ crops) are also on the horizon, with the potential to dramatically reduce drug production costs. Compared to traditional plant breeding, biotechnology can produce new varieties of plants more quickly and efficiently, and it can introduce desirable traits into plants that could not be established through conventional plant breeding techniques. First-generation GM crops that are now being grown have increased yields and/or reduced the cost of weed and pest control (including cost savings made possible by the induced simplification of some crop management activities), and have lowered the quantity of chemicals used on plants and the soil. All of this has resulted in sizeable efficiency gains, which explains enthusiastic farmers’ adoption choices despite the price premium that GM seed varieties typically command. The reduced use of pesticides and the change in the composition of herbicides used brought about by GM crops translates into substantial positive environmental benefits. The future of GM technology holds further promises of continued benefits. Novel agronomic traits such as herbicide resistance and insect resistance are expected to be extended to major food crops, such as wheat and rice. In fact, China recently commercialized Bt rice, a very significant development because this is the first GM food crop to be commercialized globally, and rice is the largest food crop in the developing part of the world. Future applications include the tackling of complex agronomic traits, including plant output traits. The second-generation GM crops will be engineered to possess desirable quality attributes (such as improved nutritional profiles leading to functional foods), and the third-generation

xxii

Introduction

GM crops will be suitable for novel uses (such as plant-made pharmaceuticals and plant-made industrial products). Appreciation of the economic potential of GM product innovations has often been overshadowed by the significant controversy that has plagued this technology from the start. Surveys of public attitudes routinely find that a sizeable segment of the population has misgivings about agricultural biotechnology. On the other hand, a strong scientific consensus is emerging that GM technology itself poses no inherent risk for human health, and a careful assessment of the data from field research and commercial cultivation indicates no environmental harm from GM crops; in fact, a number of positive environmental effects have been documented. The regulatory activism that has accompanied the diffusion of GM technology has given rise to a complex situation that is replete with obstacles for current and future GM innovations. This is particularly true for the European Union (EU), which has implemented restrictive policies that undoubtedly constrain the current status and the future potential of biotechnology. The EU and the United States have long been on a collision course when it comes to GM products. The United States first exported GM food to Europe in 1996. It was tomato puree from California, and it was voluntarily labeled as genetically engineered. The product was a big hit with consumers in Britain because it was cheaper than conventional tomato puree. However, when GM soybeans were imported into Europe later that year, there was a huge backlash from environmental groups, and the EU was then quick to introduce mandatory labeling for GM foods. Both the government and the food industry in the United States view the EU’s mandatory labeling policy as a trade barrier. The United States has rapidly adopted GM crops and is in favor of making this technology available to its own farmers and anyone else, including poor countries. Alternatively, the EU is slowing the introduction of biotech crops in Europe and elsewhere. This major dispute was aired at the World Trade Organization (WTO) and in other venues. Under the WTO’s agreement on sanitary and phytosanitary measures, nontariff barriers like an embargo on GM crops must be scientifically justified. The risks are high because soybeans and corn, and their by-products, are important U.S. agricultural exports to the EU, and GM varieties account for a large percentage of the U.S. soybean and corn crop acreage. Widespread contamination of U.S. grains due to the accidental release of unapproved GM crops such as StarLink corn and LL601 rice have also damaged trade relations between the United States and importers such as the EU, Japan, and South Korea. The topic of this volume is therefore a major global issue, not only in North-South trade (United States-Asia), but also in NorthNorth trade (United States-EU). ‘‘Asynchronous’’ GM trait approvals, coupled with the effective zero tolerance on unapproved traits, and additional restrictive GM regulations, affect research and development (R&D) decisions and could seriously deter

Introduction

xxiii

future innovation. Some developing countries, such as in Africa, are in a delicate position, and the possibility exists that the GM product controversy could seriously undermine the potential role of biotechnology in helping to keep food prices low for the very poor and in continuing to feed the world as the population approaches 9 billion by 2050. We see the discourse on biotechnology applied to food and agriculture as being at a crossroads. It is important that the full weight of economic analysis be brought to bear at this juncture to help transition to a new level of understanding. We need a more thorough picture of the actual and potential benefits of GM product innovation, as well as a clearer comprehension of possible undesirable consequences. It is also imperative that we fully understand the institutional setting of biotechnology innovation and diffusion – for example, the function of public research, the importance of private R&D investments, the role of intellectual property rights, and the scope and nature of both national and international regulations. Developing countries have a huge stake in the application of biotechnology to agriculture. They stand to gain through reduced pesticide use, higher yields, lower production costs, increased farm profits, and lower food prices. This is not to mention enhanced food security. For instance, in China, the use of agricultural pesticides has dropped sharply since the recent introduction of transgenic cotton, raising farm incomes at the same time. In India, GM cotton acres have increased sharply, raising yields and reducing pesticide use. Yet many developing countries are afraid to research and approve GM crops for fear of jeopardizing trade relations with the EU. The EU’s GM labeling regulations serve as a second line of defense against imports. The World Health Organization and several national scientific academies in Europe and around the world have judged biotech foods as safe as conventional non-GM foods. But European politicians continue to talk about environmental risk of GM crops. Given that GM crops are often environmentally friendly, it is ironic that environmental groups are leading the anti-GM charge in Europe, Asia, and elsewhere. Biotech crops reduce the use of chemicals and encourage zero-till farming, helping to conserve the soil. European agriculture is one of the heaviest users of pesticides on the globe, and a more extensive adoption of transgenic crops would sharply reduce pesticide use. In fact, a recently released report from the European Commission found that GM crops are no more risky than conventional plant breeding technologies.1 Over the last decade, a large body of applied economics work has addressed many of these questions. It is now time to take stock of the results of these efforts, and consolidate the methodological, analytical, and empirical findings. The challenge is to strengthen the consensus of what

1 European Commission, ‘‘A decade of EU-funded GMO research (2001–2010),’’ EUR 24473, Luxembourg: Publications Office of the European Union, 2010.

xxiv

Introduction

economics has to offer in the analysis of the complex issues surrounding the ongoing development of GM products for the agricultural and food sector. The task is to provide a fresh perspective on the most pressing policy questions and to help foster an intellectual climate conducive to achieving meaningful progress and lasting solutions. That is the motivation for this volume. It brings together fresh insights from top agricultural economists in the areas of consumer attitudes, environmental impacts, policy and regulation, trade, investment, food security, and development. The chapters in this volume are loosely organized by topic area as follows: first, there are four chapters setting the stage with an overview of the key issues and economic trade-offs associated with the biotechnology revolution in agriculture, highlighting both emerging applications of biotechnology along with analyses of key international development issues; second, two chapters cover public versus private R&D and commercial development of GM crops, focusing on the dominance of the sector by a small number of private firms, and how the latter price their products to U.S. farmers; third, three chapters deal with resource, legal, and renewable energy issues concerning GM crops, focusing on environmental costs and benefits, coexistence, and production of biofuels; fourth, two chapters focus on consumer concerns about GM foods and the associated issue of labeling; and finally, three chapters deal with issues covering trade-related regulations and GM crops, as well as discussion and analysis of appropriate use of the precautionary principle in regulating innovation in biotechnology. In the remainder of this introductory overview, we provide a brief summary of the various contributions in the volume. Herdt and Nelson provide a broad overview of some of the key issues described above. Their chapter surveys the new and emerging developments in biotechnology and their potential applications to agriculture in order to envision what potential new social and economic issues might arise and the associated consequences throughout the world. They begin by identifying and defining currently used rDNA-based techniques and then defining and discussing techniques that have emerged over the past five years and are now being applied to agriculture. Newer techniques that may emerge in the foreseeable future are also discussed in some detail. Herdt and Nelson then introduce the issues that may be raised by deployment of new biotech food products in the coming years. Hunger and malnutrition remain a widespread problem in the developing world. Qaim addresses the economic implications of the adoption of agricultural biotechnology by developing countries. GM crops could contribute to the world’s food supply and demand balance by increasing food availability at the global level. More importantly perhaps, when GM crops are adopted by poor farmers, they can lead to higher farm incomes. Qaim pays particular attention to the impacts of Bt cotton on millions of small-scale farmers in India, China, Argentina, South Africa, and other developing countries. Evidence from India suggests that Bt cotton is

Introduction

xxv

employment generating and contributes positively to poverty reduction and overall rural development. Studies on future GM crop applications are also reviewed by Qaim, including biofortified crops such as Golden Ricer. He discusses policy implications with a view to realizing the positive food security effects of GM crops on a wider scale. Providing an adequate amount of food to the world in 2050 is forecast to require a 70% increase in global output, and close to a doubling of output in developing countries. In their chapter, Raney and Matuschke focus on the potential of GM crops to contribute to agricultural productivity growth and poverty reduction in developing countries. Based on an analysis of case studies conducted in Asia, Africa, and Latin America, Raney and Matuschke conclude that GM crops have been beneficial to farmers in developing countries, specifically through reduced input requirements and/ or higher yields, with resulting increases in net farm incomes. However, they also find that farm-level impacts have varied considerably across regions and seasons. Importantly, availability of GM seeds and poorly functioning regulatory frameworks may have contributed to yield variability across developing countries. Raney and Matuschke conclude that greater research efforts, as well as investment in agricultural markets and associated institutions, will be necessary if GM crop technology is to be widely available and accessible to farmers in developing countries. Not only are China and India the two fastest growing emerging economies, they are also two of the world’s largest producers and consumers of food and other agricultural products. Since the 1980s, both countries’ governments and private sectors have been investing in biotechnology R&D, Chinese farmers adopting GM crops in the mid-1990s, Indian farmers following in 2000. The key GM crop planted in both countries has been Bt cotton, while China has also approved GM traits in both rice and maize. In their chapter, Pray et al. examine both the measured benefits of the adoption of Bt cotton and the future potential of other GM crops in China and India. A key contribution of their chapter is the presentation of evidence of recent changes in benefits from Bt cotton adoption in China, where pesticide use has continued to decline, and there have been spillover effects in terms of the bollworm population in all crops falling. They also suggest that adoption of Bt rice in China will reduce pesticide use as well as having a major impact on the control of borers. Analyzing the contributions of the public and private sector to R&D and innovation in GM crops is key to understanding the past and future evolution of technological change in agricultural biotechnology. Huffman’s chapter offers a wide-ranging analysis of the scientific discoveries that provided the foundation for development of GM crops, the changes in U.S. patent law that facilitated commercial application of these innovations, and changes in the structure of the research, agricultural chemicals, and seed sectors. Importantly, while the basic science for developing GM crops was undertaken in the public sector, GM traits and GM crop varieties that have

xxvi

Introduction

commercial applications have been developed almost exclusively by a small group of private firms in the U.S. private sector, with Monsanto being the clear leader in biological-event and trait-crop variety developments for cotton, soybeans, corn, and canola approved for commercial application. Given that the share of global acreage planted to GM crops remains highest in the United States, it is useful to obtain insights into what has been a contentious issue among farmers – the pricing of GM seed such as Bt corn seed. In their chapter, Stiegert, Shi, and Chavas use industrial organization methods to analyze the pricing of GM corn hybrids in the United States, with a particular focus on spatial differences in pricing between the fringe and core regions of the Corn Belt. Their research generates two key results: first, farmers in the fringe regions seem to exercise greater leverage in price negotiations, which may be because of their willingness to switch out of corn production into other crops; and, second, the exercise of market power by firms selling GM crops varies spatially. The potential impact of GM crops on the environment has been the subject of an intense debate. Wesseler, Scatasta, and Fall argue that while GM crops have a net positive effect on the environment, the regulatory response has mostly focused on negative concerns. In summarizing their chapter, they conclude that initial concerns about the negative effects of GM crops on the environment have been found almost negligible, and at the same time positive effects have been observed, for example, on habitat conservation and biodiversity. The authors conclude that policymakers need to pay more attention to documented environmental benefits as opposed to hypothetical environmental costs, and current policies toward GM crops need to be reconsidered. Because of a wide range of perceived concerns about GM crops, considerable attention has been paid to the concept of coexistence, that is, the planting of both GM and non-GM crops. A consequence of coexistence has been the development of two alternative regulatory regimes: one in which property rights for growing GM crops reside mostly with farmers growing, in the United States and Canada, and a second in which property rights reside mostly with farmers growing non-GM crops, as in the EU. The legal implications of the two systems are quite different. Under the former, farmers growing non-GM crops are responsible for ensuring their crops are non-GM, while under the latter, farmers growing GM crops are responsible for ensuring the GM-free status of those farmers growing non-GM crops. Beckmann, Soregaroli, and Wesseler show in their chapter that the two alternative property rights systems are equivalent as long as transactions costs are not prohibitively high, and using the court system is costless. However, since litigation is costly, they conclude that the property rights regime in which the GM farmer is not liable is preferable in terms of social welfare. Concerns about climate change and the scarcity of nonrenewable fossil fuels have led to a growing focus on the use of renewable energy sources such

Introduction

xxvii

as biofuels. However, following the 2008 food price crisis, concerns have been expressed that energy and global food security are not necessarily compatible and that biofuels production may actually contribute to rather than mitigate greenhouse gas emissions. The focus of the chapter by Sexton and Zilberman is on the extent to which the application of biotechnology might minimize any downside effects of biofuels production, through both yield-increasing effects and development of products using nonfood feedstocks such as trees and other perennials. Their overall conclusion is that biotechnology adoption could improve the net welfare impact of biofuels production, but that such benefits have been slowed by regulation. An issue receiving considerable attention since commercial application of GM crops has been the attitudes of consumers to the presence of GM ingredients in their food, especially in the EU. As reported by Lusk in his chapter, by 2009 there had been 51 studies of consumer attitudes toward GM foods, providing 114 estimates of their willingness to pay for such foods. Lusk also notes that, despite the plethora of studies, it is hard to distill what is actually known about consumer preferences for GM foods. Using the body of available research, Lusk proceeds by answering four questions. First, if studies show that U.S. consumers are willing to pay to avoid GM foods, why is there so little market for non-GM food? Second, if consumers are concerned about GM foods, why do they seem to know so little about biotechnology? Third, why do most economic models assume the willingness to pay for GM foods is unaffected by regulation of biotechnology and labeling? Fourth, why is there so little agreement on why U.S. and EU preferences for GM food differ? Deciding between mandatory and voluntary labeling has been an important regulatory choice with respect to GM foods. In their chapter, Golan and Kuchler analyze whether the minimal development of either GM or non-GM markets, respectively, has been due to the choice of mandatory labeling in the EU and 10 other countries, as compared to the no-GMlabeling choice made in the United States and Canada. They conclude that these regulatory choices were made for other reasons, including differences in consumer confidence about the safety of the food supply, affordability of a non-GM strategy, competition among food retailers and manufacturers, and market momentum once a choice is made between GM and non-GM. Since the introduction of biotechnology innovations, there have been many studies of their welfare effects, drawing on both partial equilibrium and computable general equilibrium (CGE) methods. In surveying this literature, Grue`re, Boue¨t, and Mevel find three main results coming out of the empirical research. First, without GM-specific trade regulations, GM crops are typically beneficial to adopting countries as well as non-adopting importing countries, although non-adopting competing countries may lose from lower world prices. Second, use of GM-specific trade regulations reduces the benefits of adoption, most notably for non-adopting countries. And third, exporters of GM crops suffer reduced benefits of adoption in the

xxviii

Introduction

presence of importer regulations on GM crops. These earlier findings are reinforced by the authors’ own application of a CGE model to GM crop adoption in Bangladesh, India, Indonesia, and the Philippines, but they also find that net importers of GM crops mostly gain through terms-of-trade effects, and segregation of non-GM crops for export markets may also be beneficial. Concerns over the safety of GM crops, as well as their environmental impact, have resulted in implementation of regulations that may affect international trade. In his chapter, Grue`re focuses on identifying the main trade impact of GM crop regulations, as well as analyzing the main motivations for supporting such regulations. Based on his examination of other results from the existing literature on GM food regulation, as well as use of a partial equilibrium trade and political economic model, Grue`re finds that in a non-GM crop producing country, trade-related regulations benefit producers but not necessarily consumers. Producer political support is necessary for implementation of regulations such as a ban on imports and mandatory labeling of GM foods, but if consumers and producers do not agree on such regulations, outside pressure groups will play the role of swing voters. Grue`re concludes that future global welfare effects of GM crops will depend on the evolution of trade-related regulations – the key challenges being to ensure new GM foods are safe for consumers, as well as to manage export risks. Introduction of GM crops has seen an extensive and ongoing debate about the most appropriate way to regulate innovations that offer a variety of potential benefits, but which also carry uncertainty about food safety and environmental risks. Specifically, much has been written in both the popular media and academic literature about the role and use of the precautionary principle. In his chapter, Randall attempts to bridge the gap between the principle of precaution and its actual application to regulation of GM crops, the overall objective of such an approach being to seek protection from any disproportionate risks associated with their introduction without unduly stifling innovation. Randall concludes that while the standard approach to risk management is appropriate for managing well-specified risks, there is scope for application of the precautionary principle to disproportionate threats. He does argue, however, that the principle must be constructed in such a way as to avoid some of the concerns raised by critics; in particular, he shows that precaution can be implemented through an iterative, sequenced decision process that takes advantage of prerelease screening and testing of GM crops in order to focus on a smaller set of cases that may present a genuine threat. Colin A. Carter GianCarlo Moschini Ian Sheldon Editors

CHAPTER 1

Biotechnology and Agriculture: Current and Emerging Applications Robert W. Herdta and Rebecca Nelsonb a

International Professor of Applied Economics and Management, Adjunct, Cornell University, Ithaca, NY, USA E-mail address: [email protected] b Associate Professor, Department of Plant Pathology and Plant-Microbe Biology, Cornell University, Ithaca, NY, USA E-mail address: [email protected]

Abstract The products of transgenic technology have captured the attention of enthusiasts and detractors, but transgenics are just one tool of agricultural biotechnology. Other applications enable scientists to understand biodiversity, to track genes through generations in breeding programs, and to move genes among closely related as well as unrelated organisms. These applications all have the potential to lead to substantial productivity gains. In this chapter we provide an introduction to basic plant genetic concepts, defining molecular markers, transgenic and cisgenic techniques. We briefly summarize the status of commercialized biotechnology applications to agriculture. We consider the likely future commercialization of products like drought tolerant crops, crops designed to improve human nutrition, pharmaceuticals from transgenic plants, biofuels, and crops for environmental remediation. We identify genomic selection as a potentially powerful new technique and conclude with our reflections on the state of agricultural biotechnology. Research at universities and other public-sector institutions, largely focused on advancing knowledge, has aroused enormous optimism about the promise of these DNA-based technologies. This in turn has led to large private-sector investments on maize, soybean, canola, and cotton, with wide adoption of the research products in about eight countries. Much has been made of the potential of biotechnology to address food needs in the low-income countries, and China, India, and Brazil have large public DNA-based crop variety development efforts. But other lower income developing countries have little capability to use these tools, even the most straightforward marker applications. Ensuring that these and other applications of biotechnology lead to products that are well adapted to Frontiers of Economics and Globalization Volume 10 ISSN: 1574-8715 DOI: 10.1108/S1574-8715(2011)0000010006

r 2011 by Emerald Group Publishing Limited. All rights reserved

2

Robert W. Herdt and Rebecca Nelson

local agriculture requires adaptive research capacity that is lacking in the lowest income, most food-insecure nations. We are less optimistic than many others that private research will fund these needs. Keywords: Transgenics, genetic engineering, marker-assisted selection, genetic mapping, drought tolerance, biofortification, plant-based pharmaceuticals, genomic selection JEL Classifications: O13, Q01, Q16, Q17, Q27 1. Introduction Genetic manipulation of agricultural crops and animals before the discovery of DNA played a critical role enabling the world’s farmers to produce ever increasing quantities of food at falling global prices over most of the past half-century, contributing substantially to global welfare. Despite that, substantial numbers of people in Africa, South Asia, and elsewhere in low-income developing countries are poor and undernourished. There, rapid population growth continues and agriculture is the primary livelihood of a majority of people. If these people are to participate in future rising global welfare, they will have to grow crops and animals with the genetic capacity for higher productivity. Many observers believe that biotechnology will play a crucial role in developing those crops and animals, even in the poorest countries. For many, ‘‘biotechnology’’ is synonymous with the production of genetically modified organisms (GMOs, also known as transgenics). While GMOs have captured the public imagination and substantial private investment, a great deal of research has been conducted and some application has been made of a wider spectrum of agricultural biotechnologies. The purpose of this chapter is to discuss a range of biotechnologies that involve genetic modification of agricultural species using DNA-based methods. We clarify some of the basic genetics and genomics concepts and terms used in the agricultural biotechnology literature, briefly summarize contribution of current commercial applications, consider prospective applications, then turn to emerging tools and their applications, and conclude with some reflections on how these innovations may contribute in low-income developing countries. 2. Genetic basics Genetic ‘‘improvement’’ – a term used by breeders to describe their craft – can be seen as comprising essentially two processes: (1) generating new genetic combinations in individual organisms and (2) selecting the most desirable individuals or groups from among the new combinations. Fig. 1 provides a schematic of crop improvement with DNA-based techniques

Biotechnology and Agriculture: Current and Emerging Applications Generate new combinations

Select among combinations

Trait based or diversity analysis

Trait based, marker aided or genomic selection

Landraces, germplasm banks, cultivars

Identify gene for desired trait

Clone gene for desired trait

Select genes or parents Recombine genes via Transgenic or cisgenic

Sexual combination Commercial farming Cross male X female

...

nth generation

rd

3 generation

st

Grow, produce seed

2nd generation

Generate whole plant

Select desired offspring

1 generation

Transfer DNA in vitro

3

Reproduce, certify, multiply seed

Select potential parents

Fig. 1.

Key steps in crop variety development, DNA-based techniques in bold.

indicated along with corresponding non DNA-based techniques. The first process, producing organisms carrying novel and potentially desirable combinations of genes, can be produced either through sexual recombination or through direct gene transfer. The second step, in which the organism(s) carrying the most desirable set of genes is identified and selected for further study or use, can be conducted using trait-based selection, marker-assisted selection (MAS), and/or genomic selection. All the DNA-based tools – molecular breeding, ‘‘biotechnology,’’ and ‘‘DNA-based techniques’’ – can be applied to plants, animals, fish, or microorganisms; our primary illustrations are from plants. Traditional or conventional breeding, which is based on sexual recombination, is largely trait based. It relies largely on traits that can be observed with the naked eye or easily measured in the field. The ‘‘phenotype’’ refers to the expressed trait(s), while the ‘‘genotype’’ refers to the underlying genetic composition. Male and female parents are typically selected based on phenotype, crossed, and the most desirable progeny are selected from subsequent generations based on phenotype. Phenotype, however, depends on the genotype, the growing environment (moisture, sunlight, nutrients, and other conditions), and the interaction between the genotype and environment. DNA-based techniques allow geneticists to more effectively dissect the effects of components of the genotype (down to the single gene, or single nucleotide of the DNA sequence) and to better understand how these interact with other genetic and environmental

4

Robert W. Herdt and Rebecca Nelson

components. This then allows them to select plants or animals carrying desirable combinations of gene variants. Although conventional breeding is largely based on phenotypic selection, it can benefit greatly from an understanding of the underlying genetic diversity and structure of the species involved. Diversity analysis, in which DNA-based information is used to understand genetic variation and population structure in germplasm, allows breeders to select diverse parents that are more likely to produce progeny with novel gene combinations and thus novel and potentially desirable traits. Diversity analysis has been extensively applied in understanding crop, livestock, and microbial populations. This has transformed the efficiency of germplasm conservation, characterization, and utilization. A gene variant is called an allele. Most organisms have two sets of chromosomes, one from each parent, with one allele of each gene on each chromosome. Different alleles result in different traits, for example, distinct eye colors. A good deal of molecular genetics research is directed at identifying genes associated with particular traits and linking the genes to particular variants in DNA sequence. A variety of techniques can be used to identify DNA variants associated with traits of interest; the detected variants are termed molecular markers.

2.1. Molecular markers A molecular marker in or near a gene of interest can be used to identify individuals containing the same allele and to assist in allele transfer or gene cloning. Early genetic markers included enzymes that were used to track resistant traits from wild species that were crossed with cultivated tomato (Rick et al., 1979; Tanksley et al., 1981; Vallejos and Tanksley, 1983). DNAbased tools proved more versatile. A series of evolving DNA-marker systems was developed, beginning with restriction fragment-length polymorphisms (RFLPs). RFLP markers were succeeded by simple sequence repeats (SSRs), also known as microsatellites, and amplified fragmentlength polymorphisms (AFLPs) (Duran et al., 2009; Jones et al., 2009), among many others. With the radically decreasing cost of DNA sequencing, single nucleotide polymorphisms (SNPs) are currently being widely used in plant genetic analysis. The ultimate molecular marker is the DNA sequence associated with the causal genetic difference. DNA sequencing is rapidly becoming more accessible based on price, speed, and analytical capacity (‘‘bioinformatics’’), enabling agricultural researchers to more efficiently identify and select for DNA variants of interest. DNA-marker technologies have been used to gain a greater understanding of the natural variation in the genetic architecture of key traits of agricultural importance, revealing the numbers, locations, and modes of action of genes. This is valuable for utilizing natural variation in breeding

Biotechnology and Agriculture: Current and Emerging Applications

5

and for understanding evolution and domestication of species. Markerassisted genetic analysis sets the stage for MAS, in which molecular markers are used to detect known genes or alleles and track their presence or absence from one generation to the next. With trait-based markers, it may take five to eight generations to purify a plant line so that it will ‘‘breed true’’ and be ready for consideration as a commercial variety; using molecular markers can reduce the number of generations needed. In genomic selection, statistical methods are used to associate a large number of DNA markers with a desirable set of traits and genes in order to select individual plants or animals. In conventional plant breeding, genes are recombined through the ‘‘natural’’ processes of meiosis and sexual combination. Human intervention occurs through the selection of the parents and progeny that are involved in this process. Pollen is introduced from the male parent onto the stigma of the female parent, fertilization occurs, and the plant is grown to produce mature seed. In direct gene-transfer techniques, or those that rely on in vitro methods (‘‘test-tube’’ or laboratory methods) rather than sexual exchange of genetic information, genes are spliced into chromosomes by any of a variety of techniques. The gene transfer takes place in cells that are then regenerated into whole plants or animals. The most desirable progeny are then selected from subsequent generations. Two general types of direct gene transfer are recognized: transgenic and cisgenic. Transgenic organisms (i.e., transgenics) are created by transferring a gene or genes from one species into a cell of another in –vitro and growing that cell into a mature seed-bearing plant using specialized techniques of tissue culture. The transferred gene may either be synthesized or found in an existing organism and cloned. Cisgenic organisms (i.e., cisgenics) are created using the same process, but the gene that is transferred comes from the same or a closely related species. Transgenic and cisgenic techniques increase plant breeding efficiency by inserting only a short segment of DNA containing a desired trait; consequently, in much-researched organisms like maize, the necessary subsequent steps of selecting plants to take to the following generations is shortened compared to conventional breeding. On the other hand, conventional breeding and various marker-assisted variants allow the breeder to manipulate alleles at tens of thousands genetic loci. Well-resourced breeding organizations take advantage of the benefits of all of these methods to generate and select individuals with allelic combinations giving superior performance.

3. Current and near-term applications 3.1. Diversity analysis Molecular markers enable biologists to describe genetic diversity at the levels of genes; genotypes and populations with the types of markers used

6

Robert W. Herdt and Rebecca Nelson

depend on the specific objectives of the work. By understanding the genetic similarity or heterogeneity at the genetic, genomic, and population levels, scientists can devise systems of molecular techniques for exploiting naturally existing diversity that can often exceed the power of genetic transformation. Similar approaches are being used to better understand pest and disease populations and beneficial organisms of relevance to agriculture. Understanding genetic diversity has been an important and productive area of research for decades, and directing molecular tools to this objective increases the contribution it can make to the ultimate objective of crop improvement (Harding et al., 1997; Park et al., 2009b; Varshney et al., 2010). DNA markers have been a fruitful tool for understanding the extent and structure of diversity in agricultural germplasm (McCouch et al., 2007; Thomson et al., 2007; Perez-Vega et al., 2009). Molecular markers have made it possible to efficiently transfer desirable traits from wild relatives to cultivated species, thereby further increasing the available diversity (McCouch et al., 2007). Markers are used to establish groupings that inform breeders’ choice of material in breeding programs, such as those aimed at taking advantage of hybrid vigor (Reif et al., 2003). Markers can also be used to establish the distinctness or identity of germplasm, which is relevant to establishing and protecting plant breeders property rights (Ibanez et al., 2009). Once a useful gene has been cloned, its DNA sequence becomes a tool for discovering ‘‘hidden’’ genetic diversity at the same locus in crop germplasm (Bhullar et al., 2009). The use of molecular markers has enabled the study of the microbial diversity as a reflection of soil health (Manici and Caputo, 2009) and a better understanding of the issue of genetic erosion (Fu and Somers, 2009; Steele et al., 2009; van Heerwaarden et al., 2009).

3.2. Products of marker-assisted selection Molecular markers for single genes of several important traits have been used to introduce traits in several crops important in low-income countries. Flooding causes frequent devastation in large areas of rice in South Asia, but plants carrying the submergence tolerance gene Sub1 can survive long periods under water (Sarkar et al., 2009). MAS was used in transferring the Sub1 gene into Swarna, and by 2009 the resulting varieties had been planted on more than 15 million hectares of rainfed lowland rice in flood-prone environments in Eastern and Central India (Ribaut et al., 2010). MAS has been used to combine different resistance genes to bacterial blight disease in rice. A plant that has 1 or 2 of the 28 resistance genes may look the same when exposed to the disease pathogen as a plant that has many such genes, but more genes are thought to provide higher levels of resistance and greater durability over time (Jeung et al., 2006). Several of the genes have been tagged with molecular markers facilitating their combination by rice-breeding programs in Asia (Jena and Mackill, 2008).

Biotechnology and Agriculture: Current and Emerging Applications

7

Molecular markers have also been used to track resistance to African maize streak virus and nematodes in soybean. Resistance to maize streak, a disease transmitted by an insect, is controlled by a single gene, but observing resistant phenotypes can be challenging because it is difficult to know whether an individual plant is resistant or has simply escaped being infected by the insect. A comparison of conventional and MAS for maize streak virus found MAS to be more cost-effective (Abalo et al., 2009). Most traits of agricultural importance, including yield and quality, vary across a continuous range from low to high. These so-called ‘‘quantitative traits’’ are controlled not by single genes but rather by genes in multiple locations around the genome, each of which has a modest effect on the trait phenotype. These genes are known as ‘‘quantitative trait loci’’ or QTLs. While progress is being made in identifying QTLs for many traits, it has proven to be difficult to efficiently select for quantitative traits using MAS in many cases (e.g., Moreau et al., 2004). Many studies have been conducted to identify QTL for quantitative disease resistance (reviewed by Poland et al., 2008), but these results have been applied in relatively few cases (reviewed by St. Claire, 2010). Many QTL associated with disease resistance have been identified in rice (e.g., Wisser et al., 2005), but as mentioned above, successful examples of MAS for disease resistance in developing countries have mostly involved combining major resistance genes (e.g., Singh et al., 2001). Molecular markers have greatly facilitated the utilization of desirable QTL from wild relatives; traitenhancing alleles may be present and utilized for crop improvement even when the wild species lacks desirable traits ((McCouch et al., 2007). Wellresourced programs are able to take advantage of these findings. QTL for yield under drought stress have been analyzed (Venuprasad et al., 2009). Using MAS, loci contributing to improved yield under drought conditions are being transferred to the popular rice variety Swarna, the same variety in which submergence tolerance has been incorporated, because both problems occur in the same fields at different times during the season.

3.3. Transgenic products Transgenic crops can be very precisely identified because the techniques used in their creation leave traces in the DNA and because companies keep records of their transgenic seed sales. Hence, data on their adoption by farmers is readily available.1 The first transgenic crop approved for commercial production was the Flavr Savr tomato, which was approved for release in 1994 but was not a commercial success (Bruening and Lyons, 2000). In 1 The data on transgenic crop area in this section come from James (2010) Global Status of Commercialized Biotech/GM crops: 2010. Ithaca, NY, International Service for the Acquisition of Agri-biotech Applications (ISAAA).

8

Robert W. Herdt and Rebecca Nelson

contrast, transgenic maize (corn), soybeans, and cotton spread rapidly in a number of countries after their approvals in the mid-1990s. By 2000, the global area of transgenics was about 44 million hectares and by 2009 reached about 134 million hectares, with about half in the United States, and most of the rest in seven other countries. Canada, with 6% of the global transgenic land area, is the only other OECD country in the top eight. Argentina and Brazil each planted about 16% of the global total area, India 6%, China 34%, and Paraguay and South Africa each planted about 2% of the global total. In 2010, transgenic soybean comprised 73% of the world’s total soybean acreage, transgenic maize comprised 27% of all maize, transgenic cotton 47% of all cotton, and transgenic canola 21% of all canola. Small areas of transgenic sugar beet, alfalfa, papaya, and a few other crops were planted. Adoption rates of transgenic soybean, cotton, and maize have been extremely rapid by historic standards, reaching their current dominant positions in less than 10 years, exceeding adoption rates of green revolution wheat and rice in Asia (Dalrymple, 1975) and of hybrid corn in the United States (Griliches, 1957). Herbicide tolerance is the most widely commercialized transgenic trait. In 2010, it was incorporated as a single gene in 62% of all transgenic crops and ‘‘stacked’’ together with another gene in 21% of all transgenics. Over half the global soybean crop contains an herbicide resistance gene. The Bacillus thuringiensis gene (‘‘Bt’’), which gives plants resistance to caterpillars (lepidopterous larva) that feed on crops, is the second most important transgenic trait, incorporated in 16% of transgenic crops as a single gene and in another 25% of transgenics in combination with another gene. Through the first decade of the 21st century, the only foods made from transgenic plants were from maize, soybeans, and canola, although transgenic eggplant and rice were in advanced stages of development in India and China. Eggplant, the most widely consumed vegetable in India, was engineered with Bt to reduce the need for pesticides to control fruit and shoot borer2 and in 2009 India’s Genetic Engineering Appraisal Committee recommended it be cleared for release. Vocal opposition delayed that release, and in February 2010 the Government of India announced it would be further delayed.3 Transgenic rice was widely tested in farmer field trials in China. In January of 2010, a strain of genetically engineered rice was approved and safety certificates were issued for it by the Ministry of Agriculture (Waltz, 2010). However, the approval process took so long and the innovation cycle is so short in China that the transgenic variety no longer had other traits that would make it attractive to farmers, so new varieties will have to be transformed before transgenic rice is widely grown.

2 On average, for a crop of 180 days, 2.34 kg/acre of active ingredients are applied in 30 sprays. See Krishan and Qaima (2007). Estimating the adoption of Bt eggplant in India: Who Benefits from public–private partnership? Food Policy, 32, 523–543. 3 The Times of India, February 9, 2010.

Biotechnology and Agriculture: Current and Emerging Applications

9

3.4. Cisgenic products In vitro transfer of genes among closely related species is of greatest interest in cases where desirable genes are known to exist in species that are vegetatively propagated and ‘‘hard to breed’’ because of their complex inheritance and/or long life cycles. Introducing disease-resistant traits through conventional breeding into such crops, which include potato, sweet potato, cassava, and most trees, can take decades while in vitro methods can take much less time. The transfer of disease resistance genes from wild relatives of potato into cultivated potato, for example, can reduce disease losses and reduce pesticide application (Jacobsen and Schouten, 2007; Park et al., 2009a) (Rommens and Kishore, 2000). Cisgenic techniques have also been used in seed-propagated crops like rice, for which researchers have identified alleles of 28 genes for resistance to rice bacterial blight disease and cloned and transferred some into rice varieties in which the alleles did not exist (Nino-Liu et al., 2006). Some have argued that cisgenics should be subjected to lower regulatory hurdles than varieties produced through transgenic means because the transferred genes come from the same species (Schouten et al., 2006; Rommens et al., 2007). Others believe that, because the gene transfer process could itself lead to changes in the structure or expression of either the gene transferred or of other gene(s) in the target genome, the same regulatory scrutiny is justified for cisgenics as for transgenics (Schubert and Williams, 2006; Akhond and Machray, 2009). 3.5. Transgenic animals Embryo transfer was developed a century ago (Hasler, 2003), and other techniques such as breeding and artificial insemination, widely used to increase milk and meat productivity, also predate the discovery of DNA. DNA-based research has been widely directed at increasing animal productivity with transgenic salmon most advanced, but not yet approved for commercial use in September 2010.4 The FDA, which has jurisdiction over transgenic animals, recognizes six categories of genetically engineered animals ‘‘based on the intended purpose of the genetic modification: (1) to enhance production or food quality traits (e.g., pigs with less environmentally deleterious wastes, faster growing fish), (2) to improve animal health (e.g., disease resistance), (3) to produce products intended for human therapeutic use (e.g., pharmaceutical products or tissues for transplantation; these GE animals are sometimes referred to as ‘‘biopharm’’ animals), (4) to enrich or enhance the animals’ interactions with humans (e.g., hypoallergenic pets), (5) to 4

http://www.businessweek.com/ap/financialnews/D9IBPPVO0.htm

10

Robert W. Herdt and Rebecca Nelson

develop animal models for human diseases (e.g., pigs as models for cardiovascular diseases), and (6) to produce industrial or consumer products (e.g., fibers for multiple uses)’’ (U.S. Department of Health and Human Services et al., 2009). Biopharm products (category 3) are just beginning to become commercially available. In February of 2009, the US Food and Drug Administration gave marketing clearance to antithrombin produced in the milk of transgenic goats, the first approval of a product from a biopharm animal for human use. The protein, to be known commercially as ATryn, is expected to reduce the cost of the product significantly compared with making the same protein in tissue culture or isolating it from collected plasma (Mary Ann Liebert Inc., 2009). Commercialized antithrombin illustrates that it is relatively easy to engineer an animal with a single gene to produce a medically useful product. As with plants, it is much more challenging to increase innate productivity of an animal per unit of feed input because the majority of important livestock traits are controlled by multiple genes (Houdebine, 2009). The accurate modification of the appropriate gene(s) to generate desired phenotypes remains problematic despite the continuous increase in scientific understanding of the functional relationship between livestock genes and production traits. Livestock genetic engineering has some additional challenges over transgenic plants. Farm animals have much longer reproductive cycles, raising and maintaining animals is more costly than plants, and the necessary techniques of cloning and embryo transfer have low-success rates. All three drive up the costs of commercialization (Faber et al., 2003; Hansel, 2003). Public concern about environmental, emotional, and ethical issues seem greater for transgenic animals than for plants. Inadvertent release of engineered microorganisms or escapes of engineered fish that may crossbreed with natural populations are difficult to prevent. Many have ethical concerns about transgenic animals. Some question the social and institutional capacity to manage and regulate transgenic animals and products from them (Committee on Defining Science-Based Concerns Associated with Products of Animal Biotechnology and Committee on Agricultural Biotechnology Health and the Environment, 2002). Although animal biotechnology research is dynamic and powerful, the day of transgenic farm animals destined for commercial use to enter the food chain, according to a recent prominent review, ‘‘remains some distance in the future’’ (Laible, 2009). 4. Prospective commercial products 4.1. Drought tolerance in crop plants Drought is widely recognized as an important constraint to agriculture in many situations, and drought resistance has been a top-rated objective of

Biotechnology and Agriculture: Current and Emerging Applications

11

conventional breeding for decades (Chapman et al., 2000). However, despite that recognition, relatively little research was done until recently because both breeders and those allocating research funds believed there was limited prospect for success. With the application of biotechnology, significant progress has been achieved in understanding plant response to drought from the molecular through the whole plant level (Chaves et al., 2003; Bradford et al., 2005), and more recently drought has become a prime target of applied agricultural biotechnology programs (Herdt, 1991). Developments in this area exemplify progress on biotechnology’s biological front, and there are now numerous efforts to combine genomic tools with traditional breeding to incorporate drought tolerance (Ishitania et al., 2004). Two broad approaches are being used (Farooq et al., 2009). One attempts to optimize phenotype – traits like deep roots, vigorous root systems, small stature, stomata control, osmoregulation, and leaf epicuticular wax (Morgan, 1999; Wu et al., 2009). The second attempts to optimize biochemical recognition of and response to water stress within the plant. Some argue that relying on phenotype alone has limited prospects for success because ‘‘internal consistency in the correlations between presence of traits and the intervening processes are rarely proven beyond doubt. Therefore, in spite of several advantages offered by the analytical approach, impact will be limited until the physiological and biochemical components of critical traits are understood’’ (Seetharama, 1995). Transgenics advocates have focused on finding ‘‘drought genes’’ that can be transferred into crops. Abscisic acid, or ABA, is prominent among the hormones associated with drought response, increasing sharply when plants dry out or are exposed to low temperatures. However, identifying an ABA receptor, ‘‘a plant cell protein that recognizes the hormone and conveys its gene-regulating orders to the nucleus, has been full of frustration and controversy’’ (Pennisi, 2009). An important lesson from such efforts is that plant responses to abiotic stresses like drought may involve hundreds of genes with the function of many still unknown (Chaves et al., 2003).5 Structural genes coding for other plant products like ‘‘mannitol, trehalose, redox proteins, and detoxifying enzymes y and regulatory genes’’ are being experimentally transferred in ‘‘wheat, maize, sugarcane, tobacco, arabidopsis, groundnut, tomato, and potato’’ in pot experiments or in contained field trails (Gosal et al., 2009). While transgenic plants are created in laboratories, the resulting candidates for varietal development must be grown in fields and evaluated for phenotypic expression of targeted transgenic traits. Despite the Rockefeller Foundation’s 1990s support for biotechnology on rice in 5 However, the search continues for single genes that have the desired effect, for example, Selvam et al., (2009) identification of a novel drought tolerance gene in Gossypium hirsutum L.cv KC3. Communications in Biometry and Crop Science, 4, 9–13.

12

Robert W. Herdt and Rebecca Nelson

Asia, progress was limited by a shortage of capacity for high-quality, science-based, repeatable field phenotyping (Normile, 1999). The more recent support of crop plant breeding in Africa by the Bill & Melinda Gates Foundation has continued the emphasis on field phenotyping. ‘‘The major point that has finally sunk in is that regardless of whether you use conventional breeding, DNA marker-assisted breeding, DNA MAS, or creation of transgenic plants, you still need quality phenotyping capacity to reach your goal!’’6 While it makes sense that using QTLs together with transgenics should lead to drought tolerance, the near-term prospects are limited. MAS is now widely deployed in wheat, but Australian workers report it has ‘‘not contributed significantly to cultivar improvement for adaptation to lowyielding environments and breeding has relied largely on direct phenotypic selection for improved performance in these difficult environments. The limited success of the physiological and molecular breeding approaches now suggests that a careful rethink is needed of our strategies in order to understand better and breed for drought tolerance’’ (Fleury et al., 2010). For maize, a meta-analysis of QTL research on drought tolerance found hundreds of possible QTLs but concluded only that they would ‘‘further facilitate the identification of candidate genes for QTL and elucidate genetic mechanisms regulating drought tolerance’’ (Hao et al., 2010). In soybean, genetic engineering for drought tolerance is still ‘‘in progress’’ despite considerable efforts directed toward identifying traits associated with it (Manavalan et al., 2009). Research on crops important in low-income developing countries like cassava, cowpea, millet, and others has identified hundreds of genes induced by drought, and progress has been made in identifying molecular markers for drought tolerance (Lopez et al., 2002; Ribaut et al., 2010; Varshney et al., 2010). Recent efforts to evaluate some of the candidate genes provide some hope that tolerance can be improved by combining these, but as yet there are no breakthroughs with crops important in such countries (Herdt et al., 2007; Xiaoa et al., 2009).

4.2. Plant genetic modification for better human nutrition In addition to raising yield, biotechnology has the potential to improve the quality of food in the developing world by eliminating problems or incorporating beneficial traits. One large, albeit as yet unproven, project is designed to create more digestible and nutritious transgenic sorghum (Botha and Viljoen, 2008). Under some circumstances, transgenic pest resistance can reduce aflatoxin contamination of maize, a problem believed 6 Personal communication from John C. O’Toole. The authors appreciate the insights and comments of John O’Toole on the entire section on drought.

Biotechnology and Agriculture: Current and Emerging Applications

13

to be of huge dimensions in the developing world (Williams et al., 2004). Aflatoxins are toxic compounds produced by fungi that can enter the maize ear when it is damaged by insects. Biotechnology could make substantial contributions to human nutrition by increasing the density of vitamins and micronutrients like iron and zinc (Paine et al., 2005). Vitamin A deficiency, affecting an estimated 400 million rice-consuming people, could be overcome by a diversified diet incorporating vegetables and fruits, but the cost of such a diet puts it out of the reach of many and changing food habits is notoriously difficult (Dawe et al., 2002). In high-income countries, staple food products like rice are routinely fortified during processing. In the developing world, rice is processed in thousands of small units, making fortification impractical. Vitamin tablet supplements are supplied by foreign assistance, but such programs are inherently temporary. For these reasons, a transgenic rice-producing beta carotene, which the human body converts into vitamin A, was identified as a high priority at the early stages of the Rockefeller Foundation rice biotechnology program (Herdt, 1991; Normile, 1999).7 Within 10 years, ‘‘proof of concept’’ had been demonstrated with the innovation dubbed ‘‘Golden Rice’’ both for its color and for its potential to alleviate diseases. Unfortunately, like many other exciting innovations, Golden Rice was publicized before a practical version was available. Its developers found support (Dawe et al., 2002; Lusk and Rozanb, 2005) but the premature publicity attracted critics of genetic engineering as well (Altieri and Rosset, 1999; Massieu and Chauvet, 2005). To demonstrate the proof of concept, the research was undertaken using patented tools under research licenses from patent holders (Kryder et al., 2000). Converting the proof of concept into a product for farmers required the inventors to get commercialization licenses from the patent holders. The companies owning the relevant intellectual property saw the opportunity to build goodwill by granting those licenses for humanitarian purposes free of charge, and they did so. The Bill & Melinda Gates Foundation provided the Golden Rice Project one of its five-year ‘‘Grand Challenges in Health’’ grants to further improve the invention.8 By 2010, Golden Rice was going through national regulatory approval processes in Bangladesh, India, Indonesia, and the Philippines (The Golden Rice 7 The Rockefeller Foundation provided the funding for the early research on Golden Rice with no financial backing from any private company Normile, D. (1999). Rockefeller to end network after 15 years of success. Science, 286, 1468–9, Toenniessen, G. (2003) Opportunities for and challenges to plant breeding adoption in developing countries. Pullman, Washington, National Agricultural Biotechnology Conference, Herdt, R. W. (1995) The potential role of biotechnology in solving food production and environmental problems in developing countries. Agriculture and Environment: Bridging Food Production and Environmental Protection in Developing Countries. American Society of Agronomy. 8 See: http://www.goldenrice.org/Content5-GCGH/GCGH1.html

14

Robert W. Herdt and Rebecca Nelson

Project, 2009). Clinical nutritional studies have demonstrated the effectiveness of the rice as a source of vitamin A for adults (Tang et al., 2009), but only after varieties are approved for release will it be possible to know whether farmers will grow them widely, whether consumers will eat them, and how much they will do to reduce disease. Sweet potato is a staple crop in Africa, but most varieties are white in color and lack the vitamin A that is present in orange-colored fleshed varieties. Because some sweet potato varieties have high levels of vitamin A, it is possible to use conventional breeding to enhance the vitamin A content of sweet potato varieties that are suited to African tastes (staple varieties must be firm and not sweet). The International Potato Center and the national programs of Uganda and Mozambique have pursued nutritionally oriented breeding programs for the past 20 years with some products released for farmer use. The HarvestPlus initiative has helped to mobilize funding for the breeding and marker work needed to create nutrient-dense varieties and is currently promoting their use to improve the nutritional composition of beans, cassava, maize, pearl millet, rice, and sweet potato, all important food sources in developing countries (Bouis et al., 2009). Some foods like peanuts, soybean, and wheat naturally contain antinutritive compounds or factors that cause allergic reactions in some people; some plant biotechnology research is directed at overcoming such allergic reactions. Companies are developing transgenic plants with increased amounts of nutritionally desirable components like lysine and methionine or reduced amounts of undesirables like trans fats. Other biotechnology applications are aimed at creating soybean and canola with modified oil composition and maize with white, waxy, hard food grade endosperm, high oil, or high amylose (Cockburn, 2004). MAS is being used to change crops in other ways that make them better foods. In peanut, for example, markers are being used to change peanut oil composition to have a longer shelf life and improved health attributes (Shi et al., 2008; Chu et al., 2009). In rice, markers are being used to track the highly valued aroma trait (Shi et al., 2006).

4.3. Pharmaceuticals from transgenic plants ‘‘Pharming’’ may be one of the most revolutionary applications of biotechnology, promising plants that produce edible vaccines or compounds that can combat various maladies (Arawaka et al., 1998; Richter et al., 2000). Conventionally, vaccines and other pharmaceuticals are produced by bacteria, yeast, or other microorganisms, but they could also be produced in plants. This might be cheaper, administering them orally would be simpler than injecting, and pharming might eliminate the need for refrigerating manufactured pharmaceuticals. The biological feasibility

Biotechnology and Agriculture: Current and Emerging Applications

15

of edible vaccines was demonstrated decades ago (Haq et al., 1995), and the capacity to generate the biological indicators of immune response in humans through an edible vaccine for hepatitis B was demonstrated in the mid-2000s (Thanavala et al., 2005). Commercialization of products is slow. More than 120 companies, universities, and research institutes are actively involved in molecular farming with plants and nearly 20 plant-derived pharmaceuticals have been submitted for clinical trials (Jefferson-Moore and Traxler, 2005). By 2009, no plant-made pharmaceutical products had made it to the market (Basaran and Rodrıguez-Cerezo, 2008), and only two plant-produced vaccine-related products had made it through all production and regulatory hurdles (Rybicki, 2009). Like genetically engineered crops, plant-made pharmaceuticals have their opponents who highlight possible negative impacts and disadvantages. The darkest scenario involves a drug or an industrial chemical entering the food chain, whether unintentionally or as a deliberate act of terrorism. Because of the challenges associated with effectively isolating production of such transgenic crops, some have proposed banning their production outside confinement facilities (Union of Concerned Scientists, 2006).

4.4. Biofuels United States law required 13 billion gallons of renewable fuels (ethanol) by 2009 and an additional 21 billion gallons of advanced biofuels by 2022 (Taheripour and Tyner, 2008). Ethanol is produced by fermenting plant materials high in sugar and then distilling the result to extract the ethanol. Ethanol is made from maize grain in the United States, while advanced biofuels are to be made of crop byproducts or other biomass (sometimes called ‘‘lignocellulosic biomass’’).9 Many plant materials might provide the lignocellulosic biomass, including agricultural byproducts like straw, maize stalks, and hulls; crops produced intentionally for conversion to energy, like switchgrass and fast-growing trees; and components of municipal solid waste like yard trash, cardboard, and waste wood. ‘‘Conversion of lignocellulosic biomass to fermentable sugars presents significant technical and economic challenges, and its success depends largely on the development of effective pretreatment, efficient enzyme conversion of pretreatment lignocellulosic substrates to fermentable sugars, and stress-tolerant microbial biocatalysts’’ (Liu et al., 2008). The process uses enzymes to attack plant biomass to produce sugars from the complex 9 A useful, relatively non-technical review of biogas processes and its challenges is contained in Wilkie, A. C. (2008) Biomethane from biomass, biowaste, and biofuels. IN WALL, J. D., HARWOOD, C. S. & DEMAIN, A. (Eds) Bioenergy. Washington, D.C, American Society for Microbiology.

16

Robert W. Herdt and Rebecca Nelson

carbohydrate polymers in plant cell walls. The sugars are then fermented and distilled into ethanol as with maize. However, existing strains of the yeast Saccharomyces cervisiae used in fermentation convert glucose but not the xylose and other sugars derived from lignocelluloses. Genetic engineering of new strains of microorganisms that ferment more of the sugars into ethanol may lead to the further increases in conversion efficiency needed to make the process economically viable (Ljungdahl et al., 2008). Another concept is to create a transgenic microbe with the capacity of converting lignocellulosic biomass directly to sugars in the absence of added enzymes, so-called consolidated bioprocessing (Lynd et al., 2008). Still more speculative is the possibility of engineering microbes to produce free hydrogen which might then be captured for use as the ‘‘ultimate clean’’ transportation fuel (Rousset and Cournac, 2008). 4.5. Environmental bioremediation Bioremediation is the elimination of natural or human-induced pollutants by living organisms. Microorganisms are key players in bioremediation applications and are seen as ‘‘environmentally friendly and cost-effective alternative to physicochemical cleanup options. However, the strategy and outcome of bioremediation in open systems or confined environments depend on a variety of physicochemical and biological factors that need to be assessed and monitored’’ (Stenuita et al., 2008). Biotechnology could be used to create plants or microorganisms with enhanced capacity to sequester pollutants or to monitor the effectiveness or otherwise of bioremediation effects. 5. Emerging DNA-based techniques 5.1. Genomic selection While MAS is powerful for identifying loci for individual desirable traits, MAS strategies have limitations when working on agronomically important complex traits like yield and drought tolerance. Breeding programs that are sufficiently well resourced to implement high-throughput marker technologies can now apply newer approaches. Genomic selection, ‘‘a brute-force and black-box procedure that exploits cheap and abundant molecular markers’’ (Bernardo and Yu, 2007) is being used to optimize the selection of desirable individuals in breeding programs. The performance of an initial test set of material (e.g., a set of plant or animal lines representing the diversity of interest to the breeding program) is evaluated in representative environments. A large set of markers (not tagging a few genes, but covering the entire genome) is determined for the test set and, using sophisticated statistical methods, breeding values are

Biotechnology and Agriculture: Current and Emerging Applications

17

calculated for all the markers. In subsequent generations, selection is based exclusively on the breeding values. Since its initial conceptualization (Meuwissen et al., 2001), this approach has been successful in initial trials in animal breeding (Luan et al., 2009; Ødega˚rd et al., 2009). Based on the results of simulations, genomic selection is expected to have profound effects on plant improvement as well (Bernardo and Yu, 2007; Heffner et al., 2010).

6. Reflections and conclusions The contrast between the increase in scientific information made possible by DNA research and the limited set of genes used in agricultural applications is striking. Science can describe the genetic composition of an organism in great detail but is still unable to use that information to predict with assurance how a specific modification will affect the organism’s performance (Yano and Tuberosa, 2009). ‘‘DNA sequencing technology is undergoing a revolution with the commercialization of second generation technologies capable of sequencing thousands of millions of nucleotide bases in each run. The data explosion resulting from this technology is likely to continue y creating new opportunities for crop improvement.’’ However, ‘‘the challenge remains to convert this mass of data into knowledge that can be applied in crop breeding programs’’ (Edwards and Batley, 2009). Despite this, optimism about practical benefits remains high among many scientists. Potential avenues for impact include (1) identification of genes that encode novel mechanisms of drought resistant from noncrop species, (2) better selection tools in plant breeding, and (3) better understanding of heterosis (hybrid vigor) (Cannon et al., 2009). The transgenic incorporation of genes for herbicide tolerance and insect resistance in maize, soybeans, cotton, and canola has been a tremendous commercial success in eight countries. The insect-resistance and herbicide tolerance genes have been combined in so-called ‘‘stacked’’ ‘‘gene’’ technology. But every living organism has tens of thousands of genes with an array of alleles, and some genes have regulatory functions that enable them to modulate the expression of other genes making innumerable potential targets for transgenic and nontransgenic technology. Single-gene and ‘‘stacked-gene’’ technology generated by transgenic or cisgenic techniques have not demonstrated the broad and flexible power shown by conventional breeding that allows vast numbers of natural genetic variants to be recombined. Applications to a wider array of crops, including crops of particular importance in poor developing countries, have been slow. Although transgenics have captured public attention, to date most genetic improvement for most crops has been achieved by exploiting natural allelic variation through conventional breeding. Natural genetic variation within crop gene pools and their sexually compatible relatives include

18

Robert W. Herdt and Rebecca Nelson

valuable gene variants that are not obvious by observing phenotype, but such genetic variation can be utilized through nontransgenic breeding methods (Tanksley and McCouch, 1997). Many scientists believe that greater progress can be made in improving crop performance through the applications of genomics tools like diversity analysis in conventional plant breeding than by using genetic engineering (Goff and Salmeron, 2004; Thro et al., 2004; Naylor and Manning, 2005). Use of DNA-based and conventional methods is not mutually exclusive, of course, and contemporary breeding in well-resourced hands involves shrewd combinations. MAS has been more widely applied, but its practical value is not universal. When it is difficult, expensive, or time-consuming to assess a trait by phenotype, or when assessment requires specific environments or developmental stages, molecular markers may be cost-effective. By signaling the presence or absence of a desired gene or set of genes for a trait not easily observed, like deep roots for drought tolerance, MAS allows plant breeders to precisely and quickly reduce the number of plants carried from one generation to the next, saving time and money. Marker-assisted backcrossing10 can increase the speed and efficiency with which single loci of strong effect can be introduced into an otherwise-desirable genotype. The application of MAS was expected to especially accelerate breeding programs in developing countries. While it has been widely applied in China and India, it is not practiced on a routine basis by public breeding programs in sub-Sahara Africa and low-income developing countries in other regions because of its cost and the ‘‘shortage of well-trained personnel, inadequate high-throughput capacity, poor phenotyping infrastructure, lack of information systems or adapted analysis tools, or simply resource-limited breeding programs’’ (Ribaut et al., 2010). Diversity analysis, MAS, and genomic selection are being used in large breeding companies that have the capacity to invest substantial funds to implement integrated molecular breeding programs to produce elite populations, lines, and hybrids of the same few crops – maize, cotton, and soybean. These companies continue a process of rapid technical and practical innovation that allow for the selection of both natural and transgenic variation. While the price-tag on this capacity is currently out of reach of most public-sector research organizations in low-income, developing countries, a recent international initiative linking scientists from a number of countries and international institutes promises to bring some of these capacities to bear on advancing the genetic improvement of important food crops in poor developing countries (Varshney et al., 2010). Expectations surrounding the application of biotechnologies in agriculture have been substantial, both for those who support and those who

10 Crossing an early generation progeny with one of its parents in order to achieve offspring with a genome identity which is closer to that of the parent.

Biotechnology and Agriculture: Current and Emerging Applications

19

oppose the technology. In this chapter, we argued that trends have deviated from the anticipated pathways in several ways. The greatest expectations and debate have centered on transgenic organisms, and the commercialization of direct gene transfer technologies has been widespread on maize and soybean in the United States and a few other countries, and relatively widespread on cotton. However, transgenic applications have been limited to a very few genes and few other crops, in part because of opposition to transgenics in some quarters. But the impact of DNA-based technologies on agriculture has been more subtle and widespread than one would see from a focus on transgenics alone. DNA markers have contributed substantially, and their use has not stirred public debate in itself, probably because they are used to understand natural diversity and to select among breeding lines, rather than to create novel genes. DNA markers were expected to transform crop and livestock breeding, but their application has been more limited in low-income countries than many practitioners anticipated. Some notable successes have been achieved, such as the efficient conversion to submergence tolerance of rice varieties that are widely grown in flood-prone areas, but few individual underutilized genes with such spectacular potential impacts have been identified and successful MAS has been limited to date. It is expected that genomic selection, using many markers covering the entire genome, will be more effective in optimizing allelic combinations in future breeding efforts. Such application depends on tremendous technical advances that have been made in genomic science, but is not dependent on the understanding of biological processes that has been the focus of the substantial public-sector genomic investment. DNA markers have enabled a much deeper and more profound understanding of the genetic diversity of crops, livestock, pests, pathogens, and beneficial organisms associated with agricultural systems. This understanding has allowed natural diversity to be better managed, used, and conserved. Successful application of most aspects of molecular breeding has been achieved in large breeding companies that can afford to integrate expensive and rapidly evolving genomic and computational technology platforms into powerful, field-based breeding programs. The public-sector programs responsible for providing crop varieties for farmers in developing countries find molecular breeding technologies even more inaccessible. Support from governments and donors may eventually allow breeding programs in developing countries to apply advanced technologies, but their success will require that both the molecular and the conventional field-based breeding programs are strong. References Abalo, G., Tongoona, P., Derera, J., Edema, R. (2009), A comparative analysis of conventional and marker-assisted selection methods in

20

Robert W. Herdt and Rebecca Nelson

breeding Maize Streak virus resistance in Maize. Crop Science 49, 509–520. Akhond, M.A.Y., Machray, G.C. (2009), Biotech crops: technologies, achievements and prospects. Euphytica 166, 47–59. Altieri, M.A., Rosset, P. (1999), Strengthening the case for why biotechnology will not help the developing world: a response to McGloughlin. Agbioforum, 2. Arawaka, T., Chong, D.K.X., Langridge, W.H.R. (1998), Efficacy of a food plant-based oral cholera toxin B subunit vaccine. Nature Biotechnology 16, 292–297. Basaran, P., Rodrıguez-Cerezo, E. (2008), Plant molecular farming: opportunities and challenges. Critical Reviews in Biotechnology 28, 153–172. Bernardo, R., Yu, J. (2007), Prospects for genomewide selection for quantitative traits in Maize. Crop Science 47, 1082–1090. Bhullar, N.K., Street, K., Mackay, M., Yahiaoui, N., Keller, B. (2009), Unlocking wheat genetic resources for the molecular identification of previously undescribed functional alleles at the Pm3 resistance locus. Proceedings of the National Academy of Sciences of the United States of America 106, 9519–9524. Botha, G.M., Viljoen, C.D. (2008), Can GM sorghum impact Africa? Trends in Biotechnology 26 (2), 64–69. Bouis, H.E., Hotz, C., McClafferty, B., Meenakshi, J.V. (2009), Biofortification: a new tool to reduce micronutrient malnutrition. 19th International Congress of Nutrition. Bangkok, Thailand, Harvest Plus a projects of IFPRI. Bradford, K.J., Van Deynze, A., Gutterson, N., Parrott, W., Strauss, S.H. (2005), Regulating transgenic crops sensibly: lessons from plant breeding, biotechnology and genomics. Nature Biotechnology 23, 439–444. Bruening, G., Lyons, J.M. (2000), The case of the FLAVR SAVR tomato. California Agriculture 54 (4), 6–7. Cannon, S.B., May, G.D., Jackson, S.A. (2009), Three sequenced legume genomes and many crop species: Rich opportunities for translational genomics. Plant Physiology 154 (3), 970–977. Chapman, S.C., Hammer, G.L., Butler, D.G., Cooper, M. (2000), Genotype by environment interactions affecting grain sorghum. III. Temporal sequences and spatial patterns in the target population of environments. Australian Journal of Agricultural Research 51 (2), 223–233. Chaves, M.M., Maroco, J.P., Pereira, J.S. (2003), Understanding plant responses to drought – from genes to the whole plant. Functional Plant Biology 30, 239–264. Chu, Y., Holbrook, C.C., Ozias-Akins, P. (2009), Two alleles of ahFAD2B control the high oleic acid trait in cultivated peanut. Crop Science 49, 2029–2036.

Biotechnology and Agriculture: Current and Emerging Applications

21

Cockburn, A. (2004), Commercial plant breeding: what is in the biotech pipeline? Journal of Commercial Biotechnology 10, 209–233. Committee on Defining Science-Based Concerns Associated with Products of Animal Biotechnology & Committee on Agricultural Biotechnology Health and the Environment, N. R. C. (2002), Animal Biotechnology: Science based Concerns. The National Academies Press, Washington, DC. Dalrymple, D.G. (1975), Measuring the green revolution: the impact of research on wheat and rice production. Washington, DC, Foreign Development Division, Economic Research Service, US Department of Agriculture. Dawe, D., Robertson, R., Unnevehr, L. (2002), Golden rice: what role could it play in alleviation of vitamin A deficiency? Food Policy 27, 541–560. Duran, C., Appleby, N., Edwards, D., Batley, J. (2009), Molecular genetic markers: discovery, applications, data storage and visualisation. Current Bioinformatics 4, 16–27. Edwards, D., Batley, J. (2009), Plant genome sequencing: Applications for crop improvement. Plant Biotechnology 8 (1), 2–9. Faber, D., Molina, J., Ohlrichs, C., Vander Zwaag, D., Ferre, L. (2003), Commercialization of animal biotechnology. Theriogenology 59, 125–138. Farooq, M., Wahid, A., Lee, D.-J., Ito, O., Siddique, K.H.M. (2009), Advances in drought resistance of rice. Critical Reviews in Plant Science 28, 199–217. Fu, Y.B., Somers, D.J. (2009), Genome-wide reduction of genetic diversity in wheat breeding. Crop Science 49, 161–168. Goff, S.A., Salmeron, J.M. (2004), Back to the future of cereals. Scientific American 291, 42–49. Gosal, S.S., Wani, S.H., Manjit S, K. (2009), Biotechnology and drought tolerance. Journal of Crop Improvement 23, 19–54. Griliches, Z. (1957), Hybrid corn: an exploration in the economics of technical change. Econometrica 25 (4), 501–522. Hansel, W. (2003), The potential for improving the growth and development of cultured farm animal oocytes. Animal Reproduction Sciences 79, 191–201. Hao, Z., Li, X., Liu, X., Xie, C., Li, M., Zhang, D., Zhang, S. (2010), Meta-analysis of constitutive and adaptive QTL for drought tolerance in maize. Euphytica 174, 165–177. Haq, T.A., Mason, H.S., Clements, J.D., Arntzen, C.S. (1995), Oral immunization with a recombinant bacterial antigen produced in transgenic plants. Science 268, 714–716. Harding, K., Benson, E.E., Clacher, K. (1997), Plant conservation biotechnology: an overview. Agro Food Industry Hi-Tech 8, 24–29. Hasler, J.F. (2003), The current status and future of commercial embryo transfer in cattle. Animal Reproduction Science 79, 245–264.

22

Robert W. Herdt and Rebecca Nelson

Heffner, E.L., Lorenz, A.J., Jannink, J.-L., Sorrells, M.E. (2010), Plant breeding with genomic selection: Gain per unit time and cost. Crop Science 50 (September–October). Herdt, R., Toenniessen, G., O’Toole, J. (2007), Plant biotechnology for developing countries. In: Evenson, R. (Ed.), Handbook of Agricultural Economics (Vol. 3). Elsiver, Amsterdam. Herdt, R.W. (1991), Research priorities for rice biotechnology. In: Khush, G., Toenniessen, G. (Eds.), Rice Biotechnology. CAB International, Wallingford, UK. Herdt, R.W. (1995), The potential role of biotechnology in solving food production and environmental problems in developing countries. Agriculture and Environment: Bridging Food Production and Environmental Protection in Developing Countries. American Society of Agronomy, Madison, WI. Houdebine, L.-M. (2009), Production of pharmaceutical proteins by transgenic animals. Comparative Immunology, Microbiology and Infectious Diseases 32, 107–121. Ibanez, J., Velez, M.D., de Andres, M.T., Borrego, J. (2009), Molecular markers for establishing distinctness in vegetatively propagated crops: a case study in grapevine. Theoretical and Applied Genetics 119, 1213– 1222. Ishitania, M., Raoa, I., Wenzlb, P., Beebea, S., Tohmea, J. (2004), Integration of genomics approach with traditional breeding towards improving abiotic stress adaptation: drought and aluminum toxicity as case studies. Field Crops Research 90, 35–45. Jacobsen, E., Schouten, H.J. (2007), Cisgenesis strongly improves introgression breeding and induced translocation breeding of plants. Trends in Biotechnology 25, 219–223. Jefferson-Moore, K.Y., and Traxler, G. (2005), Second-generation GMOs: where to from here? Agbioforum, 8. Jena, K.K., Mackill, D.J. (2008), Molecular markers and their use in marker-assisted selection in rice. Crop Science 48, 1266–1276. Jeung, J.U., Heu, S.G., Shin, M.S., Cruz, C.M.V., Jena, K.K. (2006), Dynamics of Xanthomonas oryzae pv. oryzae populations in Korea and their relationship to known bacterial blight resistance genes. Phytopathology 96, 867–875. Jones, N., Ougham, H., Thomas, H., Pasakinskiene, I. (2009), Markers and mapping revisited: finding your gene. New Phytologist 183, 935–966. Kryder, D., Kowalski, S., Krattiger, A. (2000), The intellectual and technical property components of pro-vitamin A (Golden Rice): a preliminary freedom-to-operate review. International Service for the Acquisition of Agri-biotech Applications (ISAAA), Ithaca, NY. Laible, G. (2009), Enhancing livestock through genetic engineering – recent advances and future prospects. Comparative Immunology, Microbiology and Infectious Diseases 32, 123–137.

Biotechnology and Agriculture: Current and Emerging Applications

23

Liu, Z.L., Saha, B.C., Slininger, P.J. (2008), Lignocellulosic biomass conversion to ethanol by Saccharomyces. In: Wall, J.D., Harwood, C.S., Demain, A. (Eds.), Bioenergy. American Society of Microbiology, Washington, DC. Ljungdahl, L.G., Kataeva, I.A., Uversky, V.N. (2008), Contribution of domain interactions and calcium binding to the stability of carbohydrate-active enzymes. In: Wall, J.D., Harwood, C.S., Demain, A. (Eds.), Bioenergy. American Society for Microbiology, Washington, DC. Lopez, C.G., Banowetz, G.M., Peterson, C.J., Kronstad, W.E. (2002), Wheat dehydrin accumulation in response to drought stress during anthesis. Functional Plant Biology 29, 1417–1425. Luan, T., Woolliams, J.A., Lien, S., Kent, M., Svendsen, M., Meuwissen, T.H.E. (2009), The accuracy of genomic selection in Norwegian red cattle assessed by cross validation. Genetics 183, 1119–1126. Lusk, J.L., Rozanb, A. (2005), Consumer acceptance of biotechnology and the role of second generation technologies in the USA and Europe. Trends in Biotechnology 28, 386–387. Lynd, L.R., Currie, D., Ciazza, N., herring, C., Orem, N. (2008), Consolidated bioprocessing of cellulosic biomass to ethanol using thermophilic bacteria. In: Wall, J.D., Harwood, C.S., Demain, A. (Eds.), Bioenergy. American Society of Microbiology, Washington, DC. Manavalan, L.P., Guttikonda, S.K., Tran, L.-S.P., Nguyen, H.T. (2009), Physiological and molecular approaches to improve drought resistance in soybean. Plant & Cell Physiology 50, 1260–1276. Manici, L.M., Caputo, F. (2009), Fungal community diversity and soil health in intensive potato cropping systems of the east Po valley, northern Italy. Annals of Applied Biology 155, 245–258. Mary Ann Liebert Inc. (2009), First ‘‘biopharm’’ product approved by FDA. Biotechnology Law Report 28, 167. Massieu, Y., Chauvet, M. (2005), Contesting biotechnology: crosscontinental concerns about genetically modified crops. In: Pritchard, B., Fold, N. (Eds.), Cross-continental Agro-food Chains. Structures, Actors and Dynamics in the Global Food System. Routledge, London, UK. McCouch, S.R., Sweeney, M., Li, J.M., Jiang, H., Thomson, M., Septiningsih, E., Edwards, J., Moncada, P., Xiao, J.H., Garris, A., Tai, T., Martinez, C., Tohme, J., Sugiono, M., McClung, A., Yuan, L.P., Ahn, S.N. (2007), Through the genetic bottleneck: O. rufipogon as a source of trait-enhancing alleles for O. sativa. Euphytica 154, 317–339. Meuwissen, T.H., Hayes, B.J., Goddard, M.E. (2001), Prediction of total genetic value using genome-wide dense marker maps. Genetics 157, 1819–1829. Moreau, L., Charcosset, A., Gallais, A. (2004), Experimental evaluation of several cycles of marker-assisted selection in maize. Euphytica 137 (1), 111–118.

24

Robert W. Herdt and Rebecca Nelson

Morgan, J.M. (1999), Pollen grain expression of a gene controlling differences in osmoregulation in wheat leaves: a simple breeding method. Australian Journal of Agricultural Research 50, 953–962. Naylor, R., Manning, R. (2005), Unleashing the genius of the genome to feed the developing world. Proceedings of the American Philosophical Society 149, 515–528. Nino-Liu, D.O., Ronald, P.C., Bogdanove, A.J. (2006), Xanthomonas oryzae pathovars: model pathogens of a model crop. Molecular Plant Pathology 7, 303–324. Normile, D. (1999), Rockefeller to end network after 15 years of success. Science 286, 1468–1469. Ødega˚rd, J., Yazdi, M.H., Sonesson, A.K., Meuwissen, T.H.E. (2009), Incorporating desirable genetic characteristics from an inferior into a superior population using genomic selection. Genetics 181, 737– 745. Paine, J.A., Shipton, C.A., Chaggar, S., Howells, R.M., Kennedy, M.J., Vernon, G., Wright, S.Y., Hinchliffe, E., Adams, J.L., Silvrestone, A.L. (2005), Improving the nutritional value of golden rice through increased pro-vitamin content. Nature Biotechnology 23, 482–487. Park, T.H., Vleeshouwers, V.G.A.A., Jacobsen, E., van der Vossen, E., Visser, R.G.F. (2009a), Molecular breeding for resistance to Phytophthora infestans (Mont.) de Bary in potato (Solanum tuberosum L.): a perspective of cisgenesis. Plant Breeding 128, 109–117. Park, Y.J., Lee, J.K., Kim, N.S. (2009b), Simple Sequence Repeat Polymorphisms (SSRPs) for evaluation of molecular diversity and Germplasm classification of minor crops. Molecules 14, 4546–4569. Pennisi, E. (2009), Stressed out over a stress hormone. Science 324, 1012– 1013. Perez-Vega, E., Campa, A., De la Rosa, L., Giraldez, R., Ferreira, J.J. (2009), Genetic diversity in a core collection established from the main bean Genebank in Spain. Crop Science 49, 1377–1386. Poland, J.A., Balint-Kurti, P.J., Wisser, R.J., Pratt, R., Nelson, R.J. (2008), Shades of gray: The world of quantitative disease resistance. Trends in Plant Science 14 (1), 21–29. Reif, J.C., Melchinger, A.E., Xia, X.C., Warburton, M.L., Hoisington, D.A., Vasal, S.K., Srinivasan, G., Bohn, M., Frisch, M. (2003), Genetic distance based on simple sequence repeats and heterosis in tropical maize populations. Crop Science 43, 1275–1282. Ribaut, J.-M., Vicente, M.C.d., Delannay, X. (2010), Molecular breeding in developing countries: Challenges and perspectives. Current Opinion in Plant Biology 13 (2), 213–218. Richter, L.J., Thanavala, Y., Arntzen, C.J. (2000), Production of Hepatitis B surface antegen in transgenic plants for oralimmunization. Nature Biotechnology 18, 302–305.

Biotechnology and Agriculture: Current and Emerging Applications

25

Rick, C.M., Tanksley, S.D., Fobes, J.F. (1979), Pseudoduplication in Lycopersicon pimpinellifolium. Proceedings of the National Academy of Sciences of the United States of America 76, 3435–3439. Rommens, C.M., Haring, M.A., Swords, K., Davies, H.V., Belknap, W.R. (2007), The intragenic approach as a new extension to traditional plant breeding. Trends in Plant Science 12, 397–403. Rommens, C.M., Kishore, G.M. (2000), Exploiting the full potential of disease-resistance genes for agricultural use. Current Opinion in Biotechnology 11, 120–125. Rousset, M., Cournac, L. (2008), Toward hydrogenase engineering for hydrogen production. In: Wall, J.D., Harwood, C.S., Demain, A. (Eds.), Bioenergy. American Society for Microbiology, Washington, D.C. Rybicki, E.P. (2009), Plant-produced vaccines: promise and reality. Drug Discovery Today 14, 16–24. Sarkar, R.K., Panda, D., Reddy, J.N., Patnaik, S.S.C., Mackill, D.J., Ismail, A.M. (2009), Performance of submergence tolerant rice (Oryza sativa) genotypes carrying the Sub1 quantitative trait locus under stressed and non-stressed natural field conditions. Indian Journal of Agricultural Sciences 79, 876–883. Schouten, H.J., Krens, F.A., Jacobsen, K., Jacobsen, E. (2006), Cisgenic plants are similar to traditionally bred plants – international regulations for genetically modified organisms should be altered to exempt cisgenesis. Embo Reports 7, 750–753. Schubert, D., Williams, D. (2006), ‘‘Cisgenic’’ as a product designation. Nature Biotechnology 24, 1327–1329. Seetharama, N. (1995), Biotechnology and sorghum improvement for drought and temperature stress tolerance. African Crop Science Journal 3, 223–229. Selvam, J.N., Kumaravadivel, N., Gopikrishan, A., Kuman, B.K., Ravikesavan, R., Boopathi, M. (2009), Identification of a novel drought tolerance gene in Gossypium hirsutum L.cv KC3. Communications in Biometry and Crop Science 4, 9–13. Shi, W.W., Yang, Y., Chen, S.H., Xu, M.L. (2008), Discovery of a new fragrance allele and the development of functional markers for the breeding of fragrant rice varieties. Molecular Breeding 22, 185–192. Singh, S., Sidhu, J.S., Huang, N., Vikal, Y., Li, Z., Brar, D.S., Dhaliwal, H.S., Khush, G.S. (2001), Pyramiding three bacterial blight resistance genes (xa5, xa13 and Xa21) using marker-assisted selection into indica rice cultivar PR106. TAG Theoretical and Applied Genetics 102 (6–7), 1011–1015. St.Clair, D.A. (2010), Quantitative disease resistance and quantitative resistance loci in breeding. Annual Review of Phytopathology 48, 247– 268. Steele, K.A., Gyawali, S., Joshi, K.D., Shrestha, P., Sthapit, B.R., Witcombe, J.R. (2009), Has the introduction of modern rice varieties

26

Robert W. Herdt and Rebecca Nelson

changed rice genetic diversity in a high-altitude region of Nepal? Field Crops Research 113, 24–30. Stenuita, B., Eyersa, L., Schulera, L., Agathosa, S.N., George, I. (2008), Emerging high-throughput approaches to analyze bioremediation of sites contaminated with hazardous and/or recalcitrant wastes. Biotechnology Advances 26, 561–575. Taheripour, F., Tyner, W.E. (2008), Ethanol policy – what have we learned so far? Choices, 23. Tang, G., Qin, J., Dolnikowski, G.G., Russell, R.M., Grusak, M.A. (2009), Golden Rice is an effective source of vitamin A. The American Journal of Clinical Nutrition 89, 1–8. Tanksley, D., Medina, H., Rick, C.M. (1981), The effect of isoenzyme selection on metric characters in an interspecific backcross of tomato – basis of an early screening-procedure. Theoretical and Applied Genetics 60, 291–296. Tanksley, S.D., McCouch, S.R. (1997), Seed banks and molecular maps: unlocking genetic potential from the wild. Science 277, 1063–1066. Thanavala, Y., Mahoney, M., Pal, S., Scott, A., Richter, L., Natarajan, N., Goodwin, P., Arntzen, C.J., Mason, H.S. (2005), Immunogenicity in humans of an edible vaccine for hepatitis B. PNAS 102, 3378–3382. The Golden Rice Project. (2009), Available at: http://www.goldenrice.org/ index.html. Thomson, M.J., Septiningsih, E.M., Suwardjo, F., Santoso, T.J., Silitonga, T.S., McCouch, S.R. (2007), Genetic diversity analysis of traditional and improved Indonesian rice (Oryza sativa L.) germplasm using microsatellite markers. Theoretical and Applied Genetics 114, 559–568. Thro, A.M., Parrrott, W., Udall, J.A., Beavis, W.D. (2004), Symposium: genomics and plant breeding: the experience of the initiative for future agriculture and food systems. Crop Science 44, 1893–1919. Toenniessen, G. (2003), Opportunities for and challenges to plant breeding adoption in developing countries. National Agricultural Biotechnology Conference, Pullman, Washington. U.S. Department of Health and Human Services, Food and Drug Administration, Center for Veterinary Medicine (CVM). (2009), Regulation of genetically engineered animals containing heritable recombinant DNA constructs. Biotechnology Law Review 28, 227–240. Union of Concerned Scientists. (2006), UCS Position Paper: Pharmaceutical and Industrial Crops. Union of Concerned Scientists. Vallejos, C.E., Tanksley, S.D. (1983), Segregation of isozyme markers and cold tolerance in an interspecific backcross of tomato. Theoretical and Applied Genetics 66, 241–247. van Heerwaarden, J., Hellin, J., Visser, R.F., van Eeuwijk, F.A. (2009), Estimating maize genetic erosion in modernized smallholder agriculture. Theoretical and Applied Genetics 119, 875–888.

Biotechnology and Agriculture: Current and Emerging Applications

27

Varshney, R.K., Glaszmann, J.-C., Leung, H., Ribaut, J.-M., et al. (2010), More genomic resources for less-studied crops. Trends in Biotechnology 28 (9), 452–460. Varshney, R.K., Nayak, S.N., May, G.D., Jackson, S.A. (2009), Nextgeneration sequencing technologies and their implications for crop genetics and breeding. Trends in Biotechnology 27, 522–530. Venuprasad, R., Dalid, C.O., Del Valle, M., Zhao, D., Espiritu, M., Cruz, M.T.S., Amante, M., Kumar, A., Atlin, G.N. (2009), Identification and characterization of large-effect quantitative trait loci for grain yield under lowland drought stress in rice using bulk-segregant analysis. Theoretical and Applied Genetics 120, 177–190. Waltz, E. (2010), China’s GM rice first. Nature Biotechnology 28 (8), 8. Wilkie, A.C. (2008), Biomethane from biomass, biowaste, and biofuels. In: Wall, J.D., Harwood, C.S., Demain, A. (Eds.), Bioenergy. American Society for Microbiology, Washington, DC. Williams, J.H., Phillips, T.D., Jolly, P.E., Stiles, J.K., Jolly, C.M., Aggarwal, D. (2004), Human aflatoxicosis in developing countries: a review of toxicology, exposure, potential health consequences, and interventions. American Journal of Clinical Nutrition 80, 1106–1122. Wisser, R.J., Sun, Q., Hulbert, S.H., Kresovich, S., Nelson, R.J. (2005), Identification and characterization of regions of the rice genome associated with broad-spectrum, quantitative disease resistance. Genetics 169 (4), 2277–2293. Wu, X., Shiroto, Y., Kishitani, S., Ito, Y., Toriyama, K. (2009), Enhanced heat and drought tolerance in transgenic rice seedlings overexpressing OsWRKY11 under the control of HSP101 promoter. Plant Cell Reports 28, 21–30. Xiaoa, B.-Z., Chenb, X., Xiangb, C.-B., Tanga, N., Zhanga, Q.-F., Xiong, L.-Z. (2009), Evaluation of seven function-known candidate genes for their effects on improving drought resistance of transgenic rice under field conditions. Molecular Plant 2, 73–83. Yano, M., Tuberosa, R. (2009), Genome studies and molecular genetics – from sequence to crops: genomics comes of age. Current Opinion in Plant Biology 12, 103–106.

CHAPTER 2

Genetically Modified Crops and Global Food Security Matin Qaim Department of Agricultural Economics and Rural Development, Georg-August-University of Goettingen, 37073 Goettingen, Germany E-mail address: [email protected]

Abstract Purpose – The role of genetically modified (GM) crops for food security is the subject of controversial debates. Consequently, policy-makers are unsure whether this technology is suitable for developing countries. This chapter reviews the scientific evidence. Methodology/approach – Starting from a food security definition, potential pathways of how GM crops could contribute to hunger reduction are analyzed conceptually. Furthermore, studies about the socioeconomic impacts of GM crop applications are reviewed. This includes ex post studies for present applications such as insect-resistant and herbicidetolerant crops, as well as ex ante studies for future GM technologies such as Golden Rice and drought-tolerant varieties. Findings – GM crops can raise agricultural productivity and thus contribute to better food availability. Especially when tailored to small farm conditions, GM crops can also cause income increases for the rural poor, entailing better access to food. Nutritionally enhanced, biofortified GM crops could reduce problems of micronutrient malnutrition in a costeffective way. Research limitations – The examples observable so far are still limited. Impacts also depend on the wider institutional setting. Like any technology, GM crops are not a substitute but a complement to much needed institutional and infrastructure improvement in developing countries. Social implications – The fact that available GM crops already contribute to poverty reduction and improved food security has not been widely recognized up until now.

Frontiers of Economics and Globalization Volume 10 ISSN: 1574-8715 DOI: 10.1108/S1574-8715(2011)0000010007

r 2011 by Emerald Group Publishing Limited. All rights reserved

30

Matin Qaim

Value of paper – Results presented in this chapter can contribute to a more constructive public debate, in which GM crop risks are not discussed out of the context of actual and potential benefits. Keywords: Food security, poverty, smallholder farmers, developing countries, biotechnology JEL Classifications: O13, O33, Q12, Q16, Q18

1. Introduction Globally, around 1 billion people are currently undernourished, that is, they suffer from insufficient calorie supplies. Almost all these people live in developing countries, especially in Asia and Sub-Saharan Africa (FAO, 2009). The first millennium development goal (MDG) of the United Nations foresees halving hunger by 2015. Unfortunately, this goal will not be achieved. The trend is even moving into the wrong direction: recently, not only the absolute number but also the proportion of undernourished people has risen, which is partly due to rising food prices combined with the global financial and economic crisis (von Braun, 2008).1 What are the appropriate instruments to reduce hunger and improve global food security? In this regard, the role of agricultural technology, in general, and of genetically modified (GM) crops, in particular, is the subject of controversial debates. Some consider hunger as only a distribution problem (Sharma, 2004; Holt-Gimenez et al., 2006). In their view, promoting technological progress is not an important policy approach; rather, social policies – such as improved education, health, and income redistribution – are seen as the key elements of a hunger reduction strategy. However, while the importance of social policies is undisputed, focusing on distribution alone is too shortsighted, as it neglects the fact that global food demand is increasing rapidly due to population and income growth. Over the past 10 years, growth in global demand for cereals (including for feed and biofuels) has outpaced supply; growth rates in yields of major cereals have even been declining since the 1990s (FAO, 2010; also see Figure 1). Recent international food price spikes, which were caused by various factors, have contributed to a wider public recognition of the need for more robust agricultural production increases (Godfray et al., 2010). Nonetheless, there 1 The statistics on the number of undernourished published regularly by the Food and Agriculture Organization (FAO) are officially used as one indicator to track progress toward the first MDG. It should be noted that these FAO statistics have been criticized not only for being imprecise but also for systematically overestimating the number of undernourished (e.g., Svedberg, 2002). However, different studies with detailed household level data have also shown that food price increases contribute to rising rates of hunger and poverty, at least in the short and medium run (e.g., Ivanic and Martin, 2008; Ecker and Qaim, 2011).

Genetically Modified Crops and Global Food Security

31

4.5 Rice Wheat

4 Annual growth in %

3.5 3 2.5 2 1.5 1 0.5 0 1960s

1970s

1980s

1990s

2000s

Fig. 1. Worldwide yield growth (1960–2008). Source: Own presentation based on data from FAO (2010). Note: Growth rates in global mean yields were calculated on an annual basis and then averaged over the respective time periods.

is no consensus on how this should be achieved and what role modern biotechnology could play (IAASTD, 2009; Gurian-Sherman, 2009). While some see GM crop technologies as a necessary tool for achieving long-term food security (Borlaug, 2007), others are concerned about negative economic and social consequences that these technologies could have for the poor (Sharma, 2004; FOE, 2008; Shiva, 2009). This chapter contributes to the debate by reviewing the academic literature on socioeconomic impacts of GM crops. Empirical evidence shows that this technology offers great potential to contribute to the reduction of hunger and malnutrition. Yet, concrete examples are still limited; realizing the potential on a larger scale will require more public and policy support. The rest of this chapter is structured as follows: in the next section, different potential pathways of how GM crops can improve global food security are analyzed conceptually. Then, observable impacts of GM crops that have already been commercialized are reviewed, before ex ante studies related to future GM crop applications are summarized. Subsequently, institutional and policy issues are discussed, and some conclusions are drawn.

2. GM crops and food security: potential pathways According to the FAO, food security exists when all people, at all times, have physical and economic access to sufficient, safe, and nutritious food that

32

Matin Qaim

meets their dietary needs and food preferences for an active and healthy life (FAO, 2009). This involves at least three dimensions, namely (1) physical access to sufficient food, which is a question of global and local food availability, (2) economic access to food, which is related to household income, and (3) food safety and nutritional value. In principle, GM crops can positively contribute to all three dimensions, as is explained below.

2.1. GM crops and food availability On the basis of FAO food balance sheet data, currently there is enough food available at the global level to feed the world population, at least in terms of calories. Nevertheless, there are around 1 billion people undernourished, which underlines that hunger is a serious distribution problem: whereas some people consume and waste too much food, others have too little. This phenomenon is detailed below in connection with the economic accessibility of food. However, only looking at the situation today is too static, as it neglects past developments as well as future trends. From a dynamic perspective, beyond distribution hunger is also a production problem. The fact that currently enough food is available is attributable to tremendous historical production increases. Successes in crop breeding, coupled with more irrigation and use of agrochemicals, tripled cereal yields over the past 50 years in many parts of the world, including in Asia and Latin America (FAO, 2010). These productivity gains became known as the green revolution; they outpaced population growth and helped to prevent widespread famines that had been predicted in the early 1960s (Evenson and Gollin, 2003). But food demand will further rise in the future. Through population and income growth, global demand will increase by at least 70% until 2050 (Godfray et al., 2010). Moreover, the use of biofuels soars, competing with food production for scarce natural resources, such as arable land and water (Dewbre et al., 2008). While arable land is still being expanded in some regions, soil degradation and urbanization contribute to agricultural area losses elsewhere. Total arable land can hardly be increased without causing serious environmental problems. Hence, food production increases will have to come from higher yields on the given land. Against this background, it is particularly worrisome that yield growth in major cereals has been declining over the past 20 years. While yield growth in rice and wheat was still around 3% per year in the 1980s, it has now dropped to below 1% (Figure 1). This is too little to keep pace with growth in global food demand. To sustain sufficient food availability until 2050, a minimum yield growth of 1.5% per year is required. This will only be possible through higher investments in agricultural research, including the use of new technologies. Raising agricultural productivity in a sustainable way will require a mix of different technologies, adjusted to the specific conditions in a particular

33

Genetically Modified Crops and Global Food Security

Wheat

setting. Tapping genetic knowledge will have a major role to play, because this can help reduce the strong correlation between yields and agrochemical use observed in the past, which has often led to negative environmental externalities (Huang et al., 2002; World Bank, 2007). One potential avenue is improving the effectiveness of pest control. A significant proportion of the potential world harvest is lost to weeds, animal pests, and diseases. Figure 2 shows that potential losses in some crops can reach 80% and more. A sizeable portion of these potential losses is avoided through chemical pesticides and other pest-control strategies, but 30%–40% occurs as actual damage. Actual losses are higher in developing countries than in developed countries, because pest pressure in tropical and subtropical climates is often stronger than in temperate zones (Oerke, 2006). Moreover, given more severe technical and financial constraints, pest control is often less effective in developing countries. In addition to reducing chemical pesticide use, crops with inbuilt genetic pest resistance have the potential to further reduce crop losses and thus increase effective yields. Positive yield effects of pest-resistant crops are expected to be higher in developing countries (Qaim and Zilberman, 2003). While conventional breeders also try to develop plants with pest resistance traits, GM techniques offer new opportunities because a much wider gene pool can be used. Insect and virus resistance were among the first GM traits to be commercialized in some crops, but fungal- and bacterialresistant GM crops are also approaching the end of the research pipeline (Kempken and Jung, 2010).

actual

Weeds Animal pests

potential

Cotton

Potatoes

Maize

Rice

Diseases actual potential actual potential actual potential actual potential 0

Fig. 2.

10

20

30

40 50 60 Crop losses in %

70

80

90

Global pest-related crop losses in major crops. Source: Oerke (2006).

34

Matin Qaim

Other GM traits that researchers are working on are higher plant tolerance to various abiotic stresses such as heat, drought, flood, coldness, or soil salinity (Qaim, 2009; Kempken and Jung, 2010). Such technologies could also contribute to higher and more stable yields, especially in regions affected by erratic weather conditions. Again, developing countries could benefit more than developed countries, because of higher weather variability. Moreover, especially in Africa farmers often have limited access to irrigation and other risk-reducing technologies. Abiotic stress tolerance is particularly relevant against the background of climate change. Climate change is not only associated with an increase in mean temperatures, but also with more frequent weather extremes. The first drought- and heat-tolerant GM crops are expected to be commercialized within the next five years (Kempken and Jung, 2010). In the longer run, GM techniques could also help improve nutrient efficiencies and yield potentials in crop plants. Hence, combined with conventional breeding and other innovations, GM crops could significantly raise agricultural productivity, which is important to ensure sufficient food availability for a growing world population.

2.2. GM crops and economic access to food Global and local food availability is a necessary but not a sufficient condition for food security, as the above discussion about unequal food distribution showed. Many people are too poor to have adequate economic access to food, so raising their income needs to be a central component of any food security strategy. Figure 3 shows that around 80% of the hungry people in developing countries live in rural areas, where they directly or indirectly depend on agriculture as farmers or wage laborers. There are different ways of increasing agricultural incomes and reducing rural poverty, including education, infrastructure investments, and institutional change. Yet, agricultural technology has an important role to play as well. Comprehensive analyses show that promoting the development and spread of appropriate new technologies is not only an effective but also a highly efficient way of reducing poverty, especially in Africa and Asia (Thirtle et al., 2003; Fan et al., 2005). Through the income pathway, technological progress improves economic access to food among rural households, even when the new technologies themselves may sometimes relate to non-food cash crops. In principle, GM technologies can be suitable to raise incomes in the small farm sector. Inbuilt in the seed, they are scale-neutral and relatively easy to use. Moreover, smallholders are often particularly affected by crop losses due to biotic and abiotic stress factors, because of unfavorable agroecological, financial, and technical conditions. Yield-increasing technologies can also be employment generating: in traditional production

35

Genetically Modified Crops and Global Food Security Urban poor 20%

Pastoralists, fishers, forest dependent 10%

Smallholder farmers 50%

Rural landless 20%

Fig. 3.

Who are the hungry people? Source: World Bank (2007).

systems, where most of the farming operations are performed manually, higher yields imply that more labor is hired for harvesting and related activities. Hence, the rural landless could benefit as well. Yet, whether these potentials will actually materialize also depends on a number of institutional and policy factors. For instance, even if technically possible, how can it be ensured that GM crops targeted to small farmer conditions will actually become available and accessible in developing countries? These are questions that will be addressed further below through reviewing the empirical evidence. Since a few concrete GM crop applications can already be observed in the small farm sector of developing countries, an analysis of the social effects will be instructive. 2.3. GM crops and nutritional value The third dimension of food security – as outlined in the definition above – refers to nutritional value. That is, dietary needs are broader than just food energy. While hunger and undernourishment are the results of insufficient calorie intake, the human body also needs a number of micronutrients that are contained in bigger amounts in high-value foods such as fruits, vegetables, and animal products. Since such higher-value foods are often more expensive than calorie-dense staple foods, the poor do not consume them in sufficient amounts, so that micronutrient deficiencies are widespread. Around 3 billion people are at risk of zinc deficiency, 2 billion people are anemic, many due to iron deficiency, 2 billion are iodine deficient, and 200 million are deficient in vitamin A (Stein and Qaim, 2007; WHO, 2009). Micronutrient deficiencies are responsible for severe health

36

Matin Qaim

problems, including impaired physical and cognitive development, susceptibility to infectious diseases, and higher child mortality. Reducing micronutrient malnutrition has recently been recognized as a key opportunity to promote economic development in a cost-effective way (Lomborg, 2009). Traditional interventions to address micronutrient deficiencies include food supplementation, industrial fortification, and dietary diversification programs. While all these programs have been successful in some situations, a common problem is that they are relatively expensive to implement and often do not reach poor households in remote rural areas. A new complementary strategy is biofortification, that is, the breeding of staple food crops for higher micronutrient contents (Qaim et al., 2007). While this partly builds on conventional breeding, GM approaches are particularly promising when certain micronutrients are completely absent from a crop plant or not available in sufficient amounts. A case in point is rice, where the endosperm of conventional grain does not contain any beta-carotene, which is a precursor of vitamin A. Hence, GM techniques were used to develop Golden Rice, which contains significant levels of beta-carotene (see below for further details about Golden Rice and its potential impact). While biofortified crops should not be seen as a substitute for dietary diversification and other micronutrient interventions, they could nonetheless contribute to reducing nutritional deficiencies and related health problems. This is especially true among the poor, for whom many of the other alternatives are often out of reach in the short to medium run. 3. Socioeconomic impacts of commercialized GM crops While the previous section looked at different potential pathways of how GM crops could contribute to global food security, this section focuses on the actual effects already observable in different countries. The first GM crops were commercialized in the mid-1990s in the USA and a few other countries. Since then, adoption rates have been rising rapidly. In 2009, GM crops were grown on 134 million hectares by 14 million farmers in 25 countries, including 16 developing countries (James, 2009). Yet, the portfolio of different GM crops and modified traits is still limited. Most of the commercial applications involve herbicide tolerance and insect resistance in crops such as soybean, maize, cotton, and canola. 3.1. Impacts of herbicide-tolerant crops Herbicide-tolerant (HT) crops are tolerant to certain broad-spectrum herbicides like glyphosate or glufosinate, which are more effective, less toxic, and usually cheaper than selective herbicides. HT technology is so far mostly used in soybean, maize, cotton, and canola. The dominant crop

Genetically Modified Crops and Global Food Security

37

is HT soybean, which was grown on 69 million ha in 2009, mostly in the USA, Argentina, and Brazil, but also in a number of other countries. Likewise, HT maize is cultivated primarily in North and South America, with smaller areas in South Africa and the Philippines. HT cotton is mostly cultivated in the USA, whereas HT canola is predominantly grown in Canada (James, 2009). HT adopting farmers benefit in terms of lower herbicide expenditures. Total herbicide quantities applied were reduced in some situations, but not in others. In Argentina, herbicide quantities were even increased significantly (Qaim and Traxler, 2005). This is largely because herbicide sprays were substituted for tillage. In Argentina, the share of soybean farmers using no-till almost doubled to 80% since the introduction of HT technology. Also in the USA and Canada, no-till practices expanded through HT adoption (Fernandez-Cornejo and Caswell, 2006). In terms of yields, there is no significant difference between HT and conventional crops in most cases, implying that crop losses due to weeds were effectively controlled even before the introduction of HT technology. This, however, is location-specific: where certain weeds are difficult to control with selective herbicides, the adoption of HT and the switch to broad spectrum herbicides resulted in better weed control and higher crop yields. Examples are HT soybeans in Romania and Mexico, and HT maize in Argentina (Brookes and Barfoot, 2008). Available studies show that HT technology reduces the cost of production through lower expenditures for herbicides, labor, machinery, and fuel. Yet, the innovating companies charge a technology fee on seeds, which varies between crops and countries. Several studies for HT soybean and canola in the USA and Canada showed that the fee was in a similar magnitude or sometimes higher than the average cost reduction, so that farmer profit effects were small or sometimes negative (Naseem and Pray, 2004).2 This is different in South America. While the agronomic advantages are similar, the fee charged on seeds is lower, as HT technology is not patented there. Many soybean farmers in South America even use farm-saved GM seeds. Qaim and Traxler (2005) showed for Argentina that the average profit gain through HT soybean adoption is in a magnitude of US$23 per hectare. The technology is so attractive for farmers that HT is now being used on almost 100% of the Argentine soybean area. In Brazil, adoption rates are also over 70% with a further rising trend (James, 2009). While farmers in developing countries benefit significantly from HT soybeans, most soybeans are grown on relatively large and fully mechanized farms. So far, HT crops have not been widely adopted in

2 It remains to be seen how seed prices and technology fees develop when relevant patents expire in the USA and Canada within the next few years.

38

Matin Qaim

the small farm sector. Smallholders often weed manually, so that HT crops are inappropriate, unless labor shortages or weeds that are difficult to control justify conversion to chemical practices. A case in point could be striga, a weed that can hardly be controlled manually and leads to significant yield losses in subsistence production systems of maize, sorghum, millet, and a few other crops in Sub-Saharan Africa. Overall, HT crops are attractive for many farmers from an economic perspective. There are also environmental benefits through reduced tillage operations, entailing a decrease in soil erosion, fuel use, and greenhouse gas emissions (Qaim and Traxler, 2005; Brookes and Barfoot, 2008). Nevertheless, the potentials of HT crops to contribute to global food security seem relatively limited, because yield-increasing and povertyreducing effects can only be expected in quite specific situations. 3.2. Impacts of insect-resistant crops Insect-resistant GM crops commercially grown so far involve different genes from the soil bacterium Bacillus thuringiensis (Bt) that make the plants resistant to certain lepidopteran and coleopteran pest species. The most widely used examples are Bt maize and Bt cotton. In 2009, Bt maize was grown on 35 million ha in more than 15 countries. The biggest Bt maize areas are found in the USA, Argentina, South Africa, Canada, and the Philippines. Bt cotton was grown on almost 15 million ha in 2009, mostly in India, China, and the USA, but also in a number of other countries (James, 2009). 3.2.1. Agronomic and economic effects If insect pests are effectively controlled through chemical pesticides, the main effect of switching to Bt crops will be a reduction in insecticide applications, as the genetic resistance mechanism substitutes for chemical control agents. However, as shown above (see Figure 2), there are also situations where insect pests are not effectively controlled, due to the unavailability of suitable insecticides or other technical, financial, or institutional constraints. In those situations, Bt technology adoption can help reduce crop damage and thus increase effective yields. Table 1 confirms that both insecticide-reducing and yield-increasing effects of Bt crops can be observed internationally. In conventional cotton, high amounts of chemical insecticides are normally used to control the bollworm complex, which is the main Bt target pest. Accordingly, Bt cotton adoption allows significant insecticide reductions, ranging from 30% to 80% on average. This brings about substantial environmental advantages and health benefits for farmers, farm workers, and consumers. Yield effects are also quite pronounced, especially in developing countries. In Argentina, for instance, conventional cotton farmers under-use chemical insecticides, so that insect pests are not

39

Genetically Modified Crops and Global Food Security

Table 1. Country

Average farm level effects of Bt crops

Insecticide reduction (%)

Increase in effective yield (%)

Increase in profit (US$/ha)

Bt cotton Argentina Australia China India Mexico South Africa USA

47 48 65 41 77 33 36

33 0 24 37 9 22 10

23 66 470 135 295 91 58

Bt maize Argentina Philippines South Africa Spain USA

0 5 10 63 8

9 34 11 6 5

20 53 42 70 12

Source: Qaim (2009). Notes: The results are based on data from farm surveys carried out by various research teams in the different countries between 1996 and 2009. All results are based on data from two or more growing seasons. The profit effects shown are net profits gains after payment of any technology fee or seed price premium.

effectively controlled (Qaim and de Janvry, 2005). In India and China, chemical input use is much higher, but the insecticides are not always very effective, due to low quality, resistance in pest populations, and sometimes incorrect timing of sprays (Huang et al., 2003; Qaim et al., 2006). As cotton is not a food crop, yield increases do not directly contribute to improvements in food availability. The example is interesting nonetheless, because Bt cotton is the only GM crop that is already widely used in smallholder production systems in different developing countries. Similar effects can also be expected for Bt food crops, when there is high infestation of Bt target pests. The evidence available for Bt maize confirms this prediction (Table 1). Except for Spain, where the percentage reduction in insecticide use is large, the more important result of Bt maize is an increase in effective yields. In the USA, Bt maize is mainly used against the European corn borer, which is often not controlled by chemical means (more recently commercialized Bt hybrids in the USA also provide resistance to the corn rootworm complex). In Argentina and South Africa, mean yield effects are higher, because there is more severe pest pressure. The average yield gain of 11% in South Africa shown in Table 1 refers to large commercial farms. These farms have been growing yellow Bt maize hybrids for several years. Gouse

40

Matin Qaim

et al. (2006) also analyzed data from smallholder farmers growing white Bt maize hybrids in South Africa; they found average yield gains of 32% on Bt plots. In the Philippines, average yield advantages of Bt maize are even 34%. These patterns suggest that resource-poor smallholder farmers face bigger constraints in controlling insect damage in their conventional crops. The profit effects of Bt technologies are also shown in Table 1. Bt seeds are more expensive than conventional seeds, because they are mostly sold by private companies that charge a technology fee. The fee is positively correlated with the strengths of intellectual property right (IPR) protection in a country. In all countries, Bt adopting farmers benefit financially, that is, the economic advantages associated with insecticide savings and higher effective yields more than outweigh the technology fee charged on GM seeds. Yet, the absolute gains differ remarkably between countries and crops. On average, the extra profits are higher in developing than in developed countries. Apart from agroecological and socioeconomic differences, GM seed costs are often lower in developing countries, due to weaker IPRs, seed reproduction by farmers, subsidies, or other types of government price interventions (Basu and Qaim, 2007). Moreover, profit gains are higher for Bt cotton than for Bt maize, which is partly due to lower technology fees charged by the companies in smallholder cotton environments. In addition, the pests targeted by Bt cotton are of higher economic importance than those targeted by Bt maize, although newer Bt maize events now cover a broader spectrum of target pests, which may potentially change the picture in the future. The mean values shown in Table 1 mask impact variability observed within countries. Especially during the early years of Bt cotton adoption in China and India, there were farmers that did not benefit economically, mainly because of insufficient knowledge about appropriate pesticide adjustments or the use of varieties not suitable for certain agroecological conditions (Qaim et al., 2006; Pemsl and Waibel, 2007; Grue`re et al., 2008). These initial problems were overcome, so that most cotton farmers are now highly satisfied with Bt technology. This is reflected in the rapid adoption rates. In India, around 90% of the cotton farmers have adopted Bt technology, whereas in some provinces of China all cotton is now GM. Most cotton farmers in India and China are small-scale producers, who often live near or below the poverty line. For them, financial gains of several hundred dollars per hectare through Bt adoption can improve living standards substantially, entailing better economic access to food and other basic needs. Poverty and distribution effects are analyzed more explicitly below. 3.2.2. Poverty and distribution effects Especially in China, India, and South Africa, Bt cotton is often grown in farms with less than three hectares of land. In South Africa, many smallholders grow Bt white maize as their staple food. Several studies

Genetically Modified Crops and Global Food Security

41

show that Bt technology advantages for small-scale farmers are of a similar magnitude as those of large-scale producers. In some cases, the advantages can even be bigger (Pray et al., 2001; Morse et al., 2004). However, there are only very few recent studies that have gone beyond farm profits, to analyze wider socioeconomic outcomes of GM crops, such as impacts on household income, income distribution, poverty, and rural employment. Ali and Abdulai (2010) have analyzed the effects of Bt cotton in Pakistan. Using a propensity-score matching approach, they showed that the adoption of this new technology exerts a positive and significant effect on household income and poverty reduction among cotton growers. Subramanian and Qaim (2009, 2010) developed a village social accounting matrix (SAM) and multiplier model to examine direct and indirect effects of Bt cotton adoption in India. Their results show that total household income effects of Bt cotton are US$246 per hectare higher than those of conventional cotton (Figure 4). Of these total benefits, US$135 are direct profits for cotton farmers, and US$111 are spillovers through backward and forward linkages to other local markets and sectors. That is, each dollar of direct Bt cotton benefits is associated

600 Bt Conventional

500

US$ per ha

400 300 200 100 0 All households

Extremely poor Moderately poor

Non-poor

Fig. 4. Household income effects of Bt and conventional cotton in India. Source: Qaim et al. (2009). Notes: The results show combine direct and indirect effects calculated with an SAM multiplier model. Data for this analysis come from a village census survey carried out in 2004 in Kanzara, Maharashtra, and three waves of a panel survey carried out in 2003, 2005, and 2007 in the states of Maharashtra, Karnataka, Andhra Pradesh, and Tamil Nadu. For further details of the model and data also see Subramanian and Qaim (2010). The columns for ‘‘all households’’ are the sums of the columns for the three income categories.

42

Matin Qaim

with over 80 cents of additional indirect benefits in the village economy (Qaim et al., 2009). In terms of income distribution, all types of households benefit, including those below the poverty line (Figure 4). Sixty percent of the gains accrue to the extremely and moderately poor. Hence, Bt cotton in India is poverty reducing. The technology is also net employment generating, so that landless rural households benefit as well. The employment effects have interesting gender implications: Bt cotton increases aggregate returns to labor by 42%, while the returns for hired female agricultural workers increase by 55%. This is largely due to additional labor employed for picking cotton, which is primarily a female activity in India (Subramanian et al., 2010). As is known, women’s income has a particularly positive effect for child nutrition and welfare (Quisumbing et al., 1995). These findings on social benefits in India are in stark contrast to some reports by biotech critics, who claim that Bt cotton would ruin smallholder farmers and drive them into suicide (e.g., Sharma, 2004; Shiva, 2009). However, such reports are not substantiated by reliable data. Grue`re et al. (2008) have analyzed the issue of farmer suicides in India and found no correlation with Bt cotton adoption; suicides among Indian farmers were already reported long before Bt cotton was commercialized, and the number of cases has not increased since Bt technology was released. The results summarized here on positive income and poverty reduction effects of Bt cotton in the small farm sector of Pakistan and India cannot be simply extrapolated to other countries and other GM crops, because impacts always depend on the conditions in a particular setting. Nonetheless, the fact that a first-generation GM crop like Bt cotton already contributes to poverty reduction and improved food security has not been widely recognized up till now.

4. Potential impacts of future GM crops 4.1. Crops with improved agronomic traits While Bt technology so far has mainly been used in maize and cotton, there are also other Bt crops that are likely to be commercialized soon (Romeis et al., 2008). For instance, China has recently announced the commercialization of Bt rice, while in India Bt eggplant is ready to go. Both technologies have been tested extensively in experimental stations and on farms. The available data are in line with results for Bt cotton and Bt maize: insecticidereducing and yield-increasing effects can lead to significant economic and social benefits (Huang et al., 2005; Romeis et al., 2008; Qaim, 2009). In an ex ante study for Bt eggplant in India, Krishna and Qaim (2008a) projected that the technology, which controls the eggplant fruit and shoot borer, will reduce chemical insecticide use by up to 50% and increase

43

Genetically Modified Crops and Global Food Security

yields by 40% on average. This will not only improve farmers’ profits but also lower market prices and thus improve consumer access to vegetables, with expected positive nutrition effects among the poor. Moreover, Bt eggplant will be less contaminated with pesticide residues; such residues in vegetables have become a real problem in some parts of India (Krishna and Qaim, 2008b). Despite the expected positive economic, environmental, and health effects, Bt eggplant – as the first GM food crop to be commercialized in India – has recently aroused controversial public debates. After a careful review of the biosafety and food safety data, the Genetic Engineering Approval Committee, which is the responsible authority in India, declared Bt eggplant to be safe and approved this technology in October 2009 (Kumar, 2009). However, after a series of public hearings, which were heavily influenced by anti-biotech campaigns and biased media reports, the Minister of Environment and Forests suspended the commercialization of Bt eggplant for an indefinite period of time. This example demonstrates how much the regulatory procedures, which should be science based, are influenced by subjective views of certain lobbying groups. Also for other pest-resistant GM traits that are being developed in different crops – such as fungal, virus, nematode, or bacterial resistance – pesticide-reducing and yield-increasing effects can be expected. As argued above and as already observed for Bt technologies, positive yield effects will generally be more pronounced in developing countries, where pest pressure is often higher and farmers face more severe constraints in controlling pest damage (Table 2). Especially in the non-commercial and Table 2. Region

Expected yield effects of pest-resistant GM crops in different regions Pest pressure

Developed Low to medium countries Latin America Medium (commercial) China Medium Latin America Medium (noncommercial) South and High Southeast Asia Sub-Saharan High Africa Source: Qaim and Zilberman (2003).

Availability of chemical alternatives

Adoption of chemical alternatives

Expected yield effect of GM crops

High

High

Low

Medium

High

Low to medium

Medium Low to medium

High Low

Low to medium Medium to high

Low to medium

Low to medium

High

Low

Low

High

44

Matin Qaim

semi-commercial crop sectors, where technical and economic constraints impede a more widespread use of chemicals, pest-related crop losses are often 50% and higher (Oerke, 2006). On the basis of the conditions of pest pressure and current crop protection, the biggest yield gains are expected in South and Southeast Asia and Sub-Saharan Africa. The effects of GM crops with tolerance to abiotic stresses will also be situation specific. A drought-tolerant transgenic variety can lead to substantially higher yields than conventional varieties under water stress, whereas the effect may be small when sufficient water is available. Especially in the semi-arid tropics, many small-scale farmers are operating under drought-prone conditions, so that the benefits of drought tolerance could be sizeable. In a study referring to eight low-income countries in Asia and Sub-Saharan Africa, Kostandini et al. (2009) reckon that the average yield gains of GM drought tolerance traits may be 18% in maize, 25% in wheat, and 10% in rice. This is expected to lead to annual welfare gains of US$850 million in the eight countries under study. Additional benefits of higher yield stability (variance reduction) are calculated to be US$570 million. While the development of drought-tolerant varieties is a major priority both in public and private sector crop improvement programs (Kempken and Jung, 2010), biotech researchers are also working on tolerance to other abiotic stress factors such as heat, salinity, flood, and coldness. Climate change is associated with more frequent weather extremes, so that more tolerant GM crops can help reduce the risks of crop failures and food crises. Furthermore, research is underway to develop crops with higher nutrient efficiency, especially with respect to nitrogen. Nutrient-efficient crops will reduce chemical fertilizer use and associated environmental externalities in intensive agricultural production systems, while they will contribute to yield gains in regions where fertilizers are currently underused, as is the case in large parts of Sub-Saharan Africa. Some of these traits are genetically complex, so that commercialization may not be expected in the short run. But in the medium and long run, the contribution to food security could be sizeable.

4.2. Crops with improved nutritional traits Nutritionally enhanced GM crops that researchers are working on include oilseeds with improved fatty acid profiles, crops with higher amounts of certain essential amino acids, and biofortified staples with enhanced contents of minerals and vitamins (Moschini, 2008). A well-known example of a GM biofortified crop is Golden Rice, which contains significant amounts of beta-carotene to control vitamin A deficiency (VAD). Golden Rice could become commercially available in some Asian countries starting in 2012 (Potrykus, 2008). As this technology is particularly promising from a

Genetically Modified Crops and Global Food Security

45

food security perspective, some further details about likely nutrition and health benefits are discussed in the following. VAD is a considerable public health problem in many developing countries: it affects 190 million pre-school children and 19 million pregnant women world-wide (WHO, 2009). Apart from increasing child mortality, VAD can lead to visual problems, including blindness, and it also increases the incidence of infectious diseases (UN SCN, 2004). The deficiency is most widespread in poverty households, where diets are dominated by staple foods with relatively low nutritional value. Widespread consumption of Golden Rice promises to improve the situation in rice-eating populations. Stein et al. (2008) developed a methodology for comprehensive ex ante evaluation, which they used for empirical analysis in India. India is one of the target countries for Golden Rice, because mean levels of rice consumption are relatively high, and VAD is widespread. Using a disability-adjusted life years (DALYs) approach, Stein et al. (2008) calculated the social burden associated with VAD in India.3 The combined annual mortality and morbidity burden is expressed in terms of the number of DALYs lost. The present burden of VAD, calculated based on available health statistics, is the situation without Golden Rice. In a next step, present beta-carotene intakes from nationally representative food consumption data were derived, and the likely shift in the intake distribution through future consumption of Golden Rice was established. Necessary assumptions were based on experimental data and expert estimates about the technology’s efficacy and future coverage. Higher beta-carotene intakes will improve the vitamin A status of individuals, thus reducing the incidence of adverse health outcomes. These reduced incidence rates were projected and used to re-calculate the expected remaining burden with Golden Rice. The difference in the VAD burden with and without Golden Rice is the expected impact of the technology expressed in terms of the number of DALYs saved. According to these calculations, the current annual burden of VAD in India amounts to a loss of 2.3 million DALYs, of which 2.0 million are lost due to child mortality alone. In terms of incidence numbers, more than 70,000 Indian children under the age of six die each year due to VAD. In this context, widespread consumption of Golden Rice could reduce the burden by 59%, which includes the saving of almost 40,000 lives every year (Table 3). Because the severity of VAD is negatively correlated with 3

The DALYs approach was initially developed by Murray and Lopez (1996) to quantify the burden of different diseases by combining problems of mortality and morbidity in a single index. The method was further developed by different authors to make it useful for a wide array of health and nutrition problems, including micronutrient deficiencies. It can also be used for impact evaluations and cost-effectiveness analyses of biofortified crops and other micronutrient interventions (Stein et al., 2005).

46

Matin Qaim

Table 3.

Burden of vitamin A deficiency in India and potential impact of Golden Rice

Current burden of vitamin A deficiency Number of DALYs lost each year (thousands) Number of lives lost each year (thousands)

2,328 71.6

Potential impact of Golden Rice Number of DALYs saved each year (thousands) Reduction of the DALYs burden (%) Number of lives saved each year (thousands)

1,382 59.4 39.7

Cost-effectiveness of Golden Rice and other vitamin A interventions Cost per DALY saved through Golden Rice (US$) World Bank cost-effectiveness standard for DALYs saved (US$) Cost per DALY saved through supplementation (US$) Cost per DALY saved through industrial fortification (US$)

3.1 200 134 84

Source: Stein et al. (2008). Notes: The impact estimates build on the ‘‘high impact scenario’’ in Stein et al. (2008). Given recent evidence about the high efficacy of Golden Rice (Tang et al., 2009), the assumptions in that scenario appear realistic when the technology receives public support for social marketing efforts.

income, the positive effects are most pronounced in the poorest income groups (Stein et al., 2008). While these results suggest that Golden Rice alone is unlikely to eliminate the problems of VAD, the projected improvements in public health and nutrition are huge. However, unlike available GM crops that were mostly commercialized by private companies and sold at a premium charged on seeds, Golden Rice is a humanitarian project where seeds will be distributed without a technology fee (Potrykus, 2008). Therefore, an analysis of its potential cost-effectiveness is also important. The major costs of Golden Rice are the investments in research as well as in developing, testing, and disseminating the GM technology. Dividing these costs by the number of DALYs saved, and taking into account the time when costs and benefits occur through discounting, results in the average cost per DALY saved, which is a common measure for the costeffectiveness of health interventions. According to the projections by Stein et al. (2008), the cost per DALY saved through Golden Rice is in a magnitude of US$3 (Table 3). A sensitivity analysis shows that even with much more pessimistic assumptions the cost would not rise to more than US$20 per DALY saved. These results should be compared with suitable benchmarks. The World Bank classifies health interventions as very cost-effective when their cost per DALY saved is less than US$200. This underlines that Golden Rice could be extremely cost-effective. But how does Golden Rice compare with

Genetically Modified Crops and Global Food Security

47

conventional vitamin A interventions? Scaling up food supplementation or industrial fortification programs for vitamin A in India would cost between US$84 and US$134 per DALY saved (Table 3). The major cost of these conventional interventions is not to produce the vitamin pills or food fortificants but to reach the target population in remote rural areas, which requires large investments and monitoring on a regular basis. This is different for Golden Rice: even though the initial investment for research and development is high, recurrent costs will be low, because Golden Rice seeds will spread through existing formal and informal distribution channels and can be reproduced by farmers themselves. Nonetheless, social marketing efforts will be required to explain the yellow color of the rice that is associated with beta-carotene. Furthermore, suitable strategies to convince farmers to adopt Golden Rice varieties have to be developed. A combination of beta-carotene with interesting agronomic traits in rice might be a practicable avenue. Similar effects can also be expected for other biofortified crops, containing higher amounts of iron, zinc, vitamin A, and other micronutrients (Qaim et al., 2007; Meenakshi et al., 2010). However, while this bodes well for reducing nutritional deficiencies in developing countries, biofortified crops should not be seen as a substitute for existing micronutrient interventions but as a complementary strategy. No single approach will eliminate micronutrient deficiency problems, and all interventions have their strengths and weaknesses in particular situations. While supplementation and industrial fortification might be more suitable for urban areas and feeding programs for well defined target groups, biofortified crops are likely to achieve a wider coverage, especially in rural areas. It is only in the long run that poverty reduction and economic growth may be expected to contribute to dietary diversification, which might then reduce the urgency for more specific micronutrient interventions.

5. Institutional and policy issues Most GM crops available so far were developed and commercialized by private firms. Monsanto is involved in many cases, mostly in cooperation with local seed companies. The empirical evidence reported here provides a consistent picture: developing-country farmers and consumers can benefit substantially from proprietary GM crops. The fear by many critics that GM technologies only add to the profits of multinational companies is therefore exaggerated. IPR protection influences the distribution of benefits: through strong IPRs on GM technologies, as observed in some developed countries, companies can capture a significant fraction of the overall benefits. But in most developing countries IPRs are weak, so that GM seed prices are lower and farmers capture a larger benefit share (Qaim, 2009). Strengthening IPRs could invigorate the local private seed

48

Matin Qaim

industry and accelerate innovation rates in some emerging economies, but in the least-developed countries potential advantages would probably be outweighed by disadvantages in terms of lower technology accessibility through higher seed prices. Therefore, the appropriate strength of IPR protection is country-specific. Beyond the issue of seed prices and benefit distribution, the dominance of private multinationals also has implications for the type of GM crops that emerge. The private sector develops technologies primarily for big lucrative markets. Hitherto applications concentrate on commercial crops and relatively large and economically more advanced countries. While technically feasible, it is unlikely that multinationals will commercialize GM innovations for niche markets in the least-developed world, where market failures are commonplace. Such research gaps will have to be addressed by the public sector, if biotechnology developments are not to bypass the poor. This requires an expansion of public research investments. Some of the bigger countries – like China, India, or Brazil – have public biotechnology programs and the critical mass to come up with own technologies that can complement proprietary innovations. Smaller developing countries will need more targeted external support, for instance through closer cooperation with international agricultural research centers. Also, more public–private partnerships should be sought to harness the comparative strengths of both sectors. There are numerous examples of public–private research cooperation in agricultural biotechnology, but none of these projects has yet led to a commercialized GM crop. Ex ante studies show that well-designed partnerships can be advantageous for all parties involved (Krishna and Qaim, 2007, 2008a). Still, more research is needed, in order to identify best practices for the joint development and commercialization of GM crops and issues related to IPR transfer. Against this background, the constantly rising regulatory hurdles and costs are a major stumbling block (Moschini, 2008). Kalaitzandonakes et al. (2007) have estimated the private compliance costs for regulatory approval of a new Bt or HT maize technology in one country at US$6 million to US$15 million, which is often more than the cost of actually developing the technology. Commercializing the same technology in other countries will entail additional costs. Such high regulatory costs slow down innovation rates. They also impede the commercialization of GM technologies in minor crops and small countries, as markets in such situations are not large enough to justify the fixed cost investments. And, expensive regulations are difficult to handle by small firms and public sector organizations, so that they contribute to further concentration of the agricultural biotech industry. If such lengthy and complex procedures were really necessary to regulate high-risk products, then the costs involved would be justified. But this does not seem to be the case. Since the use of genetic engineering does not entail unique risks, it is actually illogical to subject GM crops to a much higher degree of scrutiny than

Genetically Modified Crops and Global Food Security

49

conventionally bred crops (Bradford et al., 2005). The regulatory complexity observed today is rather the outcome of the politicized public debate and the lobbying success of anti-biotech interest groups (Miller and Conko, 2004). Especially with a view to the large potentials of GM crops for developing countries, some reform of the regulatory framework will be necessary, and economists have an important role in this respect in terms of quantifying costs and benefits. But also when suitable GM crop technologies are commercialized in developing countries, benefits for poor farmers and consumers will not occur automatically. A conducive institutional environment is important to promote wide and equitable access to new seed technologies. In general, well-functioning input and output markets, including efficient micro-credit schemes, will spur the process of innovation adoption. Unfortunately, such conditions first need to be established in the poorest countries of Africa and Asia, so that the GM crop impacts observed so far in China, India, and other more advanced developing countries cannot simply be extrapolated. Like any agricultural technology, GM crops are not a substitute but a complement to much needed institutional change in rural areas of developing countries.

6. Conclusion Global food security requires (1) sufficient food availability, (2) economic access to food by all, and (3) an adequate nutritional value of the diets that people consume. While GM crops are not a panacea, they can contribute to improving food security in terms of all three dimensions. Crops that are resistant to biotic or tolerant to abiotic stress factors can substantially increase effective yields and thus enhance global and local food availability. Moreover, since most of the world’s hungry people depend on agriculture as a source of income and employment, GM crops that are suitable for the small farm sector can raise the incomes of the poor and thus improve their economic access to food. And finally, biofortified crops can add nutritional value to staple food crops and thus reduce specific nutritional deficiencies in a highly cost-effective way. So far, mostly HT and Bt crops have been employed. Available impact studies show that these crops are beneficial, but they also suggest that differentiation is important. While the potentials of HT crops to contribute to food security seem to be confined to very specific situations, the positive impacts of Bt crops can be much larger. Bt cotton in particular does not only contribute to higher yields and lower insecticide use but also contribute to significant household income gains, including for farmers and rural laborers living below the poverty line. Similar effects are observed for Bt maize, and preliminary studies suggests that other pest-resistant GM food crops may also result in comparable impacts. Strikingly, farmers in developing

50

Matin Qaim

countries often benefit more than their colleagues in developed countries, which is partly due to weaker IPR protection and thus lower seed prices. But income distribution also depends on the wider institutional setting, including farmers’ access to suitable seed varieties, credit, information, and other input and output markets. Like any agricultural technology, GM crops are not a substitute but a complement to much needed institutional and infrastructure improvement. GM technologies in the research pipeline include crops that are more tolerant to temperature and water stress, more efficient in terms of soil nutrient use, or crops that contain higher amounts of vitamins and trace minerals. The benefits of such applications could be much bigger than those already observed. Against the background of a dwindling natural resource base, rapidly growing demand for food and biofuels, and widespread rural poverty, GM crops could contribute significantly to sustainable development. Despite these potentials, the public debate about GM crops remains controversial. Concerns about new risks and lobbying efforts of antibiotech groups have led to complex, costly, and unpredictable biosafety, food safety, and labeling regulations, which slow down innovation rates and lead to a bias against small countries, minor crops, small firms, and public research organizations. Overregulation has become a real threat for the further development and use of GM crops. The costs of regulation in terms of foregone benefits might be large, especially for developing countries. This is not to say that zero regulation would be desirable, but the trade-offs associated with regulation need to be considered. In the general public, the risks of GM crops seem to be overrated, while the benefits are underrated. Wider recognition of the technology’s potentials could help redirect public policy efforts towards ensuring that pro-poor outcomes can be achieved on a larger scale.

References Ali, A., Abdulai, A. (2010), The adoption of genetically modified cotton and poverty reduction in Pakistan. Journal of Agricultural Economics 61, 175–192. Basu, A.K., Qaim, M. (2007), On the adoption of genetically modified seeds in developing countries and the optimal types of government intervention. American Journal of Agricultural Economics 89, 784–804. Borlaug, N. (2007), Feeding a hungry world. Science 318, 359. Bradford, K.J., Van Deynze, A., Gutterson, N., Parrott, W., Strauss, S.H. (2005), Regulating transgenic crops sensibly: Lessons from plant breeding, biotechnology and genomics. Nature Biotechnology 23, 439–444. Brookes, G., Barfoot, P. (2008), GM Crops: Global Socioeconomic and Environmental Impacts 1996–2008. PG Economics, Dorchester.

Genetically Modified Crops and Global Food Security

51

Ecker, O., Qaim, M. (2011), Analyzing nutritional impacts of policies: An empirical study for Malawi. World Development 39, 412–428. Evenson, R.E., Gollin, D. (2003), Assessing the impact of the green revolution, 1960–2000. Science 300, 758–762. Dewbre, J., Giner, C., Thompson, W., von Lampe, M. (2008), High food commodity prices: Will they stay? Who will pay? Agricultural Economics 39 (Suppl.), 393–403. Fan, S., Chan-Kang, C., Qian, K., Krishnaiah, K. (2005), National and international agricultural research and rural poverty: The case of rice research in India and China. Agricultural Economics 33, 369–379. FAO (2009), The State of Food Insecurity in the World: Economic Crises – Impacts and Lessons Learned. Food and Agriculture Organization of the United Nations, Rome. FAO (2010), FAOSTAT Production Database. Food and Agriculture Organization of the United Nations, Rome. Fernandez-Cornejo, J., Caswell, M. (2006), The First Decade of Genetically Engineered Crops in the United States. Economic Information Bulletin 11, United States Department of Agriculture, Washington, DC. FOE. (2008), Who Benefits from GM Crops? The Rise in Pesticide Use. Agriculture and Food Issue 112, Friends of the Earth International, Amsterdam. Godfray, H.C.J., Beddington, J.R., Crite, I.R., Haddad, L., Lawrence, D., Muir, J.F., Pretty, J., Robinson, S., Thomas, S.M., Toulmin, C. (2010), Food security: The challenge of feeding 9 billion people. Science 327, 812–818. Gouse, M., Pray, C., Schimmelpfennig, D., Kirsten, J. (2006), Three seasons of subsistence insect-resistant maize in South Africa: have smallholders benefited? AgBioForum 9, 15–22. Grue`re, G.P., Mehta-Bhatt, P., Sengupta, D. (2008), Bt cotton and farmers suicides in India: reviewing the evidence. IFPRI Discussion Paper 00808, International Food Policy Research Institute, Washington, DC. Gurian-Sherman, D. (2009), Failure to Yield: Evaluating the Performance of Genetically Engineered Crops. Union of Concerned Scientists, Cambridge, MA. Holt-Gimenez, E., Altieri, M.A., Rosset, P. (2006), Ten reasons why the Rockefeller and the Bill and Melinda Gates Foundation’s Alliance for another green revolution will not solve the problems of poverty and hunger in Sub-Saharan Africa. Food First Policy Brief 12, Institute for Food and Development Policy, Oakland, CA. Huang, J., Hu, R., Pray, C., Qiao, F., Rozelle, S. (2003), Biotechnology as an alternative to chemical pesticides: A case study of Bt cotton in China. Agricultural Economics 29, 55–68. Huang, J., Hu, R., Rozelle, S., Pray, C. (2005), Insect-resistant GM rice in farmers’ fields: Assessing productivity and health effects in China. Science 308, 688–690.

52

Matin Qaim

Huang, J., Pray, C., Rozelle, S. (2002), Enhancing the crops to feed the poor. Nature 418, 678–684. IAASTD. (2009), Executive Summary of the Synthesis Report. International Assessment of Agricultural Knowledge, Science and Technology for Development, Washington, DC. Ivanic, M., Martin, W. (2008), Implications of higher global food prices for poverty in low-income countries. Agricultural Economics 39 (Suppl.), 405–416. James, C. (2009), Global Status of Commercialized Biotech/GM Crops: 2009. ISAAA Briefs 41, International Service for the Acquisition of Agri-biotech Applications, Ithaca, NY. Kalaitzandonakes, N., Alston, J.M., Bradford, K.J. (2007), Compliance costs for regulatory approval of new biotech crops. Nature Biotechnology 25, 509–511. Kempken, F., Jung, C. (Eds.) (2010), Genetic Modification of Plants. Biotechnology in Agriculture and Forestry 64. Springer, Berlin. Kostandini, G., Mills, B.F., Omamo, S.W., Wood, S. (2009), Ex ante analysis of the benefits of transgenic drought tolerance research on cereal crops in low-income countries. Agricultural Economics 40, 477–492. Krishna, V.V., Qaim, M. (2007), Estimating the adoption of Bt eggplant in India: Who benefits from public-private partnership? Food Policy 32, 523–543. Krishna, V.V., Qaim, M. (2008a), Potential impacts of Bt eggplant on economic surplus and farmers’ health in India. Agricultural Economics 38, 167–180. Krishna, V.V., Qaim, M. (2008b), Consumer attitudes toward GM food and pesticide residues in India. Review of Agricultural Economics 30, 233–251. Kumar, P.A. (2009), Bt brinjal: A pioneering push. Biotech News 4 (6), 108–111. Lomborg, B. (Ed.) (2009). Global Crises, Global Solutions (second ed. Cambridge University Press, Cambridge. Meenakshi, J.V., Johnson, N.L., Manyong, V.M., De Groote, H., Javelosa, J., Yanggen, D.R., Naher, F., Gonzalez, C., Garcı´ a, J., Meng, E. (2010), How cost-effective is biofortification in combating micronutrient malnutrition? An ex ante assessment. World Development 38, 64–75. Miller, H.I., Conko, G. (2004), The Frankenfood Myth – How Protest and Politics Threaten the Biotech Revolution. Praeger Publishers, Westport, CT. Morse, S., Bennett, R., Ismael, Y. (2004), Why Bt cotton pays for smallscale producers in South Africa. Nature Biotechnology 22, 379–380. Moschini, G. (2008), Biotechnology and the development of food markets: Retrospect and prospects. European Review of Agricultural Economics 35, 331–355.

Genetically Modified Crops and Global Food Security

53

Murray, C. J. L., Lopez, A. D. (Eds.) (1996). The Global Burden of Disease. Harvard University Press, Cambridge, MA. Naseem, A., Pray, C. (2004), Economic impact analysis of genetically modified crops. In: Christou, P., Klee, H. (Eds.), Handbook of Plant Biotechnology. Wiley, Chichester, pp. 959–991. Oerke, E.-C. (2006), Crop losses to pests. Journal of Agricultural Science 144, 31–43. Pemsl, D., Waibel, H. (2007), Assessing the profitability of different crop protection strategies in cotton: Case study results from Shandong Province, China. Agricultural Systems 95, 28–36. Potrykus, I. (2008), Golden Rice – from idea to reality. Bertebos Prize Lecture. Bertebos Conference, 7–9 September, Falkenberg, Sweden. Pray, C.E., Ma, D., Huang, J., Qiao, F. (2001), Impact of Bt cotton in China. World Development 29, 813–825. Qaim, M. (2009), The economics of genetically modified crops. Annual Review of Resource Economics 1, 665–693. Qaim, M., de Janvry, A. (2005), Bt cotton and pesticide use in Argentina: Economic and environmental effects. Environment and Development Economics 10, 179–200. Qaim, M., Stein, A.J., Meenakshi, J.V. (2007), Economics of biofortification. Agricultural Economics 37 (suppl.), 119–133. Qaim, M., Subramanian, A., Naik, G., Zilberman, D. (2006), Adoption of Bt cotton and impact variability: Insights from India. Review of Agricultural Economics 28, 48–58. Qaim, M., Subramanian, A., Sadashivappa, P. (2009), Commercialized GM crops and yield. Nature Biotechnology 27, 803–804. Qaim, M., Traxler, G. (2005), Roundup ready soybeans in Argentina: Farm level and aggregate welfare effects. Agricultural Economics 32, 73–86. Qaim, M., Zilberman, D. (2003), Yield effects of genetically modified crops in developing countries. Science 299, 900–902. Quisumbing, A.R., Brown, L.R., Feldstein, H.S., Haddad, L., Pen˜a, C. (1995), Women: The Key to Food Security. International Food Policy Research Institute, Washington, DC. Romeis, J., Shelton, A. S., Kennedy, G. G. (Eds.) (2008). Integration of Insect-Resistant Genetically Modified Crops within IPM Programs. Springer, New York. Sharma, D. (2004), GM Food and Hunger: A View from the South. Forum for Biotechnology and Food Security, New Delhi. Shiva, V. (2009), From seeds of suicide to seeds of hope: why are Indian farmers committing suicide and how can we stop this tragedy? The Huffington Post (April 28). Available at www.huffingtonpost.com/ vandana-shiva/from-seeds-of-suicide-to_b_192419.html. Retrieved in October 2010. Stein, A.J., Meenakshi, J.V., Qaim, M., Nestel, P., Sachdev, H.P.S., Bhutta, Z.A. (2005), Analyzing the health benefits of biofortified

54

Matin Qaim

staple crops by means of the DALY approach. HarvestPlus Technical Monograph 4. International Food Policy Research Institute, Washington, DC. Stein, A.J., Qaim, M. (2007), The human and economic cost of hidden hunger. Food and Nutrition Bulletin 28, 125–134. Stein, A.J., Sachdev, H.P.S., Qaim, M. (2008), Genetic engineering for the poor: Golden Rice and public health in India. World Development 36, 144–158. Subramanian, A., Kirwan, K., Pink, D., Qaim, M. (2010), GM crops and gender issues. Nature Biotechnology 28, 404–406. Subramanian, A., Qaim, M. (2009), Village-wide effects of agricultural biotechnology: The case of Bt cotton in India. World Development 37, 256–267. Subramanian, A., Qaim, M. (2010), The impact of Bt cotton on poor households in rural India. Journal of Development Studies 46, 295–311. Svedberg, P. (2002), Undernutrition overestimated. Economic Development and Cultural Change 51, 5–36. Tang, G., Qin, J., Dolnikowski, G.G., Russell, R.M., Grusak, M.A. (2009), Golden Rice is an effective source of vitamin A. American Journal of Clinical Nutrition 89, 1776–1783. Thirtle, C., Lin, L., Piesse, J. (2003), The impact of research-led agricultural productivity growth on poverty reduction in Africa, Asia and Latin America. World Development 31, 1959–1975. UN SCN. (2004), Fifth Report on the World Nutrition Situation. United Nations System, Standing Committee on Nutrition, Geneva. von Braun, J. (2008), Food and Financial Crises: Implications for Agriculture and the Poor. Food Policy Report, International Food Policy Research Institute, Washington, DC. WHO. (2009), Global Prevalence of Vitamin A Deficiency in Populations at Risk 1995–2005; WHO Global Database on Vitamin A Deficiency. World Health Organization, Geneva. World Bank. (2007), World Development Report 2008: Agriculture for Development. The World Bank, Washington, DC.

CHAPTER 3

Current and Potential Farm-Level Impacts of Genetically Modified Crops in Developing Countries$ Terri Raney and Ira Matuschke Food and Agriculture Organization of the United Nations, Rome, Italy E-mail addresses: [email protected]; [email protected]

Abstract World agriculture faces enormous challenges in the coming decades. To feed the world adequately in 2050, agricultural production in developing economies will need to nearly double. Incremental production will mainly come from increases in yields or cropping intensities. This chapter focuses on the potential of genetically modified (GM) crops to contribute to agricultural productivity growth and poverty reduction in developing economies. On the basis of a comprehensive literature review of the most recent literature, we aim to shed light on (a) whether GM crops benefit farmers in developing economies and (b) whether GM crops that are currently in the research pipeline address future challenges for agriculture. The first part of the chapter reviews farm-level impacts of GM crops in developing economies. The second part discusses the GM crop research pipeline. GM crop markets are expected to grow in the future but not to change dramatically. We conclude that GM crops benefited farmers, including resource-poor farmers, in developing economies, but benefits are location- and individual-specific. Addressing such complexities will be required to unlock technology potentials. Keywords: Farm-level impacts of genetically modified crops, global overview, research pipeline JEL Classifications: O13, O33, Q16 1. Introduction World agriculture faces enormous challenges in the coming decades: to provide higher quality diets and other products for increasingly affluent $

The views expressed in this chapter reflect the opinions of the authors and not necessary those of the Food and Agriculture Organization of the United Nations (FAO).

Frontiers of Economics and Globalization Volume 10 ISSN: 1574-8715 DOI: 10.1108/S1574-8715(2011)0000010008

r 2011 by Emerald Group Publishing Limited. All rights reserved

56

Terri Raney and Ira Matuschke

populations, to do so in ways that are environmentally sustainable, and to ensure growth opportunities for 3 billion people who will continue to rely on agriculture for their livelihoods. Feeding the world adequately in 2050 will require a 70 percent increase in global output and a near doubling in developing economies (Bruinsma, 2009). While this means that productivity growth rates will not have to be as high as in the past four decades, the incremental production requirements are considerable and need to be achieved mainly through yield increases and higher cropping intensities, because land expansions are not feasible or desirable in many countries (Bruinsma, 2009). Natural resource degradation and climate change will put additional pressures on agricultural producers. Scientists predict that rising global temperatures will make climatic conditions hotter and drier in many parts of the world; along with an increase in the frequency of extreme weather events like droughts and floods (IPCC, 2007). This will modify cropping cycles and input requirements. If adaptation measures are not implemented at farm and regional levels, agricultural output could decrease significantly (Binswanger-Mkhize, 2009). Agriculture is required not only to meet these growing demand-and supply side challenges but also to support the livelihoods of over 70 percent of the world’s poor, who continue to live in rural areas (De Janvry, 2009). Increases in agricultural productivity reduce poverty and promote broader economic growth through three main channels: raising farm incomes; stimulating the wider rural economy through higher demand for supplementary inputs, labor, and non-tradable goods and services; and boosting the purchasing power of poor urban consumers through lower food prices (FAO, 2004; World Bank, 2007). Productivity-enhancing technologies will be an important avenue to address the aforementioned challenges. However, not all technologies are equally supportive of poverty reduction. Biological technology, such as improved seeds, is often more poverty-reducing than mechanical technology because it is more scale-neutral (Evenson and Gollin, 2003). To realize the largest gains in productivity and poverty reduction, modern technologies must be embedded in supportive infrastructures and should be locally adapted and accessible to all farmers, including resource-poor farmers. This chapter focuses on the potential of genetic engineering to contribute to agricultural productivity growth in developing economies. The area under genetically modified (GM) crops rose from 1.6 million hectares in 1996 to 134 million hectares in 2009. Approximately 14 million farmers in 16 developing and 9 developed countries grew GM crops in 2009 (James, 2009). The rapid dissemination of GM crops was accompanied by a large public debate. It is argued that GM crops can contribute to yield stabilization and productivity increases owing to a stronger resistance to biotic and abiotic stresses (Borlaug, 2000). For example, Brookes and Barfoot (2009) found that GM crops significantly increased global farm output during the period 1996–2007. Qaim and Matuschke (2005)

Current and Potential Farm-Level Impacts of Genetically Modified Crops

57

and Raney (2006), using comprehensive literature reviews, showed that smallholder farmers in developing economies can share in these benefits. Nonetheless, the two studies concluded that farm-level impacts can be highly variable over time and geographical location, and they depend critically on institutional environments, infrastructures, and functioning markets (Raney, 2006). Skeptics argue that GM crops may have irreversible health and environmental impacts. Others further observe that available GM technologies are not ‘‘future-proof’’, because they fail to address the challenges of climate change and the needs of marginalized farmers (Union of Concerned Scientists, 2010). In fact, the GM crop market is currently dominated by four crops (soybean, maize, cotton, and canola) and two traits (insect-resistance and herbicide-tolerance). Research on drought or salt-tolerant staple food crops is ongoing, but these innovations have not reached the commercialization stage yet. Giving this debate and the future challenges that agriculture faces, in this chapter we aim to shed light on two questions: (i) Do GM crops benefit farmers in developing countries? and (ii) Do GM crops have the potential to address the future needs of farmers? To answer these questions, the chapter is divided into two main parts: The first part reviews farm-level impact of GM crops in developing economies. We focus on peer-reviewed studies that were published after 2004, because, owing to low data availability, early impact studies often used field trial data, single season data, or small data sets. More recent studies draw a more complex picture of farm-level impacts. In addition, earlier literature is excellently summarized in other publications. The State of Food and Agriculture 2003–2004, for example, dealt extensively with the impact of agricultural biotechnology in developing economies (FAO, 2004). The second part of this chapter provides an overview on GM crops that are currently in the research or regulatory pipeline. The chapter proceeds as follows. Section 2 reviews farm-level impacts of GM crops in Asia, Africa, and Latin America. Section 3 discusses what is in the biotechnology research pipelines, and Section 4 concludes.

2. Farm-level impacts of GM crops in developing countries: experiences from Asia, Africa, and Latin America 2.1. Asia Asia is the continent with the third largest GM crop area, following North America and Latin America. India (8.4 million hectares) and China (3.7 million hectares) dominate the region, which has a GM crop area of 12.6 million hectares. The Philippines reported a GM area of 0.5 million hectares in 2009 (James, 2009). Herring (2009) describes plantings of unapproved seeds in Viet Nam and Thailand. These seeds contain gene

58

Terri Raney and Ira Matuschke

events that have not been officially approved by the national authorities in the respective countries. Nazli et al. (2010) estimated that in 2007, 60 percent of the cotton area in Pakistan was planted with unapproved GM cotton varieties. Bt cotton is the main GM crop in Asia. Bt cotton contains a gene from the soil bacterium Bacillus thuringiensis (Bt), which makes the cotton plant resistant to cotton bollworms. In India, cotton is a major cash crop. Bt cotton was commercialized in 2002 by the private seed company Mahyco, in cooperation with Monsanto. Many Indian cotton varieties are hybrids; and by 2004, cotton hybrids covered about two-thirds of the Indian cotton area (Stone, 2011). The Bt gene was also incorporated into hybrid varieties. To ensure best crop performance, hybrids should not be farmsaved. This provides incentives for private sector companies, because it allows them to regain their investments in research and development. The technology premium that Mahyco initially charged for Bt cotton seeds was very high compared with conventional cotton seeds. Nonetheless, the Bt cotton area increased from 50,000 hectares in 2002 to 8.4 million hectares in 2009, which is about 84 percent of the total Indian cotton area. Sixtyfive percent of the total Bt cotton area was rain-fed in 2009, and the number of Bt cotton varieties available in the market increased from 3 in 2002 to 522 in 2009 (Choudhary and Gaur, 2010). As a result of the high prices for officially approved Bt seeds, the sale of illegal seeds, which are often sold loose, is also flourishing (Crost et al., 2007; Grue`re et al., 2008; Sheridan, 2009). Illegal seeds contain officially approved gene events that are backcrossed with conventional local varieties, without the seed developer’s consent. In response to high seed prices, several Indian states introduced a price cap for Bt cotton. For the 2009–2010 season, the state governments of Andhra Pradesh, Maharashtra, and Gujarat set a maximum retail price of 750 rupees per seed packet (B450 grams)1. The governments of Punjab, Haryana, and Rajasthan requested seed companies to charge less than 925 rupees per packet (The Hindu, 2010).2 Sadashivappa and Qaim (2009) analyzed seed prices, using field survey data, and found that Bt cotton prices fell from 1,600 rupees/acre in 2002–2003 to 800 rupees/ acre in 2006–2007. The authors observed that prices significantly decreased not only for officially approved seeds but also for illegal seeds. The impacts of price controls on research incentives remain to be seen.

1 One packet usually contains the optimal amount of seeds for one acre and consists of 450 gram of Bt seeds and 120 gram of conventional cotton seeds, to be planted as refuge area. 2 The governments of Andhra Pradesh, Maharashtra, and Gujarat introduced differential pricing for Bt cotton seeds: The maximum retail price (MRP) is 625 rupees/packet for Monsanto’s Bollgard-I trait, which includes a single gene event. The MRP for Monsanto’s Bollgard-II trait, which includes multiple gene events, was set at 750 rupees/packet. Governments of Punjab, Haryana, and Rajasthan asked seed companies to charge less than 750 rupees/packet for Bollgard-I and 925 rupees/packet for Bollgard-II (The Hindu, 2010).

Current and Potential Farm-Level Impacts of Genetically Modified Crops

59

A large number of studies evaluated farm-level impacts of Bt cotton in India. While early studies used field trial data (e.g., Qaim and Zilberman, 2003), more recent studies employed farm surveys for impact assessments. Table 1 summarizes selected studies published between 2004 and 2009. It differentiates between agronomic and economic effects. The table shows that Bt cotton reduced the number of insecticide applications. Even though the Bt gene makes the cotton plant resistant to cotton bollworms, it does not protect against sucking pests. Therefore, farmers may still need to spray insecticides. In addition, resistance to certain types of bollworms may be less than 100 percent, particularly in the late stages of plant growth. Bt cotton farmers also reported yield increases between 29 and 60 percent compared with farmers who did not grow Bt cotton. Yield differences were on average higher for irrigated than for nonirrigated plots (Gandhi and Namboodiri, 2006). The Bt gene does not affect yields per se. It rather reduces potential crop losses due to pests and thereby increases effective yields. Both agronomic effects – insecticide reductions and yield increases – are generally greater with higher pest infestation levels. Looking at the average economic effects, Table 1 displays that, owing to lower insecticide applications, Bt cotton reduced insecticide costs between 20 and 100 percent. Nonetheless, total costs for Bt cotton producers tended to be 8–32 percent higher than for conventional cotton farmers, which is mainly explained by higher seed prices. In addition, total labor costs were higher for Bt cotton farmers because of an increased labor demand, particularly during harvest time. Subramanian and Qaim (2009) demonstrated that higher Bt cotton yields increased the demand for harvest laborers; and since cotton picking is predominantly carried out by women in India, this benefited particularly female wage laborers. Higher total costs were offset by higher output and lower insecticide costs, thus that total revenues were higher for Bt cotton farmers compared with conventional cotton farmers. Overall, net income gains varied between 58 and 140 percent. Qaim et al. (2009) considered the distribution of direct and indirect benefits from Bt cotton between households with different levels of poverty. They found that the majority (60 percent) of benefits accrued to extremely poor and moderately poor households. Despite these overall positive effects displayed in Table 1, there is controversy over Bt cotton. Arguments are raised that the methodologies applied may not correctly capture farm-level impacts of Bt cotton (e.g., Glover, 2009). There may be selection bias, for example, farmers who are more specialized or better educated are the first to adopt a new technology. Not accounting for selection bias may thus lead to an overestimation of the impact of the technology itself. Crost et al. (2007) and Morse et al. (2007) used fixed-effects models and plot comparisons, respectively, to control for individual specific characteristics. Crost et al. (2007) demonstrated that more efficient farmers tend to be early adopters. Morse et al. (2007) found

Source

Barwale et al. (2004)

Qaim et al. (2005)

Bennett et al. (2004a)

Kambhampati et al. (2006)

Morse et al. (2007)

Crost et al. (2007)

Gandhi and Namboodiri (2006) Bennett et al. (2005) Sadashivappa and Qaim (2009)

a

Data

1,069 cotton farmers, six states, 2002 season, survey administered by Mahyco 341 cotton farmers, four states, panel data, 2002–2003 season 3496 farmers, Maharashtra state, 2002–2003 and 2003–2004 season 787 cotton farmers, Maharashtra state, 2002–2003 and 2003–2004 season 2709 cotton farmers, Maharashtra state, 2002–2003 and 2003–2004 season 157 cotton farmers, Maharashtra state, 2002–2003 and 2003–2004 season 338 cotton farmers, Maharashtra state, panel data comprising the years 2002 to 2003 694 cotton farmers, four states, 2003–2004 season 622 cotton farmers, Gujarat state, 2003–2004 season 341 cotton farmers, four states, panel data comprising the years 2002–2007

Average agronomic effects

Average economic effects

Insecticide sprays

Yields

Seed costs

Insecticide costs

Total costs

Net income

62%

þ61%

N/A

N/A

N/A

N/A

62%

þ34%

þ221%

69%

þ17%

þ69%

59%

þ51%

N/A

112%

N/A

þ58%

69%

þ54%

þ224%

111%

þ8%

þ62%

70%

þ54%

N/A

N/A

þ9%

þ62%

N/A

þ3586%a

N/A

N/A

þ1332%a

þ62144%a

N/A

þ1131%a

þ243%

15%

N/A

N/A

66%

þ47%

þ183%

44%

þ17%

þ102%

N/A

þ29%

N/A

N/A

N/A

þ132%

29%

þ40%

þ166%

24%

þ17%

þ89%

Upper and lower range, when accounting for self-selection biases (see discussion below).

Terri Raney and Ira Matuschke

Morse et al. (2005a)

Average farm-level impacts of Bt cotton in India, partial farm budgeting

60

Table 1.

Current and Potential Farm-Level Impacts of Genetically Modified Crops

61

that ‘‘the overall effect is that when comparing Bt plots of adopters and non-Bt plots of non-adopters, roughly half of the observed [yield] increase is due to a farmer effect and half to the Bt trait’’ (p. 498). Both studies concluded it is essential to account for self-selection bias when evaluating the impact of innovations. Measurement and estimation biases may also lead to incorrect measurements of farm-level impacts. They may result from small sample sizes or recalled data. Estimation biases could also result from partial farm budgeting, when inherent costs (e.g., land, family labor) are not accounted for (for a discussion, see Smale et al., 2009). Another argument is that Bt cotton impacts are highly variable across time and regions. This argument is valid and has been documented, starting from the earliest ex-post adoption studies (FAO, 2004). For example, Qaim et al. (2005), in an analysis of Bt cotton performance in four major Indian cotton growing states, found that in all states insecticide applications decreased significantly. Yet, in only three of the four states, net income gains were positive and significant. Bt cotton farmers in the fourth state, Andhra Pradesh, experienced net income losses compared with conventional cotton growers. Morse et al. (2005a) also reported district-level variations in net income gains for the state of Maharashtra. Differences in crop performance can be explained by a wide range of factors, for example, agronomic conditions, pest loads, availability of alternative pest control measures (FAO, 2004). In addition, the Bt gene may be incorporated into cotton varieties that are not sufficiently adapted to local growing conditions (Qaim et al., 2005). Another reason for the observed yield variability may be the large amount illegal seeds available in rural areas. To illustrate, Morse et al. (2005b) compared the performance of official Bt cotton hybrids in Gujarat with illegal Bt cotton hybrids; farm-saved Bt cotton hybrids, and conventional (non-Bt) cotton hybrids. Farmers in the sample referred to all three Bt cotton types as GM crops, even though the farm-saved hybrids may no longer have the hybrid vigor. Morse et al. (2005b) demonstrated that yield increases in comparison with the conventional variety were 0 percent, 14 percent, and 20–37 percent for the farm-saved, illegal, and official Bt cotton varieties, respectively. Insecticide applications were lower for all Bt cotton varieties. The authors found that net incomes were the highest for farmers growing official varieties, followed by illegal, farmsaved, and conventional seeds. The performance of different seed types, which farmers recognized as being the same, may add to the perception of a large variability in the performance of GM crops. Finally, there is controversy on the relationship between Bt cotton cultivation and farmer suicides in India (Sheridan, 2009). Grue`re et al. (2008), in a comprehensive review of available evidence, found that official statistics on farmer suicides in India vary widely. Using data from the National Crime Records Bureau, the authors found that farmer suicides increased from 13,622 in 1997 to 17,006 in 2006. In the period

62

Terri Raney and Ira Matuschke

2002–2006, that is, after the introduction of Bt cotton, the number of farmer suicides decreased, although regional differences prevailed. Grue`re et al. (2008) concluded that there was no observed causality between Bt cotton adoption and farmer suicides. These nation-wide findings were less conclusive for the state of Andhra Pradesh. As reported by Qaim et al. (2005), Bt cotton farmers in Andhra Pradesh did not experience an increase in net incomes. In their final conclusion Grue`re et al. (2008) stated that Bt cotton was not a sufficient explanations for farmer suicides in India and that root causes of suicides need to be addressed. Cotton is a major crop not only in India but also in China. Bt cotton was commercialized in China in 1997 by the private and the public sector, but public sector varieties dominate the market. In 2008, Bt cotton covered 68 percent of the total Chinese cotton area (Huang et al., 2010). Lu et al. (2010) report an adoption rate of 95 percent in northern China. Up to 300 Bt varieties were marketed in 2005 (Xu et al., 2008). Unlike in India, Chinese farmers can save Bt seeds from previous harvests, because public sector Bt varieties are open-pollinated. Xu et al. (2008) stated that 24 percent of all Bt cotton seeds are farm-saved, 20 percent from noncommercial channels, and 56 percent from commercial seed dealers. The large number from non-commercial sources induced the Chinese government to subsidize officially approved seeds to increase their adoption. To our knowledge, there are no recent peer-reviewed studies on farmlevel impacts of Bt cotton in China. Current studies, which are discussed below, rather focus on issues like insecticide use, potential insect resistance, and the evolution of secondary pests. To illustrate general research results on farm-level impacts, we pick one exemplary study: Using a dataset that comprised three years and four provinces, Pray et al. (2002) analyzed the impact of Bt cotton in China. In 1999, 283 farmers from one province were randomly selected. In 2000, the sample increased to 400 farmers from two provinces; and in 2001, 366 farmers from four provinces were interviewed. Similar to research results in India, the study reported significant reductions in insecticide applications: Over the three-year period, insecticide applications on Bt plots were 238 percent lower compared with conventional plots. This had a positive impact on farmers’ health status, because insecticide-related poisonings declined. Interestingly, despite these massive reductions, compared with conventional plots, Pray et al. (2002) observed that insecticide applications on Bt plots increased again from 1999 to 2001. Yields rose on average by 23 percent over the three-year period, with large seasonal fluctuations. Compared with their Indian counterparts, Chinese Bt cotton farmers experienced higher reductions in insecticides and lower increases in yields, because they tended to apply large quantities of pesticides before the commercialization of GM cotton. Looking at the economic effects, Pray et al. (2002) found that seed costs were on average 170 percent higher over the three-year period. Net income of Bt cotton farmers was positive in all seasons, while it was negative for conventional

Current and Potential Farm-Level Impacts of Genetically Modified Crops

63

cotton growers. In addition, Pray et al. (2001) demonstrated that smallscale farmers experienced higher gains from Bt cotton cultivation compared with farmers with larger landholdings. Recent studies predominantly focus on the ecological impact of Bt cotton in China and more specifically on increased insecticide use by Bt cotton farmers. Higher insecticide costs diminish net income gains and erode the profitability of Bt cotton cultivation. Farmers are required to spray more insecticides when cotton bollworms become resistant to the toxins in the Bt cotton plant. Cotton bollworms could become resistant, because the Chinese government does not require farmers to plant refuge areas; and Bt cotton plants generally contain only one gene event instead of two, as common in India or the United States (Huang et al., 2010). Yet, Huang et al. (2010), using farm-level data from five Chinese provinces over a six year period, found that bollworm resistance had not affected the effectiveness of Bt cotton. In addition, the authors demonstrated a negative relationship between the spread of Bt cotton and the size of bollworm populations; benefiting both Bt and conventional cotton farmers. Other studies, however, demonstrated that the number of non-target, or secondary, pests rose with the widespread adoption of Bt cotton in China. For example, Lu et al. (2010), using field trial data for 1998–2009, found a positive correlation between the spread of Bt cotton and the number of sprays against mirid bugs. Mirid bug infestations increased in cotton, but also in other crops like grapes, apples, and peaches. The authors concluded that reductions in insecticides against bollworms can increase infestation levels of non-target insect pests. These interactions need to be taken into account when evaluating the agro-ecological impact of GM crops (Lu et al., 2010). Wang et al. (2006), using a household survey of 481 farmers collected in five provinces, demonstrated that total insecticide costs for Bt and conventional cotton plots were almost equal in 2004. The authors showed that Bt farmers sprayed more against non-target pests. As a result expenditures to control non-target pests had nearly offset insecticide savings. Wang et al. (2006) concluded that farmers need to be better informed about risks associated with non-target pests to keep them in check. Pemsl and Waibel (2007), using data from 150 small-scale farmers in the Shandong province in 2002, also observed that Bt cotton farmers sprayed a large amount of other chemical insecticides. They concluded that, even with high adoption rates of Bt cotton, integrated pest management techniques could help to improve the profitability of cotton production. Moreover, the authors emphasized the necessity to provide better pest management training for farmers. 2.2. Africa Three African countries commercialized GM crops. South Africa is the country with the largest area of GM crops: 2.1 million hectares were

64

Terri Raney and Ira Matuschke

planted with GM maize, soybean, and cotton in 2009. In Egypt, GM maize covers approximately 1,000 hectares (James, 2009). Farmers in Burkina Faso grew Bt cotton on 125,000 hectares in 2009 (Vitale et al., 2010). In the following, we focus on South Africa’s experience with Bt maize and Bt cotton, because sufficient farm-level impact assessments from Egypt and Burkina Faso are not available yet. Maize is a major crop in South Africa that is mainly grown by commercial farmers. Bt maize provides protection against stalkborers that can cause significant yield losses (Gouse et al., 2006). Two types of Bt maize are cultivated in South Africa: Bt yellow maize was commercialized by Monsanto in 1998 and is mainly used as animal fodder. Bt white maize is a staple food crop and is marketed by Monsanto since 2001. In 2007– 2008, Bt yellow maize covered 55 percent of the total yellow maize area. The share of Bt white maize in the total white maize areas increased from 9 percent in 2003–2004 to 56 percent in 2007–2008 (Gouse et al., 2008, 2010). Herbicide-tolerant (HT) maize and maize that contains multiple traits (i.e., Bt and HT) are increasingly adopted by South African farmers (Gouse et al., 2010). HT maize contains a gene from the soil bacterium Agrobacterium tumefaciens that makes the plant tolerant to the broadspectrum herbicide glyphosate. Glyphosate is patented by Monsanto under the name Roundup. Most of the studies that analyzed farm-level impacts of GM maize were carried out by researchers at the University of Pretoria, South Africa. Gouse et al. (2005), for example, surveyed 33 large-scale commercial farmers of Bt yellow maize in 1999–2000 and 2000–2001. The authors found that the average yield advantage of Bt yellow maize over conventional maize was 11 percent. The authors also demonstrated that Bt yellow maize reduced insecticide costs by 163 percent and 171 percent for irrigated and non-irrigated maize plots, respectively. Net income gains were shown to be positive and significant. Gouse et al. (2006) analyzed the adoption of Bt white maize by smallholder farmers: In 2001–2002, 368 smallholder farmers in four South African provinces were interviewed. In 2002–2003, the authors surveyed 104 farmers in one province (KwaZulu Natal); and in 2003–2004, 196 farmers were interviewed in KwaZulu Natal. In all three seasons pest infestation levels were below average. In 2001–2002, 3000 farmers, who attended a training workshop, received Bt white maize seeds for free. In subsequent years farmers had to buy seeds from input dealers, and seed shortages were reported. Seed shortages were also noted for HT maize by Gouse et al. (2008). Gouse et al. (2006) found that yield differences between Bt white maize and conventional white maize were 32 percent in the first season. In the second season yield differences were only 16 percent. The authors related this to the smaller sample size and the large variation in the data. In the third season, no significant yield differences were reported, and Bt maize cultivation did not benefit adopters because of the low overall pest infestation levels (Gouse et al., 2006). Gouse et al. (2008) analyzed the

Current and Potential Farm-Level Impacts of Genetically Modified Crops

65

adoption of Bt white maize and HT white maize by 249 smallholder farmers in three districts of KwaZulu Natal in 2006–2007. The authors found that HT maize farmers achieved significantly higher yields compared with Bt maize and conventional maize farmers, respectively; with large yield variations prevailing. Looking at labor use with HT, Bt and conventional maize, the authors demonstrated that overall Bt maize required the least labor inputs. Looking at the distribution of labor inputs, Gouse et al. (2008) established that labor hours for females were lower on Bt maize plots compared with HT or conventional maize plots. For child labor the opposite was true: Children spend more time working on Bt maize plots compared with conventional or HT maize plots, respectively. In their conclusion the authors cautioned readers that results are based on a relatively small sample and represent only one season. Data are highly variable by district and should therefore be interpreted with care. To compare farm-level impacts of HT and Bt maize with more certainty, data that span several seasons are required. Cotton was produced in 5,100 hectares in South Africa in 2009–2010, of which 80 percent were irrigated. Cotton production in South Africa is dominated by large-scale farmers, and only 18 percent of the total cotton area is reported to be farmed by small-scale farmers (Cotton South Africa, 2010). Bt cotton was commercialized in 1998, and farm-level impact studies focused on the Makhatini Flats in the KwaZulu Natal province (e.g., Bennett et al., 2004b; Morse et al., 2006). Overall this province presents 8 percent of South Africa’s cotton area and 1.35 percent of the country’s total cotton production (Cotton South Africa, 2010). Cotton is mainly produced by smallholders on rainfed plots, and the number of women farmers is high. Vunisa Cotton, a private cotton company, introduced Bt cotton in the Makhatini Flats and initially provided inputs, credit, and extension advice to farmers. It also bought cotton output to finance farm credits (Morse et al., 2006). The subsequent adoption was rapid: 92 percent of all cotton farmers in the Makhatini Flats had adopted Bt cotton in 2002. Witt et al. (2006) describe this rapid uptake as supply rather than demand driven, because all inputs were provided by Vunisa Cotton. To analyze farm-level impacts of Bt cotton in the Makhatini Flats, Bennett et al. (2004b) used data from a farm survey and Vunisa Cotton that comprised three seasons (1998–1999 to 2000–2001): for the first season 1,283 farmers, the second season 441 farmers, and the third season 499 farmers. The authors found that, over the three-year period, pesticide costs were on average 56 percent lower for Bt cotton farmers compared to conventional cotton farmers. Lower insecticide applications were positively related to labor hours required to spray insecticides (Bennett et al., 2004b). Fewer insecticide sprayings benefited particularly women and children, who mainly apply insecticides. In addition to this, fewer incidences of insecticide-related poisonings were reported by Bt cotton adopters

66

Terri Raney and Ira Matuschke

(Bennett et al., 2003). Bennett et al. (2004b) also demonstrated that, over the three-year period, the average yield advantages of Bt cotton over conventional cotton was 68 percent. Yields were reported to vary widely for rainfed and irrigated crops (Hofs et al., 2006). Seed prices for Bt cotton were on average 88 percent higher, and net incomes for the seasons 1998– 1999 and 2000–2001 were on average 159 percent higher. The 1999–2000 season was characterized by particularly high pest pressures: Net incomes for conventional cotton plots were negligible, while they were positive for Bt cotton adopters. Looking at the distribution of benefits, Bennett et al. (2004b) demonstrated that smallholder farmers were able to benefit equally or more compared with large-scale farmers. Morse et al. (2006), using the same data set as Bennett et al. (2004b), analyzed the environmental impact of Bt cotton. The authors applied a Biocide Index and an Environmental Impact Quotient and concluded that insecticides sprayed on Bt cotton plots were more environmental-friendly compared with conventional cotton plots. This was related to a reduction in insecticides against bollworms and other non-target pests. Morse et al. (2006) concluded that Bt cotton performance should not only be judged on an economic or environmental basis, and they cautioned readers that an over-reliance on one company, in this case Vunisa Cotton, may increase the vulnerability of farmers to shocks. From 2003 onward – just three seasons after the introduction of Bt cotton – the production of Bt cotton fell drastically (Gouse et al., 2005). Vunisa Cotton did no longer provide inputs to farmers and withdrew altogether from the Makhatini Flats (Witt et al., 2006). Consequently, the production of Bt cotton decreased. The total cotton area in KwaZulu Natal decreased from 6,760 hectares in 2005–2006 to 400 hectares in 2009– 2010 (Cotton South Africa, 2010). Today all of KwaZulu Natal’s cotton area is planted with GM cotton (Jennifer Thomson, personal communication, July 2010). Farmers in the Makhatini Flats have few alternatives to cotton cultivation, because of the low availability of irrigation water. Improving cotton production and markets is therefore paramount. The Bt cotton experience illustrates the importance of well-functioning markets and infrastructures to unlock technology potentials (Gouse et al., 2005; Witt et al., 2006).

2.3. Latin America Brazil (21.4 million hectares), Argentina (21.3 million hectares), and Paraguay (2.2 million hectares) are the countries with the largest GM crop area in Latin America. HT soybeans are the main GM crop in Latin America; GM maize and cotton are cultivated to a lesser extent (James, 2009). Compared to the area that HT soybeans occupy in Latin America, it is surprising that only few farm-level impact studies have been published.

Current and Potential Farm-Level Impacts of Genetically Modified Crops

67

Qaim and Traxler (2005), for example, evaluated the impact of HT soybeans in Argentina. Monsanto commercialized HT soybeans in 1996, and adoption was rapid: The share of HT soybeans in the total soybean area is estimated to be 90 percent (Trigo and Cap, 2006). In Argentina HT soybeans were introduced by a multinational private seed company. Farmers in Argentina can farm-save their seeds, because they are not required to sign special contracts with the seed company. This particular institutional arrangement boosted the adoption of HT soybeans considerably. It was estimated that out of the total area of HT soybeans in 2001, 30 percent were planted with farm-saved seeds (Qaim and Traxler, 2005). Qaim and Traxler (2005), using three-year averages for a sample of 59 soybean farmers from three provinces, showed that the amount of herbicides applied was 107 percent higher on HT soybean plots compared with conventional soybean plots. Despite larger amounts of herbicides, the authors demonstrated that herbicide compositions changed: Glyphosate, applied on HT soybean plots, is classified as less toxic than herbicides used for conventional soybeans. Qaim and Traxler (2005) detected no significant yield difference between HT soybean and conventional soybean plots. This is in line with Trigo and Cap (2006) who analyzed the impacts of GM crops in Argentina over a 10-year period and concluded that, on average, there are no significant yield differences between HT and conventional soybean plots. Looking at the economic effects, Qaim and Traxler (2005) stated that HT soybean seeds were 21 percent more costly than conventional soybean seeds. Herbicide costs were reduced by 76 percent, because glyphosate was less expensive compared with other herbicides. Net income gains, over a three-year average, were on average 8 percent higher for HT soybean farmers. Researchers at the International Food Policy Research Institute recently carried out a study on the adoption of HT soybeans in Bolivia. The country commercialized HT soybeans in 2005, and HT soybeans are estimated to make up 70 percent of the total soybean production in Bolivia (Paz et al., 2009). Preliminary results from a study of 124 randomly selected soybean farmers suggested a positive yield advantage of HT soybeans over conventional soybeans. Seed costs were higher, while herbicide and labor costs3 were lower. Net incomes for HT soybean farmers were significantly higher compared with conventional soybean farmers (Paz et al., 2009). HT soybeans also provide other, less tangible, benefits. Cultivation is more flexible, because glyphosate controls for a broad spectrum of weeds and allows for a larger time window for herbicide applications (Brookes

3 Of all production costs, only labor costs were significantly different between HT and conventional soybean plots.

68

Terri Raney and Ira Matuschke

and Barfoot, 2009). In addition to this, HT soybeans can be grown in non-tillage systems that reduce labor requirements and machinery costs. No-till agriculture can also reduce soil erosion and thereby increase soil productivity (Brookes and Barfoot, 2009). As a consequence of lower labor requirements, farm household spent less of their own time in the field or hire-in less labor. This is particularly relevant in situations of labor shortages or when family labor is extensively used for cropping operations. Recent studies on HT soybean cultivation in Argentina documented glyphosate resistance in weeds, for example, Johnsongrass (Binimelis et al., 2009). If weed management practices are not changed, this could erode the profitability of HT soybean cultivation. Binimelis et al. (2009) argued that a more complex weed management system should include crop diversification, crop and herbicide rotation, and integrated weed management. Other (supplementary) ways to reduce the general over-reliance on glyphosate could be to speed up the development, testing, and regulatory approval for HT crops that are tolerant to other broad spectrum herbicides, for example, glufosinate. Similar to glyphosate the World Health Organization classified glufosinate into Toxicity class III (slightly hazardous). To our knowledge there is only one study that considered farm-level impacts of Bt cotton in Latin America (Qaim and de Janvry, 2005). Monsanto commercialized Bt cotton in Argentina in 1998, but adoption rates are low. This is related to the high technology premium that Monsanto charges. Unlike HT soybean farmers, Argentinean Bt cotton farmers cannot farm-save their seeds owing to specific contractual arrangements with the seed producing company. Using data for 299 cotton farmers in two provinces for the 1999 and 2000 growing seasons, Qaim and de Janvry (2005) demonstrated that Bt cotton decreased insecticide sprays on average by 52 percent and increased yields by 61 percent. This is in line with Bt cotton experiences in China and India. Moreover, the authors found that smallholder farmers benefited more from cultivating Bt cotton compared with large-scale farmers. In their conclusions, Qaim and de Janvry (2005) emphasized the importance of effective institutional frameworks as a precondition for technology adoption and diffusion.

3. The research pipeline The Food and Agriculture Organization (FAO, 2010) identified major research areas to address future challenges for agriculture: Current biotic stresses like pests, diseases, and weeds will need continued attention of researchers. Moreover, new diseases, like wheat black stem rust, are projected to spread faster with increased global trade. Owing to climate change, research to address abiotic stresses, like droughts or salinity, will become even more relevant. Breeding for sustainable yield increases will remain important, particularly for so-called orphan crops for which

Current and Potential Farm-Level Impacts of Genetically Modified Crops

69

potential yield gaps are large. Finally, breeding for improving the nutritional quality of food crops can be important for tackling malnutrition in developing economies (FAO, 2010). Addressing these complex challenges will require substantial investments in research, development, and extension. In the following, we discuss GM crops that are currently in the research pipeline and check whether these crops can meet the challenges of the future. 3.1. What is in the research pipeline? The area under GM crops will continue to grow in the future. It is projected that the number of countries that approve GM crops will rise to 40 countries in 2015. The majority of countries approving GM crops will be developing economies in Asia, Africa, and the Middle East (James, 2008). Depending on the adoption of GM rice, the area under GM crops could more than double, from 125 million hectares today to 300 million hectares in 2015 (James, 2008). In the short to medium run, the GM crop market will not change dramatically: New crop varieties will be varieties of maize, soybeans, and cotton. Herbicide-tolerance and insect-resistance will continue to be the main traits. Stein and Rodriguez-Cerezo (2009) provided a comprehensive overview of the global research and regulatory pipelines. Figure 1 summarizes GM events to be released in the short to medium run for six different crops. Commercialization dates should be read with caution, because regulatory processes and costs as well as consumer resistance can delay approval dates considerably. Figure 1 shows that the number of gene events is expected to increase in all crops. Soybean events, for example, will increase from 1 to 17 events. Herbicide-tolerance will remain the main trait in soybeans, and private sector companies will continue to dominate the market (Stein and 30 Number of events

25

2008

2015

20 15 10 5 0 Soybean

Maize

Rapeseed

Cotton

Rice

Potatoes

Crop

Fig. 1.

Projected increase in GM events, 2008–2015. Source: Stein and Rodriguez-Cerezo (2009).

70

Terri Raney and Ira Matuschke

Rodriguez-Cerezo, 2009). For maize the situation looks similar. Insectresistance will be the main trait in maize, and the private sector will dominated research and development. A special initiative, which is discussed below, is drought-tolerant maize that is expected to be commercialized in 2012 in the USA and in 2017 in Sub-Saharan Africa (Edmeades, 2008). The number of cotton gene events will more than double in the medium-term, and insectresistance and herbicide-tolerance will remain the main traits in cotton. The majority of events are currently developed and tested in China and India, respectively. This is in line with the general trend of biotechnology research on domestic crops being increasingly carried out within the countries (Stein and Rodriguez-Cerezo, 2009). South-South cooperation in research and development is also expected to accelerate (Dickson, 2003). Other crops that are in various stages of the research pipeline are summarized in Table 2. Even though the research pipeline is impressive, the table confirms that in the medium to long-run the market for GM crops will expand but not change dramatically. Maize, cotton, and soybeans will remain the main crops in the global GM market. Given the evidence presented in Section 1 of this chapter, we assume that farmers will be able to extract higher net incomes from cultivating these crops owing to lower input requirements and/or higher yields. However, will the future challenges for agriculture, outlined above, be sufficiently addressed? Below we consider research initiatives that aim at addressing some of the future challenges for the world’s major food crops: rice, wheat, and maize.

3.2. Genetically modified drought-tolerant maize Water scarcity and climate change have unfavorable effects on crop growing conditions, particularly in low latitude regions. Appropriate measures need to be taken to adapt to these impacts (IPCC, 2007). According to Edmeades (2008, p. 4) ‘‘as a rough rule of thumb, it has been estimated that 25 percent of losses due to drought can be eliminated by genetic improvement in drought tolerance, and a further 25 percent by application of water-conserving agronomic practices, leaving the remaining 50 percent that can only be met by irrigation’’. Among droughttolerant crops, research on maize is most advanced. The majority of the global maize area is rainfed, and yield losses in drought years can be substantial (Edmeades, 2008; Fischer et al., 2009). Monsanto and BASF are leading research in drought-tolerant maize technologies. In June 2009, the companies announced that they jointly identified a gene that makes the maize plant more resistant to abiotic stresses. Field trials, carried out in drought-prone areas of the USA, demonstrated that drought-tolerant maize had a yield advantage of 6–10 percent compared to the bestperforming, not drought-tolerant maize variety. Drought-tolerant maize is expected to be released in the USA in 2012 (Monsanto and BASF, 2009).

Current and Potential Farm-Level Impacts of Genetically Modified Crops

Table 2. Country Argentina

Research pipeline, selected crops, by developing country Crop Potato Rice Safflower Sugarcane Tomato Wheat

Bangladesh Eggplant Burkina Faso Beans Brazil

Beans Rice Sugarcane

Chile

Tomato

China

Cabbage Papaya

Trait Insect/virus resistance Herbicide tolerance, insect resistance Modified product quality Virus/insect/fungal resistance, herbicide tolerance Modified product quality/virus/ insect/fungal resistance Herbicide tolerance/fungal resistance/modified starch content Insect resistance Herbicide tolerance/fungal resistance Herbicide tolerance/fungal resistance Herbicide tolerance/insect resistance Virus/insect/fungal resistance, herbicide tolerance Modified product quality/virus/ insect/fungal resistance Herbicide/Insect resistance Virus resistance

Potato Rice

Insect/virus resistance Herbicide tolerance/insect resistance/salt tolerance

Soybean Tomato

Herbicide tolerance Modified product quality/virus/ insect/fungal resistance Herbicide tolerance/fungal resistance/modified starch content Potato Tuber Moth resistance Modified product quality/virus/ insect/fungal resistance Herbicide tolerance/fungal resistance Modified product quality/virus/ insect/fungal resistance Herbicide/Insect resistance Insect resistance Insect resistance/Herbicide tolerance Herbicide tolerance, modified composition Herbicide resistance

Wheat

Egypt

Potato Tomato

Ghana

Beans

Guatemala

Tomato

India

Cabbage Eggplant Maize

India

Millet Mustard

Status Field trials Field trials Field trials Field trials Field trials Field trial

Field trials Field trials Field trials Field trials Field trials Field trials Field trials Recommendation for commercialization Field trials Field trials, (Insectresistant rice approved) Field trials Field trials Field trial

Field trials Field trials Field trials Field trials Field trials Field trials Field trials Field trials Field trials

71

72

Terri Raney and Ira Matuschke

Table 2. (Continued ) Country

Crop Okra Potato Rice

Indonesia

Kenya

Malawi Nigeria

Pakistan Philippines

South Africa

Tanzania Tanzania

Trait

Insect resistance Insect/virus resistance Herbicide tolerance/Insect resistance/Salt tolerance Sugarcane Virus/insect/fungal resistance, herbicide tolerance Tomato Modified product quality/virus/ insect/fungal resistance Cassava Virus resistance/modified composition Rice Herbicide tolerance/Insect resistance Potato Insect/virus resistance Tomato Modified product quality/virus/ insect/fungal resistance Cassava Virus resistance/modified composition Cotton Insect resistance/Herbicide tolerance Maize Insect resistance/Herbicide tolerance Millet Herbicide tolerance, modified composition Sweet potato Herbicide tolerance/Virus and fungal resistance/modified composition Cotton Insect resistance/Herbicide tolerance Beans Herbicide tolerance/fungal resistance Cassava Virus resistance/modified composition Cotton Insect resistance/Herbicide tolerance Eggplant Insect resistance Papaya Virus resistance Rice Herbicide tolerance/Insect resistance/modified composition Tomato Multiple virus resistance Millet Herbicide tolerance, modified composition Potato Insect/virus resistance Sorghum Food composition Sugar cane Virus/insect/fungal resistance, herbicide tolerance Cotton Insect resistance/Herbicide tolerance Maize Insect resistance/Herbicide tolerance

Status Field trials Field trials Field trials Field trials Field trials Field trials Field trials Field trials Field trials Field trials Field trials Field trials Field trials Field trials

Field trials Field trials Field trials Field trials Field trials Field trials Field trials Field trials Field trials Field trials Greenhouse trials Field trials Field trials Field trials

Current and Potential Farm-Level Impacts of Genetically Modified Crops

73

Table 2. (Continued ) Country

Crop

Thailand

Tomato

Uganda

Banana Cassava Cotton Maize

Zimbabwe

Cotton Maize

West Africa

Cowpea

Trait Modified product quality/virus/ insect/fungal resistance Fungal/virus resistance Virus resistance/modified composition Insect resistance/Herbicide tolerance Insect resistance/Herbicide tolerance Insect resistance/Herbicide tolerance Insect resistance/Herbicide tolerance Insect resistance

Status Field trials Field trials Field trials Field trials Field trials Field trials Field trials Establishment of the network for genetic improvement for cowpea in Africa

Sources: Barry (2009), Eicher et al. (2006), GMO Compass (2010), Norton and Hautea (2009), Eicher et al. (2006). Notes: Most of the information in the table is from the GMO Compass, which is produced by a private consultancy company. The set-up of the database was financially supported by the European Union.

Two initiatives are currently aiming to develop drought-tolerant maize for Sub-Saharan Africa. The Water Efficient Maize for Africa (WEMA) initiative is a public–private partnership. Partners include the International Maize and Wheat Improvement Center (CIMMYT); BASF and Monsanto; the African Agricultural Technology Foundation; and National Agricultural Research Institutes and private seed companies in Kenya, Uganda, Tanzania, and South Africa (African Agricultural Technology Foundation, 2010). Kenya, under the leadership of the Kenyan Agricultural Research Institute, is expected to start the first field trials with drought-tolerant maize in 2010, and the variety is projected to be commercialized in SubSaharan Africa in 2017. The Drought-Tolerant Maize for Africa (DTMA) initiative is led by CIMMYT and the International Institute of Tropical Agriculture (IIATA) in partnership with 50 organizations. The DTMA initiative aims to ‘‘generate maize varieties with 100 percent superior drought tolerance; increase productivity under smallholder farmer conditions by 20–30 percent; and reach 30–40 million people in Sub-Saharan Africa’’ (CIMMYT, 2010). The project currently focuses on 14 countries in Sub-Saharan Africa and also attempts to address seed production and marketing in these countries.

74

Terri Raney and Ira Matuschke

While these initiatives are promising, their actual impact will depend on complex factors. For example, will the newly developed varieties be open-pollinated or hybrid maize varieties? This can have significant consequences on extension and distribution channels; seed costs; the speed of adoption; and the distribution of benefits between farmers and seed companies. In addition, as stated above, drought-tolerant varieties will need to be supplemented by improved water management techniques and irrigation facilities to achieve the highest possible gains (Edmeades, 2008).

3.3. Genetically modified rice Breeding for improved rice varieties is currently taking place in a number of countries. For example, in Bangladesh, India, Indonesia, and the Philippines, marker-assisted breeding is used to develop salt-tolerant rice varieties (Esperanza et al., 2009). Pro-Vitamin-A-enriched rice (Golden Rice) varieties are currently tested in the Philippines and are not expected to be commercialized before 2011 (Enserink, 2008). China is leading research in insect-resistant rice, and approximately 20 percent of Chinese public expenditures on agricultural biotechnology are allocated to rice research and development (Huang et al., 2008). The Bt rice plant is resistant to the rice stemborer and bacterial blight. The government approved Bt rice in December 2009, and China will be the first country to commercialize GM rice. However, the government requested production trials before Bt rice is fully commercialized (Waltz, 2010). Huang et al. (2008) analyzed farm-level impacts of Bt rice in pre-production trials, which they define as the last step before regulatory approval. The data comprised three years (2002–2004) and 17 villages in two rice-growing provinces. The authors surveyed 320 randomly selected farm households: 73 in 2002, 104 in 2003, and 143 in 2004. Of the sampled households, 192 cultivated Bt and conventional rice varieties. In total data for 584 rice plots were collected. To account for potential biases related to household characteristics, in the following we report only the results for those households that cultivated both Bt and conventional rice plots. Applying partial farm budgeting, Huang et al. (2008) found that Bt rice decreased insecticide sprayings by nearly four times. Rice yields increased by only 1 percent. The authors reported a sixfold decrease in pesticide costs, and labor days per hectare for spraying pesticides decreased by almost nine days for Bt rice adopters. Despite these positive farm trial results, there are a number of concerns associated with the commercialization of Bt rice. The demand for rice in China is expected to decrease, because consumers increasingly request protein-rich foods. In addition to this, the Bt gene has been incorporated into low-quality rice varieties, which may hamper the adoption of Bt rice (Huang et al., 2008). Moreover, inserting the Bt gene into rice hybrids may

Current and Potential Farm-Level Impacts of Genetically Modified Crops

75

influence farmers’ willingness to adopt new varieties. Finally, China is currently a net exporter of rice. In 2007 Chinese rice exports were valued at USD 28 million, which is relatively small compared to other major rice exporters. The impact of Bt rice commercialization on trade, particularly with those countries that ban imports of GM crops, remains to be seen (Huang et al., 2008). 3.4. Genetically modified wheat Research on GM wheat is currently carried out by public and private sector institutions in a number of countries. Researchers at the Chinese Academy of Agricultural Sciences, for example, develop wheat varieties that are disease and insect resistant or tolerant to drought and salinity. India also carries out research projects that aim at developing droughtand disease-resistant wheat varieties (Fox, 2009). In 2001, the Indian company Mahyco commercialized a hybrid wheat variety, which was developed for the semi-arid tropics and performs well under abiotic stress conditions (Matuschke et al., 2007). Australia is also well advanced in research on GM wheat. In 2007, field trials to test drought-tolerant wheat varieties, in the drought-prone state of Victoria, were approved by official authorities (McDonald, 2007). First trial results showed that the majority of GM varieties yielded 20 percent more than conventional wheat varieties (New Scientist, 2008). In the United States, Monsanto leads research in GM wheat. Monsanto developed herbicide-tolerant winter wheat varieties, but in 2004 the company dropped this research. The reason for this was the resistance by wheat farmers, who feared for the stability of their export markets (Fox, 2009). In 2009, the company announced that it will restart its research on GM wheat: Research will initially focus on conventional and marker-assisted breeding, but will also look at wheat varieties that are drought- and disease-resistant, and that can use nitrogen fertilizer better. It is estimated that first GM wheat varieties could be released within the next 10 years (Fox, 2009). 4. Conclusions This chapter reviewed farm-level impacts of GM crops in developing countries and examined the research pipelines in developed and developing economies. Using case studies from Asia, Africa, and Latin America, we demonstrated that GM crops are beneficial to farmers in developing economies. In general, farmers benefited significantly from the adoption of GM crops owing to reduced input requirements and/or higher yields. This led to increases in net incomes, and farmers, including smallholders, tended to benefit most from the seed technologies. Farm-level impacts, however, can vary considerably by region and season. Case studies from

76

Terri Raney and Ira Matuschke

India, for example, showed large yield variations between and within states. Studies from Africa revealed that seasonal variations have a significant impact on benefits. GM crop performance, similar to conventional crop performance, depends on a wide range of factors, like agronomic conditions, pest loads, alternative pest control measures, local adaptation of the plant, farmer’s skills, and supportive infrastructure. In addition, the availability of different GM seed types (i.e., officially approved, farmsaved, illegal) may also contribute to yield variability observed in many countries. The spread of illegal seeds was reported for all continents. This highlights that controlling the spread of GM crops is difficult, and caution should be exercised when authorizing field trials. The rapid spread of illegal seeds also emphasizes the importance of functioning regulatory frameworks, which should ensure that new crops are safe for human and animal health and for the environment. Regulatory frameworks should also stimulate innovative research by being cost-efficient and transparent. To date, a number of countries have not set up functioning regulatory systems; and even if frameworks are in place, a lack of technical and managerial capacities may prevent the implementation of regulations (Barry, 2009; Falck-Zepeda et al., 2009). In addition to the regulatory environment, the innovative strength of the agricultural research system and functioning rural infrastructures were shown to affect the speed of adoption and the distribution of benefits between farmers and seed companies (Raney, 2006). These factors are highly relevant to unlock technology potentials and to ensure that all farmers are able to access productivity-enhancing technologies. Furthermore, the case studies also highlighted the significance of providing information and training on modern technologies. Communication should be increased, for example, on refuge areas, the emergence of non-target pests, possible glyphosate resistance, and farm practices, like integrated pest and water management. In the short to medium term, research on GM crops in developed and developing countries will remain focused on commercial crops. Nonetheless, large developing economies, like China and India, emerge as new leaders in the market for GM crop research, and South-South cooperation in research and development is projected to increase. Many technologies that address future challenges for agriculture are currently in the research pipelines. Greater research efforts and large-scale investments in agricultural markets and regulatory and institutional frameworks are required to make these technologies available and accessible for farmers in developing countries. Acknowledgments The authors thank Alexander J. Stein for his very useful comments and James Edge for editorial support.

Current and Potential Farm-Level Impacts of Genetically Modified Crops

77

References African Agricultural Technology Foundation. (2010), Water efficient maize for Africa (WEMA) – Project Description. Available at www.aatf-africa.org/userfiles/WEMA-brief.pdf. Barry, G. (2009), The GM crop pipeline – contributing to sustainability of global food, feed and fibre productivity. Presentation at the Fourth International Conference on coexistence between genetically modified (GM) and non-GM based agricultural supply chains, Melbourne. Barwale, R.B., Gadwal, V.R., Zehr, U., Zehr, B. (2004), Prospects for Bt cotton technology in India. AgBioForum 7 (1&2), 23–26. Bennett, R., Buthelezi, T.J., Ismael, Y., Morse, S. (2003), Bt cotton, pesticides, labour and health: A case study of smallholder farmers in the Makhathini Flats, Republic of South Africa. Outlook on Agriculture 32 (2), 123–128. Bennett, R.M., Ismael, Z., Kambhampati, U., Morse, S. (2004a), Economic impact of genetically modified cotton in India. AgBioForum 7 (3), 96–100. Bennett, R., Ismael, Y., Morse, S., Shankar, B. (2004b), Reductions in insecticide use from adoption of Bt cotton in South Africa: Impacts on economic performance and toxic load to the environment. Journal of Agricultural Science 142, 665–674. Bennett, R., Ismael, Y., Morse, S. (2005), Explaining contradictory evidence regarding impacts of genetically modified crops in developing countries. Varietal performance of transgenic cotton in India. Journal of Agricultural Science 143, 35–41. Binimelis, R., Pengue, W., Monterosso, I. (2009), ‘‘Transgenic treadmill’’: Responses to the emergence and spread of glyphosate-resistant Johnsongrass in Argentina. Geoforum 40, 623–633. Binswanger-Mkhize, H.P. (2009), Challenges and opportunities for African agriculture and food security: High food prices, climate change, population growth, and HIV and AIDS. Paper presented at the Expert Meeting on How to Feed the World in 2050, Food and Agriculture Organization of the United Nations, Rome. Borlaug, N. (2000), Ending world hunger: The promise of biotechnology and the threat of antiscience zealotry. Plant Physiology 124 (2), 487–490. Brookes, G., Barfoot, P. (2009), Global impact of biotech crops: Income and population effects, 1996–2006. AgBioForum 12 (2), 184–208. Bruinsma, J. (2009), The resource outlook to 2050: By how much do land, water and crop yields need to increase by 2050? Paper presented at the Expert Meeting on How to Feed the World in 2050, Food and Agriculture Organization of the United Nations, Rome. Choudhary, B., Gaur, K. (2010), Bt Cotton in India: A country profile. ISAAA Series of Biotech Crop Profiles, International Service for the Acquisition of Agri-Biotech Applications, Ithaca, NY.

78

Terri Raney and Ira Matuschke

International Maize and Wheat Improvement Center (CIMMYT). (2010), The drought tolerant maize for Africa initiative. Background. Available at http://dtma.cimmyt.org/index.php/background. Cotton South Africa. (2010), Cotton South Africa corporate website. Available at www.cottonsa.org.za. Retrieved on October 17, 2010. Crost, B., Shankar, B., Bennett, R., Morse, S. (2007), Bias from farmer self-selection in genetically modified crop productivity estimates: Evidence from Indian data. Journal of Agricultural Economics 58 (1), 24–36. de Janvry, A. (2009), Agriculture for development. Toward a new paradigm and guidelines for success. A sequel to the World Development Report 2008. Presentation at the High Level Expert Forum on How to Feed the World in 2050, Food and Agriculture Organization of the United Nations, Rome. Dickson, D. (2003), South-South collaboration picks up steam. Available at www.scidev.net/en/editorials/southsouth-collaboration-picks-up-steam. html. Retrieved on January 11, 2010. Edmeades, G.O. (2008), Drought tolerance in maize: An emerging reality. A Feature. In: James C. (Ed.), Global status of commercialized biotech/ GM crops: 2008. ISAAA Briefs No. 39, International Service for the Acquisition of Agri-Biotech Applications, Ithaca, NY. Eicher, C.K., Maredia, K., Sithole-Niang, I. (2006), Crop biotechnology and the African farmer. Food Policy 31, 504–527. Enserink, M. (2008), Though lessons from Golden Rice. Science 320, 468–471. Esperanza, V.-L., Alpuerto, B., Norton, G.W., Alwang, J., Ismail, A.M. (2009), Economic impact analysis of marker-assisted breeding for tolerance to salinity and phosphorous deficiency in rice. Review of Agricultural Economics 31 (4), 779–792. Evenson, R.E., Gollin, D. (2003), Assessing the impact of the Green Revolution, 1960 to 2000. Science 300 (5620), 758–762. Falck-Zepeda, J., Falconi, C., Sampaio-Amstalden, M.J., Solleiro Rebolledo, J.L., Trigo, E., Vera´stegui, J. (2009), La Biotecnologı´a Agropecuaria en Ame´rica Latina. Una Visio´n Cuantitativa, IFPRI Discussion Paper 00860SP, International Food Policy Research Institute, Washington, DC. Food and Agriculture Organization of the United Nations (FAO). (2004), The State of Food and Agriculture. Agricultural biotechnology. Meeting the needs of the poor?. FAO, Rome. Food and Agriculture Organization of the United Nations (FAO). (2010), Current status and options for crop biotechnologies in developing countries. Background paper prepared for FAO International Technical Conference on Agricultural biotechnologies in developing countries. Guadalajara. Available at www.fao.org/fileadmin/user_upload/abdc/ documents/crop.pdf.

Current and Potential Farm-Level Impacts of Genetically Modified Crops

79

Fischer, R.A., Byerlee, D., Edmeades, G.O. (2009), Can technology deliver on the yield challenge to 2050? Paper presented at the Expert Meeting on How to Feed the World in 2050, Food and Agriculture Organization of the United Nations, Rome. Fox, J.L. (2009), Whatever happened to GM wheat? Nature Biotechnology 27 (11), 974–976. Gandhi, V.P., Namboodiri, N.V. (2006), The adoption and economics of Bt Cotton in India: Preliminary results from a study. Working Paper No. 2006-09, Indian Institute of Management, Ahmedabad. Glover, D. (2009), Undying promise: Agricultural biotechnology’s pro-poor narrative, Ten years on. STEPS Working Paper 15, STEPS Centre, Brighton. GMO compass. (2010), Plant database. Available at www.gmo-compass.org/eng/database/plants/. Gouse, M., Kirsten, J., Shankar, B., Thirtle, C. (2005), Bt cotton in KwaZulu Natal: technological triumph but institutional failure. AgBiotechNet 7, 1–7. Gouse, M., Pray, C., Schimmelpfennig, D., Kirsten, J. (2006), Three seasons of subsistence insect-resistant maize in South Africa: Have smallholders benefited? AgBioForum 9 (1), 15–22. Gouse, M., Piesse, J., Poulton, C., Thirtle, C. (2008), Efficiency and employment effects of GM maize in KwaZulu Natal. Paper presented at the Harvard International Institute for Advanced Studies Workshop on: African Agricultural Development, Improving African Agriculture for Accelerated Growth, Accra. Gouse, M., Kirsten, J., Piesse, J., Thirtle, C., and Poulton, C. (2010), Insect resistant and herbicide tolerant maize adoption by South African smallholder farmers – making sense of seven years of research. Paper presented at the 14th ICABR Conference on Bioeconomy governance: Policy, environmental and health regulation, and public investments in research, Ravello. Grue`re, G.P., Mehta-Bhatt, P., Sengupta, D. (2008), Bt cotton and farmer suicides in India. Reviewing the evidence. IFPRI Discussion Paper No. 00808, International Food Policy Research Institute, Washington, DC. Herring, R. (2009), Persistent narratives: Why is the ‘‘Failure of Bt cotton in India’’ story still with us? AgBioForum 12 (1), 14–22. Hofs, J.L., Hau, B., Marais, D., Fok, M. (2006), Boll distribution patterns in Bt and non-Bt cotton cultivars. II. Study on small-scale farming systems in South Africa. Field Crops Research 98, 210–215. Huang, J., Hu, R., Rozelle, S., Pray, C. (2008), Genetically modified rice, yields, and pesticides: Assessing farm-level productivity effects in China. Economic Development and Cultural Change 56 (2), 241–263. Huang, J., Mi, J., Lin, H., Wang, Z., Chen, R., Hu, R., Rozelle, S., Pray, C. (2010), A decade of Bt cotton in Chinese fields: Assessing the direct

80

Terri Raney and Ira Matuschke

effects and indirect externalities of Bt cotton adoption in China. Science China Life Sciences 53 (8), 981–991. Intergovernmental Panel on Climate Change (IPCC) (2007), Climate Change: Impacts, adaptation and vulnerability. In: Parry, M.L., Canziani, O.F., Palutikof, J.P., van der Linden, P.J., Hanson, C.E. (Eds.), Contribution of Working Group II to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change. Cambridge University Press, Cambridge. James, C. (2008), Global status of commercialized Biotech/GM crops: 2009, ISAAA Briefs No. 39, International Service for the Acquisition of AgriBiotech Applications, Ithaca, NY. James, C. (2009), Global status of commercialized Biotech/GM crops: 2009, ISAAA Briefs No. 41, International Service for the Acquisition of AgriBiotech Applications, Ithaca, NY. Kambhampati, U., Morse, S., Bennett, R., Ismael, Y. (2006), Farm-level performance of genetically modified cotton. A frontier analysis of cotton production in Maharashtra. Outlook on Agriculture 35 (4), 291–297. Lu, Y., Wu, K., Jiang, Y., Xia, B., Li, P., Feng, H., Wyckhuys, K.A.G., Guo, Y. (2010), Mirid bug outbreaks in multiple crops correlated with wide-scale adoption of Bt cotton in China. Science 328, 1151–1154. Matuschke, I., Mishra, R.R., Qaim, M. (2007), Adoption and impact of hybrid wheat in India. World Development 38 (8), 1422–1435. McDonald, K. (2007), GM Wheat trial approved. Drought-proof wheat lines will be trialed at two sites in Victoria. Available at www.lifescientist.com. au/article/186628/gm_wheat_trial_approved/. Monsanto and BASF. (2009), Monsanto und BASF legen Entdeckung eines Gens offen, das Mais vor Trockenheit schu¨tzt. BASF/Monsanto Press Release. Available at www.basf.com/group/corporate/de/function/ conversions:/publish/content/products-and-industries/biotechnology/ images/BASF_P_09-274_Gen_Entdeckung.pdf. Morse, S., Bennett, R.M., Ismael, Y. (2005a), Genetically modified insect resistance in cotton: some farm-level economic impacts in India. Crop Protection 24, 433–440. Morse, S., Bennett, R., Ismael, Y. (2005b), Comparing the performance of official and unofficial genetically modified cotton in India. AgBioForum 8 (1), 1–6. Morse, S., Bennett, R., Ismael, Y. (2006), Environmental impact of genetically modified cotton in South Africa. Agriculture, Ecosystems and Environment 117, 277–289. Morse, S., Bennett, R., Ismael, Y. (2007), Isolating the ‘farmer’ effect as a component of the advantage of growing genetically modified varieties in developing countries: a Bt cotton case study from Jalgaon, India. Journal of Agricultural Science 145, 491–500. Nazli, H., Sarker, T., Meilke, K., Orden, D. (2010), Economic performance of Bt cotton varieties in Pakistan. Paper presented at the

Current and Potential Farm-Level Impacts of Genetically Modified Crops

81

Annual Meeting of the Agricultural and Applied Economics Association, Denver, CO. New Scientist. (2008), Drought-resistant wheat beats Australian heat. New Scientist 2662, June 28. Norton, G.W., Hautea, D.M. (2009), Projected impacts of agricultural biotechnologies for fruits and vegetables in the Philippines and Indonesia. ISAAA, Ithaca, NY. Paz, R., Fernandez, W., Zambrano, P., Falck-Zepeda, J. (2009), GM soybeans in Bolivia. Presentation made at the 13th ICABR Conference on The Emerging Bio-Economy, Ravello. Pemsl, D., Waibel, H. (2007), Assessing the profitability of different crop protection strategies in cotton: case study results from Shadong Province, China. Agricultural Systems 95, 28–36. Pray, C., Ma, D., Huang, J., Qiao, F. (2001), Impact of Bt cotton in China. World Development 29 (5), 813–825. Pray, C.E., Huang, J., Hu, R., Rozelle, S. (2002), Five years of Bt cotton in China-the benefits continue. The Plant Journal 31 (4), 423–430. Qaim, M., de Janvry, A. (2005), Bt cotton and pesticide use in Argentina: economic and environmental effects. Environment and Development Economics 10, 179–200. Qaim, M., Matuschke, I. (2005), Impacts of genetically modified crops in developing countries: a survey. Quarterly Journal of International Agriculture 44 (3), 207–227. Qaim, M., Subramanian, A., Naik, G., Zilberman, D. (2005), Adoption of Bt cotton and impact variability: insights from India. Review of Agricultural Economics 28 (1), 48–58. Qaim, M., Subramanian, A., Sadashivappa, P. (2009), Commercialized GM crops and yield. Nature Biotechnology 27, 803–804. Qaim, M., Traxler, G. (2005), Roundup Ready soybeans in Argentina: farm level and aggregate welfare effects. Agricultural Economics 32, 73–86. Qaim, M., Zilberman, D. (2003), Yield effects of genetically modified crops in developing countries. Science 299 (5608), 900–902. Raney, T. (2006), Economic impact of transgenic crops in developing countries. Current Opinion in Biotechnology 17, 1–5. Sadashivappa, P., Qaim, M. (2009), Bt Cotton in India: development of benefits and the role of government seed price interventions. AgBioForum 12 (2), 172–183. Sheridan, C. (2009), Doubts surround link between Bt cotton failure and farmer suicide. Nature Biotechnology 27 (1), 9–10. Smale, M., Zambrano, P., Grue`re, G., Falck-Zepeda, J., Matuschke, I., Horna, D., Nagarajan, L., Yerramareddy, I., Jones, H. (2009), Measuring the economic impacts of transgenic crops in developing agriculture during the first decade. Approaches, findings and future directions. Food Policy Review 10, 1–103.

82

Terri Raney and Ira Matuschke

Stein, A.J., Rodriguez-Cerezo, E. (2009), The global pipeline of new GM crops. Implications of asynchronous approval for international trade. JRC Technical Report EUR 23486 EN. Office for Official Publications of the European Communities, Luxembourg. Stone, G.D. (2011), Field versus Farm in Warangal: Bt cotton, higher yields and larger questions. World Development 39 (2), 387–398. Subramanian, A., Qaim, M. (2009), Village-wide effects of agricultural biotechnology: The case of Bt cotton in India. World Development 37 (1), 256–267. The Hindu. (2010), Seed price controls may impact Bt cotton availability. Available at http://www.thehindubusinessline.com/2010/05/22/stories/ 2010052251791700.htm. Trigo, E.J., Cap, E.J. (2006), Ten years of genetically modified crops in Argentine agriculture. Argentine Council for Information and Development of Biotechnology (ArgenBio), Buenos Aires. Union of Concerned Scientists. (2010), Impacts of genetic engineering. Available at http://www.ucsusa.org/food_and_agriculture/science_and_ impacts/impacts_genetic_engineering/impacts-of-genetic.html. Vitale, J., Ouattarra, M., Vognan, G., Traore, O. (2010), Burkina Faso. Bollgard II. Socio-economic Study: Outcomes from 2009 Field Survey. Presentation made at the 14th ICABR Conference on Bioeconomy governance: Policy, environmental and health regulation, and public investments in research, Ravello. Wang, S., Just, D.R., Pinstrup-Andersen, P. (2006), Tarnishing silver bullets: Bt technology adoption, bounded rationality and the outbreak of secondary pest infestations in China. Paper presented at the American Agricultural Economics Association Annual Meeting, Long Beach, CA. Waltz, E. (2010), In brief, China’s GM rice first. Nature Biotechnology 28 (1), 8. Witt, H., Patel, R., Schnurr, M. (2006), Can the poor help GM crops? Technology, representation and cotton in the Makhatini Flats, South Africa. Review of African Political Economy 109, 497–513. World Bank. (2007), World Development Report 2008. Agriculture for development. World Bank, Washington, DC. Xu, N., Fok, M., Bai, L., Zhou, Z. (2008), Effectiveness and chemical pest control of Bt-cotton in the Yangtze River Valley, China. Crop Protection 27, 1269–1276.

CHAPTER 4

The Impact of Bt Cotton and the Potential Impact of Biotechnology on Other Crops in China and India Carl E. Praya, Latha Nagarajana, Jikun Huangb, Ruifa Hub and Bharat Ramaswamic a Department of Agriculture, Food, and Resource Economics, School of Environmental and Biological Sciences, Rutgers University, New Brunswick, NJ 08901-8520, USA E-mail addresses: [email protected]; [email protected] b Center for Chinese Agricultural Policy, Chinese Academy of Sciences, Institute of Geographical Sciences and Natural Resources Research, Anwai, Beijing 100101, China E-mail addresses: [email protected]; [email protected] c Planning Unit, Indian Statistical Institute, New Delhi 110 016, India E-mail address: [email protected]

Abstract Since the 1980s agricultural biotech investments by the public sector have increased substantially in both China and India. In the last two decades there has also been a dramatic increase in private section investment in agricultural biotechnology particularly in India. The promise of major benefits of Bt cotton identified in early socioeconomic studies of Bt cotton has proven to be true. Bt cotton has spread to at least 66% and 85% of total cotton areas of China and India, respectively – wherever bollworm is a major problem. Bt cotton continues to control bollworm in both countries, and farmers continue as major beneficiaries rather than biotech or seed companies. The major impacts have been yield increases in India and reduced pesticides consumption in China. In China, evidence also suggests that Bt cotton has suppressed the bollworm population so that non-Bt cotton growers and producers of other crops that are susceptible to bollworm are also benefitting. The chapter also provides evidence that in the near future Bt rice and Bt eggplant could have major positive impacts by reducing pesticide use and farmers’ exposure to chemical pesticides and increasing yields. Both crops were approved for commercial production by government biosafety regulators, but are not yet available for commercial cultivation. Keywords: Agricultural biotechnology, Bt cotton, GMOs, India, China JEL Classification: Q1, Q15, Q16, Q18, Q19 Frontiers of Economics and Globalization Volume 10 ISSN: 1574-8715 DOI: 10.1108/S1574-8715(2011)0000010009

r 2011 by Emerald Group Publishing Limited. All rights reserved

84

Carl E. Pray et al.

1. Introduction China and India are two of the world’s largest producers and consumers of food and other agricultural products. Since the 1980s, both the Chinese and Indian governments have heavily invested in agricultural biotechnology research, with the Chinese government making more public-sector investment than the Indian government and the private sector in India exceeding China’s private sector. Chinese farmers began adopting genetically modified (GM) crops in the mid-1990s with Indian farmers following in 2000. In spite of R&D investments in various crops in both countries, one crop, Bt cotton, covers almost all areas in China and India in cultivation by a GM crop. Recently, China approved GM traits in two major food crops, rice and maize, and India nearly approved its first GM food crop, Bt eggplant. None of these crops are being grown by Chinese or Indian farmers. The experiences of India and China with GM crops are not only particularly important for Asia where few other countries grow GM crops (Philippines is the only exception) but also for other parts of the world where small farmers produce most crops. Comparing the experiences of China and India allows us to see how the same technology – Bt cotton – can have different impacts because of differences in technology policies, regulatory institutions, agricultural conditions and levels of development. Early studies of Bt cotton’s impacts in China (Pray et al., 2001) and India (Qaim and Zilberman, 2003) found increases in yield per hectare, especially in India, reductions in pesticide use, especially in China, and positive health impacts, in both countries, due to reduced pesticide exposure. In recent years, little has been published on impacts of GM crops, leading to the question: have Chinese and Indian farmers continued to benefit? Our paper addresses this question. It also provides empirical evidence for potential impacts of Bt rice and Bt eggplant, the next GM traits in the commercialization pipeline. This chapter finds that the promise of major benefits of Bt cotton identified in early studies has proven true. Bt cotton has spread to all areas of China and India where bollworm is a problem pest. Bt cotton continues to control bollworm in both countries, and farmers continue as major beneficiaries rather than biotech or seed companies. In China, evidence also suggests that Bt cotton has suppressed bollworm population to the extent that growers and producers of other crops also susceptible to bollworm are benefitting. This chapter also finds that GM rice and GM eggplant could have major positive impacts – increasing farmers’ profits by reducing pesticide use and reducing their exposure to chemical pesticides. Both crops were approved for commercial production by government biosafety regulators, but adoption of GM eggplant has been held up by the Indian Minister of the Environment and specific Bt rice hybrids are still being evaluated by the Chinese Ministry of Agriculture.

Impact of Bt Cotton

85

The chapter is organized as follows. It first reviews recent investments in biotechnology in China and India and spread of GM crops in the two countries. It then reviews impact studies, examines evidence available on potential impact of GM food crops in the two countries and finally summarizes results. 2. Investments in agricultural biotechnology 2.1. Public investments in Chinese agricultural biotech The Chinese government began investing in agricultural biotechnology research in the mid-1980s, with funding rapidly increasing since 1996 (Figure 1). Basic biotechnology research got a major boost from the ‘‘863’’ program funded by the Ministry of Science and Technology in March 1986, and, 10 years later, the 1997 ‘‘973’’ program further increased basic and applied biology research (Huang and Wang, 2002). Of total agrobiotechnology research investment, nearly 60% was in plant biotechnology, with the remainder allocated to animal and microorganism research. Since 2000, expenditure has increased even more rapidly, reaching about US $200 million in 2003. Unpublished survey data collected by the Center for Chinese Agricultural Policy in 2010estimates expenditures of at least US$1.2 billion for agricultural biotech research in 2009 (based on a survey of government agencies). Private-sector agricultural biotechnology research is small relative to the public sector. Some large Chinese seed companies, such as Origin Seed and

Million Yuan (2003 Prices)

1800 1600

Plants

1400

Agriculture

1200 1000 800 600 400

0

1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003

200

Fig. 1. Agricultural biotech research investment in China (1986–2003). Source: Huang et al. (2005). Note: Investments are calculated at 2003 year base prices equivalent terms. The conversion factor is, therefore, 1.65 billion yuan ¼ US$200 million using the 2003 market exchange rate.

86

Carl E. Pray et al.

Denghai Seeds, have established biotechnology laboratories in recent years. A few multinational seed and biotechnology companies have conducted applied biotechnology research in China. For example, Monsanto, in collaboration with Delta & Pineland and Chinese government research institutes, bred and tested Bt cotton varieties in the 1990s. This research was abandoned after 2000 as a result of low profits from Bt cotton due to lack of enforceable patents and trademarks. As a result, events such as Bollgard II (BG II), the most popular Bt cotton in India, and RR Flex cotton (stacked Bts plus herbicide-tolerant genes) have not entered into the Chinese biosafety regulatory process.1 Recently, foreign firms and local private firms have begun reinvesting in biotechnology innovation and research (Table 1), believing it possible to appropriate some gains from R&D as IPRs appear to be strengthening. The Plant Variety Protection Law passed in 1997 was gradually broadened to include all major crops including cotton in 2005. Government enforcement of patents, trademarks and PVP also appears to be improving. In interviews with the authors (Beijing, August 2009), seed company executives have stated they can charge high enough prices for hybrid maize to make investments in breeding profitable. Further, the growing sophistication of Chinese seed companies in their use of IPRs and contracts makes it easier for foreign companies to do business in China. Finally, the recent approval of GM rice and maize opens up food crops for GMOs for the first time. Factors in place for some time – the numbers of research institutes, the skills of Chinese scientists and the huge potential market – have also attracted international R&D investments. Foreign companies, with Chinese partners, have begun working on BG II, other Bt genes and herbicide-tolerant products for cotton, maize and other crops. In recent years, Monsanto, DuPont and Syngenta have invested in basic biological research in Beijing and Shanghai. Origin, a Chinese seed company, has introduced a GM maize event for high phytase developed by a Chinese government research organization.

2.2. Agricultural biotech R&D investments in India The establishment of the Department of Biotechnology (DBT) in 1985– 1986, under the Ministry of Science and Technology, marks the beginning of major public-sector biotechnology investment in India – with a substantial amount of this investment focused on agricultural biotechnology. Public-sector agencies involved in plant genomic research and crop biotechnology include the Indian branch of the International Center of 1 An ‘‘event’’ is defined as a specific set of genes that have been placed in specific plant background material. So there are many Bt cotton events that consist of different types of Bts in one background or the same Bts in different backgrounds.

87

Impact of Bt Cotton

Table 1.

Emerging public–private partnerships in agbiotechnology research in China (2009)

International company

Chinese biotech institute

Commodity focus

Pioneer/ DuPont

Peking University

Rice

Dow

Rice

Bayer Crop Sciences Monsantoa

China National Rice Research Institute National Institute of Biological Sciences Peking-Yale Joint Center for Plant Molecular Genetics and Ag-Biotechnology Institute of Genetics and Developmental Biology, Anhui Rice Research Institute, China Agricultural University China National Rice Research Institute Biotech research lab, Beijing

Syngentaa

Biotech lab, Beijing

$65 million in 5 years, 200 scientists, and technicians by 2010

Novozymea Mendel Geneticsa

Own lab in Beijing Research facility, Southern China

BASF Monsanto

Syngenta

Research description Stress, efficient nutrition utilization

Corn, soybean, rice

Yield genes

Gift for scholarships

Plant biotech

Corn, soybean, wheat, sugar beet, and sugar cane

Novel genes for agronomic traits

Rice

Hybrids

Company crops

Genomics and bioinformatics Yield, pest and drought resistance, and biomass conversion for biofuels Bioenzymes Miscanthus for biofuels

Selection for yield, disease resistance

Source: Author interviews with companies in 2009 Beijing or company websites. In-house research facility headquartered in China by the respective firms.

a

Genetic Engineering and Biotechnology, the Department of Science and Technology (DST), the Indian Council of Agricultural Research (ICAR) and the National Center for Plant Genome Research (NCPGR), established in New Delhi in 1998. From 1992 to 2002, public investment in crop biotechnology research more than tripled as the DBT increased its Five Year Plan expenditure from $40 million in the 8th plan (1992–1997) to $150 million in the 10th plan (2002–2007) (Rengasamy and Elumalai, 2009). Although estimates of India’s total R&D expenditures in agricultural biotechnology across relevant agencies are unreliable, James (2008) estimates that India’s public-sector investments in crop biotechnology R&D have totaled about $1.5 billion over the last five years, or $300 million per year.

88

Carl E. Pray et al.

Numerous private Indian seed companies along with subsidiaries of multinational companies have also heavily invested in crop biotechnology research beginning in the late 1990s – with private-sector investments (biotech and conventional breeding) estimated at about $200 million (Choudhary and Gaur, 2009). Thus, total investment in crop biotechnology is estimated at about $500 million a year.2 Private firms in India invest in crop biotechnology generally in two ways: (1) through investments in new R&D infrastructure (laboratories, green houses, and field testing facilities) and (2) by expanding existing human resources, skills, and commercial bases into new geographical areas of India. The DST has identified more than 150 private companies, research institutions, and laboratories engaged in research activities related to transgenic agriculture, tissue culture, biopesticides, biofertilizer, animal biotech, food and nutrition, and biofuels (DST Annual Report, 2008). An additional 40 firms use biotechnology tools to produce biofertilizer and biopesticides.

2.3. Adoption of biotechnology 2.3.1. Adoption of GM crops in China China approved GM cotton and petunia for commercial production in 1997 and tomato and sweet pepper soon thereafter in 1998 (Table 2). Then, after a long lag, Bt poplars were approved in 2005 and virusresistant papaya in 2006. The most recent crops approved for commercial production are Bt rice and high-phytase maize (both in 2009) with seed being available to farmers in about 2013 or 2014. Almost all GM crop biotechnologies approved and now commercially used in China were developed by the Chinese public sector – the only exceptions being Monsanto’s Bt gene and Bt cotton hybrids commercialized in China through joint ventures with Chinese and foreign biotech companies. Although GM papaya is grown extensively in southern China and GM sweet pepper and tomato are grown in small areas throughout several regions, Bt cotton is China’s only major GM field crop. First introduced in the northeastern cotton zone along the Yellow River (1997) and then spreading south into the Yangtze River region, Bt cotton now almost completely covers both regions. However, it has not made headway in the other main cotton-growing region, irrigated regions of the desert in Xinjiang Province where bollworm is not a major pest. Figure 2 shows the rapid spread of Bt cotton since its introduction in 1997 to its peak in 2004 where it covered nearly 70% of China’s total cotton production area. Adoption has 2 More recent unpublished estimates by Pray and Nagarajan suggest that $200 million may be too high and $100 million per year is probably closer to the amount actually spent by private firms engaged in crop biotechnology research in India.

89

Impact of Bt Cotton

Table 2.

GM crops approved for commercial production in China

Crop

Year

Trait

Cotton Petunia Sweet pepper/pepper Tomato

1997 1997 1998 1998

Poplar trees Papaya Rice Maize

2005 2006 2009 2009

Bt (Cry-1A), Bt (Cry-1Ac)þCPTi CHS (modified flower color) CMV-CP (Virus resistance) EFE-anti (Delayed ripening) CMV-CP (Virus resistance) Bt (Cry-1Ac) PRSV (Virus resistance) Bt (Cry-1A) High-phytase gene

Source: Compiled by R. Hu and J. Huang, CCAP, Beijing (2010).

(In thousand Hectares)

(%)

4000.0 3500.0

70 Area Ha.

Bt % 60

3000.0

50

2500.0 40 2000.0 30 1500.0 20

1000.0

10

500.0 0.0

0 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007

Fig. 2.

Bt cotton adoption in China, 1997–2008. Source: CCAP (2008). Note: About 7.1 million farmers adopted Bt cotton in 2008.

slowed since 2004 because farmers in all areas where bollworm is a major pest now use Bt cotton. No GM cotton varieties have been developed for Xinjiang. Hundreds of Bt varieties and hybrids are approved for use in China, with most areas covered with Bt varieties rather than hybrids. 2.4. Adoption of GM cotton in India In contrast to China, only one GM crop – Bt cotton – has been approved and introduced for commercial production in India. The Government of

90

Carl E. Pray et al.

India approved Bt cotton for commercial release in 2002, although it was first introduced illegally in Gujarat around 2000. In 2002, Indian farmers grew only about 50,000 hectares of Bt cotton, but adoption increased rapidly over the next few years (see Figure 3 and Table 4) so that by 2008 7.6 million acres were planted in Bt cotton, representing 82% of all cotton planted in that year (Figure 3). By the 2009–2010 cropping season, the area under Bt cotton cultivation is expected to reach nearly 90% of total areas planted. The states of Gujarat, Maharashtra, Andhra Pradesh, Madhya Pradesh, Punjab, and Haryana account for more than 71% of total cotton production in India and farmers in these areas are major users of Bt cottonseed.

2.4.1. Illegal Bt introduction and spread Farmers used Bt cottonseeds in India before the first official Bt hybrids were approved in 2002. During 2001, the Bt cotton hybrid NB-151 of the NavBharat Company was cultivated on more than 4000 hectares in Gujarat State. This hybrid had not undergone testing and trials mandated by biosafety regulations and had not been approved by the Genetic Engineering Approval Committee (GEAC) – hence the term ‘‘illegal seeds.’’ Though GEAC recommended that the NB-151 cotton crop be destroyed, farmer opposition prevented this. As a result, illegal Bt cottonseeds were multiplied and sold under various names on a growing black market in different Indian states (Sadashivappa and Qaim, 2009). Illegal Bt seeds were priced between 800 and 1200 Indian rupees (US$18–27) per packet of 450 g compared to Rs. 1,600 (US $36.45) for

12.0 10.0 Total cotton area (Mill Ha) 8.0 Bt cotton area (Mill Ha) 6.0 Illegal Bt Area 4.0 Total Bt area

2.0

0 01

09 20

20

08

-2

-2

00

00 -2

07

06

9

8

7 00

6 00

-2 20

20

05

-2

-2 04

20

20

4

00

3

00 -2

00 -2

03 20

02 20

5

0.0

Fig. 3. Trends under Bt cotton adoption in India (2002–2003 to 2009–2010). Source: 2002–2006, Singh (2007); 2007–2010 Illegal Bt area is estimated by Indian seed industry sources in personal communication with the authors.

Impact of Bt Cotton

91

official Bt cottonseeds (Murugkar et al., 2007).3 In 2004–2005, illegal Bt seeds reached an estimated 800,000 hectares (Pray et al., 2005). On average, illegal Bt hybrids generated higher profits than conventional cotton hybrids but lower profits than legal Bt hybrids (Bennett et al., 2005). The approval and wider availability of legal Bt cotton hybrids from multiple sources in 2002 combined with price controls implemented in 2006 significantly reduced use of illegal Bt seeds. 2.4.2. Legal Bt technology and status Monsanto released in March 2003 three Bt cotton hybrids with BG I trait for cultivation – with GEAC approval and in collaboration with its Indian partner, Maharashtra Hybrid Seed Company (Mahyco). Marketed by a 50–50 joint venture called Mahyco Monsanto Biotech (MMB), the Bt hybrids were sold commercially to farmers in the central and southern zones. A good monsoon season in that year increased the popularity of insect-resistant cotton among farmers. MMB licensed the Bt gene to regional seed companies that were market leaders in their locations and that were selling popular hybrids. These regional companies incorporated the Bt gene into their own hybrid varieties and began selling them after meeting necessary regulations. In May 2006, MMB produced hybrids with two Bt genes, BG II. Also in 2006, two domestic seed companies – JK AgriGenetics Ltd and Nath Seeds Ltd – released approved events of Bt cotton. JK AgriGenetics developed ‘‘Event 1’’ featuring the Cry1Ac gene sourced from the Indian Institute of Technology (IIT), Kharagapur. ‘‘Vishwanath’’ by Nath Seeds contained a fusion Cry1Ac/Cry1Ab Bt gene from Biocentury Transgene Technology Company (BTCC). This Bt gene was developed at the Chinese Academy of Agricultural Sciences (CAAS). By 2008, 30 seed companies were producing 274 Bt cultivars across 9 states (Natesh and Bhan, 2009). Notably, the first indigenous Bt cotton variety, Bikaneri Narma, was granted approval in 2008. This was the first GM crop developed by the Indian public sector – the Central Institute of Cotton Research, Nagpur and the University of Agricultural Sciences, Dharwad, Three varieties were commercially released for the 2009 crop season. Also by 2009, Metahelix, a biotechnology firm in Bangalore, was granted approval of its event. Three new events carrying Bt genes are currently undergoing extensive field testing. These proposed events cover broad spectrum insecticidal properties. For example, Monsanto’s RRF and Dow Agro’s Bt events carry herbicide and insecticidal tolerance. By the end of 2009, six Bt events were approved for the cotton crop alone, including two Monsanto events, a Chinese event and three domestic firm events (Table 3). Since their first official approval in 2002, the number of Bt-based hybrid cultivars has increased exponentially (Table 3). Between the years 3

One packet of seed is equivalent to 450 g of seed, suffice to plant 1 acre of cotton.

92

Table 3.

Carl E. Pray et al.

Bt cotton events approved and under trials for cultivation in India (2009)

Event name

Bollgard I (IR) Bollgard II (IR) Event 1(IR) GFM Cry1A(IR) CICR Event(IR)

Source

Monsanto Monsanto IIT, Kharagapur/JK AgriGenetics Chinese Academy of Sciences Nagpur/University of Agric. Sciences, Dharwad Metahelix JK AgriGenetics

9124(IR) Event 1þEvent 24 (IR) Widestrike (HTþIR) Dow Agro Roundup Ready Flex Bt (IRþHT)

Monsanto

Genes

Year of approval

Number of cultivarsa

cry1Ac 2002 cry1Ac and cry2Ab 2006 Truncated cry1Ac 2006

200þ 300þ 38

cry1Abþcry1Ac

2006

69

Truncated cry1Ac

2008

3

Synthetic Cry1C 2009 cry1Ac and cry1EC Pending approval cry1Ac and cry1F Pending approval Cry 1Ac, cry2Ab, Pending CP4EPSPS approval

2 NA NA NA

Source: APCoAB, 2009; Indian GMO Research Information System (IGMORIS) website, 2010. IR: insect resistance; HT: herbicide tolerance; NA: not available. a Cultivars approved till May 2010.

2002–2005, four companies (Mahyco, Rasi, Ankur, and Nuziveedu) released around 20 Bt-based cotton hybrids. In 2006, 62 Bt cotton hybrids were approved for planting, with two more events and a few more companies entering the market. By mid 2007, 111 Bt cotton hybrids were approved for commercial cultivation, and in 2008 the number of commercially released hybrids reached 278. The largest number of hybrids has been developed using MON 15985 and MON 531, totaling nearly 91% of acreage under Bt cotton (Francis Kanoi CCTK, 2009–2010). According to GEAC (2009), most released cultivars contained events from Monsanto (around 95%), with JK AgriGenetics Event-1 (2–3%) and Chinese Academy of Sciences-based events (4–5%) sharing the remainder of the market. By May 2010, 600 Bt cotton hybrids and varieties had been approved, and currently 33 companies are developing Bt cotton hybrids and varieties (IGMORIS, 2010). 3. Empirical studies on the impact of adoption of Bt cotton 3.1. Impact of Bt cotton adoption in China Empirical studies of GM traits in developing countries started with the Pray et al. (2001) study of Bt cotton using 1999 and 2000 production

Impact of Bt Cotton

93

surveys. This study was the first in a series conducted on Bt cotton with the CCAP. Other studies have been conducted by Pemsl et al. (2005); Pemsl (2006) and Wang et al. (2008). CCAP surveys, conducted in 1999, 2000, 2001, 2004, 2006, and 2007, are the only surveys available after 2004. They encompass a random sample of up to 500 farmers in the northeast (Yellow River) and central (Yangtze River) cotton zones.4 3.1.1. Agronomic and economic impacts Figures 4, 5, 6 and 7 summarize agronomic and economic impacts of Bt cotton adoption in CCAP field samples. Figure 4 shows that mean yields of Bt cotton were higher than conventional varieties in all years except 2004 when there was no statistically significant difference. Differences between Bt and non-Bt are reported after 2004, but they are somewhat less reliable because the number of plots where non-Bt was grown is scarce. In 2006, there were only 14 plots and in 2007 only 4. Figure 4 also shows that Bt cotton yields have remained high in recent years, with annual yield variations and variations in differences between Bt and non-Bt largely due to variations in weather and severity of pest attack (Hu et al., 2009; Huang et al., 2010). Because surveys in China were conducted over six years (from 1999 to 2007), it is possible to investigate two other potential impacts: (1) decline in Bt efficacy as a result of being backcrossed into more varieties by numerous public- and private-sector plant breeders or as a result of the development of bollworm resistance to Bt and (2) growth of secondary pests into major pest problems. Chinese data do not support the hypothesis of declining Bt efficacy. Aggregate cotton yields continue to rise in China suggesting that Bt cotton also continues to do well. Using historical data beginning in the 1950s, Figure 5 shows a gradual increase in cotton yields until the late 1970s. A yield jump in the late 1970s coincides with the 1978 introduction of the household responsibility system and continues to 2002 when a period of declining yields began as bollworms developed resistance to chemical pesticides. Yield growth and decline in yield volatility after 1995 correspond with the introduction and spread of Bt cotton. Early studies of three provinces in our sample in northern China found that Bt cotton reduced pesticide use by 35.7 kg per hectare, or a reduction of 55% of pesticide use in the entire sample between 1999 and 2001 (Huang et al., 2002). Henan is the only province in the survey where farmers cultivated some non-Bt fields as late as 2006 and 2007. Figure 6 shows that insecticide use against bollworm in Bt cotton fields in Henan has been less than 10 kg per hectare for the entire period except 2000, and 4 Xinjiang Province, China’s third major cotton-growing region, was not surveyed as cotton bollworm is not a significant pest and little Bt cotton is used.

94

Carl E. Pray et al.

4000 3500 3000 2500 Bt Non-Bt

2000 1500 1000 500 0 1999

2000

2001

2004

2006

2007

Fig. 4. Yields of Bt cotton versus conventional cotton in China (kg/ha). Sources: Data from 1999 to 2001 in the figure (Huang and Wang, 2002); R. Hu and J. Huang, CCAP for the years 2004–2007. Note: In 2006, only 14 farmers and in 2007 only 4 farmers reported growing non-Bt cotton in their plots.

1.400 1.200 1.000 0.800 0.600 0.400 0.200 0.000 1950 1955 1960 1965 1970 1975 1980 1985 1990 1995 2000 2005

Fig. 5.

Cotton yields in China (mt/ha) 1950–2008. Source: Compiled by R. Hu and J. Huang, CCAP, 2010.

that spraying for bollworms on non-Bt cotton fields has also declined dramatically since 1999. This supports the hypothesis that bollworm populations in the entire area have declined. Econometric modeling of pesticide use in all fields surveyed between 2001 and 2007 also supports this hypothesis of declining bollworm infestation over time (Huang et al., 2010). Although Bt cottonseed prices were higher than conventional cultivars during the survey years, differences in seed costs were offset by reductions in expenditures on pesticides and labor, due in large part to reductions in number of required sprays. This resulted in overall decreases in production costs

95

Impact of Bt Cotton 90 80 70

Bt variety pesticides for all insects Non-Bt variety pesticides for all insects Bt variety pesticides only for bollworm Non-Bt variety pesticides only for bollworm

Kg/ha

60 50 40 30 20 10 0 1999

2000

2001

2004

2006

2007

Fig. 6. Cotton pesticide use (kg/ha) by sample households in Henan Province, China (1999–2007). Source: Data from Huang et al., 2010. Note: In 2006, only 14 farmers and in 2007 only 4 farmers reported growing non-Bt cotton in their plots.

16000 Bt

14000

Non-Bt

12000 10000 8000 6000 4000 2000 0 1999

2000

2001

2004

2006

2007

Fig. 7. Net Revenues (RMB yuan current prices) from BT versus non-BT cotton among surveyed villages in China. Source: Data from 1999 to 2001 in the figure (Huang et al., 2002). Unpublished data collected by authors for the years 2004, 2006, and 2007. Note: In 2006, only 14 farmers and in 2007 only 4 farmers reported growing non-Bt cotton in their plots. for Bt cotton, as compared to non-Bt cotton, and increased net revenue (Huang et al., 2002). As shown in Figure 7, net revenue from Bt crops exceeded net revenue from conventional cotton in all surveyed farm households. In some villages studied, farmers reported increased levels of mirids, which had been only a minor pest when high levels of broad spectrum

96

Carl E. Pray et al.

pesticides were used before the adoption of Bt cotton. A recent study (Wang et al., 2008), measuring farm-level pattern of insecticide use from 1999 to 2006, shows increased insecticide use to control mirids between 2001 and 2004. This increase does not continue in most sampled villages after 2004. Increased insecticide use to control secondary pests is lower than reduction in total insecticide use due to Bt cotton adoption. Further econometric analyses show that fluctuation in mirid infestation is largely related to local temperature and rainfall (Huang et al., 2010). A new study by biologists in Science (Lu et al., 2010) confirms farmers’ observations that decreased pesticide use due to Bt cotton has led to an increase in mirids. Lu et al. (2010) argue that improved pest management strategies, such as carefully integrating Bt with other pesticides and combined with improved cultural practices, are needed to control secondary pests. This coincides with the findings of Pemsl and Waibel (2007) that to better realize potential benefits of Bt technology farmers must be trained in the use of other control measures (especially cultural practices and chemical pesticides).This is supported by results of onfarm trials conducted in Hubei Province (2002) that showed that use of non-Bt cotton combined with farmer training in integrated pest management (IPM) using the Farmer Field School (FFS) approach equaled the economic performance of Bt cotton with additional pesticide use (Yang et al., 2005). 3.1.2. Health effects Many authors have speculated about positive and negative health impacts of GM crops on farmers, but only one study (Hossain et al., 2004) provides statistical evidence linking GM adoption and use to farmer health. The impact of pesticide on health of exposed people includes both immediate sickness and long-term effects on the nervous system, which lead to sickness many years after initial exposure. Ideally, evidence of acute toxicity and potential long-term impacts is identified through physician and hospital medical records but resources for this study was limited. This study surveyed farmers who were asked if they felt sick during the season they were growing cotton. If so, how sick did they get and what were their symptoms (dizziness, nausea, headaches) and did they visit a doctor or hospital as a result? Data from the same surveys northeastern China 1999, 2000, and 2001 are depicted in Figures 4, 6, and 7. These data were pooled, and used to estimate a two-stage econometric model. The study shows that the amount of pesticide farmers sprayed on cotton was a major factor influencing whether farmers reported feeling sick or not. The more pesticide sprayed the higher probability that farmers would report feeling sick. To ensure that results were not biased by other factors that could lead to sickness, we included in the survey a number of farmer

Impact of Bt Cotton

97

characteristics that could affect susceptibility to poisoning such as age and preexisting health problems. As expected farmers who were already unhealthy had a higher probability of being poisoned and farmers with higher education had a lower probability of being poisoned. Other characteristics of farmers that we could measure such as age did not have a significant impact. Poisoning was linked to Bt cotton through impact of Bt on pesticide use. The study (and many others) found that Bt cotton adoption reduced pesticide use dramatically even after controlling for weather, incidence of serious pest attacks, and pesticide prices. Taken together, the impact of Bt on reducing pesticide use and the impact of the reduced pesticide use on poisonings indicates that adoption of Bt cotton can substantially reduce risk and incidence of pesticide poisoning. 3.2. Impact of adoption of Bt cotton in India A range of field studies assessing economic performance of Bt cotton in India revealed that farmers have benefited from adopting Bt cotton technology through increased yields and reduced pesticide costs. Although Bt technology does not target increased yield, substantial yield increases are attributed to decreased pest damages. In spite of higher costs of Bt cottonseed, reduced pesticide use, and reduced costs associated with pesticide use, offsets increased expenditures on seed. Additional spillover benefits include improved quality of life due to increased income and better health due to less pesticide exposure (ASSOCHAM Survey, 2007). 3.2.1. Reduction in insecticide use, increase in aggregate yields The Qaim and Zilberman (2003) study comparing Mahyco Bt hybrids with the same hybrid without Bt in three major cotton-growing states of India (Tamil Nadu, Maharashtra, and Madhya Pradesh) is the most carefully designed study of Bt trait impacts. Comparing Bt hybrids with genetically identical hybrids except for the Bt gene, the study found that Bt hybrids were sprayed three times less (70%) against bollworms than non-Bt hybrids and local varieties, and that yield increased by 80–87%. A study funded by Mahyco (Barwale et al., 2004) documented results of a 1,069 farmer survey in six states during the 2002 season. According to this report, Bt cotton increased yields by 42% and reduced pesticide use by 57%. In their assessment of Bt cotton performance among farmers in Maharashtra during the 2002 and 2003 cropping seasons, Bennett et al. (2004) found a significant reduction in pesticide expenditure, 72% in 2002, and 83% in 2003. Seed costs were higher; however, increased yields of Bt cotton of 45% in 2002 and 63% in 2003 over non-Bt cotton compensated for higher seed costs. Similar results were reported by Bennett et al. (2006)

98

Carl E. Pray et al.

from a survey conducted in Maharashtra, Gujarat, Madya Pradesh, and Karnataka. In Maharashtra, Bambawale et al. (2004) conducted a participatory field trial with MECH-162 variety Bt cotton and a conventional variety/hybrid using IPM techniques with both crops. Their results showed that IPM in Bt cotton was most effective with only 11% damage. Non-Bt hybrids using IPM had as much as 33% damage. Seed cotton yield was also higher in Bt cotton hybrids by 300 kg compared to non-Bt varieties and hybrids. A Front-Line Demonstrations (FLD) study on cotton conducted by the ICAR collected details from 1,200 Bt demonstration and farmer plots across 11 states in the 2005–2006 growing season. Results confirmed that Bt cotton hybrids registered a net yield increase of 33.7% over non-Bt hybrids and a 73.8% increase over open-pollinated cotton varieties (OPV). In their study of 694 growers from four major cotton-growing states of Gujarat, Maharashtra, Andhra Pradesh, and Tamil Nadu, Gandhi and Namboodiri (2006) found significantly higher yields of Bt cotton with reduced pesticide costs under both irrigated and rain-fed conditions. One of the few analyses examining Bt technology performance as late as 2006–2007 is the study conducted by Sadashivappa and Qaim (2009). Covering the first five years of Bt adoption in India and using three rounds of survey data between 2002–2003 and 2006–2007, they document reductions in pesticide use at around 30% and increased yields of 40% among Bt growers. Subramanian and Qaim (2009) also analyzed villagelevel welfare and distribution effects of Bt cotton adoption, documenting that, in addition to yield gains and decreased pesticide costs, the region as a whole showed improved aggregate employment, especially for hired female agricultural laborers, and increased household income among cotton growers. Each additional hectare of Bt cotton was shown to produce 82% higher aggregate incomes than obtained from conventional cotton. Farm-level studies are substantiated by aggregate cotton production data in India. After the release of the first commercial hybrids in 1970s, cotton yields showed marginal improvement, due to both public- and private-sector research. Yet, overall or aggregate yields only increased significantly with increased adoption of Bt cotton since 2002 (Figure 8). Prior to Bt cotton, India had one of the lowest cotton yields in the world – 308 kg per hectare in 2001–2002. The global average for cotton production is 788 kg per hectare (USDA-FAS, 2007). As shown in Table 4, the average yield of Bt cotton has increased to 560 kg per hectare in 2007–2008 as compared to 300 kg per hectare in 2001–2002. Currently, India accounts for 25% of the global area under cotton cultivation, around 33.4 million hectares. However, in terms of production, India accounts for only 20% of world production, due to lower productivity per hectare.

99

Impact of Bt Cotton 700 600 500 Bt Cotton

400 300

First Commercial Hybrid (H-4)

200 100

1950-51 1952-53 1954-55 1956-57 1958-59 1960-61 1962-63 1964-65 1966-67 1968-69 1970-71 1972-73 1974-75 1976-77 1978-79 1980-81 1982-83 1984-85 1986-87 1988-89 1990-91 1992-93 1994-95 1996-97 1998-99 2000-01 2002-03 2004-05 2006-07 2008-09

0

Fig. 8. Cotton yields in India (kg/ha) 1950–51 to 2008–2009. Source: Department of Agriculture, Ministry of Agriculture, New Delhi (2009).

Table 4. Year

Area, yield, and seed sales of Bt cotton in India

Total cotton area Million ha

2002–2003 2003–2004 2004–2005 2005–2006 2006–2007 2007–2008 2008–2009 2009–2010b

7.7 7.6 8.8 8.8 9.1 9.6 9.4 9.5

Bt cotton area

Yield

Bt seed packets solda

Illegal Bt area

Million

Million ha

Million ha

% to total

Kg/ha

0.05 0.1 0.5 1.3 3.8 6.3 7.6 8.4

0.7 1.3 5.7 14.7 41.6 65.7 81.1 88.8

302 404 470 478 521 560 526 575

0.1 0.2 1.3 3.1 4.0 16.0 27.0 30.0

0.03 0.1 0.6 1.2 2.0 1.8 1.6 1.3a

Source: James (2008); Ministry of Agriculture, GOI (2009). a Each packet sold is equivalent to 450 g of Bt cottonseed. b Estimates based on personal communication with seed firms (2009–2010 only).

3.2.2. Increased farm income and spillovers Reduced insecticide use combined with significant yield increases due to lower crop losses has resulted in considerable gains in farm-level profit. Profit differences between Bt and non-Bt cotton have increased over time – from US $49.23 per acre in 2002–2003 to US$66.97 in 2006–2007

100

Carl E. Pray et al.

(Sadashivappa and Qaim, 2009). Bennett et al. (2006), using 2002 and 2003 surveys, also conclude that Bt growers received higher gross margin – US $1,157 per hectare for Bt growers compared to US$665.4 per hectare for nonBt growers, even after account for seed cost and varying cotton prices. Another study involving 150 Bt cotton growers during the 2003 season in Maharashtra also reported a 79.2% higher profit from Bt cotton cultivation, compared to non-Bt cultivation under irrigated conditions (Vaidya, 2005). Gandhi and Namboodiri (2006), in their survey across four major cotton-growing states cited above, found that profits per hectare of Bt cotton cultivation ranged from US$ 347 to US$729, while non-Bt cotton profits ranged from US$123 to US$ 414 per hectare. The consulting firm ASSOCHAM (2007) conducted a study across 23 districts in 6 states and reported increases in net revenue of US$ 175 per acre associated with Bt adoption. This means Bt growers earned on average a 64% higher income per acre than conventional growers. Other measurable farm-level benefits from Bt cotton production are important. At the farm level, in addition to improved yields and higher incomes per acre, 87% of Bt cotton farmers reported better lifestyles, 84% reported improved peace of mind due to risk reduction, 72% invested more in their children’s education, and 67% repaid debts (ASSOCHAM Survey, 2007). A survey conducted by Indicus Analytics (2007) across 9000 farmers in eight states also found positive health impacts and increased investments in education among farm households growing Bt cotton. Additionally, the cotton industry has captured benefits. Qaim (2003) projected surplus gains from Bt cotton at $315 million for 2005. Of this, farmers captured two thirds while biotech and seed firms garnered the remainder. Bt cotton in India is commercialized in hybrids, so use of farm-saved seeds is low. Thus, the private sector profits from selling the GM cotton hybrids are higher than in China. A 2004 study conducted in Gujarat by Kambhampati et al. (2005) reveals that the textile industry, ginners, and textile manufacturers also benefitted from improved fiber quality due to less insect damage. When analyzing aggregate effects of increased production and trade of Bt cotton, the world market must also be considered. The Frisvold and Reeves (2007) paper is the first to consider concurrent impacts of Bt adoption in India and globally. Examining effects of Bt cotton production on world and Indian cotton prices at 2005 adoption levels, they estimated a global increase in total factor productivity (TFP) at around 3.3%, with 0.9% and 0.7% increases in textile and apparel production, respectively. They concluded that while Bt adoption in India led to a more than US$ 200 billion gain in India, increased worldwide production led to a 3% decline in world cotton prices. Anderson et al. (2008) estimate that widespread adoption of Bt cotton in India and other South Asian countries will result in additional regional welfare gains on the order of $1 billion per year. In India, cotton exports increased from 0.05 million bales in 2002–2003 to 8.5 million bales in 2007–2008, with earnings increasing from US$10.4

Impact of Bt Cotton

101

million in 2002–2003 to US$2.2 billion by 2007–2008. During the same period, cotton imports decreased from 2.5 million to 0.7 million bales. Cotton textile exports also increased in value from US$3.4 billion in 2002– 2003 to US$4.7 billion in 2007–2008 (CCI, 2009). Although partly a result of increased yields, export increases are generally attributed to changes in domestic and international agricultural trade regulations. Some groups lost profits because of Bt cotton and maize adoption. For example, insecticide-producing companies and distributors competing with Bt varieties had reduced profits because of declines in demand for chemical pesticides. In India, aggregate pesticide use on cotton has declined. Traditionally, cotton production has required significant insecticide use. Thus, with reductions in use in the Bt cotton crop, total pesticide use in India has declined from 47,020 MT in 2001–2002 to 37,959 in 2006–2007 (James, 2008). The real value of insecticide use for bollworm management was down from $147 million in 1998 to $65 million in 2006 (ISAAA, 2009). If profit margin is estimated at 20% of sales, this represents $16 million in lost profits. 3.2.3. Impact of seed and royalty price controls and farmers’ benefits5 India presents a unique case study as the first time a government body imposed pricing regulations on Bt cotton hybrid seeds. When Bt cotton hybrids were first approved for release in 2002, MMB held the only Bt genes approved for commercialization in India and firms were required to license technology from MMB. However, in 2006, the government of Andhra Pradesh petitioned the Monopolies and Restrictive Trade Practices Commission (MRTPC) to reduce seed prices. The Commission agreed and MMB appealed to the Supreme Court. Meanwhile, the Andhra Pradesh government negotiated with seed companies to set the price of hybrid Bt cottonseed at US$18 for a 450 g packet, including technology fee. This price was less than half of MMB’s price of US $29 per 450 g packet. Soon, other state governments adopted the same pricing policy, with price caps spreading to important cotton-growing states throughout the country including Maharashtra, Gujarat, Tamil Nadu, Karnataka, Madhya Pradesh, and West Bengal. Even domestic firms with their own Bt events such as Nath Seeds and JK AgriGenetics sell hybrid seeds at the mandated price of US$18 per 450 g packet. MMB currently sells BG II seeds at US$23 per 450 g packet. Price controls, introduced mid-2006, are partially responsible for increased sales of Bt cottonseeds (see Figure 3 and Table 4). Adoption of Bt cotton cultivars soared from 28% of cultivated area in 2005 to 63% in 2006. During the 2007–2008 season, demand for Bt cottonseed packets 5 An in-depth study of how these price controls affect biotech and seed company profits and their incentive to conduct research and innovate is forthcoming in Pray and Nagarajan (2011).

102

Carl E. Pray et al.

almost quadrupled (16 million packets were sold), covering more than 90% of cultivated area under hybrid cotton. Part of this increase in Bt seed sales is also a result of increased availability of locally adapted Bt hybrids that had gone through the regulatory process by 2006. Introduction of BG II and new Bt events from JK AgriGenetics and Nath seeds also increased farmer choice. State governments also now regulate the cottonseed trade by penalizing illegal seed suppliers through heavy fines and punishments to ensure tested and approved varieties reach farmers (Nath Seeds, 2008– 2009; Rasi Seeds, 2010, personal communication). Prices controls increased farmer share of benefits from Bt cotton adoption and reduced share accruing to seed companies and biotech trait providers (see Table 5). The price of Bt seed before price control was three times that of conventional seeds. By 2006–2007, the average Bt seed price declined by 68% (Sadashivappa and Qaim, 2009), resulting in increased seed demand, increased seed sales, and expansion of area under Bt cotton cultivation. Although seed demand and sales increased with price control, companies made less profit per unit sold and less total profit. Qaim (2003) projected US$315 million in Bt cotton surplus gains for India in 2005. Of this, farmers captured two thirds while the rest accrued to biotech and seed firms. Table 6 presents total revenue (net) realized by all stakeholders in the Bt seed value chain (excluding consumers) since Bt’s introduction in 2002. We estimate that farm-level profit share has increased substantially by farmer adoption of Bt cotton. Farm profits make up nearly 85–90% of the total revenue earned by the Bt cotton industry, including technology provider and seed-firm profits.6 Table 5 also shows the dramatic impact of price control on providers of Bt genes and seed companies licensing the gene. Profit share was as high as 28% and 16% for seed firms and MMB prior to 2006–2007. This declined immediately after imposition of price control. Revenue earned by seed firms was especially affected (falling from 27% to 2%), perhaps due to two reasons – a reduction in seed prices by nearly 50–60% combined with increased cost of seed production by 35–40%. Royalties paid before price controls were as high as US$ 40 per packet of Bt cottonseeds in 2002–2003. This was reduced to US$ 9 per packet with the onset of price control and in 2009–2010 royalties went down to as low as US$1 per packet for some Bt cottonseeds. Rao (2008) suggests that short-term benefits from current policies may outweigh potential losses from forgoing technology in the long term, but 6 The fact farmers were the major beneficiaries of Bt cotton suggests that biotechnology has not forced farmers deep into debt for the benefit of the biotech and seed companies and that there is little support for the anti-GM groups have tried to link Bt cotton, with increased debt and ultimately to farmers suicides. This conclusion is further supported by a recent study that shows that debt and suicides are not related to Bt cotton but are caused by a series of other factors (Grue`re et al., 2008).

103

Impact of Bt Cotton

Table 5. Year

2002–2003 2003–2004 2004–2005 2005–2006 2006–2007d 2007–2008 2008–2009 2009–2010

Bt cotton adoption and net revenue realized by farmers, seed firms, and technology providers Bt packets sold

Bt cotton area

Net revenue Bt cotton (Farmersa þFirmsbþMMBc)

Million

Million ha

Million Indian rupees

0.07 0.23 1.30 3.13 4.00 16.00 27.00 30.00

0.05 0.10 0.50 1.30 3.80 6.30 7.60 8.40

474.10 1,110.90 5,975.00 15,113.30 26,762.00 45,550.90 56,761.60 63,434.40

Share of stakeholders in net revenue (%)

Farmers

71.20 60.80 56.50 58.10 95.80 93.40 90.40 89.40

Seed firms

18.20 24.80 27.70 26.90 1.90 3.10 4.20 4.40

Technology provider (MMB) 10.60 14.40 15.80 15.00 2.30 3.60 5.40 6.30

Source: Authors’ calculations based on information provided by industry sources on total number of seed packets sold. a Net revenues for farm households were calculated based on field-based studies conducted by Qaim et al. (2006) in various years. The net revenue assumptions for the years 2002–2003 to 2006–2007 were based on Qaim et al. (2006) farm-level survey results; for 2009–2010, based on Francis-Kanoi 2010. b Net revenue of seed firms ¼ Bt seed sales price – Technology provider trait fee (MMB fee) – cost of seed production (that includes revenue shared with actors in distribution channels) times the number of packets (450 g) sold. The cost of seed production incurred by seed firms assumed indifferent for BG 1 and BG 2. c Technology providers’(MMB) revenue is calculated from their share in total trait value. Of the total revenue, Monsanto shares 50% of revenue with their domestic partner, Mahyco. d Price controls were imposed in three states from the 2006 to 2007 season.

Table 6.

GMO crops in pipeline in China

Crop

Trait/Institution

Rice

Bt – Chinese Academy of Science (Bt/sck rice) Disease-resistant rice – Chinese Academy of Agricultural Sciences (Xa21) Bt rice – Zhejiang University (cry1Ab) Bt – Chinese Agricultural University Herbicide tolerance – Henan Agricultural University Herbicide tolerance – Chinese Academy of Agricultural Sciences Disease and pest resistance – Chinese Academy of Agricultural Sciences

Maize Wheat Soybean Tomato, rapeseed, chili, and cabbage Source: Huang (2010).

104

Carl E. Pray et al.

results are sensitive to assumptions about future technology foregone and the discount rate. Herbicide-tolerant (HT) and drought-tolerant (DT) varieties are examples of cost-saving technologies that may benefit farmers. If a few or all of these varieties are delayed from entering the market, welfare losses to farmers could result. Limiting technology fees by imposing price controls transfers benefit from technology owners to technology consumer (in this case farmers) in the short-term (Rao, 2008), but long-term consequences remain unclear. 4. Potential impact of biotechnology in other crops 4.1. GM crops in pipeline in China As mentioned above, Bt rice and high-phytase maize have been approved for cultivation. Hybrids in which these events are used will be approved by the cultivar registration system (a requirement for all new cultivars whether transgenic or not) within the next two years. With the approval of these crops, a number of new transgenic events are likely to be released. Table 6 lists GMO crops currently in the regulatory pipeline. High-phytase corn will primarily impact the livestock industry by increasing phosphorus and micronutrient availability to the maize plant, decreasing phosphorous pollution of ground water, and decreasing cost of feedstuffs. No studies have yet been published quantifying impact. The CAAS has perhaps the highest global investment in biotech wheat. It is developing a wide range of traits such as resistance to yellow mosaic virus, head scab, powdery mildew, and insect. A wheat line with resistance to yellow mosaic virus is expected to be commercially available by 2015. Henan Agricultural University is also developing sprouting-tolerant wheat to overcome the 20% loss in production due to early sprouting. This will be commercially available by 2012 or 2013 (ISAAA, 2010). 4.1.1. Impact of GM rice In China, GM rice (or any other GM crop) must be grown under farmer conditions in extensive preproduction trials before approved for commercial cultivation. As a result, Huang et al. (2005, 2008) were able to measure impact of Bt rice varieties on yields, pesticide use, and health in two provinces in southern China. Results from these studies provide evidence of the positive impacts of insect-resistant GM rice: increased yields, reductions in pesticide use, and improvements in farmer health. Insectresistant GM rice produced yields 6–9% higher than conventional varieties, with an 80% reduction in pesticide use and a concomitant reduction in adverse health impacts of pesticides (Huang et al., 2005). Huang et al. (2008) expanded on the 2005 analysis to include another year and more sophisticated econometric techniques. To measure

Impact of Bt Cotton

105

economic impact of GM rice on yield and pesticide use in preproduction trials, they conducted from 2002 to 2004 extensive farm-level surveys across 320 households growing two GM hybrids: GM II-Youming 86 and GM Xianyou 63. Sampled respondents consisted of farm households growing non-GM hybrids, either alone or with GM rice for comparison. Data from all sampled households demonstrated that GM rice farmers applied pesticide less than one time per season (0.6 times) as compared to 3.7 times per season by non-GM growers. Reduction in pesticide use also decreased labor among GM growers, with GM growers using only 1.4 labor days per hectare for spraying versus the conventional 10.1 labor days per hectare. The study further estimated that among GM rice growers, the point estimates of yields were higher than those for non-GM rice growers (although not significant at 5%). Adoption of GM rice in preproduction villages showed increased yields of rice by 9–12% compared to control villages. At the individual household level, pesticide reductions were as high as 85 to 90% among GM rice growers. Though yield effects were not large with GM rice, yield variance was significantly reduced. It is important to note that GM rice adoption led to large reductions in pesticide use without diminishing yields. If it is assumed that GM rice would be equally effective across a larger part of China (especially in stem borer infested areas), potential gains to China’s economy could be as large as US$ 4.2 billion (Huang et al., 2004). 4.2. GM crops research in India In India, public-sector institutes are currently conducting biotech research on more than 20 crops, focusing their research on four traits: (1) insect and disease resistance; (2) tolerance to abiotic stresses (drought and cold); (3) saline resistance; and (4) fortification. For example, scientists at the Indian Agricultural Research Institute in New Delhi are pursuing several transgenic wheat and rice projects, seeking drought- and disease-resistant cultivars. In the private sector, more than 35 companies are actively engaged in GM crop research. Mahyco leads the research with more than 10 transgenic lines (Table 7) and the Mahyco Research Center is also engaged in GM wheat cultivar research in collaboration with its multinational company partner, Monsanto. 4.2.1. Bt eggplant Bt eggplant is the first GM food crop in India, the crop closest to being approved for cultivation and commercialization and the crop with the most research on its economic impacts. Traditionally, cotton required more insecticide sprays of any field crop. Eggplant (and Indian chili) is its equivalent among vegetables. Fruit and shoot borer (FSB) alone damages

106

Carl E. Pray et al.

Table 7. Crop

GMO research in India (2006–2009)

Number of firms with own events

Number of firms testing GM cultivars

Public

Private

Eggplant Cabbage Castor Cauliflower Corn Cotton Peanut/pigeon pea Okra Potato Rice

3 0 1 0 0 2 1 0 3 7

3 2 0 2 3 5 0 4 0 9

14 7 – 7 4 37 3 9 – 13

Sorghum Wheat Tomato

1 1 4

0 1 2

4 – 10

Trait focus

Insect resistance Insect resistance Insect resistance Insect resistance Insect resistance/herbicide tolerance Insect resistance/herbicide tolerance Virus resistance/drought tolerance Insect resistance Disease resistance Pest/disease/drought resistance fortified food Insect resistance Biotic and abiotic resistance Pest/disease/drought resistance

Source: Indian GMO Research Information System (IGMORIS) Website, Hosted by Department of Biotechnology, Ministry of Science and Technology, India. Private firms include the research on transgenic rice by two nongovernmental organizations.

by 48–86% of the eggplant fruit and reduces yield by 50–60%.7 Given the success of Bt technology in cotton, the same technology has been adapted for Bt eggplant. Mahyco, a leading Indian seed company, developed Bt eggplant by inserting the same Cry1Ac used in Bt cotton. The Bt eggplant event EE1 was developed through a public–private partnership under the aegis of Cornell University’s USAID-sponsored Agriculture Biotechnology Support Project. The Bt technology available with Mahyco has been transferred (free of cost) to Tamil Nadu Agriculture University (Coimbatore), the University of Agricultural Sciences (Dharwad), and the Indian Institute of Vegetable Research (Varanasi). Mishra (2003) estimated potential welfare benefits from Bt eggplant adoption in India at US$ 422 million, with consumers gaining 57% of these benefits. Mahyco conducted multilocal field trials during the 2004– 2005 cropping season to compare Bt eggplant hybrids with non-Bt in different agro-climatic regions. Their results suggest a 45% reduction in insecticide usage on the Bt plots (2.82 kg/acre).The major impact of Bt technology was in yield, with the mean yield of Bt eggplant at 2.2 t per acre compared to non-Bt hybrids at 1.02 t per acre (Krishna and Qaim, 2007). 7 Some studies estimate that FSB is responsible for losses of up to 60–70%. See Jeyanthi and Kombairaju, 2005; Kolady and Lesser, 2006.

Impact of Bt Cotton

107

Using contingent valuation techniques, Krishna and Qaim (2007, 2008) report net benefits of US$ 370 to US$440 per acre from Bt eggplant hybrid adoption. Cornell University’s Agricultural Biotechnology Support Project II (ABSP II) estimates that Bt eggplant offers resource poor farmers in India significant benefit, including:  45% reduction in insecticide sprays, with implications for human health and the environment production costs  117% increase in yield with implications for more affordable vegetables for consumers  US$411 million per annum increase in net benefits to Indian eggplant farmers and consumers at the national level (ABSP II, 2007; James, 2007). Further, Krishna and Qaim (2008) project that Bt eggplant in India may produce farmer health benefits worth approximately $4 billion per year. An ex-ante assessment conducted by Ramasamy et al. (2007) also estimated net benefits from Bt eggplant cultivation by Indian farmers and consumers in the range of US$25–142 million per annum assuming only 10% adoption of Bt eggplant in the first year of commercialization. In spite of encouraging results from various farm-level trials conducted by both public and private firms and in spite of approval in October 2009 after lengthy review by GEAC, which includes experts from the Ministry of Environment, the Ministry of Environment put approval for Bt eggplant on hold pending further consultations. Losers from this decision include small vegetable farmers who could reduce production costs and their exposure to pesticides, consumers who also would also reduce exposure to pesticide residue in the eggplants they consume, and farmers and seed and biotech companies hoping this would open the door for GM food crops. Winners are pesticide companies, anti-GM groups, and consumers who fear biotechnology more than they fear pesticides. This decision also reflects the political clout of various players. Vegetable farmers have little political clout. Some may worry that adoption of GM eggplant would reduce consumer demand. Commercial farmers and seed companies do have political clout, explaining why the Ministry of Agriculture publicly argued in favor of allowing commercialization of Bt eggplant. The pesticide industry and anti-GM groups are very well organized. Urban consumers, who may be more aware of potential problems from pesticide use, are inundated with Indian newspaper reports about potential problems of GM food crops. However, little is reported about pesticide residues in vegetables. The pesticide industry influences events quietly while anti-GM groups work noisily. The Ministry of the Environment and Forests is the final biosafety regulatory authority and seems to have decided it is

108

Carl E. Pray et al.

more politically expedient to side with anti-GM groups and urban consumers than with small farmers. However, the debate continues.

5. Conclusions The earliest studies of Bt cotton impacts were conducted in China using data collected from a random sampling of farmers in northern China beginning in 1999. Using means comparisons and econometric analysis, these studies found small increases in yields, major reductions in pesticide use, and increased profits for farmers adopting Bt cotton. In addition, Bt cotton adoption led to reduced pesticide use, which resulted in farmers reporting fewer pesticide-related- illnesses. In the early years, farmers captured almost the entire economic surplus from Bt cotton adoption. Suppliers of genetic traits and seeds made limited amounts of money because seeds were not hybrids and were quickly copied by farmers and other seed companies. Also in these years, limited benefits were transferred to consumers through lower cotton prices because the Government of China procured most of the cotton crop at a government-established price. In India, results of Bt adoption were different. Introduction of insect resistance had a significant impact on yields, with increases of 40–80% as farmers in India did not have good pest control available to them. Reduction in pesticide use for bollworm control was also substantial but less than in China. Like Chinese farmers, Indian farmers increased their net incomes despite higher seed prices. Indian seed and biotech firms had more ways to appropriate benefit from the technology embodied in the seed than did Chinese companies. Indian farmers typically use hybrid seed and, until 2006, the Indian government only permitted one company to supply a Bt gene. However, farmers in India captured two-thirds of the social surplus generated by Bt cotton adoption even in the early years before price controls were mandated. Perhaps this chapter’s most important contribution is new evidence presented on recent changes in benefits from Bt cotton adoption. In China, CCAP economists have found that pesticide use for bollworm in Bt cotton has continued to decline up to 2007 when their last study was conducted. This is consistent with findings by entomologists (Wu et al., 2008) that the bollworm population in all crops has declined because of Bt cotton. This suggests positive externalities for other crops such as maize and vegetables that had been sprayed extensively for bollworm but now have less damage and require fewer sprays. As yet, no outbreaks of Bt-resistant bollworms have been reported in China. CCAP economists have also found that in some villages a minor pest, mirids, has become an increasing problem since Bt cotton was introduced, seemingly due to the decline in broad spectrum pesticides

Impact of Bt Cotton

109

previously used to control bollworms (Lu et al., 2010). The benefits from reducing pesticide sprays for bollworm outweigh costs of increased spraying for mirids.8 Chinese farmers rather than biotech or seed companies continue to be Bt cotton’s main beneficiary as seed prices remain low because IPR enforcement is still weak and most seed used is varietal, not hybrid. Indian farmers now obtain a greater share of benefits from Bt cotton. State government policies increased farmer benefit at the expense of the seed and biotechnology industry. In both India and China, Bt cotton has spread to all areas where bollworm is a major pest, in India about 90% of the cotton area and in China about 70–80%. The area under Bt cotton is likely to remain the same until new superior traits are introduced. Thus, the development and commercialization of new GM crops is the most likely avenue for increased benefit from crop biotechnology in the near future. The approval of Bt rice in China and evidence of its efficacy in controlling borers and reducing pesticide suggests it will contribute significantly to China’s growth. Bt eggplant also has potential to transform vegetable production in India and elsewhere. However, it is now in regulatory limbo and may not be soon cultivated. Economists and plant scientists must continue to measure impact of Bt cotton to potentially identify ways to use Bt more effectively and to reduce further the use of chemical pesticides. For example, Arizona is using Bt and other forms of pest control in a coordinated program to eradicate pink bollworm (NCC, 2001). Continued research can also identify new problems farmers face from changes in pests and weather. In addition, comparative studies of the impacts on health and the environment of GM and chemical pesticides could be useful for decision makers. Finally, studies of the new GM traits and crops in regulatory trials or have recently adopted by farmers could help farmers and governments determine that traits best fit farmer needs.

Acknowledgments We would like to thank the two anonymous reviewers for their helpful comments. We would like to thank Ms. Judith Killen, for helping us to edit and revise the manuscript. Funding support from the Bill and Melinda Gates Foundation (BMGF) and Economic Research Service at USDA is gratefully acknowledged. 8 In India, there have been reports that pink bollworms resistant to Bollgard I were found in a small area of Gujarat. However this report has not discussed the economic damage caused by such attack in the affected areas. Pink bollworm is not the major bollworm pest in Gujarat or elsewhere in India, and Monsanto reports that Bollgard II controls pink bollworm.

110

Carl E. Pray et al.

References ABSP-II (Agricultural Biotechnology Support Project –II) (2007), Bt Eggplant Insect Resistance Management Strategy for the Eggplant Fruit and Shoot Borer. Cornell University, Ithaca, NY, Available at www.sathguru.com/absp2/irmstrategy.pdf. Anderson, K., Valenzuela, E., Jackson, L.A. (2008), Recent and prospective adoption of genetically modified cotton: a global computable general equilibrium analysis of economic impacts. Economic Development and Cultural Change 56, 265–296. ASSOCHAM – Association of Chambers of Commerce and Industry of India. (2007), Economic benefits of Bt cotton cultivation in India. In: Bt Cotton Farming in India. New Delhi, India. Available at http:// monsanto.mediaroom.com/index.php?s¼43&item¼508 Bambawale, O.M., Singh, A., Sharma, O.P., Bhosle, B.B., Lavekar, R.C., Dhandapani, A., Kanwar, V., Tanwar, R.K., Rathod, K.S., Patange, N.R., Pawar, V.M. (2004), Performance of Bt cotton (MECH-162) under integrated pest management in farmers’ participatory field trial in Nanded district, Central India. Current Science 86 (12), 1628–1633. Barwale, R.B., Gadwal, V.R., Zehr, U., Zehr, B. (2004), Prospects for Bt cotton technology in India. AgBioForum 7, 23–26. Bennett, R.M., Ismael, Y., Kambhampati, U., Morse, S. (2004), Economic impact of genetically modified cotton in India. AgBioForum 7, 96–100. Bennett, R.M., Ismael, Y., Morse, S. (2005), Explaining contradictory evidence regarding impacts of genetically modified crops in developing countries: varietal performance of transgenic cotton in India. Journal of Agricultural Science 143, 35–41. Bennett, R., Kambhampati, U., Morse, S., Ismail, Y. (2006), Farm-level economic performance of genetically modified cotton in Maharashtra, India. Review of Agricultural Economics 28, 50–71. CCI (Cotton Corporation of India). (2009), National Cotton Scenario, Ministry of Textile. Available at http://www.cotcorp.gov.in/national_ cotton.asp Choudhary, B., Gaur, K. (2009), Agri-biotech in India: a new surge. Biotech News 4 (2), 30–33, Department of Biotechnology, Ministry of Science and Technology, Government of India. DST – Department of Science and Technology Annual Report. (2008), Ministry of Science and Technology, Technology Bhavan, New Delhi. Francis Kanoi Agri-Inputs Marketing Research Data Bank. (2010), Cotton crop track: a syndicated report on seeds. Francis-Kanoi Marketing Research, Chennai, India. Frisvold, G.B., Reeves, J.M. (2007), Economy-wide impacts of Bt Cotton. Proceedings of the Beltwide Cotton Conference, January. Gandhi, V., Namboodiri, N.V. (2006), The adoption and economics of Bt cotton in India: preliminary results from a study. Indian Institute of

Impact of Bt Cotton

111

Management (IIM), Ahmedabad, India. Working Paper No. 200609-04. GEAC – Genetic Engineering Approval Committee. (2009), Available at http://www.envfor.nic.in/divisions/csurv/geac/decision-jan-91.htm. Hossain, F., Pray, C.E., Lu, Y., Huang, J., Fan, C., Hu, R. (2004), GM cotton and farmer’s health in China: an econometric analysis of the relationship between pesticide poisoning and GM cotton use in China. International Journal of Occupational and Environmental Health 10 (3), 307–314, Available at http://ipts.jrc.ec.europa.eu/publications/pub.cfm? id¼2199. Hu, R., Pray, C.E., Huang, J., Rozelle, S., Fan, C., Zhang, C. (2009), Reforming intellectual property rights and the Bt cotton seed industry in China: who benefits from policy reform?. Research Policy 38, 793–801. Huang, J. (2010), Biotech and approval of GM rice and maize in China and its implications. Presented at the 14th ICABR conference on bioeconomy governance: policy, environmental and health regulations, and public investments in research (June 16–18), Ravello, Italy. Huang, J., Hu, R., Rozelle, S., Pray, C.E. (2005), Insect-resistant GM rice in farmers’ fields: assessing productivity and health effects in China. Science 308, 688–690. Huang, J., Hu, R., Rozelle, S., Pray, C.E. (2008), Genetically modified rice, yields, and pesticides: assessing farm-level productivity effects in China. Economic Development and Cultural Change 56, 241–263. Huang, J., Hu, R., Rozelle, S., Qiao, F., Pray, C.E. (2002), Transgenic varieties and productivity of smallholder cotton farmers in China. Australian Journal of Agricultural and Resource Economics 46 (3), 367–387. Huang, J., Hu, R., van Meijl, H., van Tongeren, F. (2004), Biotechnology boosts to crop productivity in China: trade and welfare implications. Journal of Development Economics 75 (1), 27–54. Huang, J., Mi, J.W., Lin, H., Wang, Z., Chen, R., Hu, R., Rozelle, S., Pray, C.E. (2010), A decade of Bt cotton in Chinese fields: assessing the direct effects and indirect externalities of Bt cotton adoption in China. Science China Life Sciences 53, 981–991. doi: 10.1007/s11427010-4036-y. Huang, J., Wang, Q. (2002), Agricultural biotechnology development and policy in China. AgBioForum 5 (4), 122–135. Grue`re, G.P., Mehta-Bhatt, P., Sengupta, D. (2008), Bt cotton and farmer suicides in India. IFPRI Discussion Paper 00808. Environment and Production Technology Division, International Food Policy Research Institute, Washington, DC. IGMORIS – Indian GMO Research Information Service. (2010), Available at http://igmoris.nic.in/commercial_approved.asp

112

Carl E. Pray et al.

Indicus Analytics. (2007), Socio-economic appraisal of Bt cotton cultivation in India. Indicus Analytics Study. International Service for the Acquisition of Agribiotech Applications (ISAAA). (2009). Global status of commercialized biotech/GM crops: The first fourteen years, 1996 to 2009. ISAAA – International Service for the Acquisition of Agri-Biotech Applications. (2010), Biotech wheat. Brief No. 38, August. SEAsiaCenter, Metro Manila, Philippines. James, C. (2007), Global status of commercialized biotech/GM crops. ISAAA Brief No. 37. International Service for the Acquisition of AgriBiotech Applications, Ithaca, NY. James, C. (2008), Global status of commercialized biotech/GM crops. ISAAA Brief No. 39. International Service for the Acquisition of AgriBiotech Applications, Ithaca, NY. Jeyanthi, H., Kombairaju, S. (2005), Pesticide use in vegetable crops: frequency, intensity and determinant factors. Agricultural Economics Research Review 18, 209–221. Kambhampati, U., Morse, S., Bennett, R., Ismael, Y. (2005), Perceptions of the impacts of genetically modified cotton varieties: a case study of the cotton industry in Gujarat, India. AgBioForum 8 (2&3), 161–171. Kolady, D.E., Lesser, W. (2006), Who adopts what kind of technologies? The case of Bt eggplant in India. AgBioForum 9 (2), 94–103. Krishna, V.V., Qaim, M. (2007), Estimating the adoption of Bt eggplant in India: Who Benefits from public-private partnership?. Food Policy 32 (5&6), 523–543. Krishna, V.V., Qaim, M. (2008), Potential impacts of Bt eggplant on economic surplus and farmers’ health in India. Agricultural Economics 38 (2), 167–180. Lu, Y., Wu, K., Jiang, Y., Xia, B., Li, P., Feng, H., Wyckhuys, K.A.G., Guo, Y. (2010), Mirid bug outbreaks in multiple crops correlated with wide-scale adoption of Bt cotton in China. Science 328, 1151–1153. Mishra, S. (2003), An ex-ante economic impact assessment of Bt eggplant in Bangladesh, the Philippines and India. Unpublished Master’s Thesis, Virginia Tech University, USA. Murugkar, M., Ramaswami, B., Shelar, M. (2007), Competition and monopoly in the Indian cotton seed market. Economic and Political Weekly 62 (37), 3781–3789. Natesh, S., Bhan, M.K. (2009), Biotechnology sector in India: strengths, limitations, remedies and outlook. Current Science 97 (2), 157–169. NCC – National Cotton Council. (2001), Pink bollworm eradication: a window of opportunity. Available at http://www.cotton.org/tech/ pest/bollworm/loader.cfm?csModule¼security/getfile&pageid¼10771. Accessed in January, 2001.

Impact of Bt Cotton

113

Pemsl, D. (2006), Economics of agricultural biotechnology in crop protection in developing countries – the case of Bt-cotton in Shandong Province, China. Pesticide Policy Project Publication Series, Special Issue No. 11. University of Hannover, Germany. Pemsl, D., Waibel, H. (2007), Assessing the profitability of different crop protection strategies in cotton: case study results from Shandong Province, China. Agricultural Systems 95 (1-3), 28–36. Pemsl, D., Waibel, H., Gutierrez, A.P. (2005), Why do some Bt-cotton farmers in China continue to use high levels of pesticides?. International Journal of Agricultural Sustainability 3 (1), 44–56. Pray, C.E., Bengali, P., Ramaswami, B. (2005), The cost of regulation: the India experience. Quarterly Journal of International Agriculture 44 (3), 267–289. Pray, C.E., Nagarajan, L. (2011), Price controls and biotechnology innovation: are state government policies reducing research and innovation by the ag biotech industry in India? Agbioforum, 13 (4), 297–307 Pray, C.E., Huang, J., Ma, D., Qiao, F. (2001), Impact of Bt cotton in China. World Development 29, 813–825. Qaim, M. (2003), Bt cotton in India: field trial results and economic projections. World Development 31 (12), 2115–2127. Qaim, M., Subramanian, A., Naik, G., Zilberman, D. (2006), Adoption of Bt cotton and impact variability: insights from India. Review of Agricultural Economics 28 (1), 48–58. Qaim, M., Zilberman, D. (2003), Yield effects of genetically modified crops in developing countries. Science 299, 900–902. Ramasamy, C, Selvaraju, K.N., Norton, G.W., Vijayaraghavan, K. (2007), Economic and environmental benefits and costs of transgenic crops: ex-ante assessment. Working Paper –CARDS, Department of Agricultural Economics, Tamil Nadu Agricultural University, Coimbatore, India. Rao, S.S. (2008), Economic impact of price controls and regulations: case of GM cotton in India. Unpublished MS Thesis, Department of Agricultural, Food and Resource Economics, Rutgers, The State University of New Jersey, New Brunswick, NJ. Rengasamy, N., Elumalai, K. (2009), Market opportunities and challenges for agri-biotech products in India. Agricultural Economics Research Review 22, 471–481. Sadashivappa, P., Qaim, M. (2009), Bt cotton in India: development of benefits and the role of government seed price interventions. AgBioForum 12 (2), 172–183. Singh, S.K. (2007), India: cotton and products annual 2007. GAIN Report Number IN7041. United States Department of Agriculture (USDA), Washington, DC.

114

Carl E. Pray et al.

Subramanian, A., Qaim, M. (2009), Village-wide effects of agricultural biotechnology: the case of Bt cotton in India. World Development 37, 256–267. Vaidya, A. (2005), Monsanto’s cotton has deficiencies: study. The Times of India. Available at http://timesofindia.indiatimes.com/city/pune/ Monsantos-cotton-has-deficiencies-study/articleshow/1132562.cms. Wang, G., Wu, Y., Gao, W., Fok, M., Liang, W. (2008), Impact of Bt cotton on the farmer’s livelihood system in China. International Cotton Conference, Rationales and evolutions of cotton policies in main producing countries. ISSCRI International Conference (May 13–17), Montpellier, France. Wu, K.M., Lu, Y.H., Feng, H.Q., Jiang, Y., Zhao, Z.J. (2008), Suppression of cotton bollworm in multiple crops in China in areas with Bt toxin-containing cotton. Science 321 (5896), 1676–1678. Yang, P., Iles, M., Yan, S., Joliffe, F. (2005), Farmers’ knowledge, perceptions and practices in transgenic Bt cotton in small producer systems in Northern China. Crop Protection 24 (3), 229–239.

CHAPTER 5

Contributions of Public and Private R&D to Biotechnology Innovation$ Wallace E. Huffman Department of Economics, Iowa State University, Ames, IA 50011, USA E-mail address: [email protected]

Abstract Purpose – The objective of this chapter is to examine and provide new perspectives on the contributions of public and private R&D to biotech crop improvement. Methodology/approach – The chapter examines a set of topics that have affected the way that research is undertaken on plant germplasm improvement and how it has changed with the genetically modified (GM) trait revolution. Findings – Although the basic science providing the foundations for GM crops was undertaken in the public sector, GM traits and GM crop varieties have been developed almost exclusively by the private sector. The biotech events leading to GM traits are currently being developed largely by five companies – all having ties to both the chemical and the seed industries. The GM crop revolution started in North American in 1996 and has spread slowly to the largest developing countries that have large agricultural sectors, including Argentina, China, Brazil, and India, but not to Europe or Japan. Practical implication – To shed new light on the economic reasons for private sector dominance in GM crop varietal development in selected crops but not in others. Social implication – Shows how GM traits have contributed to technical change and declining real food prices. Keywords: Crop biotechnology, R&D, public, private, corn, soybean, cotton, genetic modification, funding JEL Classifications: O3, Q16, Q10

$

The author is C.F. Curtiss Distinguished Professor of Agriculture and Life Sciences and Professor of Economics, Iowa State University. The project was supported by the Iowa Agriculture and Home Economics Experiment Station.

Frontiers of Economics and Globalization Volume 10 ISSN: 1574-8715 DOI: 10.1108/S1574-8715(2011)0000010010

r 2011 by Emerald Group Publishing Limited. All rights reserved

116

Wallace E Huffman

1. Introduction Growth in population and per capita incomes more than doubled the demand for food worldwide during the second half of the 20th century, and this may double again by the mid-21st century (Rainey, 2004). Significant increases in scientific and technical effort are needed in the world’s poorest countries to transition to sustainable agricultural production. The public sector agricultural experiment stations emerged in the late 19th century as an institution for advancing scientific knowledge and technology for agricultural production (Huffman and Evenson, 2006). In contrast, developed countries in the late 20th century have witnessed a dramatic increase in private sector agricultural research capacity. The invention of hybrid corn and the protection of intellectual property (IP) provided new opportunities for private investment in the development and marketing of seed and plant breeding research (Griliches, 1960; Fernandez-Cornejo, 2004; Huffman and Evenson, 2006).1 Advances in self-pollinated crops generated a demand for institutional innovation to protect IP of inventors and scientists, first in the form of breeders’ rights and more recently in the form of patent protection for transgenic ally modified organisms. By the end of the century, the private sector emerged as the primary supplier of new field crop technologies to farmers in developed countries (Fernandez-Cornejo, 2004; Huffman and Evenson, 2006). A recent OECD report (OECD, 2009) provides an assessment of future prospects for crop yield improvement in corn, soybean, wheat, and potato in developed countries. Advances in modern biotechnology have opened up dramatic new possibilities for agricultural biotechnology. The first patent on a living organism was issued in 1980 for genetically engineered bacterium; yet, the number of agricultural biotechnology patents issued by the US Patent Office before 1985 was less than 200 per year across all agricultural fields (see Figure 1). The US Patent Office extended patents to plants in 1985. Agricultural biotechnology patents then increased, from 500 in 1994, to 1,600 in 1999. The rate peaked in 2001 at 1,800 before declining over the next four years. The issue rate of agricultural biotechnology patents increased in 2006, but declined again in 2007 (Figure 1). In 1996, GM cotton and corn with insect-resistant (IR) Bt genes were first marketed to US farmers, and the first commercial soybean and cotton varieties were planted with herbicide tolerant (HT) genes. Subsequently, HT canola was marketed to farmers in Canada and HT soybean and Bt corn in Argentina. GM crops have since been marketed to farmers in Brazil, China, India, South Africa, and Australia. Genetically engineered or modified plants have emerged as a major new technology for pest control by farmers.

1 It became necessary to seek patent protection on hybrid corn varieties when GM traits were incorporated into them. The protection is on the trait(s).

Contributions of Public and Private R&D to Biotechnology Innovation

117

2000 1800 1600 1400

No. of ABP US ABP

1200 1000 800 600 400 200 0 1976 1978 1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002 2004 2006

Fig. 1. The number of agricultural biotechnology patents issued, 1976–2007. Notes: ABP, number of agricultural biotechnology patents issued by the US Patent Office to all assignees; US ABP, number of agricultural biotechnology patents issued to assignees who are registered in the United States. Source: Zhao (2011) In 2009, 330 million acres of GM crops were grown by farmers worldwide (James, 2010). In roughly half a century, science advanced from the first knowledge of the structure of deoxyribonucleic acid (DNA) to successful commercial applications of genetic engineering to plants (and animals). The objective of this chapter is to examine and provide new perspectives on the contributions of public and private R&D to biotech innovations in crop varieties. This is accomplished by examining a set of topics that have affected the way that research is undertaken on plant germplasm improvement and how it has changed with the GM trait revolution. The GM revolution started in North America and spread slowly to the largest developing countries but not to Europe or Japan.2 The chapter unfolds in the nine following sections. 2. Some background Commercial development and sale of seed to farmers emerged in the 20th century. However, sales were small until the tools and scientific knowledge advanced sufficiently to make yield improvement in crop varieties. Hybrid seed corn, first marked in the United States in the late 1920s, made annual seed sales dependable for seed companies because a hybrid cannot reproduce itself, thereby eliminating farmer-saved seed. By 1934, hybrid seed corn companies in Iowa had developed double-cross hybrids that were adapted to the area and had higher expected yields of 10%–15% than the open pollinated varieties that they replaced. Griliches (1960) showed that it took only four years for farm-level adoption in Iowa to increase by 2 For more information on innovating through science and technology, including biotechnology, for poor countries, see reports by FAO (Rainey, 2004) and the World Bank (2008, pp. 158–179).

118

Wallace E Huffman

80 percentage points. However, as one moved farther from the center of the Corn Belt, adapted hybrids were made available later, the expected yield gain was lower, and farmers planted fewer acres of corn per farm, which greatly reduced profitability of hybrids to farmers on the fringes of the Corn Belt (Griliches, 1957). Hence, adoption started later and proceeded at a slower pace outside of the Corn Belt. Subsequently, hybrid corn was transferred to South Africa and Argentina. As late as 1970, 61% of inbred lines used in US commercial hybrid seed corn varieties of private seed companies were from the public sector. However, over the next 15 years, the direct use of public in-bred lines in commercial hybrids declined dramatically to only 22%, with most of the decline occurring over 1979–1984 (Huffman and Evenson, 1993, p. 160). When public universities withdrew from inbred line development, many small hybrid seed corn companies were left with no source of good inbred corn lines. These companies then pursued primarily two sources of inbreds. First, new ‘‘foundation’’ seed companies emerged. These companies, such as Holden Foundation Seeds, specialized in germplasm collection and inbred line development and placed these lines for sale. Second, some small companies established a small breeding program around one or two corn breeders. Although the foundation seed companies provided a lifeline for small hybrid seed corn companies, at least one foundation seed company used unethical practices to obtain enriched germplasm for its inbred line inventory. Holden Foundation seeds was sued by Pioneer Hi-Bred for pirating inbred lines from their fields and road spillage and thereby violating a trade secret of Pioneer’s (Pioneer Hi-Bred International v. Holden Foundation Seeds Inc., 1994), and Pioneer was awarded $45 million by the federal courts. By 1980, total US farm seed expenditures reached $8.637 billion (in 2009 dollars), which represented 2.7% of US total farm expenditures. Real expenditures were relatively unchanged over 1980–1995, but seed expenditures as a share of total farm expenditures increased to 3.0% in 1990 and 3.25% in 1995. Over 1995–2009, total seed expenditures increased steadily to $15.50 billion, a compound rate of increase of 4.9% per year. Likewise, the share of seed expenditures in total farm expenditures increased to 5.4%. Hence, during the early era of GM field crops sales, seed expenditures and the share of seed expenditures in total farm expenditures have increased significantly. 3. Important scientific discoveries providing the foundation for GM crops The general concept of heredity, containing the concepts of nucleic acid molecules organized into functional units called genes, was developed by the end of the 19th century. DNA was seen as the organic substance of heredity and the material of which genes was composed, providing the chemical information to direct the synthesis of cell proteins. Giants in the development of the foundations to modern genetics include Charles Darwin and Gregory Mendel (Table 1). Darwin developed the theory of natural selection, a process

Contributions of Public and Private R&D to Biotechnology Innovation

Table 1.

119

Important scientific and other events providing the foundation for GM crops

Year

Discoverer-innovator and/or major event

1859

Charles Darwin, British scientist and early evolutionary biologist, published On the Origin of the Species, establishing the theory of natural selection (Grace 1997, p. 6–9). Gregor Mendel, an Austria-Hungarian Catholic monk and scientist, published Experiments with Plant Hybrids, which outlined the principles of heredity (Alcamo 2001, p. 2–6). Johnann Miescher, Swiss-German scientists with an advanced degree, makes first chemical analysis of nucleic acid and related studies (Alcamo 2001, 9–11). Archibald Garrod, British physician, early biochemist and later Professor of Medicine at Oxford University, speculates that genes consist of instructions for making proteins (Alcamo 2001, pp. 42–43) Thomas Hunt Morgan, American, Ph.D. Johns Hopkins, geneticist, professor at Columbia University and Cal Tech, establishes that genes are located on chromosomes and are the mechanical basis of heredity. Discoveries were basis for modern genetics. Nobel Prize in Medicine, 1933 (Alcamo 2001, p. 8) George Beadle, American, Ph.D. Cornell, geneticist and Professor at Stanford and Cal Tech, and Edward Tatum, American, Ph.D., University of Wisconsin, biochemist at Stanford and Yale University, established the relationship between genes and enzymes, discovered that genes encode proteins or enzymes, and showed that that one gene makes one enzyme, proving the ‘‘one-gene-one-enzyme’’ hypothesis. Nobel Prize in Medicine 1958 (Alcamo, 2001, pp. 43–44) Oswald Avery, Canadian born American Physician, MD degree, one of first molecular biologists, discovered that DNA is the material of which genes and chromosomes are made – proving that Griffith’s ‘‘transformation principle’’ is DNA (Alcamo 2001, pp. 12–15) James Watson, American, Ph.D. genetics, University of Indiana, and Francis Crick, British, advanced training in physics and biology, molecular biologist, correctly formulated the structure of the DNA molecule as a double helix, and shortly thereafter showed how it replicates. Nobel Prize in Medicine 1962 (Alcamo, 2001, pp. 31–32) Plant Variety Protection Act of 1970 provided intellectual property protection to sexually reproducible plants, extended to F1 hybrids and tuber propagated plants in 1994 and to soup vegetables in 2001. Variety must be stable in successive generations. Developer has 20 years of protection. Farmers can save own seed for planting, but no resale (Huffman and Evenson, 2006) Stanley Cohen, American, MD, University of Penn, geneticist, Professor at Stanford, and Herbert Boyer, Ph.D., University Pittsburg, biochemist, University of California-San Francisco, discovered recombinant DNA, a method by which genetic material could be cut into small pieces and inserted into another species. This was the start of genetic engineering (Alcamo, 2001, pp. 84–89) Stanford University applied for a patent on the Cohen and Boyer gene splicing technique, which made moving genes from one species to another relatively easy. Patent expired Dec 1997, with total licensing revenue of $255 million (Alcamo, 2001, pp. 84–89) US Court of Customs and Patent Appeals approves patent on a genetically engineered bacterium – the first patent for a living (and transgenic) organism (Diamond v. Chakrabarty) US Patent and Trademark Office decided in Ex parte Hibberd (1985) that patent protection extends to plants, including hybrids and other plants. Use of biotech methods facilitated proof of novelty

1866

1869 1902

1910

1941

1944

1953

1970

1973

1980

1980

1985

120

Wallace E Huffman

where heritable traits become more frequent in a population over successive generations. In some cases, a new, specialized population emerged and is designated a new species. Gregory Mendel was the first to discover the principles of heredity by undertaking experiments on garden peas and conducted statistical analyses that revealed the patterns of inheritance. Mendel’s greatest contribution was the discovery of a predictable mechanism by which inherited characteristics move from parents to offspring. However, Mendel’s findings were far ahead of contemporary scholars and were overlooked for decades (Alcamo, 2001, p. 6). Research by Miescher in 1869 and Garrod in 1902 provided the first chemical analysis of nucleic acid and the proposition that genes consist of instructions for making proteins, respectively. In 1919, Thomas Morgan established that genes are located on chromosomes and that they are the mechanical basis of heredity, which is the basis of modern genetics (Table 1). George Beatle and Edward Tatum established the relationship between genes and enzymes or proteins and provided evidence for the ‘‘one-gene-one-enzyme/protein’’ hypothesis. By the mid-1940s, research on DNA was progressing steadily. Avery in 1944 discovered that DNA is in fact the material of which genes and chromosomes are made, and James Watson and Francis Crick made the landmark discovery (1953) of the structure of the DNA molecule – a double helix – and shortly thereafter showed how it replicates (Alcamo, 2001). Hence, genes on chromosomes provide the genetic instructions – blueprints – used in the development and functioning of all living organisms, except for a few special viruses. In the 1970s, scientists were searching for a method for transferring genes across species. Stanley Cohen and Herbert Boyer discovered recombinant DNA (1973), a method by which genetic material could be cut into small pieces and inserted into another species. This technique provided the basis for transgenic GM organisms and the GM revolution that followed (Alcamo, 2001). In 1980, Stanford University applied for a patent on the Cohen–Boyer gene splicing technique (Table 1).3 4. New intellectual property rights that facilitated innovation in plants The issuance of intellectual property rights (IPRs) is primarily a prerogative of national governments, and these rights hold only within its borders. However, there are some important international agreements dealing with IP in plants. One of the oldest international IPR agreements, the Paris Convention for the Protection of Industrial Property of 1883, seeks to harmonize patent regimes among signatory countries. The convention 3 The patent generated $255 million in patent licensing revenue by the time that the patent expired in 1997.

Contributions of Public and Private R&D to Biotechnology Innovation

121

provides members only limited property rights protection for innovation of plant varieties and biological processes for plant production. The International Convention for the Protection of New Varieties of Plants was adopted in Paris in 1961. This Convention established the International Union for the Protection of New Varieties of Plants (UPOV), and its objective was to provide IP rights to breeders of new plants. It provided protection to breeders of new plant varieties who belong to member countries. These ‘‘breeder’s rights’’ were amended in 1972, 1978, and 1991. The 1991 revision expanded protection to address new issues in agricultural biotechnology. In the United States, the UPOV Convention was implemented by the Plant Variety Protection Act of 1970. A Plant Variety Protection Certificates (PVPCs) is awarded to breeders who develop new sexually propagated plants. The PVPC for a crop variety gives the innovator the right to reproduce, sell commercially, or license his variety for a fixed number of years – initially 18, and then in 1994, the protection was extended to 20 years. To be eligible for a PVPC, a variety must be new and distinct from other varieties (novel), genetically uniform, and stable through successive generations.4 The applicant, however, is not required to disclose the scientific nature of his IP or his innovation. The protected variety is available for research uses, that is, the study of the nature of the innovation, and for the development of new varieties, and farmers who plant a protected variety can save part of their harvested seed for their own later plantings. However, they cannot re-sale seed of protected varieties within UPOV countries. See Huffman and Evenson (2006) and Fernandez-Cornejo (2006). In 1980, the US Court of Customs and Patent Appeals (in Diamond v. Chakrabarty) approved a patent on a genetically engineered bacterium. This was the first patent of a transgenic organism. Then, in the US Patent Office decision of ex parte Hibberd, patenting was extended to plants. The use of biotech methods is an aid in proving novelty of traits/products seeking a patent.5 All developed countries have IPRs in the form of plant breeders’ rights and patents and are members of the UPOV. Hence, plant breeders in these countries have strong protection of their rights to innovations. However, lowand middle-income countries do not have such rights nor are they members of the UPOV. Hence, when new plant varieties are developed in high-income countries and moved to low- or middle-income countries, the inventor generally has no IPR protection there. Without IP protection, innovators have little incentive to make new varieties available in these countries. 4 To be eligible for a PVPC, a variety must be new; not sold or propagated for more than one year. 5 Patenting of plants and plant varieties is allowed in the United States, Canada, Japan, and a few other countries. However, laboratory tools required to genetically engineer plants, including techniques to insert genes, are widely patentable in developed countries (UNCTA, 2006).

122

Wallace E Huffman

What is the record on PVPC issued in the United States? Over 1971–1986, 805 PVPCs, or an average of 50 per year, were issued on field crop varieties to breeders; 146 on corn and 503 on soybean varieties. Over 1987–1998, there were 1,326 PVPCs issued on all field crops, or an average of 110 per year; 383 on corn and 428 on soybean varieties. A larger share of these PVPCs were issued to private companies and organizations (Huffman and Evenson, 1993, p. 145, 2006, p. 163). Hence, many PVPCs were issued on field crops before the first plant patent was issued. 5. Innovation in GM traits for major crops Cotton, corn, and soybeans are three crops that the private sector targeted for GM trait development. Cotton is a crop where the boll and budworm complex have been difficult to control with commercial pesticides, and resistance had grown to the existing commercial pesticides that were also toxic to the environment and field workers. Soybeans are not competitive against weeds, and weed control with conventional herbicides and hand weeding were only partially successful and expensive. In corn, infestations of insects that target the stock and roots were uncertain events, but when they occurred, they caused significant yield reductions. Finding biotech solutions to these problems were difficult. A high priority of the crop biotechnology industry was the development of a new and effective biological insect control or a trait that when inserted into a plant conferred IR against harmful and difficult-to-control insects. The industry turned to bacteria that occur naturally in the soil, to Bacillus thuringiensis (Bt, 2010) in particular. Its crystals or powder form had been used since 1920 by organic farmers to help control insects. However the biotech industry was interested in inserting the Bt gene into a plant and transforming the plant into an environmentally friendly powerhouse producing insecticide to protect itself (NRC, 2010). Several advantages exist for Bt crop varieties. First, the level of toxin expressed can be very high, thus delivering a lethal dosage to target insects. Second, a plant produces the toxin throughout its life, and the toxin is expressed relatively uniformly throughout all plant parts. Hence, only those insects that feed on the plant are affected, for example, cotton boll and budworms, various types of corn borers and earworms, and corn rootworms.6 Third, the toxin expression can be modulated by using tissue-specific promoters and replaces the use of synthetic pesticides in the environment. Fourth, the Bt toxin expressed in plants is not toxic to humans or animals. In 1991, scientists at Monsanto 6 Bt produces spores that form the crystal protein insecticide d-endotoxins. The protein toxin is active against species of the orders Lepidoptera, Diptera, Coleoptera, Hymenoptera, and nematodes. When these insects ingest toxin laden crystals, chemicals in their digestive track activate the toxin. It inserts into the insect’s gut cell membrane and dissolves it and eventually causes death of the insect (Wikipedia).

Contributions of Public and Private R&D to Biotechnology Innovation

123

successfully inserted the soil bacteria, Bacillus thuringiensis, into cotton plants to induce Bt resistance to the cotton bollworm-budworm complex. HT occurs naturally in some plants, for example, imidazolinone and sulfonylurea herbicides, but can also result from insertion of a particular GM trait that creates resistance. The molecular basis for this selectivity arises from herbicide detoxification, target enzyme insensitivity, lack of herbicide uptake, or translocation. However, gene transfer methods are available, for example, using the Agrobacterium tumefaciens. Agrobacterium-mediated gene transfers have been highly successful for gene introduction and expression in several plant families, for example, cotton, canola, sugar beet, and tomato. For crops that are unresponsive to Agrobacterium-based transformation, physical methods of DNA delivery, for example, gene-gun blasting, are possible. Glyphosate tolerance is the most common source of HT. Glyphosate is a broad spectrum, nonselective, postemergence herbicide, and it is highly effective against the majority of annual and perennial grasses and broadleaf weeds. Also, glyphosate has favorable environmental features such as rapid soil deactivation and degradation to natural products, little or no toxicity to nonplant life forms, and minimum soil mobility (Kishore et al., 1992; NRC, 2010). Glyphosate is a systemic herbicide that is rapidly transported from the foliar tissue to the metabolically active regions of the shoot and root tips of a plant. Within these tissues, glyphosate inhibits the biosynthesis of amino acids. To induce glyphosate resistance, it is necessary to block the enzyme inhibited by glyphosate, called EPSPS. When tolerance exists, the toxicity of a plant can be reduced by a factor of several thousand, and hence, the plant may exhibit no negative effects of direct contact with glyphosate. Moreover, introduction of a particular HT trait has important implications for the company that manufactures and sells this herbicide. As GM traits were developed and new crop varieties carrying these traits were improved, the US agency with the responsibility of receiving and approving requests for field testing, that is, for releasing genetically modified organisms (GMOs) into the environment, was the USDA’s Animal and Plant Inspection Service (APHIS).7 In the late 1980s, few requests and approvals

7 The role of APHIS in biotechnology regulation is as follows. APHIS uses the term ‘‘biotechnology’’ to mean the use of recombinant DNA technology, or genetic engineering to modify living organisms, and it regulates certain genetically engineered organisms that may pose a risk to plants or animals. In addition, APHIS participates in programs that use biotechnology to indentify and control plant and animal pests. In particular, APHIS’s Biotechnology Regulatory Services regulates the introduction – importation, interstate movement, and release into the environment – of genetically engineered organisms that may pose a risk to plant health. APHIS’s Veterinary Services National Center for Import-Export regulates the input, export, and interstate movement of all animals and animal productions – tissues, blood, and semen – including those that are genetically engineered. In addition, if a plant is engineered to produce a substance that ‘‘prevents, destroys, repels, or mitigates a pest,’’ it is considered to be a pesticide, and then is subject to regulation by the Environmental Protection Agency (EPA).

124

Wallace E Huffman

occurred, but they started increasing rapidly over 1990–1995. Annual applications peaked in 1998 with 1,206 and annual approvals peaked in 2002 with 1,194 (Figure 2). Over the period 1985–2010, 20,500 or roughly 93% of applications were approved. Most applications to APHIS for approval to field-test GM traits and varieties over 1985–2010 involved major crops: corn with 7,030 applications approved, followed by soybeans (1,763), cotton (969), potato (832), tomato (658), wheat (421), alfalfa (409), tobacco (381), rapeseed (290), and rice (266) (see USDA, APHIS, 2011). The total number of phenotyped traits numbered 20,498; 29.3% with HT, 22.3% with IR, 18.5% with agronomic properties, 7.2% with virus resistance, 5.4% with fungal resistance, 1.0% with bacterial resistance, and 16.2% with other traits (Figure 3). Hence, a broader set of GM traits have been approved for field testing than HT and IR, but commercialization in crop varieties has been limited to HT and IR traits, first as single traits, HT or IR, but later as stacked traits, for example, HT and Bt. In the European Union, the number of GM field trials approved for testing lagged far lower than in the United States. EU approvals of GM field trials started roughly in 1991 and rose rapidly up to about 255 in 1997 (Welters, 2006). However, in 1998, the European Commission placed a moratorium on approvals for commercializing GM crop varieties. This moratorium signaled to the seed and biotech companies that the EU was going to take an unfriendly stand on GMOs, and requests and approvals for field trials dropped dramatically over 1999–2002. In the EU, fewer than 100 GM field trials were approved over 2001–2006. Although the

Release Counts 1500

(Permits and Notifications)

1000

1083 1071 983 925

1194

763 711 579 612

954 893 864924877 813 754 658 519

500 301 160 51 4 11 11 16 30

90

19 8 19 5 8 19 6 8 19 7 8 19 8 8 19 9 9 19 0 9 19 1 92 19 9 19 3 9 19 4 95 19 9 19 6 9 19 7 98 19 9 20 9 00 20 0 20 1 0 20 2 03 20 0 20 4 05 20 0 20 6 0 20 7 08 20 0 20 9 1 20 0 11

0

Fig. 2.

Number of APHIS-approved field trials of gm crops, 1985–2011. Source: USDA, APHIS (2011).

125

Contributions of Public and Private R&D to Biotechnology Innovation Number of Approved Releases by Phenotype Category (Permits and Notifications)

8000

6225 6000 4687 4132 4000

1767

2000

1629

1502

1157 211

135 e ta is

es R e od

ria

at

te

em

ac N

R

-N

-B BR

nc

e ta

nc

e lR

es

is es lR ga

un -F FR

is

ta

nc ta es R

s iru -V VR

nc

e

e is

rG ar ke

-M

AP

-A

gr

M

G

om on

en

er th -O O

ic

O

Pr op

ta is es R

ct se -In IR

er tie

e nc

ce ra n le To de ci bi er H TH

Fig. 3.

s

0

Frequency of traits in APHIS-approved GM field trials, 1985–2010. Source: USDA, APHIS (2011)

European Commission removed the moratorium on approving field testing of GM crop varieties in 2006, and the number of approved field trials doubled from the previous year, many of the experiments have been disrupted. These disruptive forces have been common in Europe, and private companies have largely moved their GM field trials out of Europe. With the relatively favorable GM climate in the United States, including the extension of patents to plants in 1985 and APHIS’s approval of many GM field trials, patenting activity on varieties for soybean and corn rose dramatically. Early patenting activity for GM corn varieties was higher than for soybean, but both rates were low through the 1990–1994 period (Moschini, 2010). However, over 1995–1999, more patents were issued than for PVPCs on corn and soybeans. Both PVPC and patents issued were higher in 2000–2004 than in the previous five years. Over 2005–2009, the rate of issuance of patents for corn and soybean varieties were roughly twice as high as for the previous five-year period. In contrast, PVPCs issues, which are generally viewed as a weaker IP right, dropped off dramatically.

6. Adoption of GM crop varieties GM crop varieties with HT and Bt traits first became commercially available to farmers in 1996. By 2000, 108 million acres (44.2 million

126

Wallace E Huffman

hectares) of biotech crops were grown; only five countries planting 500,000 acres (0.2 hectares) or more (James, 2001). These five countries were the United States, Argentina, China, South Africa, and Australia; the United States accounting for 65% of the total. By 2009, the global area planted to GM crops had greatly expanded – to 330 million acres (134 million hectares) with 12 countries planting 0.2 hectares or more (James, 2010). These countries were the United States, Brazil, Argentina, India, Canada, China, Paraguay, South Africa, Uruguay, Bolivia, Philippines, and Australia (Table 2). The acreage remains concentrated, 47% in the United States and another 45% in the other top four countries – Brazil, Argentina, India, and Canada. Hence, the United States is the dominant country for adoption of GM field crops. The initial set of commercially successful GM crops was limited to soybean, corn, and cotton. However, by 2000, significant GM canola was planted. In 2009, the list of GM crops in the United States had expanded to include sugar beet but also a small area in alfalfa, papaya, and squash. In China and India, the major GM crop is cotton. China also has small area in GM poplar, tomato, papaya, and sweet pepper (James, 2010). In the United States, GM varieties are sold at a premium price, generally including a technology fee and signed agreement that the farmer will not save seed for future use or sale. GM cotton varieties containing Bt got off to a fast start; 18% of the US cotton acreage was plant to Bt cotton in 1996 (Figure 4). Herbicide tolerant soybean adoption started more

Table 2.

Global area of biotech crops in 2009, by major producing countrya

Rank

Country

Area (Mil. Hectars)

1

USA

64.0

2 3 4 5 6

Brazil Argentina India Canada China

21.4 21.3 8.4 8.2 3.7

7 8 9 10 11 12

Paraguay South Africa Uruguay Bolivia Philippines Australia

2.2 2.1 0.8 0.8 0.5 0.2

a

Biotech Crops Soybean, maize, cotton, canola, sugar beet (very small area of squash, papaya, alfalfa) Soybean, maize, cotton Soybean, maize, cotton Cotton Canola, maize, soybean, sugar beet Cotton (very small area of tomato, poplar, papaya, sweet pepper) Soybean Maize, soybean, cotton Soybean, maize Soybean Maize Cotton, canola

Other countries with small amounts: Burkina Faso, Spain, Mexico, Chile, Colombia, Honduras, Czech Republic, Portugal, Romania, Poland, Costa Rica, Egypt, and Slovakia, largely producing GM maize.

127

Contributions of Public and Private R&D to Biotechnology Innovation 100 HT soybeans

90

93

Percent of acres

80 70

78 73 63 70

HT cotton

60 50

Bt cotton

40 30

Bt corn

20

HT corn

10 10

09

20

08

20

07

20

20

06

05

20

04

20

03

20

02

20

20

01

00

20

99

20

98

19

97

19

19

19

96

0

Fig. 4. Percent adoption of GM corn, cotton, and soybean varieties in the United Staes, by trait type, 1996–2010. Source: USDA (2010).

slowly than Bt cotton, but after one year; HT soybean acreage was the leading GM crop. Except for a small retreat in 2000, the adoption of HT soybeans increased steadily, and since 2005, the adoption rate has been greater than 90% of the US soybean acreage. In 1998, HT cotton was planted on a larger share of the US cotton acreage than Bt cotton. By 2001, 55% of US cotton was HT, and this slowly increased to 78% in 2010. GM corn started slowly in the United States; in 2000, only 10% of planted corn acreage was to HT and 20% to Bt corn. The rate of adoption of GM corn has been most rapid in the last half of the decade, and in 2010, 70% of US corn acreage has HT and 73% has the Bt trait. Figure 5 displays the adoption of HT soybean in Brazil and HT canola in Canada and of Bt cotton adoption in China and India. Over 1996–2002, the Brazilian government claimed to be GM-free and did not acknowledge that GM HT soybean varieties were being smuggled into the country from Argentina. However, two Presidential decrees, one in 2003 and another in 2004, approved the planting of farmer-saved biotech soybean seed by Brazilian farmers for 2003/2004 and 2004/2005. In 2005, the Brazilian Congress passed a Biosafety Bill for GM crops, and its first major action was to approve commercially certified RRssoybean seed for sale to farmers. However, farmers had to wait one additional year because of initial unavailability of GM soybean seed. The official estimates of the adoption of GM soybean increased from 15% in 2003 to 75% of acres planted in 2009 (James, 2010, p. 28). In Canada, HT canola became commercially available to farmers in 1996, and within five years of

60

Percent Adoption

Percent Adoption

70

50 40 30 20 10 0 2003

2004

2005

2006

2007

2008

100 90 80 70 60 50 40 30 20 10 0

2009

128

80

1995

60

Percent Adoption

Percent Adoption

70

50 40 30 20 10 0 1999

2001

2003

2005

Bt Cotton Adoption in China

Fig. 5.

1999

2001

2003

2005

2007

2009

HT Canola Adoption in Canada

80

1997

1997

2007

2009

Wallace E Huffman

HT Soybean Adoption in Brazil

100 90 80 70 60 50 40 30 20 10 0 2003

2004

2005

2006

2007

2008

2009

Bt Cotton Adoption in India

Percent adoption of major biotech crops in Brazil, Canada, China, and India. Source: James (2010).

Contributions of Public and Private R&D to Biotechnology Innovation

129

introduction, adoption rose to 50%. The rate of adoption of GM canola increased more slowly thereafter, but reached 90þ% by 2009. Bt cotton first became available in China in 1997, but availability of seed increased slowly to 2000. However, over 2000–2004 adoption increased much more rapidly – going from 15% to 65% of the Chinese cotton acreage. The adoption rate stalled at 65%. In India, Bt cotton became available in 2003, and the rate of adoption, availability of seed increased slowly to 2005. However, over 2003–2009, Bt cotton adoption increased from 15% of the cotton acreage to 85%. Both China and India have sourced Bt traits for cotton varieties from Monsanto. Crop varieties with stacking of two or more GM traits have become common. Over 2004–2008, stacked traits in US hybrid seed corn increased from 6% of acreage to 40% (USDA, 2010).8 For example, hybrid corn varieties containing HT and BT for corn borer and rootworm, a triple stack, has been available to farmers since 2008. During 2010, a limited quantity of a new eight-transgene-stack hybrid corn variety was planted by US farmers, and large-scale marketing in 2011 is occurring. This SmartStaxt hybrid corn technologies, which have an intricate transgenic pesticide control mechanism, is the result of an elaborate cross-licensing arrangement (Marra et al., 2010) – Monsanto, Dow AgroSciences, and Bayer CropScience are sharing traits and germplasm for the new hybrid corn varieties. With its multiple modes of HT and IR protection, SmartStaxt varieties have extensive durability to evolving pest resistance and more complete control of pests.9 Moreover, refuge requirements are greatly reduced.10,11 In addition, the Federal Crop Insurance program recognizes the risk-reduction and yield-increasing benefits of the SmartStaxt.12 7. Pricing and benefit distribution from a GM trait It is insightful to apply a little economics to the benefits from an innovation (or sequence of innovations) leading to the widespread 8

Stacking of HT and IR traits in cotton seed has exceeded 20% of the US planted acreage since 2000 (USDA, 2010). 9 The development required major advance in gene transformation technology; the eight genes are first concentrated in one soil bacteria and then inserted as one capsule into a particular corn inbred line. This insures that multiple traits will be inserted in the same location for maximum complementary impact and not randomly located. 10 The standard refuge requirement has been 20% of GM corn acreage in the Corn Belt but will be reduced to only 5% with SmartStaxs technology. 11 Marra et al. (2010) estimate the anticipated increase Corn Belt farmers’ profits from adopting SmartStaxs hybrid corn varieties at $400 million per year and added nonmonetary value to farmers of $360 million per year. 12 In addition, biotech seed companies are also developing new crop varieties that improve tolerance to stress, for example, heat and drought, or that provide direct benefits directly to consumers.

130

Wallace E Huffman

adoption of a GM trait in a particular species. With a patent, the innovator obtains up to 20 years of controlled use of the innovation. Ignoring for the moment the cost of a new GM crop variety, the widespread adoption of the new variety is assumed to reduce the cost of production of a commodity. Moreover, we can show that an optimal price exists on a new GM crop variety and that it covers the marginal cost of the GM seed, including the trait, and an income stream to reward R&D outlays associated with the innovation. Consider the Bt trait for IR due to the European corn borer in hybrid seed corn. In Figure 6, upper panel, the US demand for corn for grain (in bushels/year) is represented by the downward sloping demand curve DQDQu. The pre-trait US supply of corn for grain is represented by P*0S0. The pre-Bt trait price of corn for grain in $s is P*0 and the equilibrium quantity is Q0. Now assume that the introduction and wide-scale adoption of Bt hybrid seed corn lowers the cost of corn production, net of the cost of the technology fee, such that the new supply curve for corn for grain is

PQ

DQ a

P*0

S0 e

P*2 = P1 + Re′

S2 b

P1

0

Q0*

S1 D′Q Q

Q2*

RX a′

DX Re′

f

c′

mc 0

e′

X0

Xe

MCX0 MRX

b′

X

Fig. 6. The US market for corn (Q), derived demand for cost-saving Bt hybrid seed corn (X), optimal pricing, and benefit distribution.

Contributions of Public and Private R&D to Biotechnology Innovation

131

P1S1.13 The area P*0abP1 represents the potential benefits or surplus due to widespread use of Bt hybrid seed corn. Given that the expected production of corn is directly linked to the amount of Bt hybrid seed corn planted, for example, roughly 190 seed kernels planted per bushel of corn harvested (Elmore and Abendroth, 2008), the above area can be used to trace out the derived demand for the services of Bt hybrid seed corn. This is represented by the kinked demand curve Dxaubu in the lower panel of Figure 6 (the rental rate for the Bt trait in $s, Rx, is on the vertical axis and X (in units of 190 Bt hybrid seed kernels) is on the horizontal axis). The marginal revenue curve for X is the discontinuous curve DxaufMRx. Lets assume that the marginal cost (for seed production, marketing, and a technology fee) of a unit (190 kernels/year) of Bt seed corn is constant at mc, then marginal revenue equals marginal cost at cu with Xe units of Bt hybrid seed corn sold. At Xe, the optimal price per unit of X is at eu on the demand curve or is R0e . This rental price and quantity for X yields an annual profit on the Bt hybrid seed corn represented by the area Re0 eucumc. This profit provides an annual income stream to pay for the research and development, testing and obtaining regulatory approval for Bt hybrid seed corn. What does the optimal pricing of the Bt trait imply for the market price of corn (for grain) in the upper panel? Given the optimal price for the Bt trait in seed corn of 0Re0 in the lower panel, this amount must be added vertically to P1 in the upper panel, which yields a new perfectly elastic supply curve for corn (grain) at P2 (upper panel). At this price, the new equilibrium quantity of corn for grain is Q*2. The net result of the widespread adoption of Bt seed corn is a reduction in the price of corn to consumers from P*0 to P*2, which generates consumer surplus of P0 ae P2 . Given the assumption of a perfectly elastic supply curve of corn (grain) produced by farmers, they do not receive any of the long-run surplus associated with the successful adoption of Bt seed corn.14 8. A major transformation of research and the seed industry Historically, the farm seed industry in developed countries focused on plant germplasm improvement, developing and testing new crop varieties and marketing the most successful to farmers. With bioengineering, it became possible to take traits or biological events, for example, glycophaste tolerance and Bt toxicity, from other species, for example, 13 This assumes that the expected loss in yield due to the European corn borer is uniform across US corn-growing areas. 14 However, in the short run, farmers will not adopt BT seed corn unless it increases their expected profits and hence must obtain some of the benefits (see Huffman and Evenson, 2006, p. 260). See Falck-Zepeda et al. (2000) for a discussion of the early rent distribution in Bt cotton.

132

Wallace E Huffman

soil bacteria, and transfer them into superior field crop germplasm to create a new and potentially successful GM crop variety. However, achieving Bt toxicity in plant parts that were sufficient to kill target insects was a difficult task, especially for the cotton bollworm-budworm complex and the European corn borer. Moreover, research in this area required sophisticated research laboratories, scientists trained in gene-splicing or gene-gun blasting techniques, and Ph.D. degrees in molecular biology. This type of training did not exist in the global seed industry in 1990. Moreover, the new research required a significant scale of operation to be successful, and this scale of operation could not be supported by smalland medium-sized seed companies. Although chemical companies had large sophisticated research laboratories, they were unfamiliar with germplasm enhancement and plant breeding. However, the biotech revolution did open new opportunities for small biotech companies to develop plant traits, which might be licensed to other companies. A few large chemical companies soon acquired these promising businesses. Also, crop varieties containing HT and IR required deregulation by the Environmental Protection Agency (EPA) (see Berwald et al., 2006).15 Seed companies did not have prior experience with this type of deregulation. Figuring out the type of data, research, and information needed for this activity was costly and time consuming. Although chemical companies did not know anything about genetic improvement of crop varieties, those that had been developing and marketing agricultural chemicals for application on plants were familiar with the US regulator process on plant-pest protectants and the requirements of the EPA. They also had both legal and scientific expertise that could facilitate obtaining regulatory approval of new plants that produced HT and IR. Hence, new opportunities for chemical companies in the seed industry were born in the GM revolution. As recently as 1970, most seed companies were independent. Shortly thereafter, a trend toward acquisition, mergers of small seed firms by larger companies emerged. For example, more than 50 seed companies were acquired by pharmaceutical, petrochemical, and food firms after IP for plant materials was strengthened in the United States in the 1970s and 1980s. Acquiring companies sought strong, well-developed small- and medium-sized seed companies, anticipating increased profits. At the same time, agricultural chemical companies were looking for new ways to market their chemicals, and some recognized the usefulness of their legal and scientific for obtaining deregulation of GM crops. Starting in the 1980s, chemical companies entered the US market for ‘‘traits’’ and later for germplasm owned by seed companies. The big players in this 15 The EPA has regulatory oversight over agricultural pesticides, which include crops with ‘‘plant-incorporated protectants.’’ Plant-incorporated protectant is the EPA’s term for pesticidal substances produced by plants and the genetic material necessary for the plant to produce such substances, made possible through the use of biotechnology.

Contributions of Public and Private R&D to Biotechnology Innovation

133

process were large corporations, many of them multination at conglomerates that possessed the resources and scientific personnel to achieve scale economics in R&D to support the regulatory process associated with herbicide-tolerant and insect-resistant crop varieties. During the 1980s, new developments in biotechnology created an additional incentive for firms to increase their R&D capacity and expand further, either as a larger conventional or as a foundation seed company. As the first biotech cotton, soybean and corn varieties began large-scale testing in the 1990s, the structure of the US seed industry entered a new phase of transformation. The industry reorganized through extensive mergers, acquisitions, and joint ventures as companies sought to achieve the genetic material (germplasm and GM traits), scientific manpower, and economies of scale. Companies with significant tenure in the US seed market that survived this round of restructuring were Pioneer Hi-Bred, Asgrow, and Sandoz. Even big companies faced tough decisions. For example, chemical and new agricultural biotech companies could develop GM traits and license them to existing seed companies. Alternatively, these companies could purchase seed companies owning important germplasm and market their GM traits directly. Existing large seed companies could pay for license to use GM traits developed by the biotech and chemical companies, or they could invest heavily in the new science of biotechnology and attempt to develop good GM traits and insert them into their crop varieties. Some scope economies were apparent; for example, once a biological event is developed creating a GM trait for one particular variety of crop or species, the US regulatory process permitted it to be inserted into other crop varieties of the same species. However, if that event is inserted into another species, the whole regulatory process must start over again. Added complexities arise when more than one event is stacked into a crop variety, and the EPA requirements are best described as ‘‘negotiated’’. Also, the potential exists to export the technology to other countries, but this generally requires that the trait be inserted into locally superior crop varieties. Moreover, all countries have their own regulatory process. Some firms evolved into ‘‘life science’’ companies organized around agricultural chemicals, seeds, pharmaceuticals, and advancing biotechnology. Monsanto, Novartis, and AgrEvo gained significant market share in the 1990s through such activity. Some of these life science companies later divested themselves of their agricultural operations at the turn of the century. A biotech event is a genetic method for obtaining a GM trait, for example, HT, IR, IRþHT, for example, MON810 is an event that inserts B. thuringiensis into a particular location in the corn germplasm to induce European corn borer resistance (James, 2010, pp. 263–279). Companies that recently achieved regulatory approval for marketing GM traits in corn, soybean, cotton, or canola varieties are Bayer CropScience, Dow AgroSciences, Monsanto, Pioneer-DuPont, and Syngenta. Bayer CropScience has

134

Wallace E Huffman

been particularly successful developing HT varieties for corn, soybean, and cotton. Dow AgroSciences has been successful with developing IR and IRþHT varieties in cotton and corn. Monsanto has been successful developing soybean, canola, corn, and cotton varieties that are HT; cotton and corn varieties that are IR; and cotton and corn varieties that are HIþIR. Pioneer-DuPont have a few successes in HT soybean and IRþHT corn varieties. Syngenta has had success with IR cotton varieties and HTþIR corn varieties. Monsanto is the inventor of 21 event-trait-crop varieties across corn, cotton, soybean, and canola approved for planting by farmers in the United States over 1995–2009. In contrast, Bayer CropScience is the inventor for 11 and Dow AgroSciences, Syngenta, and Pioneer account for 6 or 7 (including coinvention; see James, 2010, pp. 263–279). However, in the marketing of GM seeds, Monsanto and Pioneer-DuPont are the leaders, and Syngenta is most likely the third. A bite of history can provide some useful insights into the evolution of private sector research and seed sales. Monsanto and Pioneer-DuPont provide dramatic contrasts, one starting with a long history in the chemical industry and the other starting with a long history in the hybrid seed corn business. Syngenta is quite a new company formed from the merger of the agricultural businesses of companies with strong ties to the chemical industry. Monsanto, founded as a chemical company in 1901, started producing and marketing agricultural chemicals in 1945, including the herbicide 2,4D. In 1960, it established an Agricultural Division, and important new chemical herbicides were released in 1964 (Ramrods), 1968 (Lassos), and 1976 (Roundups). In 1975, Monsanto established a cell biology research program in its Agricultural Division to take advance of new biotech advances, and in 1981, a molecular biology group was set up; biotechnology is firmly established as part of its research mission. In 1982, scientists working for Monsanto were the first to genetically modify a plant cell, and in 1984, Monsanto established a new Life Sciences Research Center in Chesterfield, MO (Monsanto, 2010). In 1987, Monsanto conducted the first US field trials of plants with a GM trait. In 1991, Monsanto scientists were the first to insert the Bt gene into the cotton plant and to induce resistance to the cotton bollworm-budworm complex. Through 1995, Monsanto invested heavily in GM trait development but had acquired only one seed company – Jacob Hartz Seeds, a soybean seed company. Hence, Monsanto’s strategy up to this point is best described as scientific discovery and development of GM traits with a goal of licensing them to existing seed companies.16 For example, the early GM hybrid corn varieties marketed by Pioneer Hi-Bred containing traits for HT and Bt were obtained through a licensing arrangement between 16 By 2004, scientists at Monsanto had a library of over 1,000 different soil bacteria that might be useful in GM trait development. Mycogen, an agricultural biotech firm in San Diego, CA, also developed a large library of soil bacteria early that could be used in GM trait development.

Contributions of Public and Private R&D to Biotechnology Innovation

135

Pioneer and Monsanto on event MON809. However, both companies quickly decided to diversify their GM marketing potential. Monsanto chose to purchase diverse companies that owned elite germplasm for major crops, and Pioneer chose to acquire GM traits from non-Monsanto sources, and to start developing some of its own GM traits. Pioneer is widely perceived as being far behind Monsanto in GM trait development. In 1997, Monsanto bought Asgrow Seed Company (soybean) from ELM, Calgene, a biotech company, Holden Foundation Seeds, and Corn States Hybrid Service (Figure 7). Monsanto paid $1.2 billion for the latter two companies that owned large amounts of corn germplasm and supplied highquality foundation seed to the seed corn industry. In 1998, Monsanto further extended its access to elite corn germplasm by purchasing the second largest US seed corn company, DeKalb Genetics Corporation, for $3.7 billion and the international seed business of Cargill for $1.4 (FernandezCornejo, 2004). In 1998, it set a goal of obtaining access to large amounts of US cotton germplasm and attempted to acquire Delta & Pine Land, which had roughly 70% of the US cotton seed market. This deal was called off in 1999. In March 2000, Monsanto (the original company) merged with Pharmacia & Upjohn, a multinational pharmaceuticals giant. The agricultural part of the merger retained the Monsanto name, but the pharmaceutical and related side operations were under the Pharmacia Corporation. After a partial initial public offering of Monsanto stock in October 2000, Monsanto was relaunched with partial Pharmacia ownership. However, in August 2002, Pharmacia spun off Monsanto as an independent agricultural chemical and seed company (Monsanto, 2010). In 2006, Monsanto finally acquired its large cotton germplasm base with the purchase of Delta & Pineland for $1.5 billion. In 2005 and 2008, Monsanto reinvested in the vegetable seed industry by purchasing Seminis, the world’s largest vegetable seed company for $1.4 billion, and de Ruier, a major Dutch vegetable seed company, for $0.8 billion. Also, in 2007 and 2008, Monsanto purchased two Brazilian seed companies – Agroeste Sementis and Aly Paticipacoes. Agroestes is a cotton seed company, and Aly is a sugar cane seed company. These latter two acquisitions strengthened Monsanto’s foothold in the South American seed market. Monsanto is now a company that is a result of mergers and acquisitions of roughly 30 former seed and biotech companies, which occurred over a relatively short period (see Figure 7). After stumbling in some of its early GM varietal introductions, Monsanto has become a behemoth in the US market for GM hybrid corn and soybeans and cotton seed. Its major competitors in the GM hybrid corn and soybean seed markets are Pioneer Hi-Bred International, a DuPont Company, and Syngenta. The Hi-Bred Corn Company of Des Moines, Iowa, was renamed the Pioneer Hi-Bred Corn Company in 1935 and changed its name to Pioneer Hi-Bred International, Inc. in 1970. Seeing new potential in the US soybean

Empressas La Moderna (ELM) (Mex.)

Asgrow 1995

Farmer’s Hybrid O’s Gold 1983

Associated Seeds 1972

Asgrow $.22 bn 1997

Calgene $.06 bn 1997

Crow’ Hybrid Ecogen 1996

Channel Bio $0.12 bn 2004

Midwest Seed Genetics NC+Hybrids

Corn State’s Hybrid 1997

Stoneville Pedigreed 1987

DeKalb $3.7 bu 1998

Desert Cotton Research And Development 1988

Cargil (Int’l) $1.4 bn 1998

Plant Genetics 1989

Monsanto

Farmer’s Hybrid 1983

Delta & Pineland $1.5 bn 2006

Jacob Harz Seed Company 1996

Agroeste Sementis (Br) $0.10 2007

Stoneville Pedigreed $.092 bn 1999

Plant Breeding Institute (UK) 4.325 bn 1998

Bioseeds International 1990 Pharmacia 20001 Monsanto 2002

Agruceres(Br) 1997

Seminis $1.4 bn 2005

Emergent Genetics 1999

Br = Brazil.

1

Monsanto Company became an agricultural subsidiary of Pharmacia Corporation in April 2000. Monsanto became completely separate and independent from Pharmacia on August 13, 2002. 2 Formed in November 1995 by the merger of Pharmacia Akiebolag and the Upjohn Company, prior to this point, Upjohn had owned Asgrow solely since 1968. Sources: Fernandez-Cornejo (2004); various Monsanto Annual Financial Reports.

The merger and acquisition tree of Monsanto.

De Ruiter $0.8 bn 2007 Aly Paticipacoes Ltd (Br) $0.29 bn 2008

Wallace E Huffman

Farmcraft 1965

Fig. 7.

Heartland Hybrids

Holden’s Foundation Seed $1.2 bn 1997

Asgrow 1968

Bn = billion

Fielder’s Choice Direct

Landec’ 0.07 bn 2006

Agracetius $.15 bn 1996

Monsoy (Br) 1996

Upjohn & Pharmacia2

Jacob Hartz Seed Company 1983

136

Hybritech 1982

Contributions of Public and Private R&D to Biotechnology Innovation

137

seed market, Pioneer purchased Peterson Seed Company in 1973 (Fernandez-Cornejo, 2004). During this era, Pioneer followed a strategy of developing hybrid seed corn varieties that were well tailored to local geo-climatic conditions, and hence, it marketed many more corn hybrids than its major competitor, DeKalb Genetics. This policy translated into a very large corn germplasm inventory for Pioneer going into the GM corn era. In 1992, Pioneer paid $450,000 to Monsanto for right to use the Roundup Readys (RR) trait in its research operation (Pioneer Hi-Bred, 2010). In 1993, it also paid roughly $38 million to Monsanto for a Bt gene that was effective against European corn borers. In 1995, Pioneer tried to expand its access to bacteria that might be used to create IR. It formed a memorandum of understanding with Mycogen Corporaton, San Diego, CA. Pioneer agreed to invest $51 million in Mycogen, $30 million through the purchase of $3 million Mycogen shares, and the remainder in new technology investment. Pioneer was interested in access to Mycogen’s library of soil bacteria that could be used for inserting GM traits into field crops, including Bt. The main purpose of the collaboration was to speed up IR trait development in field crops (PR Newswire, 1995). Moreover, Pioneer was to obtain access to all Bt crop protection technology and related technologies owned by Mycogen, or developed by the collaboration, for 10 years after the start of new joint development program. However, both companies were to market Bt traits in their own seed products, and no proprietary seed lines were to be shared (New York Times, 1995).17 In 1997 DuPont, a large US chemical company, acquired a 20% stake in Pioneer, and the two companies formed a joint venture – Optimum Quality Grains LLC – and in 1999, DuPont purchased the remaining shares of Pioneer. Although DuPont had legal and scientific expertise in deregulating agricultural chemicals, which was useful for preparing requests to deregulate new GM crops, they did not understand the US seed industry, and therefore, benefits of the merger have been slowly unfolding. Pioneer soon sold one million of its Mycogen shares to Dow AgroSciences, and in 1998, it sold the remaining two million shares to Dow (New York Times, 1998). Recently Pioneer-DuPont has been investing heavily in research to develop new GM traits on its own, but also with some continuing, sometimes stormy arrangements with Monsanto. Syngenta is a decade-old company. In 2000, Novartis and Astra Zeneca merged their agribusinesses to form Syngenta. Novartis was founded in 17 Pioneer dominated the US hybrid seed corn market in the mid-1990s, but as Monsanto acquired corn germplasm and seed corn companies in the mid-1990s, including distribution systems, and inserted its GM traits into new varieties, it grew in strength relative to Pioneer. Currently, Monsanto’s has a slightly larger share of the of the US hybrid seed corn market relative to Pioneer-DuPoint.

138

Wallace E Huffman

1996 through a merger of two large chemical companies, Sandoz and Ciba. However, in 1975, Sandoz started to diversify and acquire seed companies with the purchases of Rogers Brother, a vegetable seed company, and in 1976 Northrup King, a significant producer of US hybrid corn and soybean seeds. Later acquisitions included Zaadunie (Dutch, 1980), McNair (1980), Stauffer Seeds (1986), Fredonia (flowers and vegetables, 1988), Cokers Pedigrees (cotton and soybean, 1988), Hilleshog (1989), and Vaughan’s Seeds (1989) (see McDougall, 2008). Ciba, also a chemical company, acquired Funk Seed International in 1974, an early US Midwestern hybrid seed corn company, and established a special biotechnology unit in 1980. In addition, Novartis acquired four seed companies: Eridania Beghin-Say (1999), Agritrading (1998), CC Benoist (1998/2001), and Maisadour Semences (1998). Furthermore, Syngenta made a major addition in corn and soybean germplasm with the acquisition of Garst Seeds (Advanta NAFA corn and soybean seed) in 2004. Other acquisitions were Dia-Engel (Japanese flowering plant and vegetables) and Golden Harvest (corn and soybean seed) in 2004, Emergent Genetics Vegetable and Conrad Fafad (lawn and garden seeds) in 2006, Zeraim Gadera (2007), and Fischer (2007). Also, in 2005, Syngenta entered into a three-year collaboration with Hubei Biopesticide Engineering Research Center (China) for development of novel crop protection agents. In 2008, Syngenta acquired Argentine seeds company SPS (soybean and corn) and flower businesses of US Yoder Brothers and Goldsmith Seeds. In 2010, Syngenta and EMBRAPA, the Brazilian Agricultural Research Corporation, started a partnership to improve crop quality and yields in their field crops, and in 2009, Syngenta made its first direct investment in a US biotech company called Metabolon. It completed a $600 million capacity expansion program in 2010 for the production of crop protection production. Hence, through acquisitions and its own research, Syngenta has become a major crop biotech company.

9. More evidence on private and public R&D investments in crop improvement and biotechnology Historically, the large seed companies of the United States have been leaders in varietal development internationally for hybrid corn. Starting in the 1980s, the private sector moved strongly into varietal development for cotton and soybean in the United States. Small-grain varietal development remains a public sector activity in North America, and CIMMYT/IRRI have provided improved germplasm at minimal cost to the national agricultural research systems of developing countries (Morris and Ekasingh, 2002). Very limited information is available on the investments in scientists and resources for germplasm enhancement, crop improvement, and GM trait development, especially outside the United States. However, some

Contributions of Public and Private R&D to Biotechnology Innovation

139

information is available for the United States, and given the influence of the United States on the development of GM crop varieties, the data are insightful. Except for the CGIAR system, little information is available elsewhere. However, GM crops have made little headway in the CGIAR system. Public sector research in collaboration with international crop biotech companies has developed GM crop varieties for China and India.18 Brazil is new on the crop biotech scene, but seems to be progressing rapidly. At the dawn of the GM crop era, Frey undertook a survey of the US public and private plant breeding institutions. He reports (Frey, 1996) a total of 2,205 scientists years (SYs) employed in plant breeding research across the US public and private sectors in 1994, of which 1,499 (68%) were in the private sector and 706 (32%) in the public sector (Table 3).19 In corn research, 545 total SYs were allocated to plant breeding and 93.5% of them were in the private sector. In soybean research, only 156 SYs were allocated to plant breeding, and 65% of them were in the private sector. In cotton research, 134 total SYs were invested, and 77% were in the private sector. Wheat, however, is a major self-pollinated crop where hybrids have not made major inroads. A total of 130 SY’s were invested there, but only 41% were in the private sector. For fruit and vegetable crops, a total of 213 SYs were invested in plant breeding research, and 78% were in the private sector. Hence, in the crops that had GM potential, the private sector had taken a dominant role in plant breeding research by the mid-1990s. In an attempt to replicate Frey’s earlier survey, Traxler and Frey undertook a later survey in 2002. Traxler et al. (2005) report a total of 2,063 SYs invested in all types of plant breeding research in the United States in 2001, 69% of them were invested by the private sector. Hence, the share had risen a little compared to the 1994 survey. They asked about investments in biotechnology R&D focused on plant varietal improvement and reported 677 SYs invested in this type of research – 70% of it invested by the private sector (Table 4). However, most of the data reported in Traxler et al. (2005) are for US public plant breeding activities because of a low participation rate by the the private seed companies in the 2002 survey. However, over the intervening years, the US seed industry was undergoing major consolidation and merger with the agricultural chemical industry. 18 Also, the Chinese Academy of Agricultural Sciences has developed two-event IR cotton varieties that are approved for farmers in Northern China. Huazhong Agricultural University has developed one Bt rice variety to protect against the rice borer insect, and it is approved for the final field testing stage. In India, Mahyco collaborated with Monsanto to develop twoevent IR cotton varieties for farmers. To the extent that there are crops with GM traits approved for farmers to plant elsewhere, they are tied to the private sector research of GM seed companies operating in the United States. 19 The cost per scientist year differs across private seed companies being largest for the largest seed companies. In 1994, the average cost of an SY was about $290,000 in the largest seed companies and ranged down to $150–215,000 for small seed companies. At the USDA-ARS and SAES, the average cost of an SY was about $295,000.

140

Table 3.

Wallace E Huffman

Number of scientist years (SYs) devoted to plant breeding, public and private, by crop, 1994 Public Sector

Crop/Crop category

Corn Soybeans Cotton Wheat Other cereal crops Other grain legumes Other fiber crops Forage Fruit vegetable Other crops Total

Number of SY employed

Share of total for the crop (%)

Private sector Number of SY employed

Share of total for the crop (%)

Total Number of SY employed

Share of total SY (%)

35 55 31 76 77 26

6.48 35.01 22.94 58.63 35.48 50.98

510 101 103 54 139 25

93.52 64.99 77.06 41.37 64.06 49.02

545 156 134 130 217 51

24.72 7.07 6.09 5.91 9.84 2.31

2 71 46 287 706

100.00 58.20 21.60 45.27

0 51 167 348 1,499

0.00 41.80 78.40 54.89

2 122 213 634 2,205

0.09 5.53 9.66 28.75

Source: Frey (1996).

Traxler et al. reported the number of public sector SYs invested in US plant breeding across all cereal crops were 184.5 SYs in 1994 and 2001 (Table 5). There were only 35 public SYs invested in breeding corn varieties in 1994 and 36% or 47.5 SYs more than in 2001. Wheat is a crop where the public sector has been a major developer of new crop varieties, and there were a total of 77 SYs invested in 1994 and a slightly small 72.8 in 2001. A majority of the SYs were in the state agricultural experiment station (SAES). Across all legume crops, plant breeding decreased by a small percentage between 1994 and 2001. A total of 54.6 SYs were invested in public soybean breeding research in 1994 and a slightly larger 60.2 SYs in 2001. For cotton, a total of 30.9 SYs were invested by US public institutions in 1994, and a slightly small 29.4 SYs were invested in 2001. Public soybean breeding research is largely in the SAES, but public cotton research is more equally split between the USDA-ARS and the SAES (Table 5).20 In 2009, Monsanto and Pioneer, the two US largest seed and trait developers and suppliers, invested roughly $1.6 billion on R&D or 11% of seed and trait sales in R&D for improving these products. Syngenta, Dow AgroScience, and Bayer may invest another roughly $750 million in GM traits and varietal development. This amount has grown rapidly (constant dollars) over the past 25 years, reflecting new opportunities in biotech 20 The single commercially successful public sector GM crop has been GM papaya that is resistant to the ringspot virus.

Contributions of Public and Private R&D to Biotechnology Innovation

141

Table 4. Numbers and percentages of plant breeding SYs devoted to plant breeding research, germplasm enhancement, cultivar development, and biotechnology, SAES, USDA (including ARS & PMCs) and private industry, 2001 Category

SAES

Plant breeding research Germplasm enhancement Cultivar development Biotechnology R&D Totals

85 70 144 121 420

ARS/USDA 20% 17% 34% 29% 100%

138

64%

80 218

36% 100%

Private Industry 180 96 673 476 1425

13% 7% 47% 33% 100%

Source: Traxler et al. (2005).

crops, new IP rights, and new biotech method for field crops. However, private sector biotech and crop improvement research continues to build from a scientific base that includes basic or general science research of the public sector, for example, the USDA’s Agricultural Research Service and SAES. For example, consider public sector research in biochemistry and biophysics, molecular biology, microbiology, physiology, and genetics.21 In 1984, total expenditures on these five fields of science were $595.3 million (constant 2000 prices), and 30.8% of total SAES research expenditures were on the abovementioned five biological science fields. Still, molecular biology research accounted for only 1.3% of SAES total research expenditures in 1984. Research allocated to genetics, as a share of the total expenditures, was unchanged at 11.3% (see Huffman and Evenson, 1993). For 1994, 2004, and 2008, Table 6 reports total public agricultural research expenditures (USDA-ARS, SAES, and other cooperating institutions) on the abovementioned five fields of science and total expenditures (all in constant 2000 prices). For this larger set of institutions and in 1994, a total $1.17 billion or 31.2% of the total public agricultural research expenditures were allocated to the five biological science fields. Now, molecular biology research accounts for a much larger, 5.6% of total public agricultural research expenditures, and genetics accounts for a smaller 9.0%. Hence, molecular-based research increased and statisticsbased genetic research declined. In 2004, the public agricultural research institutions allocated $1.39 billion to the five fields of biological sciences or 34.5% to public agricultural research expenditures. Molecular biology now accounts for 7.3% of the total public agricultural research

21 These expenditures have a plant, animal, or a general focus. Data are not easily obtainable on total public agricultural research expenditures by field of science in 1969 and 1984.

142

Table 5.

Wallace E Huffman

Numbers of SY’s devoted to public plant breeding activities, by employer and crop categories, 1994 & 2001

Crop Category

Types of crops

Cereal Corn Wheat Fiber Cotton Forage Grain legume Soybean Oilseed Fruits & vegetables Root & tuber Sugar Temperate fruit & nut Tropical fruit & nut Lawn & turf Ornamental Leafy, bulbous stems Medicinal, spice & special crops Stimulant (tobacco) Miscellaneous Total

Number of SYs Employed By SAES

SAES

2001

1994

% ARS/ ARS/ Change USDA USDA 1994– 2001

2001

1994

% Change 1994– 2001

Total public

Sector 2001

Sector 1994

% Change 1994– 2001

124.0 29.2 54.7 19.8 18.8 26.2 56.2 43.3 20.1 18.8 30.5 4.4 33.4 8.2 15.7 38.9 5.8 2.6

155.0 27.1 64.5 20.0 19.2 38.0 67.0 45.0 24.0 38.0 45.0 4.0 50.0 10.0 15.0 18.0 16.0 6.0

20 8 15 1 2 31 16 4 16 50 32 10 33 18 4 116 64 56

61.3 18.3 18.1 10.6 10.6 26.5 22.1 16.9 6.1 7.3 10.9 15.8 18.4 4.5 0.1 24.9 7.1 1.0

34.0 8.2 12.0 13.0 11.7 33.0 14.0 9.6 6.0 8.0 12.0 15.0 23.0 6.0 0.0 5.0 2.0 4.0

80% 124% 51% 19% 9% 20% 58% 76% 1% 9% 9% 5% 20% 25% – 398% 255% 75%

185.3 47.5 72.8 30.4 29.4 52.8 78.2 60.2 26.1 26.1 41.4 20.2 51.8 12.7 15.8 63.8 12.9 3.6

189.0 35.0 77.0 33.0 30.9 71.0 81.0 54.6 30.0 46.0 57.0 19.0 73.0 16.0 15.0 23.0 18.0 10.0

2 36 6 8 5 26 3 10 13 43 27 6 29 21 5 177 28 65

10.7 4.4 419.7

13.0 9.0 528.0

18 51 21

1.0 0.5 218.0

2.0 0.0 177.0

50%

11.7 4.9 637.7

15.0 9.0 705.0

22 46 10

23%

Source: Traxler et al. (2005).

expenditures and genetics accounts for 11.0%. In 2008, funding for molecular biology research was only slightly lower than for genetics research in public sector agricultural research. Hence, through the 1980s, research in the US public agricultural research system was lagging in the field of molecular biology research and other basic biological sciences; by 2008, these institutions were on frontier. Among other countries, China is the leading investor in public agricultural research. James (2010, p. 112) reports that in 2009, China invested 1% of agricultural sector GDP in agricultural R&D. With an economy-wide GDP of $4.9 trillion and agricultural sector GDP being 10.6% of total GDP, this translates into $5.2 billion per year (in exchange rate converted units). China has collaborated with Monsanto on developing GM cotton and soybean varieties and with Monsanto and Syngenta on GM corn varieties, but only GM cotton varieties are currently available to farmers (James, 2010, pp. 266–267). However, Monsanto struggles with its big company image there.

143

Contributions of Public and Private R&D to Biotechnology Innovation

Table 6.

Total public expenditures on basic biological science (constant 2000 dl and % distribution), 1969–2008a

Fields of Science

1969b $mil

Biochemistry Biophysics Molecular Biology Microbiology Physiology Genetics Subtotal Total all areas

95.0 6.1

1984b

%

$mil

1994

%

$mil

2004

%

$mil

2008

%

$mil

%

8.0 127.5

6.6

217.0

5.8

168.5

4.2

166.4

4.3

0.5

1.3

210.9

5.6

296.6

7.3

350.1

9.1

25.8

33.8 2.9 68.2 3.5 165.1 4.4 298.4 7.4 267.7 7.0 100.5 8.5 141.9 7.3 235.6 6.3 186.1 4.6 176.1 4.6 127.3 10.7 231.9 11.3 337.3 9.0 444.9 11.0 397.8 10.3 (362.7) (23.0) (595.3) (30.8) (1,165.9) (31.2) (1,394.5) (34.5) (1,368.1) (35.3) 1,185.4 100.0 1933.1 100.0 3,740.5 100.0 4,044.5 100.0 3,849.4 100.0

Source: 1969 and 1984 data are from Huffman and Evenson (1993, p. 112); 1994, 2004 and 2008 data are from USDA-CRIS, Funding Summaries (Table E). See USDA. a Includes USDA, SAES, Veterinary Medicine Colleges, and other public institutions. b Data are only for SAES system.

Since 2008, high-level Chinese government officials have spoken strongly in support of bringing big science to bear on future food needs, and in November, 2009, China increased its public sector investment in crop biotechnology by $3.5 billion, spread over the next 15 years, or an increase of roughly $235 million per year (James, 2010, p. 125).22 In addition to GM cotton, they have a stated objective of making GM soybean, corn, rice, and wheat available to farmers relatively soon. These new GM crops are part of an aggressive plan to increase food production over the next two decades to meet growing food demand. Only rough estimates of expenditures on agricultural research exist for China. Single and double Bt-gene hybrid cotton varieties have been very successful since introduction in 2002. Public and private sector investments in crop biotechnology may be roughly about $300 million and $200 million per year, respectively (James, 2010, p. 82). India’s highly successful Bt cotton varieties have heavily used biotech traits from Monsanto, s Bollgard I and BollgardsII, that are inserted into hybrid cotton varieties that are sold to farmers. In 2009, a publicly bred Bt cotton variety was released for sale to farmers. Indian research is also under way on GM rice, eggplant, and corn (James, 2010).

22 In 2008, Bayer CropScience signed a memorandum with the Chinese Academic of Agricultural Sciences for joint development of new crop traits and global marketing of new agricultural products. In 2009, Monsanto established a new crop biotech research center in Zhongguancun (Bejing) to strengthen links between Monsanto scientists and the Chinese Research Institute in plant biotech and genomics.

144

Wallace E Huffman

Brazil did not officially permit farmers to plant GM crop varieties until 2003, and over 1996–2002 claimed that it was GM-free, although it did approve an HT event developed by Monsanto for soybean in 1998. Hence, crop biotech research in Brazil is a recent phenomenon, and GM events for cotton and corn started in 2005 and 2008, respectively (James, 2010, pp. 204–205). The country has set a goal of investing in crop biotech roughly $700 million per year over the next 15 years (James, 2010, p. 45).

10. Conclusions Except for hybrid seed corn that was discovered and widely commercialized in the United States starting in 1930s, private seed sales of other crops were modest relative to the amount of farmer-saved seed through the 1970s. This began to change with the enactment of breeders’ rights in developed countries, including the US Plant Variety Protection Act (1970). In the mid-1990s, GM cotton, corn, soybean, and canola varieties were patented and first sold to farmers. Biotech methods have accelerated the transfer of genes across species; frequently from soil bacteria into cotton, corn, soybean, and canola to instill HT and IR. The science of cells, proteins, and enzymes were advanced by scholars in the late 19th and first half of the 20th century. Moreover, the discovery of the structure of DNA and how it replicates was a landmark discovery in 1953. Building upon these advances, Cohen and Boyer discovered recombinant DNA, a method for cutting and splicing genes taken from one species into another species in 1973. This set the stage for biotech plant-varietal development in the late 1980s and early 1990s and commercial releases starting in 1996. New transgenic GM cotton, soybean, corn, and canola varieties have been successful in the United States and in some other countries. All major advances in science leading to GM crops were performed in the not-for-profit sector, and the United States has been the hotbed of field testing and new introductions of GM crop varieties. As recently as 1970, most seed companies were small independents, but starting about 1970s, major structural change in the US seed industry began to emerge. For example, in the United States more than 50 seed companies were acquired by pharmaceutical, petrochemical, and food firms following the passage of new property rights in the 1970s and 1980s. Acquiring companies sought out strong, well-developed, small- and medium-sized seed companies, anticipating that they would be able to increase profits. At the same time, agricultural chemical companies were exploring new markets for their chemicals, and some recognized that their legal and scientific expertise acquired with deregulating chemical pesticides were useful for deregulation of GM crops. Starting in the 1980s and 1990s, chemical companies entered the US market for ‘‘traits’’ and later for germplasm owned by seed companies. The big players in this process were

Contributions of Public and Private R&D to Biotechnology Innovation

145

large corporations, many of them multinational conglomerates, who possessed the resources and scientific personnel to achieve scale economics in R&D, and to support the regulatory process associated with HT and IR crop varieties. In the United States, Monsanto placed an early emphasis on GM trait development and anticipated licensing them to other companies. Pioneer had acquired a large library of corn germplasm and had obtained access to significant soybean germplasm through a major acquisition and set out to obtain licenses to use GM traits developed by others, including Monsanto. However, by 2000, both of these companies had reversed courses. Monsanto then purchased large amounts of corn, cotton, and soybean germplasm so as to be able to market its traits directly in its own plant germplasm. Pioneer started looking for new partners in GM trait development and slowly started building expertise in GM trait development. Monsanto remains the clear leader in biological event trait crop variety developments for cotton, soybean, corn, and canola that are approved for commercial sale to farmers. Bayer CropScience is second, and Syngenta, Dow AgroScience, and Pioneer-DuPont have been less important in new trait development. However, in hybrid corn variety sales, Monsanto and Pioneer-DuPont are the leaders, and in soybean sales, Pioneer-DuPont is the leader. Monsanto’s biotech events for GM traits have been successfully transferred to crop varieties grown by farmers in Argentina, Brazil, China, and India. Chinese scientists have been successful in developing a few new GM crop varieties. Also, China seems set to make rapid advances in approving new GM crops for planting by farmers, including GM corn, rice, and wheat. India, with the assistance of Monsanto, has developed GM cotton varieties that are widely sold there.

References Alcamo, I.E. (2001), DNA Technology, 2nd ed. Harcourt Academic Press, San Diego, CA. Bacillus thuringiensis. (2010), Wikipedia, 2010. Available at http:// en.wikipedia.org/wiki/Bacillus_thuringiensis. Berwald, D., Matten, S., Widawsky, D. (2006), Economic Analysis an regulating pesticide biotechnology at the U.S. Environmental Protection Agency. In: Just, R.E., Alston, J.M., Zilberman, D. (Eds.), Regulating Agricultural Biotechnology: Economics and Policy. Springer ScienceþBusiness Media, Inc, New York, pp. 21–35. Diamond v. Chakrabarty. (1980), 447 U.S. 303. Elmore, R., Abendroth, L. (2008), Seeding rates in relation to maximum yield and seed costs. Iowa State University Extension, May. Available at http://www.agronext.iastate.edu/corn/production/management/planting/ seedcosts.html.

146

Wallace E Huffman

Ex parte Hibberd, 227 U.S.P.Q. 443 (Bd. Pat. App. & Interferences 1985). Falck-Zepeda, J., Traxler, G., Nelson, R. (2000), Surplus distribution from the introduction of a biotechnology innovation. American Journal of Agricultural Economics 82, 360–369. Fernandez-Cornejo, J. (2004), The seed industry in U.S. agriculture: an exploration of data and information on crop see markets, regulation, industry structure, and research and development. USDA, ERS Agricultural Information Bulletin No. 786. Fernandez-Cornejo, J., Caswell, M. (2006), The first decade of genetically engineered crops in the United States. USDA-ERS, Economic Information Bulletin, No. 11, April. Frey, K.J. (1996), National plant breeding study-1: human and financial resources devoted to plant breeding research and development in the United States in 1994. Special Report 98. Ames, IA: Agricultural and Home Economics Experiment Station, Iowa State University. Grace, E.S. (1997), Biotechnology Unzipped: Promises and Reality. Joseph Henry Press, Washington, DC. Griliches, Z. (1957), Hybrid corn: an exploration in the economics of technical change. Econometrica 25, 501–522. Griliches, Z. (1960), Hybrid corn and the economics of innovation. Science 132, 275–280. Huffman, W.E., Evenson, R.E. (1993), Science for Agriculture: A LongTerm Perspective, 2nd ed. Iowa State University Press, Ames, IA. Huffman, W.E., Evenson, R.E. (2006), Science for Agriculture: A LongTerm Perspective, 2nd ed. Blackwell Publishing, Ames, IA. James, C. (2001), Global review of commercialized transgenic crops: 2000. ISAAA Briefs, No. 23, Cornell University, Ithaca, NY. James, C. (2010), Global review of commercialized transgenic crops: 2000. ISAAA Briefs, No. 41, ISAAA, Ithaca, NY. Kishore, G.M., Padgette, S.R., Fraley, R.T. (1992), History of herbicidetolerant crops, methods of development and current state of the art – Emphasis on glyphosate tolerance. Weed Technology 6, 626–634. Marra, M.C., Piggott, N.E., Goodwin, B.K. (2010), The anticipated value of SmartStaxt for US corn growers. AgBioForum 13, 1–12. McDougall, P. (2008). Evolution of the leading seed companies. Available at http://www.phillipsmcdougall.com. Monsanto. (2010). A company history. Available at http://www.monsanto. com/whoweare/Pages/monsanto-history.aspx. Morris, M., Ekasingh, B. (2002), Plant breeding research in developing countries: what roles for the public and private sectors. In: Byerlee, D., Echeverria, R.G. (Eds.), Agricultural Research Policy in an Era of Privatization. CABI, New York, pp. 199–225. Moschini, G. (2010), Competition issues in the seed industry and the role of intellectual property. Choices 25(2nd Quarter). Available at http:// www.choicesmagazine.org/magazine/print.php?article¼120.

Contributions of Public and Private R&D to Biotechnology Innovation

147

New York Times. (1995). Investment in Mycogen. New York Times, September 19. New York Times. (1998). Company news: Seed company sells biotechnology state to Dow. New York Times, March 8. NRC, Committee on the Impact of Biotechnology on Farm-Level Economics and Sustainability (2010), Impact of Genetically Engineered Crops on Farm Sustainability in the United States. The National Academies Press, Washington, DC. OECD (Organization for Economic Co-operation and Development). (2009), Technology and innovation in world agriculture: prospects for 2010–2019. Paris, France. Pioneer Hi-Bred. Wikipedia. (2010), Available at http://en.wikipedia.org/ wiki/Pioneer_Hi-Bred. Pioneer Hi-Bred International v. Holden Foundation Seeds. (1994), Nos. 92-3292, 92-3556, United States Court of Appeals, Eighth Circuit, July. PR Newswire. (1995), Pioneer Hi-Bred and Mycogen sign Memorandum of Understanding to Pursue Agreement. PR Newswire, September 18. Rainey, T. (Ed.) (2004). The State of Food and Agriculture: Agricultural Biotechnology. Food and Agriculture Organization of the United Nations (FAO), Rome, Italy. Traxler, G., Acquaye, A.K.A., Frey, K., Thro, A.M. (2005), Public sector plant breeding resources in the US: Study results from the year 2001, November. Available at http://www.csrees.usda.gov/nea/plants/pdf/ plant_report.pdf UNCTA (United Nations Conference on Trade and Development). (2006), Tracing the trend towards market concentration: the case of agricultural input industry, April. Available at http://www.unctad.org/ en/docs/ditccom200516_en.pdf. USDA, APHIS. (2011), Information system for biotechnology: summary charts. Available at http://www.isb.vt.edu/release-summary-data.aspx. USDA, ERS. (2010), Adoption of genetically engineered crops in the U.S.: corn varieties. Available at http://www.ers.usda.gov/Data/ BiotechCrops/ExtentofAdoptionTable1.htm. Welters, P. (2006), Globalization of European biotechnology: commercialization of Agbiotech products despite political and legal restrictions. In: Eagesham, A., Hardy, R.W.F. (Eds.), Agricultural Biotechnology: Economic Growth through New Products, Partnerships and Workforce Development. National Agricultural Biotechnology Council, Ithaca, NY, pp. 89–95. World Bank. (2008), Agricultural development. Washington, DC: World Development Report. Zhao, Y. (2011), Agricultural biotechnology patent identification and analysis. Master’s Thesis, Cornell University.

CHAPTER 6

Spatial Pricing of Genetically Modified Hybrid Corn Seeds Kyle W. Stiegert, Guanming Shi and Jean-Paul Chavas Department of Agricultural and Applied Economics, University of Wisconsin-Madison, Madison, WI 53706, USA E-mail addresses: [email protected]; [email protected]; [email protected]

Abstract Objective – The current biotechnology revolution has been associated with newly developed genetic modifications (GM) that offer new prospects for increasing agricultural productivity. This has stimulated a rapid adoption of GM corn hybrids by U.S. farmers. Yet, there is concern about the structure of competition among biotech firms that own patents over GM traits. This chapter evaluates the spatial differences in pricing of biotech corn hybrids, with a focus on the fringe versus core regions of the U.S. Corn Belt. Methods – The analysis examines how local conditions and market concentrations affect the pricing of GM corn hybrids in different locations. Results – We find evidence of more extensive subadditive pricing in the fringe region. We also examine how both own- and cross-market concentrations affect prices across regions. For GM hybrids, the results show that market power is generally more prevalent in the core region compared to the fringe. Conclusions – The evidence shows that the pricing of GM corn hybrids varies across space. The observed pricing schemes benefit farmers more in the fringe than in the core region of the Corn Belt. Keywords: Spatial pricing, biotechnology, corn hybrids JEL Classifications: L13, L4, L65 1. Introduction Over the last decade, biotechnology has had a major impact on agriculture. By 2009, acreage planted in genetically modified (GM) crops grew to a total of 330 million acres in 25 countries on six continents Frontiers of Economics and Globalization Volume 10 ISSN: 1574-8715 DOI: 10.1108/S1574-8715(2011)0000010011

r 2011 by Emerald Group Publishing Limited. All rights reserved

150

Kyle W. Stiegert et al.

(James, 2009). The United States is the largest adopter with 158 million acres, followed by Brazil, Argentina, India, Canada, and China. These six countries account for 96% of the global GM cropland (James, 2009). The development and rapid adoption of patented transgenic seeds have led to major changes in the hybrid corn seed market. Over the past 15 years, the hybrid corn seed market has evolved from providing farmers with a simple menu of conventionally bred corn hybrids to a market of hybrid seeds containing numerous combinations of genetically altered traits designed to offer specific on-board services to the plant. These traits arrive from patented technologies, which are owned by a few large biotech firms (Monsanto, Dow, Bayer, Syngenta, BASF, and DuPont). The issue of pricing in the U.S. agricultural biotechnology seed markets has drawn considerable attention from a wide array of interested parties. This includes farmer groups, legal scholars, the U.S. Department of Agriculture, and the U.S. Department of Justice. Concerns have been raised about the growing concentration of large agricultural biotech firms and expansion of these firms into corn, soybean, and cottonseed markets. Table 1 presents 4-firm concentration ratios (CR4) of the U.S. corn and soybean acres from 2000 to 2007.1 Over the last decade, concentration in U.S. seed markets for both corn and soybean has been high and is rising, reflecting imperfectly competitive markets. This trend is not overly surprising given that the GM seeds are patented by the biotech firms that developed them (see Moschini and Lapan (1997) for an excellent discussion of the legal protections allowed for the research and development of transgenic seed traits). Yet, high concentration is an important public policy issue. Current concerns include possible adverse effects of imperfect competition in the seed market on farm profitability, and the potential impact of further horizontal or vertical consolidation in the seed industry. Different corn hybrid seeds are sold in different regions, depending on their adaptation to local agro-climatic conditions. This means that market shares of firms selling conventional and/or GM seeds to farmers vary both over time and across space. This raises the following questions: How do prices of patented seeds vary across space? And how strategic pricing of biotech and seed firms vary across space? Addressing these questions is the main motivation for this chapter. In general, the value of biotech traits to farmers is expected to be regionspecific and related to production costs and yields. Pest and weed infestations, temperature, and rainfall all vary by region, leading farmers to weigh the value of on-board traits through their unique production situation. Conventional and GM seeds are developed for submarkets with similar agro-climactic growing conditions. For example, corn hybrid

1

The concentration indices are calculated from the survey data discussed below.

Spatial Pricing of Genetically Modified Hybrid Corn Seeds

Table 1. Year

2000 2001 2002 2003 2004 2005 2006 2007

151

Concentration ratios of four largest seed firms in the national market Corn

Soybeans

CR4

CR4

59.73% 60.28% 58.55% 57.94% 56.97% 68.15% 69.39% 71.82%

50.5% 48.7% 47.7% 48.8% 49.7% 48.7% 53.1% 55.1%

selection is made according to the length of the growing season. This leads to a north–south pattern in the distribution of hybrids in the U.S. Corn Belt. Regional hybrids that tolerate increased seeding rates tend to offer superior yields, which may also alter the on-board trait values. Farmers try to select hybrids and choose seeding rates, GM traits, and pesticide and herbicide application strategies that maximize profit potential while limiting risk. The price of hybrid seeds has a role in this choice. In addition, the farm decision to purchase corn hybrids depends on other crop choices. Specifically, farmers in regions with numerous options for producing other crops (i.e., wheat, cotton, soybeans, etc.) may be more price-sensitive to hybrid corn seed prices. The oligopoly structure of the hybrid corn seed market along with considerable spatial heterogeneity in U.S. corn production provides the basis for the research reported in this chapter. This chapter investigates how seed pricing varies across regions. We also study the factors that may contribute to spatial price discrimination in the corn seed market across space. This includes evaluating how the exercise of market power differs in different regions. Our model draws upon the conceptual framework developed by Shi, Chavas, and Stiegert (SCS, 2010a) that considers hybrid corn pricing through a hedonic model. Allowing for product differentiation, the hedonic structure allows prices to vary depending on the genetic characteristic of each seed. The analysis distinguishes between conventional seeds, seeds with stand-alone GM traits, and seeds with differing combinations of stacked traits (e.g., herbicide tolerance, rootworm resistance, and corn borer resistance). The SCS model also includes the effects of imperfect competition on prices. As an extension of the standard Herfindahl–Hirschman index (HHI), SCS proposed measures of market concentrations given by generalized Herfindahl–Hirschman indices (GHHI). The GHHIs distinguish between own-market concentration

152

Kyle W. Stiegert et al.

effects and cross-market concentration effects. In a multiproduct Cournot setting, the GHHIs provide a basis for investigating empirically how the exercise of market power affects the pricing of differentiated products. This chapter relies on annual farm-level survey data of seed transactions in the United States over the period 2000–2007. Using the SCS approach, it investigates how the exercise of market power affects corn seed prices in the U.S. Corn Belt, and documents how these effects vary across space. The chapter is organized as follows. Section 2 presents a review of the literature. The SCS conceptual model is summarized in Section 3. The data are briefly discussed in Section 4. The econometric analysis is presented in Section 5, followed by the estimated results in Section 6. Economic implications of the research are evaluated in Section 7. Finally, Section 8 contains the summary comments and directions for future research.

2. Literature review The rapidly advancing markets for biotech seeds and associated public policy issues have spawned much research in the past decade. Both supply side and demand side issues have been examined.2 Several studies have shown that farmers can gain significantly from the adoption of improved biotech seeds due to cost reductions, yield increases, and income benefits, even in presence of monopoly power by biotech firms and/or seed companies (e.g., Kalaitzandonakes, 1999; Marra et al., 2002; Price et al., 2003; Jefferson-Moore and Traxler, 2005; Qaim and Traxler, 2005).3 Qaim and de Janvry (2003) analyzed the GM cotton adoption under monopoly pricing in Argentina and suggest that the monopolistic pricing slows down the adoption process among farmers. Liu (2008) found evidence that risk aversion tends to delay the farm adoption decision for GM cotton. More recently, Useche et al. (2009) analyzed the adoption of GM corn at the upper Midwest. They document the role of various GM traits (herbicide savings, insecticide savings, yield improvements, and labor savings) and their implications for farm technology adoption. Biotech and seed firms may not charge the same prices to all farmers, that is, they can implement some form of price discrimination. Gouse et al. (2004) have observed that while small-scale farmers pay significantly less for seed than larger farmers, such differences vary by region. Acquaye and Traxler (2005) demonstrate that even though price discrimination is often considered to be an unwanted market distortion, it may increase total welfare by increasing total output and by making goods available in 2

See Moschini (2008) for a discussion of issues on the demand for products using GM food ingredients. 3 Sunding and Zilberman (2001) provided a good review of the literature on general topic of technology adoption in agriculture.

Spatial Pricing of Genetically Modified Hybrid Corn Seeds

153

markets where they would not appear otherwise. Specifically, using data from the introduction of Bt cotton and assuming third-degree price discrimination, Acquaye and Traxler (2005) find that total welfare increases from price discrimination (compared to uniform pricing), the monopoly’s gains exceeding farmers’ aggregate losses. In addition, the welfare effect on farmers may vary across regions and across farm types (e.g., depending on the farm GM adoption decisions). Recent studies by Shi et al. (2010a, 2010b) and Shi et al. (2011) have examined the pricing of biotech seeds in the United States. One commonly observed feature in these studies is the presence of subadditive pricing in the bundling of traits for biotech seeds.4 In the Shi et al. (2010b) study, some limited evidence of superadditive and component pricing was observed for some firms for specific type of seeds. Subadditive pricing suggests that the process of integrating GM traits into seeds may involve scope economies leading to lower costs that are passed to farmers in the form of prices below component pricing. It may also involve strategic consideration regarding the rate of adoption in GM technologies and finite life of patented genes. Different products exhibiting different demand elasticities can face different pricing rules under imperfect competition. When spatial arbitrage is not possible (as in the case of GM traited seeds sold under contract), these effects may play out differently in different regions. The study by Shi et al. (2010a) also evaluated the role of market structure in corn seed pricing. They found evidence of market power effects, with higher market concentration contributing to higher seed prices. However, these effects were only present in conventional and herbicide tolerant hybrids. They also found the potential for significant welfare gains through cross-product market power effects, which were particularly evident in the market for insect-resistant traited hybrids. For the cotton market, Shi et al. (2011) found that a period involving a merger led to higher prices while an earlier period in which entry occurred led to lower prices. The results follow a classic interpretation that increased/ decreased concentration leads to higher/lower prices.

3. The model The model in this chapter follows from Shi et al. (2010a). It is briefly summarized in this section. Suppose there are N firms, producing a set of M outputs. The vector of output quantity for each firm is: n yn  ðyn1 ; :::; ynm ; :::; ynM Þ 2 <M þ , where ym is the mth output quantity 4 In our context, subadditive bundle pricing means that sum of price increases associated with adding multiple GM traits to a seed is less than the sum value of the same traits when they appear individually in single-traited seeds.

154

Kyle W. Stiegert et al.

P P produced by the nth firm. The profit of the nth firm is: m ½pm ð i yi Þynm   P C n ðyn Þ; where pm ð i yi Þ represents price-dependent demand for the mth output and C n ðyn Þdenotes the nth firm’s total cost of producing yn . Assuming a Cournot game5 and under differentiability, the profit maximizing decision of the nth firm for the mth output ynm satisfies X @p @C n k n y  n  0, (1a) pm þ k @yn k @ym m ynm  0,

(1b)

  X @p @C n n k n y  n ym ¼ 0. pm þ k @yn k @ym m

(1c)

Equation (1c) is the complementary slackness condition, which holds whether the mth output is produced by the nth firm (ynm 40) or not (ynm ¼ 0). This is important for our analysis because it allows each firm to enter/exit and offer different products in each market. Assume  linear demand for each product and constant marginal costs, as the market share of the @C n ðyn Þ @ynm ¼ cm . Define S nm ¼ ynm Y m 2 ½0; 1P nth firm for the mth product, where Y m  n ynm W0 is the aggregate output level of the mth product. Dividing Equation (1c) by Ym and summing across all firms yield ! X X X i i pm ¼ c m  akm S k S m Y k ¼ cm  akm H km Y k , (2) k

i



k

@pk @ynm

where akm ¼ defines the cross-market impactPof change in quantity of good m on the price of good k, andH km  i S ik S im ; is the GHHI of market concentration. When k ¼ m, Hmm is the traditional HHI. Under a downward sloping demand for good m, we have amm o0, meaning that an increase in the traditional HHI for the mth good is associated with higher prices. As a rule of thumb, H mm 40:18 represents a threshold that suggests markets are highly concentrated (Whinston, 2006). Equation (2) shows how the GHHI is interpreted for a multiproduct market. When products k and m are substitutes in demand, then akm ¼  @pk @ynm o0, and increases in the H km is associated with an increases in pm (Hicks, 1939). In our context, we might view a seed with one patented herbicide tolerance trait as a close substitute for seed with a different herbicide tolerant trait. Alternatively, when products are complements in demand, then akm 40and increases in the Hkm is associated with a decrease in pm. Complementarities between biotech seeds are possible. For example, controlling the rootworm with a traited seed is likely to 5 The single period Cournot assumption can be justified on the grounds that each year, hybrid seeds are sold from a stock of seeds that were produced in the previous growing season.

Spatial Pricing of Genetically Modified Hybrid Corn Seeds

155

strengthen the plant leading to healthier corn plants. This can open the door for greater damages from corn borer infestation. Farmers planting seeds with rootworm protection may subsequently develop a complementary demand for seeds with corn borer protection. This shows how both own-market concentration (as measured by Hmm) and cross-market concentration (as measured by Hkm) can affect pricing under imperfect competition. Equation (2) is a pricing equation for the mth product. The market shares of each firm’s seed, used to calculate the Hkms, represent a core strategy dimension for firms P in quantity competition. Equation (2) includes the term M m ¼  k akm H km Y k , which represents the deviation from marginal cost pricing for the mth good. Note that HkmA[0, 1] with Hkm-0 under perfect competition and Hkm ¼ 1 under monopoly. It follows that Mm-0 under competitive markets, and that Mm is expected to depart from 0 under imperfect competition. For the mth good, this departure can be measured by the Lerner index Lm 

pm  c m M m ¼ . pm pm

(3)

Equation (3) shows that the Lerner index Lm, which reflects the departure from marginal cost pricing for the mth good, is given by the ratio Mm/pm. As such, Lm ¼ Mm/pm provides a convenient relative measure of the impact of imperfect competition on pricing. Equations (2) and (3) will be used in our empirical analysis below.

4. Data This study is one from a recent line of research that uses proprietary survey data of farmer seed transactions in the United States over the period 2000– 2007. The data are collected by DMRKYNETIC (hereafter dmrk), St. Louis, MO. The dmrk survey involves a stratified sample of farmers throughout the United States, conducted every year. The samples are stratified to ‘‘oversample’’ producers with large acreage. The survey data are collected using computer-assisted telephone interviewing. The dmrk survey reports the seed price paid by the farmer, the time of purchase, the volume purchased, the planned acreage, the specific seed GM traits, the intended end use, and other useful information. Our analysis examines corn seed pricing in two geographically separate regions of the U.S. Corn Belt. The two regions are the core of the Corn Belt where corn is usually the predominant crop; and the fringe region. The fringe is best described as a corn-producing region that is less productive than the core and with more competing crops grown than in the core. The two regions are shown in Figure 1.

156 Kyle W. Stiegert et al.

Fig. 1.

Map of core and fringe regions of the U.S. Corn Belt.

157

Spatial Pricing of Genetically Modified Hybrid Corn Seeds 80% 70% 60% 50% 40% 30% 20% 10% 0% 2000

2001

2002

2003

2004

2005

2006

2007

Year Conv share

Fig. 2.

Single share

Double Share

Triple share

Quad share

Evolution of relative U.S. acreage planted in conventional and GM corn, 2000–2007.

Using the dmrk data, Figure 2 illustrates the proportions of acreage planted in GM corn over the last decades. It shows a significant decline in conventional hybrids, the initial rise then decline of single-trait GM hybrids, and newly advancing stacked GM hybrids. Our study uses much of this information to evaluate pricing strategies in the fringe and core regions of the U.S. Corn Belt. 5. Estimation Our analysis of regional differences uses Equation (2) to measure the determinants of hybrid corn seed prices in each region. Since seeds are often developed under specific agronomic conditions that vary across geographic regions, we define the local market at the USDA crop reporting districts (CRD) level. We introduce a series of binary variables to capture the specific on-board GM traits present in each hybrid seed transaction. In Equation (2), the price p represents the net seed price paid by farmers (in $ per bag).6 Consider the hedonic model in which the determinants of the price p for a hybrid seed containing possible combinations of characteristics based on the following binary terms: K 1 ¼ No GM traits : conventional hybrid; K 2 ¼ Insect resistance; European corn borer ðECBÞ; 6 Corn hybrids are often sold at a list price less a discount available at the point of sale. In this study, we use the after discount net price. The technological fee associated with the biotech trait is already included in the seed price.

158

Kyle W. Stiegert et al.

K 3 ¼ Insect resistance; rootworm ðRWÞ; K 4 ¼ Herbicide tolerance 1 ðHT1Þ:7 Using the above binary terms, we specify the empirical counterpart to Equation (2) as: p¼bþ

5 X i¼1

þ

di K i þ

5 X 5 X

dij K ij þ

j¼iþ1 i¼1 5 5 5 X 5 X X X

5 5 X 5 X X z¼jþ1 j¼iþ1 i¼1

dijz K ijz (4)

dijzr K ijzr þ uX þ e,

r¼zþ1 z¼jþ1 j¼iþ1 i¼1

where X is a vector of other relevant covariates, and e is an error term with mean zero and constant variance. Each of the Ki binary terms in Equation (4) was described earlier. They signify that a specific trait is present in a single-traited hybrid seed or in a stacked hybrid seed. We introduce additional binary terms to allow for nonlinear pricing in stacked hybrids. In Equation (4), Kij is a binary variable for double-stacking the ith and jth GM traits. Similarly, K ijz andK ijzr are binary terms for triple-stacked and quadruple-stacked GM hybrids.8 Because the underlying single-trait binaries are turned on in the presence of stacking, this means that the coefficients dijz ; and dijzr in Equation (4) measure the effects of bundling on seed price relative to standard component pricing. If they are positive (negative), the model reveals superadditive (subadditive) pricing in the stacked trait market. When component pricing prevails, they are not different from zero, and the price of the hybrid is just the sum of the value of its genetic components. Following Equation (2), the role of market structure is introduced in Equation we let di ¼ d 0i þ d i H ii ; where H ii  P n n (4) in two ways. First, n S S is the traditional HHI (S being the market share of the nth firm in i n i i the market for the ith trait) measuring market concentration in the ith P P P P market. Second, we let b ¼ b0 þ 5j¼iþ1 5i¼1 bij H ij K i þ 5j¼iþ1 5i¼1 bji P n n H ji K j , where H ij  H ji  n S i S j , allowing the cross-GHHIs to affect prices across markets. With this specification, the coefficient of the traditional HHI, d i a0, would reflect market power related to the ith trait, while the coefficient of the GHHI,bij a0 and bji a0 would reflect the exercise of market power across traits. Furthermore, since the exercise of market 7 Our data suggest that farmers purchase seeds with two herbicide tolerance traits inserted. We interpret the stacking system as implying that farmers view these two traits as being differentiated, and label them as HT1 and HT2. P PP PPP PPPP 8 The Ks in Equation (4) satisfy K i  K ij  2 K ijz  3 K ijzr ¼ 1, implying that they are perfectly collinear with the intercept. Thus, we set d1 ¼ 0 in Equation (4) to capture the price of conventional seeds in the intercept. The other d parameters relect price differences relative to conventional seeds.

Spatial Pricing of Genetically Modified Hybrid Corn Seeds

159

power across markets may not necessarily be symmetric, we may expect the coefficient of the GHHI to be different when it is associated with the ith market (bij ) versus the jth market (bji ). We will test the null hypothesis of symmetric restriction: bij ¼ bji , later to see whether we need to impose the symmetric restriction in some of the markets. And from Equations (2) to (3), it follows that the market power P component of the ith seed price in Equation (4) is given by M i ¼ d i H ii K i þ jai bij H ij K j . This will provide a basis for evaluating the effects of imperfect competition across regions (see Section 7 below). The relevant covariates in X in Equation (4) include variables that capture price impacts due to the specific location of the farm within estimated region, a time trend, each farm’s total corn acreage, binary terms covering the range of how each purchase was sourced, and several binary terms to capture the impacts from entry and exit of traited hybrids in various CRDs. Location is represented by state binary variables, along with the longitude and latitude of the county where the farm is located. These variables capture spatial heterogeneity in farming systems and agroclimatic conditions for each regression. If the pricing of seed correlates to a farmers’ willingness to pay, then agro-climatic conditions such as regional soil quality, rainfall, and length of the growing season should impact prices. This may be more pronounced in the fringe regions where corn production conditions vary more than in the core region. Including the spatial variables does two things. First, it allows for a deeper analysis of intra-regional spatial issues. Second, it controls for these factors when comparing the structural market parameters between core and fringe regions. The latitude and longitude variables are specified in both linear and quadratic forms, which allows for a possibly nonlinearity in their effects. Farm acreage captures possible price discrimination effects related to farm size. The aggregate market share of GM hybrids has increased significantly during the years of our study (see Figure 2) while conventional hybrids have consistently declined in use. While some GM hybrids have gained in market share, other have lost market share. Some types of GM hybrids, particularly the stacked hybrids, were not available in all CRDs in all periods. Using the same pretests as did SCS, we found for both regressions evidence of the endogeneity in the Hs, and evidence of heteroscedastic disturbances in the error terms. As a result, following SCS, the model is estimated by instrumental variable estimation and a weighting scheme is used to correct for heteroscedasticity. Additionally, binary terms were introduced to control for either when specific GM hybrids were not yet available, or when they had exited a CRD. The endogeneity of the Hs comes from the fact that prices and quantities are jointly determined as parts of firms’ strategies. The standard

160

Kyle W. Stiegert et al.

solution for endogenous regressors is to estimate the model using an instrumental variable procedure: two-stage least squares (2SLS). Because of production lags, decisions on seed production must be made at least one year ahead of the time when they are sold to farmers. This means that lagged quantities are part of the information set available to seed/biotech firms at the time they make quantity decisions. This suggests that lagged quantities are good candidates for instruments for the Hs. On that basis, we used as instruments one-period lagged value of the market size Ys and the own- or cross-HHIs. To evaluate the validity of these instruments, the models were subjected to tests of overidentification. The Hansen test is not statistically significant in either region indicating that our choice for instruments satisfies the required orthogonality conditions. A series of additional tests suggest that that neither the regressions for the fringe or core regions suffer from weak instruments. As we already discussed, farmers use a menu of hybrids each growing season. In our data, it was common to observe a farmer make purchases of four or more varieties. We consider these purchasing patterns when making the corrections for heteroscedasticity. Specifically, we anticipate a common variance among purchases by the same farmer in a given year. This suggests the use heteroscedastic-robust standard errors with clustering at the farm level. Summary statistics of the data used for each region are presented in a comparative fashion in Tables 2 and 3. Table 2 shows that average prices of conventional hybrids in the core region are higher in all years compared to the average prices for conventional hybrids sold in the fringe region. This same pattern is generally present in the markets for insect-resistant single-trait hybrids and for triple- and quadruple-stacked hybrids. At the annual average level, the regional pricing differences for double-stacked hybrids are quite small. For single-trait herbicide tolerant hybrids, the tendency is for higher average prices in the fringe region. These pricing patterns can be explained along several lines. The supplies of GM and conventionally produced hybrids are determined by seed and biotech firms in the previous crop year, which means they are likely trying to meet anticipated demand for specific types of seed at profitable prices. When supplies are large (small) relative to the actual demand, prices are likely adjusted lower (higher). As discussed earlier in conjunction with subadditive pricing, our aggregate price differences across regions may be in part driven by strategic factors by oligopoly seed/biotech firms. Pricing is likely to impact the rate of adoption, which affects the aggregate revenue and profitability of a GM technology over the patent life. If fundamentally different products are sold under different demand elasticity scenario, different pricing patterns are likely to emerge along spatial and product lines. Table 2 contains other summary statistics for the core and fringe regions. Notice that average farm size is larger in the core. Though not

161

Spatial Pricing of Genetically Modified Hybrid Corn Seeds

Table 2.

Average nominal net price for different seeds ($ per bag), 2000–2007

Year Conventional

IR single

HT single

Double

Triple

Quadruple

Core Fringe Core Fringe Core Fringe Core Fringe Core Fringe Core Fringe 2000 2001 2002 2003 2004 2005 2006 2007 Total

81.82 83.04 84.02 86.10 88.56 89.97 93.42 95.91 86.87

79.86 80.78 81.77 83.97 86.04 88.29 91.73 93.26 84.46

104.36 108.15 107.66 107.78 111.01 107.24 113.59 115.00 109.14

100.51 105.32 104.16 105.13 109.95 107.28 112.80 110.89 106.49

86.23 88.30 86.78 94.14 98.59 102.88 113.20 119.28 103.10

Table 3. Variable

Number of observationsa,b Core

Price ($) Farm size (acre) Longitude Latitude H11 H22 H33 H44 H12 H13 H14 H23 H24 H34

Fringe

91.19 92.42 92.24 96.44 101.16 105.92 113.68 117.65 104.85

95.60 102.20 103.43 112.19 115.89 116.94 125.34 126.76 121.44

98.66 100.52 104.93 108.84 114.30 116.18 124.59 125.80 118.42

105.07 115.02 96.06 83.3 116.68 123.52 144.64 137.08 137.62

95.87 99.45 91.45 96.11 108.09 127.68 144.76 134.41 135.85

NA NA NA NA NA NA 151.04 144.21 144.55

NA NA NA NA NA NA 133.14 143.29 142.91

Summary statistics

Mean

Standard deviation

Minimum

Core Fringe Core Fringe Core

Fringe

Maximum

Core Fringe

69267 13242

70140 17030

101.36 97.90 23.51 23.59 8 553.02 440.08 605.75 568.73 5

3 5

200 230 10300 15500

13242 13242 176 176 112 176 176 112 176 112 176 112

17030 17030 464 464 192 464 440 192 440 192 456 192

92.24 41.31 0.174 0.798 0.924 0.689 0.100 0.088 0.057 0.814 0.479 0.639

80.75 36.71 0.084 0.337 0.976 0.500 9.90E-05 0.011 5.52E-04 0.311 0.032 0.171

98.07 43.38 0.429 1 1 0.945 0.470 0.184 0.170 0.991 0.816 0.921

91.09 42.03 0.230 0.810 0.9986 0.767 0.095 0.124 0.082 0.813 0.586 0.867

3.28 1.24 0.092 0.163 0.139 0.155 0.093 0.048 0.049 0.155 0.229 0.265

5.63 2.40 0.133 0.183 0.005 0.159 0.084 0.108 0.086 0.185 0.295 0.214

86.65 38.01 0.076 0.430 0.533 0.500 0.022 0.023 0.005 0.501 0.053 0.056

103.76 46.98 0.836 1 1 1 0.431 0.444 0.443 1 1 1

a

The whole data contain 1,39,407 observations from 80 CRDs spanning 8 years (2000–2007). Each farm purchases multiple seeds; therefore, the number of observations for farm size is the total count of farms per year. The longitude and latitude information is based on the county level measurement for each farm. b For the market concentration measurements Hs, we only report the summary statistics of those nonzeros at the CRD level; therefore, the number of observations is at most 22  8 ¼ 176 (core), or 58  8 ¼ 464 (fringe).

shown, the coefficient of variation for farm size suggests that more variability exists in the fringe. This is consistent with the idea that farms in the fringe region are likely to have different cropping strategies than the relatively more homogeneous farms in core region. The traditional HHIs

162

Kyle W. Stiegert et al.

(the Hiis) suggest that the biotech hybrid markets are extremely concentrated and that the core and fringe regions are not that much different. There is plenty of variation across CRDs within each region; therefore, our regressions should be able to unravel any differences in way market power emerges across the regions. The concentration rate for conventional hybrids (H11) indicates that the fringe region is considerably more concentrated than the core and is well above the 0.18 threshold suggested by the Department of Justice for a highly concentrated region. Finally, we note a great deal of variation in some of the cross-GHHIs (i.e., the Hijs ) in the fringe versus the core and no obvious pattern emerging.

6. Econometric results The econometric results from the 2SLS estimation of Equation (4) are presented in Table 4. On the basis of a simple Wald test, we imposed a symmetry restriction for the parameters of the cross-GHHIs for conventional and corn borer hybrids (e.g., b13 ¼ b31 ). The table is split into three sections. The respective sections contain parameter estimates for the core region regression, the fringe region regression, and then a test of whether the regression coefficients are statistically different between the two regressions. Our discussion follows sequentially through each part. In the next section, we use simulation methods to evaluate the effects of changing market structures in the north fringe, south fringe, and core region of the U.S. Corn Belt. The first section of Table 4 presents the binary terms that capture the price impacts of individual and stacked GM traits. The first set of coefficients shows that prices vary significantly both across seed types and across regions. The first four rows report the direct effect on price due to the presence of a trait in a GM hybrid. These coefficients are measured against the base price of conventional seed. Thus, for example, when the European corn borer trait is present in a GM seed, the impact is to raise price by $19.36/ bag in the core region and $21.13/bag in the fringe region. While both values represent statistically significant price increases above the conventional hybrid seed price, they are not statistically different from each other. Many coefficients of Kij and Kijz are negative and statistically significant in both regions. This is evidence of ‘‘subadditive pricing,’’ which means that price premiums of GM hybrids tend to increase at a decreasing rate as additional traits are added. Such a price pattern possibly reflects economies of scope in the development and production of GM hybrid seed types. It means that farmers who want access to multiple patented genes can do so at a lower cost than if the market had component pricing or superadditive pricing. In the core region, there is one case of superadditive pricing (K35) and two cases of component pricing (K25, K235). In the fringe region, all stacked seeds followed subadditive pricing.

Table 4. Dependent variable: price ($/bag)

2SLS regression with robust standard errorsa,b Core

Fringe

Difference

Coefficient Robust z Coefficient Robust z Coefficient tstatistics statistics ratio Characteristic and bundling effects, benchmark is K1: conventional seed 19.36*** 5.54 21.13*** 3.81 K2 (ECB) 30.30*** 5.32 113.8** 2.21 K3 (RW) 3.87 0.48 7.94 1.28 K4 (HT1) 5.58*** 4.10 13.90*** 7.37 K5 (HT2) ** 4.63 2.46 25.17*** 3.04 K23 9.65*** 6.67 17.36*** 8.37 K24 0.44 0.31 11.04*** 5.82 K25 7.53*** 3.95 26.56*** 4.76 K34 *** 9.22 4.16 23.45** 2.47 K35 20.16*** 6.60 50.44*** 4.75 K234 4.68 1.53 49.35*** 3.07 K235 15.97*** 5.77 30.12*** 7.85 K245 ** 9.26 2.31 36.97*** 3.57 K345 21.9*** 5.16 67.17*** 4.46 K2345 Market concentration effects 15.80*** 3.21 11.05*** 3.33 H11(conventional) ** 7.04 2.30 5.19 0.74 H22(ECB) 76.96*** 5.08 31.91 0.64 H33(RW) 9.57 0.77 3.72 0.45 H44(HT1) 33.01*** 2.97 135.62** 2.00 H12 on conventional 2.58 0.21 1.85 0.12 H21 on ECB trait 12.27 0.48 43.75 1.52 H13 on conventional seed/RW trait 1.24 115.41* 1.69 H14 on conventional seed 28.52 80.82*** 2.60 23.84* 1.66 H41 on HT1 trait 9.23*** 3.24 1.58 0.54 H23 on ECB trait 55.98*** 4.55 45.85*** 2.77 H32 on RW trait 5.44 1.55 1.82 0.45 H24 on ECB trait 11.96*** 2.55 8.52** 2.05 H42 on HT1 trait 32.51*** 3.49 15.77 1.43 H34 on RW trait 0.42 0.11 9.08*** 2.90 H43 on HT1 trait Other variables Post-exit1 2.95 0.61 Pre-entry2 3.64 0.90 Pre-entry3 2.49 1.33 2.53 1.29 Pre-entry4 3.97 1.16 0.001*** 5.35 Total farm corn acreage 0.001*** 2.03 (1,000 acres) Longitude 1.09 1.61 0.65* 1.66 2.04 Longitude squared 0.03 1.22 0.03** 2.44 Latitude 0.75 0.71 1.58** Latitude squared 0.07 0.70 0.15*** 3.04 7.18 2.16*** 8.49 Year 1.91*** *** 5.46 69.03*** 16.92 Constant 72.73 No. of observations 62,141 61718 a

1.77 0.27 83.50 1.61 11.81 1.16 8.32*** 3.58 20.54** 2.42 7.71*** 3.05 10.60*** 4.48 19.03*** 3.23 32.67*** 3.35 30.28*** 2.74 44.67*** 2.73 14.15*** 2.99 27.71*** 2.50 45.32*** 2.90 4.74 0.80 1.85 0.24 108.87** 2.08 5.85 0.39 102.60 1.49 0.73 0.04 56.03 1.46 86.90 1.21 56.98* 1.66 10.81*** 2.64 10.13 0.49 3.62 0.68 3.44 0.55 16.75 1.16 8.66* 1.77

Statistical significance: * at the 10% level, ** at the 5% level, *** at the 1% level. To save space, results for the location effects and purchase source effects are not reported here.

b

164

Kyle W. Stiegert et al.

Importantly, the coefficients of Kij and Kijz differ significantly across regions: they tend to be more negative in the fringe region as compared to the core (see first and third columns of Table 4). This indicates that the extent of subadditive pricing of corn hybrids is stronger in the fringe. It represents major finding and provides evidence of spatial price discrimination. One plausible explanation is that corn operations in the fringe regions are less productive leaving farmers with a lower willingness to pay for premium priced GM hybrids. If faced with high hybrid prices, farmers in the fringe may also have more options allowing them to switch to different crops. Turning next to the coefficients on market concentrations, interesting and key findings emerge. The first four rows in the second section of Table 4 contain the partial effects emanating from the traditional Herfindahl indices (Hii).9 Three of the four own-HHIs in the core region have a positive and significant effect on price. In contrast, only the conventional hybrid own-HHI is significant in the fringe region. While being positive and significant in both regions, the own-HHI effect on conventional hybrids is found to be similar across regions: the difference in coefficients is not significant. Table 4 shows that a one-point increase in H11 in the core (fringe) region is associated with a $15.80 ($11.05) per bag increase in the price of conventional hybrids. For RW seeds, the own-HHI coefficient is significantly higher in the core than in the fringe region. In general, the results presented in Table 4 indicate that the prospects for own-market concentration to increase GM seed prices are higher in the core region. From our discussion in Section 3, this indicates that the demand for GM seeds tends to be more inelastic in the core than in the fringe region of the Corn Belt. Again, this may be due to two facts. First, corn productivity tends to be higher in the core region. Second, farmers may have greater options to plant different crops in the fringe region. Table 4 reports three cross-GHHI terms that measure the impact of other product concentration on conventional hybrid prices (H12, H13, H14). Of these three terms, only H12 is significant in both regions. Additionally, both are positive suggesting that the coefficients on H12 reflect substitutes in demand. From the discussion in Section 3, this means that hybrid seed companies may be able to extract additional rent from the conventional hybrid market where there is jointly greater concentration in the corn borer trait market and in the conventional market. Interestingly, the reverse is not true: the coefficient on H21 is small and insignificant in both regions. The coefficients on H13 are insignificant in both regions. Due to the symmetry restriction, the coefficient on the H13 term also measures 9 Because HT2 is exclusive to a single firm and the firm does not operate across market, the own-HHI (with value 1 when present) and all cross-HHIs involving HT2 (with value zero) are dropped from the analysis.

Spatial Pricing of Genetically Modified Hybrid Corn Seeds

165

the price impact from concentration in the conventional market on the price of hybrids with rootworm protection. We found that the impact of higher concentrations in the HT1 and conventional markets (H14) contributes to lowering the price for conventional seed, but only for the fringe region. This suggests that a higher concentration in the HT1 GM hybrid market puts downward pressure on conventional hybrid prices. Our simulations presented in the next section provide additional insights on this point. Interestingly, the results are quite different when evaluating the same cross-GHHI’s impact on the HT1 market price. Here, the higher joint concentration in these markets leads to higher prices in the HT1 GM hybrid market in both core and fringe region. They are statistically different in each region with a much larger price impact in the core compared to the fringe. Our remaining cross-GHHI terms provide several key results for these markets. The relationship between corn borer and rootworm traits is largely complementary, leading to lower prices as the markets become jointly concentrated. This result is more prevalent in the core region, particularly for the corn borer market. There is evidence of a substitution relationship in the cross-market effects between corn borer and HT1 traits on the HT1 market price. This means prices tend to rise in the HT1 markets when the joined concentration rises. For the cross-market effects between rootworm and HT1 (i.e., H34 and H43), we observe a complementary effect on the core region for rootworm market while there is a substitution effect in the fringe region for the HT1 market.

7. Implications for spatial pricing Using the econometric estimates in Table 4, this section evaluates how the effects of market power on prices vary across regions. This is done through two sets of simulations. First, we evaluate the Lerner index given in Equation (3). Second, we evaluate how changing market structure affects pricing both over time and over space. We consider three regions of the Corn Belt: the core, the southern-fringe, and the northern-fringe region. Selected measures of market concentrations are presented in Table 5. They are presented for the years 2003 and 2006.10 Table 5 shows that levels of market concentration can vary across both time and space. One striking difference across regions is shown in the first row involving the own-HHIs for conventional hybrids. Note that the levels of concentration are considerably higher in the two fringe regions for conventional seeds and 10 We chose 2003 and 2006 for simulation purpose for three reasons: (1) the RW trait was not present in the market prior to 2003; (2) the total acreage of 2003 and 2006 are similar, thus avoiding any potential size effects; (3) the market concentrations demonstrate fair amount of changes between these two years (see Table 5).

166

Kyle W. Stiegert et al.

Table 5.

Own- and cross-market concentration in 2003 and 2006

Name

H11 H22 H33 H44 H12 H13 H14 H23 H24 H34

(conventional) (ECB) (RW) (HT1)

Core

Southern fringe

2003

2006

0.147 0.841 1.0 0.684 0.065 0.086 0.048 0.897 0.566 0.644

0.214 0.607 0.835 0.836 0.095 0.101 0.104 0.667 0.666 0.797

2003 0.263 0.785 1.0 0.779 0.051 0.041 0.045 0.771 0.519 0.658

2006 0.313 0.732 0.934 0.755 0.079 0.085 0.076 0.777 0.678 0.787

Northern fringe 2003 0.238 0.651 1.0 0.879 0.063 0.109 0.084 0.818 0.699 0.947

2006 0.276 0.676 0.941 0.941 0.115 0.124 0.126 0.742 0.744 0.862

that the core region’s HHI has increased substantially between 2003 and 2006. For GM hybrids, the own-HHIs display some variability and are all suggestive of very high rates of concentration. Note finally that crossHHIs involving conventional hybrids are small while the cross-HHIs exclusive to GM hybrids are all large. We evaluated the Lerner indices given in Equation (3). Expressed in percentage (i.e., L  1 0 0), the results are presented in Table 6 for the year 2006. Table 6 shows that the Lerner index is around 11% for conventional seeds in the fringe region. It is statistically significant. It means that market power contributes to an 11% increase in conventional seed prices in the fringe. In contrast, the Lerner index for conventional seeds is not statistically different from zero in the core. Interestingly, this spatial difference is not due to the own-HHI effects: from Table 4, the H11 effects are not statistically different across regions. It means that, for conventional seeds, the higher Lerner index reported in Table 6 is due to more concentrated markets in the fringe (see Table 5). In this case, spatial differences in market concentration contribute to higher conventional seed prices in the fringe region. In addition, the Lerner indices reported in Table 6 indicate that market power does not have a statistically significant effect on the price of ECB and RW seeds in any region. This contrasts with the Lerner indices of other GM seeds reported in Table 6. For HT1, ECB-HT1, RW-HT1, and ECB-RW-HT1, the Lerner indices are positive and statistically significant in the core: they vary from 17% for RW-HT1 to 24% for HT1. This indicates that market power contributes to significant price increases (between 17 and 22%) in the core region of the Corn Belt. For the fringe region, these effects are also present for HT1 and ECB-HT1. Note, however, that the HT1 Lerner indices in the fringe regions are 20%–25% smaller than the measured impact in the core region. Additionally, Lerner

167

Spatial Pricing of Genetically Modified Hybrid Corn Seeds

Table 6. Name

Core Lerner index

Conventional ECB RW HT1 ECB-RW ECB-HT1 RW-HT1 ECB-RW-HT1 a

Simulated Lerner indices, 2006a

2.67 1.75 1.82 24.19*** 0.23 22.03*** 17.26** 17.46***

Statistical significance:

*

SE t-ratio

Southern fringe Lerner index

1.98 1.35 10.81*** 3.09 0.57 5.68 4.92 0.37 66.73 7.72 3.13 17.87*** 4.14 0.06 55.47 6.45 3.41 20.02*** 7.90 2.19 47.27 6.74 2.59 38.68 at the 10% level,

**

Northern fringe

SE.

t-ratio

Lerner index

SE

t-ratio

2.67 6.05 43.66 4.52 35.14 5.83 40.85 33.22

4.05 0.94 1.53 3.95 1.58 3.44 1.16 1.16

11.23*** 5.53 60.16 19.62*** 49.87 22.27*** 40.89 33.93

2.87 5.96 38.06 5.16 30.36 6.13 36.78 30.42

3.912 0.93 1.58 3.80 1.64 3.64 1.11 1.12

at the 5% level,

***

at the 1% level.

indices are not statistically significant for RW-HT1 and ECB-RW-HT1 in the fringe. Thus, the evidence suggests that for several GM seeds, the exercise of market power is less pronounced in the fringe region. This reflects in part the differences in market concentrations reported in Table 5. And as discussed above, this also reflects different patterns of substitution/ complementarity across regions. Finally, we evaluate how the role of market power can change over space. Using our econometric estimates reported in Table 4, we evaluated seed prices in three regions of the Corn Belt: the core, the fringe-North and the fringe-South. Recall from Equations (2) to (4) that the market power P component of the ith price pi, is given by M i ¼ d i H ii K i þ jai bij H ij K j , which is the sum ofPtwo terms: the own-market effect d i H ii K i and the cross-market effect jai bij H ij K j . Under market conditions observed in 2003 and 2006, we simulated prices for each region under two scenarios: scenario A held the cross-market effect constant at its 2003 level while the own-market concentrations changed to 2006 level; scenario B allowed both own- and cross-market effects to change. The simulated regional prices under each scenario are reported in Table 7, where part A captures only the own-HHI effects, while part B captures both own- and cross-GHHI effects. This provides useful information on the spatial and temporal factors influencing the exercise of market power. Table 7 shows that major differences exist across regions. Part A shows that changes in own-market concentrations induced a significant increase in the price of conventional seeds everywhere, and a significant decline in the price of single-traited GM seeds but only in the core. As noted above, these simulated results neglect the cross-GHHI effects. Part B, which captures both own-HHI and cross-GHHI effects, underscores the importance of cross-market effects in the analysis. Part B shows that including the cross-market impact reduces the effects of changing market concentrations on conventional seeds in the core, up to a point where the

168

Kyle W. Stiegert et al.

Table 7.

Simulated effects of changing market concentrationsa

Name

Core 2003

2006

Southern fringe

Difference 2003

Part A: Own-HHI effects Conventional 91.93 92.99 1.06*** 88.74 ECB 110.12 108.5 1.62** 109.69 RW 124.81 110.12 14.69*** 114.6 HT1 103.41 104.92 1.51 104.54 Part B: Own- and cross-product HHI effects Conventional 91.93 92.14 0.21 88.74 ECB 110.12 111.22 1.1 109.69 RW 124.81 117.92 6.89*** 114.6 HT1 103.41 110.7 7.29*** 104.54 ECB-RW 140.37 134.58 5.78** 117.15 ECB-HT1 114.02 122.40 8.39*** 114.56 RW-HT1 130.81 131.21 0.40 110.37 a

2006

Northern fringe

Difference 2003

2006

Difference

89.3 0.56*** 109.45 0.25 116.53 1.93 104.44 0.10

89.49 110.39 113.97 110.92

89.91 110.51 117.42 111.18

0.42*** 0.12 3.45 0.26

2.69*** 0.11 1.84 3.17*** 1.94 3.28*** 5.01

89.49 110.39 113.97 110.92 116.32 120.75 115.24

92.7 110.56 119.38 111.82 121.89 121.80 121.54

3.2*** 0.16 5.41 0.89 5.57 1.05 6.30

91.43 109.8 116.44 107.71 119.09 117.84 115.38

Statistical significance is noted by * at the 10% level,

**

at the 5% level,

***

at the 1% level.

effect is no longer statistically significant. In the core, Table 7 also shows that cross-market impacts reduce the (negative) effect on RW price but make the HT1 effect positive and significant (þ$7.3/bag). And in the fringe region, cross-market effects contribute positively to enhancing the price of conventional seeds: þ$2.69/bag in the fringe-South and þ$3.2/bag in the fringe-North. Interestingly, such positive effects are found to be stronger in the fringe-South (about þ$3.2/bag) than in the fringe-North for both HT1 and ECB-HT1. This documents significant differences in pricing across regions. These results show the usefulness of our analysis. For example, the simulation of alternative market structures could be used to evaluate the likely effects of a merger in one or more of the regions. This could be done by calculating the expected post-merger GHHI, and then simulate the price impacts using the regression estimates. 8. Concluding remarks The research in this chapter focuses on seed pricing and the effects of market concentration for corn hybrid seeds in the U.S. Corn Belt. Specifically, we evaluate pricing differences of hybrid corn seeds in the core and fringe regions of the U.S. Corn Belt. The model considers the spatial differences in the bundling of traits, regional-specific oligopoly structures, and other differences in drawing inferences about each market for hybrid corn seeds. We study the factors that may contribute to spatial price discrimination.

Spatial Pricing of Genetically Modified Hybrid Corn Seeds

169

For each region, differentiated hybrid corn seed pricing is modeled in a hedonic structure that depends on the patented genetic modifications present in each hybrid. This includes no modifications (i.e., conventional hybrids), herbicide tolerance, corn rootworm resistance, European corn borer resistance, and hybrids with stacked traits. Using the GHHI structure developed in SCS (2010a), the model also includes the effects of own- and cross-market concentration on prices. The analysis is conducted using survey data of seed transactions in the United States over the period 2000–2007. Our analysis generated several useful results on how space affects corn seed prices. First, we find statistically and economically greater levels of subadditive pricing in the fringe region of the Corn Belt. These results may reflect that, in the fringe region, corn productivity is lower and/or farmers have greater leverage in price negotiations (compared to farmers in the core region). Such leverage could be due to a greater willingness to switch out of corn production in favor of other crops and/or to lower reservation prices for corn hybrids owing to lower returns in corn production relative to the core region. Second, this study documented how the exercise of market power varies across space. Two sets of factors played a role. First, market concentration varied across space. For example, we showed how higher concentration in the conventional seed market contributed to greater price enhancement for conventional seeds in the fringe of the Corn Belt (compared to the core). Second, our econometric analysis showed that the effects of market concentration on prices varied across space. We attribute these differences to different demand elasticities and different patterns of substitution/ complementarity between seed types. The first effects were captured in our analysis by traditional HHI measures. The second effects were measured by our GHHI capturing cross-market effects. A simulation exercise illustrated how these effects translate into impact on seed prices. Using a Lerner index, we estimated that the exercise of market power was not statistically significant for ECB and RW seeds in any region. We also found price-enhancing effects of market power to be significant for HT1 and ECB-HT1 in each region. And we found price-enhancing effects to be significant for RW-HT1 and ECB-RW-HT1 but only in the core. Our analysis of spatial price differences could be extended in several ways. First, our spatial delineation of the core and fringe regions was somewhat arbitrary. Additional analysis could be conducted to evaluate empirically theories of spatial competition (i.e., Lo¨sch, Greenhut–Ohta, Hotelling–Smithies) in a manner similar to Mulligan and Fik (1994). Second, there is likely to be useful statistical information across spatial regions, which could be exploited using a joint estimation approach. Finally, with panel data of seed purchases by select farmers, the demand side could be jointly estimated with our supply side model to yield a broader understanding of how pricing strategies impact adoption patterns.

170

Kyle W. Stiegert et al.

Acknowledgments This research was funded in part by USDA-NRI grant #144-QS50, a USDA Hatch grant, and from funding through the Food System Research Group. The authors acknowledge research assistance support from Narek Sahakyan.

References Acquaye, A.K.A., Traxler, G. (2005), Monopoly power, price discrimination and access to biotechnology innovations. AgBioForum 8, 127–133. Gouse, M., Pray, C., Schimmelpfennig, D.E. (2004), The distribution of benefits from Bt cotton adoption in South Africa. AgBioForum 7, 187–194. Hicks, J.R. (1939), Value and Capital: An Inquiry into some Fundamental Principles of Economic Theory. Clarendon Press, Oxford. James, C. (2009), Global status of commercialized biotech/GM crops: 2009. Brief No. 41-2009. International Service for the Acquisition of Agri-Biotech Applications, Ithaca, NY. Jefferson-Moore, K.Y., Traxler, G. (2005), Second-generation GMOs: where to from here? AgBioForum 8, 143–150. Kalaitzandonakes, N. (1999), A farm-level perspective on agrobiotechnology: how much value and for whom. AgBioForum 2, 61–64. Liu, E.M. (2008), Time to change what to sow: risk preferences and technology adoption decisions of cotton farmers in China. Working paper No. 526, Princeton University, Princeton, NJ. Marra, M.C., Pardey, P.G., Alston, J.M. (2002), The payoffs to transgenic field crops: an assessment of the evidence. AgBioForum 5, 43–50. Moschini, G. (2008), Biotechnology and the development of food markets: retrospect and prospects. European Review of Agricultural Economics 35 (4), 331–355. Moschini, G., Lapan, H. (1997), Intellectual property rights and the welfare effects of agricultural R&D. American Journal of Agricultural Economics 79 (4), 1229–1242. Mulligan, G.F., Fik, T.J. (1994), Price and location conjectures in medium and long-run spatial competition models. Journal of Regional Science 34 (2), 179–198. Price, G.K., Lin, W., Falck-Zepeda, J.B., Fernandez-Cornejo, J. (2003), Size and distribution of market benefits from adopting biotech crops. Technical Bulletin No. 1906. Economic Research Service, USDA, Washington, DC. Qaim, M., De Janvry, A. (2003), Genetically modified crops, corporate pricing strategies, and farmers’ adoption: the case of Bt cotton in Argentina. American Journal of Agricultural Economics 85 (5), 814–828.

Spatial Pricing of Genetically Modified Hybrid Corn Seeds

171

Qaim, M., Traxler, G. (2005), Roundup ready soybeans in Argentina: farm level and aggregate welfare effects. Agricultural Economics 32, 73–86. Shi, G., Chavas, J.P., Stiegert, K. (2010a), An analysis of the pricing of traits in the U.S. corn seed market. American Journal of Agricultural Economics. doi: 10.1093/ajae/aaq063. Shi, G., Chavas, J.P., Stiegert, K. (2010b), Bundling and bundle pricing: the case of the corn seed market. Food System Research Group Working Paper FSRG2010-07. Available at http://www.aae.wisc.edu/ fsrg/publications/wp2010-07.pdf. Shi, G., Stiegert, K., Chavas, J.P. (2011), An analysis of bundle pricing in horizontal and vertical markets: the case of the U.S. cottonseed market. Agricultural Economics, forthcoming. Sunding, D., Zilberman, D. (2001), The agricultural innovations process: Research and technology adoption in a changing agricultural sector. In: Gardner, B.L., Rausser, G.C. (Eds.), Handbook of Agricultural Economics. Elsevier, Vol. 1A, Ch. 4, pp. 207–261. Useche, P., Barham, B., Foltz, J. (2009), Integrating technology traits and producer heterogeneity: a mixed-multinomial model of genetically modified corn adoption. American Journal of Agricultural Economics 91, 444–461. Whinston, M.D. (2006), Lectures in Antitrust Economics. MIT Press, 2006, Cambridge, MA.

CHAPTER 7

The Environmental Benefits and Costs of Genetically Modified (GM) Crops Justus Wesselera, Sara Scatastab and El Hadji Fallc a

Center of Life and Food Sciences Weihenstephan, Technische Universita¨t Mu¨nchen, Weihenstephaner Steig 22, 85354 Freising, Germany E-mail address: [email protected] b Rural Development Theory and Policy, Universita¨t Hohenheim, Wollgrasweg 45, 70593 Stuttgart, Germany E-mail address: [email protected] c UNDP, 5 Boulevard de l’Est, Point E, Dakar, Senegal E-mail address: [email protected]

Abstract The widespread introduction of genetically modified (GM) crops may change the effect of agriculture on the environment. The magnitude and direction of expected effects are still being hotly debated, and the interests served in this discussion arena are often far from those of science and social welfare maximization. This chapter proposes that GM crops have net positive environmental effects, while regulatory responses focus mainly on environmental concerns, giving an unbalanced picture of the regulatory context. This unbalance supports the hypothesis that environmental concerns about GM crops have been politically instrumentalized and that more attention should be paid to regulatory responses considering the environmental benefits of this technology. It is also argued that a number of environmental effects have not yet been quantified and more research is needed in this direction. Keywords: Biodiversity, environmental cost-benefit-analysis, externalities, genetically modified crops, pesticide use JEL Classifications: O32, Q16, Q18, Q28, Q5

1. Introduction The cultivation of agricultural crops has effects on the environment, including changes in the pollution of water resources due to changes in Frontiers of Economics and Globalization Volume 10 ISSN: 1574-8715 DOI: 10.1108/S1574-8715(2011)0000010012

r 2011 by Emerald Group Publishing Limited. All rights reserved

174

Justus Wesseler et al.

pesticide or fertilizer use, changes in biodiversity and agrobiodiversity, changes in the emission of greenhouse gases (GHGs) through cultivation of soils and burning of fossil fuels, and changes in soil erosion by wind and water. The introduction of genetically modified (GM) crops is changing the environmental footprint of agriculture. Early concerns about substantial, negative implications for the environment from GM crops (Krimsky and Wrubel, 1996; Kendall et al., 1997) have not been confirmed, but have triggered a number of regulatory responses and are one of the major reasons put forward by the EU (EU Environment Council, 1999) for its quasi moratorium on genetically modified organisms (GMOs) as well as the implementation of the Cartagena Protocol on Biosafety to the Convention on Biological Conservation. The economic implications of these concerns are far-reaching and complex. Some authors have argued that these measures are not so much generated out of concern for the environment, but are rather due to fear of ‘‘Big Pharma’’ (Winston, 2002) and are embedded in the political economy of regulation (Graff et al., 2009) or international public and private aid (Paarlberg, 2008; Herring, 2009), where environmental concerns are instrumentalized to achieve other objectives. While environmental concerns around the introduction of GM crops have been partially addressed and assessed in several reviews (see, e.g., Fontes et al., 2002; Mellon and Rissler, 2003; Benbrook, 2009), the literature providing an overview of the actual and potential benefits of GM crops is sparse. A notable exception is the National Research Council Report (2010), but the report concentrates on U.S. agriculture. With this chapter we want to fill this gap and address different kinds of environmental affects, paying attention to the benefit side of the story. In this contribution we will focus on a number of effects of GM crops on pesticide and fertilizer use, on the emission of GHGs, and on soil erosion as well as agrobiodiversity and biodiversity. Special attention will be paid to the environmental benefits of herbicide-resistant (HR) crops, about which some recent studies have shown greater environmental benefits than mentioned in earlier studies. We show that many environmental benefits have not yet been assessed from an environmental economics point of view and argue that environmental economists should start paying more attention to the strong evidence available on the environmental benefits of GM crops, in order to improve the regulatory understanding of this technology. The contribution is organized by first establishing the theoretical framework in Section 2. Section 3 presents and discusses the evidence on environmental benefits of GM crops. In Section 4, the environmental issues are discussed in the wider policy context of new technologies.

The Environmental Benefits and Costs of GM Crops

175

2. Theoretical framework for assessing the environmental benefits of GM crops Environmental impacts have socioeconomic, temporal and ecological dimensions. From a socioeconomic point of view, such impacts can be broadly categorized based on whether they are private or external. When impacts are private, their socioeconomic value will be captured in market transactions, and technology adoption will be at the social optimum. The same might not hold for externalities, the presence of which can be seen as an effect of market failure that justifies government intervention aimed at reaching the socially optimal amount of technology adoption, as has been the case with GM crops. Environmental impacts have a temporal dimension: they can be short term or long term, actual or potential, and reversible or irreversible. As technology adoption also has a temporal dimension – it can be immediate or postponed – the temporal dimension of environmental impacts plays a key role in technology adoption decisions and regulatory responses (Mooney and Klein, 1999; Morel et al., 2003; Demont et al., 2004; Ervin and Welsh, 2005; Wesseler et al., 2007; Wesseler, 2009). Environmental impacts have an ecological dimension characterized by the environmental media (soil, water, air, climate, vegetation, and biota) and the ecological function (regulation, habitat, production, and information) affected (de Groot et al., 2002). Changes in GHG emissions, for example, can affect the climate regulating function. Changes in insecticide and herbicide use can have an effect on soil productivity and the regulation function of vegetation and biota. Environmental impacts of agrobiotechnology adoption are caused by changes in agronomic practices and cropping systems through input substitution or changes in input use efficiency. In particular, an increase in input use efficiency may yield a conversion of saved resources to other uses (Kalaitzandonakes, 2003). In this respect, environmental impacts of agriculture biotechnology can be considered to be direct when stemming from input substitution or changes in input use efficiency. Indirect impacts of the technology derive instead from the conversion of saved resources to other uses. Further, the chain of agricultural biotechnology development is important. The public and the private sectors invest resources in the development and use of knowledge to produce agricultural crops with new traits. Those new crops are sold to farmers who plant them and sell the harvest to the downstream sector, which further manufactures the products until they finally reach the end consumer via retailers. Important feedback mechanisms are needed to be included in this chain. Consumers, often via advocacy groups, send signals via retailers about their food preferences back to the farm sector as well as directly to technology providers, who also receive them from the farm sector as well.

176

Justus Wesseler et al.

Additionally, information between agents in the chain influences whether and how a new GM crop and derived food products will be successfully introduced and, hence, their environmental impact. The rules and regulations that national governments and international organizations use to govern the release of GM crops also influence the behavior of agents within the chain and, consequently, environmental impacts. Such rules and regulations do not appear out of the blue; in fact, they are made by humans who act in their own interest. The political economy of deciding about rules and regulations adds another dimension of complexity, making the issue inherently dynamic (Kealey, 1998; Shleifer, 2010). Despite the complexity and dynamics, looking at the environmental benefits and costs of GM crops at the farm level and investigating welfare implications in a comparatively static way is a good starting point and will help to clarify a number of issues that are also relevant for assessing the empirical evidence regarding environmental benefits and costs. Assessing environmental benefits and costs can best be done by differentiating between private and external ones as well as differentiating between irreversible and reversible ones (Mooney and Klein, 1999; Demont et al., 2004; Ervin and Welsh, 2005; Wesseler et al., 2007). Within a competitive market, along the supply chain – including technology providers and farm-level producers on the supply side and processors, retailers and final consumers on the demand side – the marginal benefits of GM crops are derived from the sum of the individual willingnessto-pay, while the marginal private costs of supply are derived form the sum of the individual supply. The difference between marginal private benefits and marginal private costs result in the net marginal benefits (NMB). Environmental benefits and costs not yet internalized (e.g., via regulations) by private behavior are considered to be external. The social marginal net benefits (SMNB) of GM crops are the sum of the individual marginal benefits and costs. The social marginal externality can either be positive or negative, depending on the GM crop and the externalities considered. In most cases where a new GM crop is planted, it does not fully dominate existing non-GM crops and some non-GM crops may still continue to be cultivated. There are several reasons, the main one being that a GM variety may not provide additional benefits for all growers of a crop. For example, European Corn Borer-controlling Bt maize varieties will provide additional benefits in areas where the pest is a problem, but not in others (Goldberger et al., 2005; Wesseler et al., 2007). Another decrease of benefits occurs when a market for non-GM crops emerges with a price markup (Hurley et al., 2004a, 2004b) as demonstrated by a number of consumers studies (Rousu et al., 2007; Dannenberg et al., 2009), signaled via labeling (Scatasta et al., 2007; Gruere et al., 2009), and realized for products such as milk from cows raised with non-GM feed (MVS Milchvermarktungs GmBH, 2009).

The Environmental Benefits and Costs of GM Crops

177

Assessment of the environmental benefits and costs of GM crops needs to consider that non-GM crop production also has external environmental benefits and costs. Further, most of the external effects of agricultural production are internalized through a number of policies, such as regulations on pesticide use, fertilizer use, and so on (Oskam et al., 2010). This internalization does not imply that the environmental costs of non-GM agricultural production are zero, but rather that they are close to the social optimal level when ignoring dynamic aspects.1 This has implications for assessing the environmental benefits and costs of GM crops: firstly, the additional external benefits and costs of GM crops in comparison to non-GM crops should be considered when assessing whether or not an environmental effect should be considered a benefit or cost. This is a nontrivial issue in the debate about the regulation of GM crops, as the discussion about HR crops illustrates (Bonny, 2008; see also Section 3). Another implication is that external costs of GM crops do not per se justify a ban on them, if social welfare maximization is the objective. If the external costs of the GM crop are higher than those of the non-GM crop, a ban might be considered, but an alternative to a ban might include specific regulations addressing the externality, such as refuge areas to control the buildup of pest resistance. Decisions about optimal regulatory responses should depend on comparison of the benefits and costs of alternative strategies to address externalities (Arrow et al., 1996; Coase, 2006). Furthermore, in cases where GM crops reduce external effects, i.e., provide an external benefit, subsidizing the introduction of GM crops might be justified. As Section 3 seeks to demonstrate, this is exactly what the empirical evidence is suggesting.

3. Environmental benefits and costs of GM crops When GM crops were introduced in the mid-nineties, a number of concerns were raised regarding their environmental safety (Krimsky and Wrubel, 1996; Kendall et al., 1997). By and large, however, until now the net environmental effects of GM crops have been considered to be positive. A range of environmental benefits have been identified for GM crops, including positive effects on input use, indirect effects on ecosystems, and effects on GHG emissions, which will be discussed in the following sections. The most important, but least assessed, indirect effect on the environment that we start with here is that on land use through gains in productivity, measured in yield per hectare. 1 It may be debatable whether or not current regulations sufficiently address externality issues. This will differ country by country, region by region. But what can be said is that the marginal net benefits of non-GM crops are, among other things, a result of responses by producers to the regulatory environment they face, including regulations addressing environmental issues.

178

Justus Wesseler et al.

3.1. Yield effects of GM crops The adoption of GM crops has the potential to reduce inputs such as inorganic fertilizers and pesticides (Bennett et al., 2004a, 2004b). GM crops with resistance to insects and herbicides can substantially simplify crop management and reduce crop losses, leading to increased yields (e.g., Pray et al., 2002, 2011; Nickson, 2005). Such yield increases due to GM crops can also have positive land-use effects, often reducing pressure on protected land, in particular, in countries where the enforcement of regulations is weak. Increased resistance to fungal and viral diseases expected from future GM crops are expected to result in further efficiency advantages compared to non-GM crops. Development of GM technology to introduce genes conferring tolerance to abiotic stresses such as drought or inundation, extremes of heat or cold, salinity, aluminum, and heavy metals are likely to enable marginal land to become more productive and may facilitate the remediation of polluted soils (Czako et al., 2005; Uchida et al., 2005). The multiplication of GM crop varieties carrying such traits may increase farmers’ management possibilities and capacities to cope with these and other environmental problems (Dunwell and Ford, 2005; World Bank, 2007; Sexton and Zilberman, 2011). Therefore, GM technology holds out further hope of increasing the productivity of agricultural land (FAO, 2004). Using a Ricardian rent model, Brookes and Barfoot (2009) calculate substantial productivity gains for GM corn, cotton, oilseed rape, and soybeans. Similarly, Qaim (2011) reports substantial productivity gains from GM crops at the global level as well as their important contribution toward food security. While the productivity gains reported thus far are substantial and positive welfare effects have been identified, assessments regarding implications for land use and, in particular, habitat conservation have been lacking. In many developing countries, habitat loss is related to the need for additional land for agriculture production, resulting in wetlands, rainforests, and other protected areas being converted for crop production. Such pressure on valuable habitats is directly linked with demographic and socioeconomic development. Demographic development puts pressure on agricultural land to increase the supply of food in certain regions. While, in general, there is enough food being produced worldwide to nourish the world population, the unequal distribution of food supply and food demand creates pressure on important habitats in regions where food supply remains particularly scarce (United Nations, 2008). Socioeconomic development puts pressure on agricultural land to adapt the quality of food supply to higher standards of living (e.g., higher supply of foods with a higher protein content, such as meat) and to meet the everrising demand for new sources of energy. Pressure on agricultural land increases the opportunity cost of land uses other than agriculture,

The Environmental Benefits and Costs of GM Crops

179

translating into a threat for those uses not directly contributing to economic activities, such as natural habitats. Reducing this pressure requires, on the one hand, improvements in the distribution of food supply and increased food production in areas where food supply is extremely scarce, on the other. Increasing agricultural productivity is key to reducing pressure on habitats for two reasons: first, while redistribution of food may help to cope with an increase in population level, it will only partially address the demand for higher living standards, including changes in the demand for certain foods, such as rising demand for meat per capita; second, with the world population increasing by about one-third by 2050 (doubling in Africa), the number of undernourished people expected to be about 600 million by 2015 (United Nations, 2008), and world energy demand increasing 50% by 2030 (Energy Information Administration, 2009), increasing the carrying capacity of agriculture (the number of people being fed by one hectare of land) has become indispensable and a key issue for sustainable growth. For the past 12 years, GM plants have increased outputs per hectare. While most of the genetically engineered crops being produced, such as HR soybeans and most of the Bt corn, have been used as feed within animal production, some crops, such as Bt corn or soybean oil, have increased the amount of food being produced per unit of land in key regions such as Southern Africa and parts of Latin America (Qaim, 2009). Additionally, increasing efficiency in feed production with genetically engineered crops may help to increase animal production without increasing pressure on natural habitats. In any particular case, the impact on the overall area needed for feed production will depend on the size of the efficiency gain relative to the increase in demand for animal products. The same line of argument holds for the production of cotton and other GM crops. Bt cotton, for example, has lowered production costs for cotton and led to increased amounts of cotton produced per unit of land. This increase in efficiency reduces the amount of land needed for the same amount of cotton, but it also increases the competitiveness of cotton with respect to other crops, thereby increasing the amount of land allocated to cotton. Again, the net effect on the overall area allocated to cotton is an empirical question, depending on the size of the efficiency gain relative to an increase in the demand for cotton. In both cases, if the demand for the final product is such that the demand for the input (crop) to production is highly elastic, the overall land used for agricultural production of such an input may stay the same or even decrease. As a result, GM crops can contribute toward a reduction in pressure on natural habitats from agricultural land uses. The existence and magnitude of this contribution is an empirical question which has not yet been addressed. However, the evidence on productivity gains summarized by several authors (Brookes and Barfoot, 2009; Qaim, 2009; Carpenter, 2010;

180

Justus Wesseler et al.

Pray et al., 2011) indicate that such benefits are present and can be expected to be quite substantial. While increases in productivity can generate substantial indirect benefits which have not yet been adequately addressed quantitatively, the direct environmental implications of changes in pesticide use induced by GM crops have been relatively well studied, as we demonstrate in the next section. 3.2. Pesticide use effects of GM crops Pesticides in agriculture are used to control pests and nontarget plants. The reduction of pesticide applications is a major direct benefit of GM crop cultivation: reducing farmers’ exposure to chemicals (Hossain et al., 2004; Huang et al., 2005) and lowering pesticide residues in food and feed crops, while also releasing less chemicals into the environment and potentially increasing on-farm diversity in insects and pollinators (Nickson, 2005). Additionally, pest resistance through the protection genetic engineering confers can reduce the level of mycotoxins in food and feed crops (Wu, 2006). Although decreased pesticide use has emerged as one of the most important direct impacts of GM crops on the environment (Kleter and Kuiper, 2005), the problem of how to measure this impact has still not been solved. One commonly used indicator is the Environmental Impact Quotient (EIQ), which includes the impact of pesticides on the environment, on farm workers, and on consumers. Application of the EIQ to HR soybeans indicates an overall positive environmental impact from them over non-HR soybeans. In a review, Kleter et al. (2007) compare conventional and GM oilseed rape in the United States, calculating that for the GM variety the application of pesticide active ingredients was 30% lower, the total EI per hectare was 42% lower, the ecological impact was 39% lower, and the farmer impact was 54% lower. Brookes and Barfoot (2008) use the EIQ methodology to compute and compare EIQ values for conventional and GM crops at the national level, finding that EIQ values decreased by 15.4%. In their analysis of HR canola in North America, the amount of active chemical ingredients applied to canola decreased by 7.9 million kg, or 12.6%. As Smyth et al. (2011a) note, the study by Brookes and Barfoot assumed that the highest application rate was used in all instances, creating the potential for an overestimation of active ingredient application and thus underestimating the decline in usage and the net overall benefit. Recently, Gusta et al. (2011) and Smyth et al. (2011a, 2011b) investigated the environmental effects of HR canola in Canada, the adoption of which has changed weed control practices. Farmers’ have shifted from 2

The EIQ has been weighted with the application rate and the area (Smyth et al., 2011b).

The Environmental Benefits and Costs of GM Crops

181

soil-incorporated to foliar-applied postemergent herbicides. A majority of the farmers, more than 60% of the respondents, also reported a simplification of weed management using HR canola. As a result of these changes, the environmental impact of canola production in Canada – calculated based on a modified EIQ2 – dropped by 59% between 1995 and 2006. While the EIQ has been used in many studies, it, as well as other indicators, has some shortcomings that are addressed by Kleter and Kuiper (2005). The EIQ does not, for example, consider temporal aspects, which can be important for measuring phenomena such as the effect on water reservoirs of a continuous use of glyphosate on HR crops or changes in insecticide use caused by pest resistance. Such long-term effects also pose problems for environmental risk assessment in general (European Food Safety Authority [EFSA] Panel on Genetically Modified Organisms, 2010). Concerns have been raised in the literature about the control of volunteer canola (e.g., Ellstrand, 2001). According to the study by Smyth et al. (2011b), only 8% of the HR canola-growing farmers mentioned volunteer canola as a problem and ranked it either fourth or fifth after other weed problems. Of those 8%, only 35% reported that additional effort to control volunteer canola was needed, supporting previous results by Beckie et al. (2006) and the Serecon Management Consulting (2005) that canola volunteers are not a serious problem in general. Nevertheless, 9% of the respondents reported an increase in yield loss caused by volunteer canola over a 10-year period from 1995 to 2006, but control costs for volunteer canola have been calculated to be less than C$3 per hectare (Gusta et al., 2011). Another concern is the selection pressure caused by intensive use of glyphosate on nontarget flora. Glyphosate resistance has recently been reported for Amaranthus palmeri (Gaines et al., 2010). Until 2007, 13 glyphosate-resistant (GR) weeds had been reported worldwide (Service, 2007) posing a medium- to long-tem threat on the use of HR technology (Bonny, 2008). As weed-resistant pest resistance has become an emerging issue and resistance of pests to toxins expressed in Bt crops have been reported for a number of cases (Frisvold and Reeves 2010). Frisvold and Reeves (2010) has pointed out, however, that the emergence of weed as well as pest resistance can potentially be addressed by appropriate crop management strategies. Wu et al. (2008) report for China a two- to seven-fold suppression of the cotton bollworm population in areas where Bt cotton has been introduced since the late 1990s. This suppression has not only reduced pest damages in non-Bt cotton fields as well as other crops damaged by the cotton bollworm, but also resulted in a host-preference change of the cotton bollworm (Jongsma et al., 2010). Kuosmanen et al. (2006) report that Btcotton planting farmers in China do not necessarily give up the use of insecticides completely. One reason for this might be that of seed quality, as has also been reported for India (Herring, 2009), or the problem of secondary pests, as mentioned by Pemsl et al. (2008) and recently reported

182

Justus Wesseler et al.

on by Lu et al. (2010). Similarly, Carrie`re et al. (2003) report suppression of the pink bollworm by Bt cotton in Arizona. The implications of targetpest suppression for resistance management and pest management in general are not yet well understood and need further investigation (Carrie`re et al., 2003; Jongsma et al., 2010). Whether the suppression of pest damages in non-Bt cotton fields has changed pesticide use on other crops – as reported for the cases of HR canola (Gusta et al., 2011) and HR soybeans (Brookes and Barfoot, 2009) – has not been mentioned by Wu et al. (2008). Gusta et al. (2011) have measured the spillover benefits of HR canola cultivation on herbicide applications of the following crop. While such benefits had already been mentioned in the report of Serecon Management Consulting (2005), the authors of this newer study were able to quantify them. Fifty four percent of the respondents reported a spillover benefit worth about C$37 per hectare, with a calculated average annual benefit between about C$12 and C$20 per hectare of HR canola in Canada, meaning overall between C$67 and C$110 million for the 2007 crop.

3.3. Fertilizer use and GM crops Fertilizer is used in agricultural production to increase yields and crop quality by providing plants with three major nutrients: nitrogen, phosphorus and potassium. Secondary nutrients such as calcium and magnesium, or micronutrients such us iron, zinc and copper may also be provided, but in this section we focus on impacts of GM crops on nitrogen, phosphorus and potassium use and use efficiency while simplification in management practices may also result in lower on-farm fuel consumption. Fertilizer can be divided into two broad categories: organic and inorganic. The former is composed of organic matter derived from animals or plants, while the latter is composed of chemicals or minerals derived from nonrenewable resources. The type and amount of fertilizer applied on agricultural fields strongly depends not only on soil and crop characteristics, but also on economic factors, such as fertilizer prices. While in Western Europe and the United States the fertilizer application rate is about 250 kg/ha, it amounts to only 73 kg/ha and 9 kg/ha in Latin America and in Sub-Saharan Africa, respectively (Molden, 2007). Although fertilizer use has increased enormously in the past five decades – U.S. fertilizer use in 2007 was 200 times higher than in 1960, according to data from the USDA (2009) – fertilizer demand is expected to increase even further. Tenkorang and Lowenberg-DeBoer (2008) forecast a global increase of fertilizer demand in 2030 to levels 1.5 times higher than in 2005. Such an increase is a source of great concern, due to expected associated negative impacts on air, water and soil quality (Tilman et al., 2002).

The Environmental Benefits and Costs of GM Crops

183

Fertilizer use, organic and inorganic, may increase GHG emissions, eutrophication in water bodies, pollution of drinking water and soil acidification. Especially Nitrogen-rich compounds have been found to play a very important role in generating these effects. Environmental impacts and damage costs from synthetic nitrogen in Europe have been investigated by von Blottnitz et al. (2006), who estimate damage costs due to global warming caused by N2O and CO2 emissions from fertilizer production and N2O emissions from fertilized agricultural field to be in the order of 0.3 h/KGN (i.e., 60% of the current fertilizer market price). Based on contingent valuation studies eliciting the willingness-to-pay to restore ecological conditions in the Baltic Sea, Gren (2001) estimates the marginal benefits of reduced eutrophication due to a reduction in nitrogen loads to be in the range of 1 to 19 h/KGN. Fishman et al. (2009) estimate treatment costs of drinking water in the coastal aquifer of Israel to be in the range of 0.6 to 0.95 h/KgN, depending on the amount of water pumped for irrigation purposes (1,400–2,300 m3/year). In a greenhouse experiment with maize as experimental plants, Rodriguez et al. (2008) find that Nitrogen fertilization in the form of urea and UAN at rates of 100 and 200 KgN/ha significantly increase a Typic Argiudol soil pH from 5.9 to 6.2. In addition, Nitrogen fertilizer production is highly energy intensive and relies almost exclusively on nonrenewable energy sources, such as natural gas. There are several ways to define nitrogen use efficiency (NUE); here we follow Moose and Below (2008) and consider agronomic NUE, that is the ratio of crop yield to nitrogen fertilizer supplied. According to this definition, an improvement in NUE arises from the following changes: an increase in crop yield for the same amount of N fertilizer applied, a decrease in N fertilizer applied for the same crop yield, or both. The contribution of agricultural biotechnology to NUE improvements is exclusively indirect, brought about through yield-improving traits such as pest and herbicide resistance. For example, reduced damage to the root system of GM maize resistant to Corn Rootworm can lead to greater N uptake. Drought tolerant crops may also exhibit increased N uptake and N utilization (Moose and Below, 2008). Negative impacts on NUE are also possible. Zablotowicz and Reddy (2007) investigate the impact of GR soybean on N fixation and assimilation. Glyphosate is toxic to the soybean nitrogen-fixing symbiont, Bradyrhizobium japonicum, and because adoption of GR soybean increases glyphosate use, N fixation and assimilation maybe affected. As soybean is typically assumed to obtain enough N through symbiotic N2 fixation, and N fertilization is not part of traditional nutrient management practices for this crop, it is particularly important to identify such effects. In a threeyear field study (2002–2004), Zablotowicz and Reddy found only slight effects at label use rates, but a consistent reduction of N fixation and

184

Justus Wesseler et al.

assimilation at above-label use rates, meaning when glyphosate is used in excessive amounts. Whether adoption of GR soybean induces excessive use of glyphosate remains to be proven. Trigo and Cap (2003), for example, found fertilizer use in Argentina – one of the world’s largest adopters of GR soybean, with 16 million ha in 2007–2008 – still below risk levels in 2003 and below use rates in other countries. In addition, the authors note that GR soybean adoption has increased the agricultural area under no-till practices exponentially, with a potential positive effect on NUE (Rao and Dao, 1996). Concerns have also arisen with respect to impacts of GM crops on soil microbes, because these are involved in several processes influencing nutrient cycling. No empirical evidence to back up these concerns scientifically has been found yet for insect-resistant maize (Saxena and Stotzky, 2001; Al-Deeb et al., 2003; Motavalli et al., 2004). Donegan et al. (1995) find that insect-resistant cotton significantly stimulates growth of culturable bacteria and fungi with accompanying changes in substrate utilization. Evidence of alterations in community fatty acids, communitylevel physiological profile, taxonomic diversity of the root-associated community, diversity of Rhizobium leguminosarum and of variation in Pseudomonas populations can be found for HR rapeseed (Siciliano et al., 1998; Siciliano & Germida, 1999; Gyamfi et al., 2002). In GR soybean the incidence of Fusarium (soilborne pathogen) on the root system was greater within one week after the application of glyphosate as compared to the conventional isoline (Kremer et al., 2000), though the effect of these impacts on NUE has not yet been quantifiable.3 In the light of the empirical evidence just shown for cotton, soybean and rapeseed, it is not possible to conclude that yield gains due to genetic modification directly result in NUE improvement. Several studies investigating effects of GM crops at the farm or society levels consider the technology to be neutral in terms of fertilizer use (see, e.g., Qaim and Traxler, 2005).

3.4. Environmental safety issues of GM crops The transfer of pest-resistant and HR traits to weedy species and the persistence of feral crop plants carrying these traits raise issues about their impacts on the agricultural environment. Environmental safety issues focus on the direct or indirect effects of GM crops on nontarget organisms and the transfer of GM traits to populations of wild plants (FAO, 2003). Gene flow via pollen to a wild or cultivated plant generates the transfer of GM traits from crops to wild relatives and non-GM crops. Seed escaping during harvest, transportation or processing can also enable the 3 For a review of similar evidence for crops other than the once considered in this chapter, see Mina et al. (2008).

The Environmental Benefits and Costs of GM Crops

185

establishment of feral crop populations expressing the GM trait that can even facilitate further gene flow between crop plants and wild relatives (Dunwell and Ford, 2005). The impact of a particular transgene in the wild is dependent on several factors. If the GM trait confers a selective advantage over wild plants, then persistence and introgression of this trait into wild or weedy populations is more likely (Jenczewski et al., 2003). If the trait confers a physiological disadvantage, then selection pressure is against the trait and individuals containing the transgene will be competed out of the population. For these reasons, different GM traits have the potential to cause different environmental and agronomic impacts (Dunwell and Ford, 2005). Traits such as inherent resistance to insect, fungal and viral infection will undoubtedly confer an advantage over plants lacking these traits. This increased fitness may lead to an increase in transgene frequency in the wild, which may have further downstream effects in terms of the dynamics of insect populations and the organisms that predate upon them. The extent of the benefit will depend upon the severity of the infection (Pilson and Prendeville, 2004). Movement of herbicide-tolerance traits into wild populations will only confer an advantage where herbicides are applied. The physiological effort of sustaining the trait in the absence of herbicide selection may by costly in the longer term (Snow et al., 1999; Gueritaine et al., 2002), resulting in selection against these plants. Essentially, the processes of genetic drift in the population will determine the fate of traits that confer no benefit (Pilson and Prendeville, 2004). The movement of GM traits conferring selective advantage into wild plant populations has the potential to reduce the number or diversity of wild plants and so alter the ecological structure of communities. Wild relatives may in effect become extinct as a result of swamping by competitive plants and repeated hybridization (Dunwell and Ford, 2005). Specific traits such as drought and salt tolerance may allow plants carrying these transgenes to invade new habitats and out-compete native plants, leading to unwanted ecological change (Dunwell and Ford, 2005). Van de Wiel et al. (2005) point out the high variations of results in studies on gene flows, making it difficult to get a consistent view about their implications for the environment. Regional aspects seem, however, to be very important in quantifying the magnitude of gene flow (Scatasta, 2005). Gilligan et al. (2005) provide further evidence about the spatial importance of planting GM crops. The theoretical framework they present allows consideration of the spatial and temporal dynamics of gene movement. By using the case of oilseed rape, they show that stochastic models are far more important than deterministic models of gene movement. The resulting probability distributions about local persistence of novel genes provide important information for an environmental risk assessment of GM crops. The model indicates the scope for reducing

186

Justus Wesseler et al.

environmental risk by introducing novel genes that are spatially explicit, providing a challenge for plant breeders. The effects of Bt crops on nontarget pests has been a continuous concern among scientists. Wolfenbarger et al. (2008) conducted a metaanalysis on the effects of Bt crops on functional guilds of nontarget arthropods. They could not find uniform negative or positive effects when comparing Bt crops with their non-GM counterparts, treated without any additional insecticides. Some species-specific effects have been identified, but when the non-GM counterpart has been controlled with insecticides, Bt crops exhibit a higher abundance of nontarget arthropods. The effect of Bt maize pollen on nontarget Lepidoptera in Europe has been estimated to be extremely low. Perry et al (2010) calculated mortality rates in the worst-case scenario of less than one individual per 1,572 (one per 5,000 at the median) individuals for butterflies and less than one individual per 392 (one per 4,366 at the median) individuals for moths. This is hardly a relevant biodiversity effect, even when not considering the effects of insecticide use otherwise. Similarly, Bartsch et al. (2010) conclude that so far no negative environmental impact of Cry1Ab expressing Bt maize has been reported. A´lvarez-Alfageme et al. (2010) point out that previous results showing a toxic effect of Cry1Ab and Cry3Bb on ladybirds feeding on maize could not be replicated and were most likely based on poor study design and procedures.

3.5. Tillage and GHG emission effects The emission of GHG from agriculture has become an important issue within the debate on GHG emission reduction. Agriculture has been estimated to contribute to about 15% of annual global GHG emissions. According to a UN (2008) report, however, GM crops can provide a solution toward reducing these. One important contribution is via an increase in reduced- or zero-tillage systems through the adoption of HR crops (Ward et al., 2002; Frisvold et al., 2009). The savings reported in the literature for diesel under reduced-tillage systems are about 37 liters per hectare for the United States (Griffith and Parsons, 1980). The standard conversion rate for a liter of diesel in kg of CO2 is about 2.63 (Defra, 2007), resulting in about 97 kg of CO2 emission reduction per hectare and year. Demont et al. (2004) calculate savings in diesel use of about 1.43 liters per hectare for HR sugar beets in Europe as a result of savings in pesticide applications. Koga et al. (2003) calculated energy savings of reduced-tillage systems in Japan at about 47.51 ha1. The differences here indicate that the main gain from fuel savings in HR crops is a result of the adoption of reduced-tillage systems and not due to fuel savings in pesticide application.

The Environmental Benefits and Costs of GM Crops

187

In addition, fields planted with HR crops require less tillage between crops to manage weeds and, as a result, no-tillage and conservation tillage practices may reduce soil erosion (Fawcett and Towery, 2003; Nickson, 2005). Early studies have reported a reduction in soil erosion of up to 90% for the United States (Baker and Johnson, 1979). Reduced tillage also increases soil organic matter and improves soil structure and water-holding capacity. Water losses can be reduced by up to 50% during dry years and the flow of meteorological water be reduced by about 30% (Karlen and Sharpley, 1994). Further, pesticide runoff can be reduced significantly (Baker, 1990; Waibel and Fleischer, 1998). Improved water-holding capacity reduces soil-nutrient losses, resulting in higher soil-nutrient content (Blevins et al., 1983, Karlen, 1995) and improved above- and below-ground water quality, caused by reduced nitrogen emissions (Wheatley et al., 1995). A survey among HR canola-growing farmers in Canada (Smyth et al., 2011a) confirms higher soil moisture content and less erosion problems. Reduced tillage also has a positive effect on the biodiversity of soil microorganisms and above-ground fauna biodiversity. Higher bird populations have been reported in areas with reduced tillage (Best, 1985; Castrale, 1985; Basore et al., 1986) as well as positive effects on populations of small mammals (Basore et al., 1986) and benthic invertebrate communities (Barton and Farmer, 1997). While reduced tillage provides important environmental benefits, these might be reduced through the use of glyphosate or other broadspectrum herbicides in reduced-tillage systems. Nevertheless, the U.S. Environmental Protection Agency considers glyphosate to have only a minimal toxicity toward mammals, fish, and invertebrates (US EPA, 1996). Studies by Phillips (2003), Beckie, et al. (2006), and Kleter et al. (2007) found correlations between adoption of HR oilseed rape and adoption of zero-tillage systems. According to Smyth et al. (2011a), the amount of tillage operations among farmers growing HR canola in Canada dropped by more than 70%: from on average 2.73 passes to 0.74 passes. The authors calculate an annual value of carbon sequestration of about C$2.36 million from reduced and zero tillage in comparison to conventional tillage and an amount of about 470,000t carbon sequestered. The results of Smyth et al. (2011a) cannot directly be transferred to other areas where farmers grow HR crops in combination with reduced- or zerotillage systems. While less soil disturbance reduces the decomposition of soil organic matter and soil erosion, which often reduce carbon losses, this does not always have to be the case (Smith et al., 2008). The timeline of soil carbon sequestration also needs to be considered, as soil organic carbon (SOC) is expected to reach a new equilibrium when moving from one tillage practice to another and additional carbon sequestration ceases or decreases substantially. According to Smith et al. (2008), the most appropriate approach would be to measure changes in SOC through a change in land management practices, though this would require a well documented land-use history that is often not available. Nevertheless, the contribution of HR crops to the

188

Justus Wesseler et al.

adoption of reduced- and zero-tillage systems (Frisvold et al., 2009) cannot be ignored, nor can the indirect overall net environmental benefits of the change in tillage systems induced by the adoption of HR crops.

4. Discussion and conclusions Regarding the environmental benefits and costs side of GM crop introduction, early concerns about severe negative implications on the environment have not materialized. While some negative effects on nontarget pests and plants have been observed, overall negative impacts on biodiversity have not been confirmed. The effect of HR crops on the biodiversity of nontarget flora has been negative. This is not surprising, as the objective of the technology is to reduce nontarget flora within the field, contributing to other environmental benefits such as more efficient use of nutrients, increasing yields, as well as a higher quality of the harvested product, reducing postharvest energy costs such as cleaning and drying. These are aspects that policy instruments such as the European Food Safety Authority (EFSA) recommendation for the cultivation of HR maize NK603 do not consider (EFSA, 2009). The positive effects of HR crops in combination with the adoption of reduced- and zero-tillage systems and the productivity gains of GM crops on biodiversity and habitat conservation have so far received little attention in the literature. Similarly, the spillover effects on nontarget flora control on following crops has only recently been addressed in Canada, but has also been reported for Argentina, and warrants additional investigation. Studies assessing the changes in herbicide and insecticide use due to the growing of GM crops show a decrease in EIQ. An increase in EIQ from growing GM crops has not yet been reported. This is not surprising, considering available GM traits that either control insect pests and substitute insecticide use or are herbicide-resistant and substitute specific herbicides with broad-spectrum herbicides that are more environmental friendly. The long-term effects of an increase in the use of broad-spectrum herbicides combined with an increase in the planting of HR crops thus deserves further attention. The challenge is comparing a number of chemical compounds in the environment which may all be below the chemical compound-specific tolerance level with a single chemical compound that even may have passed the chemical specific-tolerance level, as in the case of glyphosate. While the environmental benefits of insect-resistant crops have been documented, resistance management has recently become more important. The change in host plant preference of the cotton bollworm, as observed in China, and the recently reported resistance of the pink bollworm to Bt cotton in India (Sharma, 2010) challenge resistance management. On the

The Environmental Benefits and Costs of GM Crops

189

one hand, this is important from an environmental point of view as a gain for future environmental benefits of GM crops, on the other hand, the suppression of the cotton bollworm in multiple crops may generate additional external benefits that have not yet been investigated. In summary, early concerns about the negative effects of GM crops on the environment have proven to be almost negligible, while a number of positive environmental effects have been observed, namely effects on habitat conservation, biodiversity, spillover effects on following crops, and pest suppression on multiple crops, though these have not yet been assessed from an environmental economics point of view, with the exception of HR canola in Canada. The documentation of the positive environmental benefits of GM crops disproves the hypothesis that the net environmental effects of GM crops would be negative and supports the arguments of Graff et al. (2009), Herring (2009), Paarlberg (2008), and Winston (2002) that environmental concerns are instrumentalized in order to achieve other non-environmental objectives. The empirical evidence regarding the environmental benefits of GM crops calls for a change in research focus such that environmental economists as well as policy makers pay more attention to the documented environmental benefits rather than to the hypothetical environmental costs. Policy makers also need to reconsider their policies toward GM crops in light of these documented environmental net benefits and their contribution to reducing the ecological footprint of agriculture. Finally it should be mentioned that, while this chapter has focused on secondary environmental effects of GM crops whose primary purpose is to improve crop yields or crop production efficiency, GM crops exist whose primary goal is to reduce environmental pollution. For example, Indian mustard has been genetically engineered to reduce selenium pollution at rates three to four times higher than wild relatives (Ban˜uelos et al., 2005). Genetically engineered Eastern Cottonwood has been designed and has been shown to have a high potential for in situ Mercury remediation from soils (Che et al., 2003). At the light of these and other important uses society can make of this technology to reduce ever-increasing environmental problems, policy makers need to start considering concerns from the scientific community about potential positive environmental effects of GM crops and GM plants that are not materializing because of regulatory burdens.

References Al-Deeb, M.A., Wilde, G.E., Blair, J.M., Todd, T.C. (2003), Effect of Bt corn for corn rootworm control on non-target soil micro arthropods and nematodes. Environmental Entomology 32 (4), 859–865.

190

Justus Wesseler et al.

A´lvarez-Alfageme, F., Bigler, F., Romeis, J. (2010), Laboratory toxicity studies demonstrate no adverse effects of Cry1Ab and Cry3Bb1 to larvae of Adalia bipunctata (Coleoptera: Coccinellidae): the importance of study design. Transgenic Research, doi:10.1007/s11248-010-9430-5. Arrow, K.J., Cropper, M.L., Eads, G.C., Hahn, R.W., Lave, L.B., Noll, R.G., Portney, P.R., Russell, M., Schmalensee, R., Smith, V.K., Stavins, R.N. (1996), Is there a role for benefit-cost analysis in environmental, health, and safety regulation?. Science 272, 221–222. Baker, D.B. (1990), Herbicide concentrations in Ohio’s drinking water supplies: a quantitative assessment. In: Weigmann, D.L. (Ed.), Pesticides in the Next Decade: The Challenges Ahead. Virginia Water Resources Center and VPI, Blacksburg, VA. Baker, J.L., Johnson, H.P. (1979), The effect of tillage system on pesticides in runoff from small watersheds. Transactions, American Society of Agricultural Engineers 22, 554–559. Ban˜uelos, G., Terry, N., LeDuc, D.L., Pilon-Smits, E.A.H., Mackey, B. (2005), Field trial of transgenic Indian mustard plants shows enhanced phytoremediation of selenium-contaminated sediment. Environmental Science & Technology 39, 1771–1777. Barton, D.R., Farmer, M.E.D. (1997), The effects of conservation tillage practices on benthic invertebrate communities in Headwater Streams in Southwestern Ontario, Canada. Environmental Pollution 96 (2), 207– 215. Bartsch, D., Devos, Y., Hails, R., Kiss, J., Krogh, P.H., Mestdagh, S., Nuti, M., Sessitsch, A., Sweet, J., Gathmann, A. (2010), Environmental impact of genetically modified maize expressing Cry1 proteins. In: Kempken, F., Jung, C. (Eds.), Genetic Modification of Plants. Springer, Berlin, pp. 575–611. Basore, N.S., Best, L.B., Wooley, J.B., Jr (1986), Bird nesting in Iowa notillage and tilled cropland. Journal of Wildlife Management 50, 19–28. Beckie, H.J., Harker, K.N., Hall, L.M., Warwick, S.I., Le´ge`re, A., Sikkema, P.H., Clayton, G.W., Thomas, A.G., Leeson, J.Y., Se´guinSwartz, G., Simard, M.-J. (2006), A decade of herbicide-resistant crops in Canada. Canadian Journal of Plant Science 86 (4), 1243–1264. Benbrook, C. (2009), Impacts of genetically engineered crops on pesticide use in the United States: the first thirteen years. On-Line Report. Available at http://www.ucsusa.org/. Bennett, R., Ismael, Y., Morse, S., Shankar, B. (2004a), Reductions in insecticide use from adoption of Bt cotton in South Africa: impacts on economic performances and toxic load to the environment. Journal of Agricultural Sciences 142, 665–674. Bennett, R., Phipps, R., Strange, A., Grey, P. (2004b), Environmental and human health impacts of growing genetically modified herbicidetolerant sugar beet: a life-cycle assessment. Plant Biotechnology Journal 2, 273–278.

The Environmental Benefits and Costs of GM Crops

191

Best, L.B. (1995), Impacts of tillage practices on wildlife habitat and populations. Farming for a Better Environment. Soil and Water Conservation Society, Ankeny, IA. Blevins, R.L., Thomas, G.W., Smith, M.S., Frye, W.W., Cornelius, P.L. (1983), Changes in soil properties after 10 years of continuous no-tilled and conventionally tilled corn. Soil and Tillage Research 3, 135–146. Bonny, S. (2008), Genetically modified glyphosate-tolerant soybean in the USA: adoption factors, impacts and prospects. A review. Agronomy for Sustainable Development 28 (1), 21–32. Brookes, G., Barfoot, P. (2008), Global impact of biotech crops: socioeconomic and environmental effects, 1996–2006. AgBioForum 11 (1), 21–38. Brookes, G., Barfoot, P. (2009), Global impact of biotech crops: income and production effects 1996–2007. AgBioForum 12 (2), 184–208. Carpenter, J.E. (2010), Peer-reviewed surveys indicate positive impact of commercialized GM crops. Nature Biotechnology 28 (4), 319–321. Carrie`re, Y., Ellers-Kirk, C., Sisterson, M., Antilla, L., Whitlow, M., Dennehy, T.J., Tabashnik, T.E. (2003), Long-term regional suppression of pink bollworm by Bacillus thuringiensis cotton. Proceedings of the National Academy of Sciences of the United States of America 100 (4), 1519–1523. Castrale, J.S. (1985), Responses of wildlife to various tillage conditions. Transactions of the North American Wildlife and Natural Resources Conference 50, 142–156. Che, D., Meagher, R.B., Heaton, A.C.P., Lima, A., Rugh, C.L., Merkle, S.A. (2003), Expression of mercuric ion reductases in Eastern Cottonwood (Populus deltoides) confers mercuric ion reduction and resistance. Plant Biotechnology Journal 1, 311–319. Coase, R. (2006), The conduct of economics: the example of fisher body and general motors. Journal of Economics & Management Strategy 15 (2), 255–278. Czako, M., Feng, X., He, Y., Liang, D., Marton, L. (2005), Genetic modification of wetland grasses for phytoremediation. Zeitschrift fu¨r Naturforschung 60c, 285–291. Dannenberg, A., Scatasta, S., Sturm, B. (2009), Keine Chance fu¨r genetisch vera¨nderte Lebensmittel in Deutschland? Eine experimentelle Analyse von Zahlungsbereitschaften. Perspektiven der Wirtschaftspolitik 10 (2), 214–234. Defra (2007), Guidelines to Defra’s GHG conversion factors for company reporting. DEFRA, London. Available at http://www.defra.gov.uk/ environment/business/reporting/pdf/conversion-factors.pdf. de Groot, R.S., Wilson, M.A., Boumans, R.M.J. (2002), A typology for the classification, description and valuation of ecosystem functions, goods and services. Ecological Economics 41, 393–408.

192

Justus Wesseler et al.

Demont, M., Wesseler, J., Tollens, E. (2004), Biodiversity versus transgenic sugar beets – the one Euro question. European Review of Agricultural Economics 31 (1), 1–18. Donegan, K.K., Palm, C.J., Fieland, V.J., Porteous, L.A., Ganio, L.M., Schaller, D.L., Bucao, L.Q., Seidler, R.J. (1995), Changes in levels, species and DNA fingerprints of soil microorganisms associated with cotton expressing the Bacillus thuringiensis var. kurstaki endotoxin. Applied Soil Ecology 2, 111–124. Dunwell, J.M., Ford, C.S. (2005), Technologies for biological containment of GM and non-GM crops. Defra Contract CPEC 47, DEFRA, London. EFSA Panel on Genetically Modified Organisms (2009), Scientific Opinion of the Panel on Genetically Modified Organisms on applications (EFSA-GMO-NL-2005-22 and EFSA-GMO-RX-NK603) for the placing on the market of the genetically modified glyphosate tolerant maize NK603 for cultivation, food and feed uses and import and processing, and for renewal of the authorisation of maize NK603 as existing product. The EFSA Journal 1137, 1–50. EFSA Panel on Genetically Modified Organisms (2010), Guidance on the environmental risk assessment of genetically modified plants. Draft. European Food Safety Authority, Parma, Italy. Available at http:// www.efsa.europa.eu/en/consultations/call/gmo100305a-ax1.pdf. Ellstrand, N.C. (2001), When transgenes wander, should we worry?. Plant Physiology 125, 1543–1545. Energy Information Administration (2009). International energy outlook 2009. Office of Integrated Analysis and Forecasting, U.S. Department of Energy, Washington, DC. Ervin, D., Welsh, R. (2005), Environmental effects of genetically modified crops: differentiated risk assessment and management. In: Wesseler, J. (Ed.), Environmental Costs and Benefits of Transgenic Crops. Wageningen UR Frontis Series (vol. 7), Springer, Dordrecht, pp. 7–30. EU – Environmental Council (1999), 2194th Council meeting, Luxembourg, June 24–25, 1999. C/99/203/. Available at http://europa.eu/rapid/ pressReleasesAction.do?reference¼PRES/99/203&format¼HTML& aged¼0&lg¼fi&guiLanguage¼en. Fawcett, R., Towery, D. (2003), Conservation tillage and plant biotechnology: how new technologies can improve the environment by reducing the need to plow. Conservation Technology Information Center (CTIC), Purdue University, West Lafayette. Available at http://www.whybiotech.com/resources/tps/ConservationTillageandPlant Biotechnology.pdf. Fishman, Y., Becker, N., Schechter, M. (2009), Treatment versus prevention of nitrogen fertilizer pollution: an inter-sectoral externality policy model. Natural Resource Modelling 22 (1), 137–171.

The Environmental Benefits and Costs of GM Crops

193

Fontes, E.M.G., Pires, C.S.S., Sujii, E.R., Panizzi, A.R. (2002), The environmental effects of genetically modified crops resistant to insects. Neotropical Entomology 31 (4), 497–513. Food and Agriculture Organization (FAO) (2003). Report of the FAO expert consultation on the environmental effects of genetically modified crops, June 16–18, 2003. Food and Agriculture Organization of the United Nations, Rome. Food and Agriculture Organization (FAO) (2004): Agricultural biotechnology: meeting the needs of the poor? The state of food and agriculture 2003–04. Food and Agriculture Organization of the United Nations, Rome. Frisvold, G.B., Reeves, J.M. (2010), Resistance management and sustainable use of agricultural biotechnology. AgBioForum 13 (4), 343–359. Frisvold, G.B., Boor, A., Reeves, J.M. (2009), Simultaneous diffusion of herbicide resistant cotton and conservation Tillage. AgBioForum 12 (3&4), 249–257. Gaines, T.A., Zhang, W., Wang, D., Bukun, B., Chisholm, S.T., Shaner, D.L., Nissen, S.J., Patzoldt, W.L., Tranel, P.J., Culpepper, A.S., Grey, T.L., Webster, T.M., Vencill, W.K., Sammons, R.D., Jiang, J., Preston, C., Leach, J.E., Westra, P. (2010), Gene amplification confers glyphosate resistance in Amaranthus palmeri. Proceedings of the National Academy of Sciences of the United States of America 107 (3), 1029–1034. Gilligan, C., Claessen, D., van den Bosch, F. (2005), Spatial and temporal dynamics of gene movements arising from deployment of transgenic crops. In: Wesseler, J. (Ed.), Environmental Costs and Benefits of Transgenic Crops. Wageningen UR Frontis Series (vol. 7), Springer, Dordrecht, pp. 143–162. Goldberger, J., Merrill, J., Hurley, T. (2005), Bt corn farmer compliance with insect resistance management requirements in Minnesota and Wisconsin. AgBioForum 8 (2&3), 151–160. Graff, G.D., Hochman, G., Zilberman, D. (2009), The political economy of agricultural biotechnology policies. AgBioForum 12 (1), 34–46. Gren, I.M. (2001), International versus national actions against nitrogen pollution of the Baltic Sea. Environmental and Resource Economics 20 (1), 41–59. Griffith, D.R., Parsons, S.D. (1980), Energy requirements for various tillage-planting systems. NCR Extension Publication NCR-202-W, Purdue University, West Lafayette. Gruere, G.P., Carter, C.A., Farzin, Y.H. (2009), Explaining international differences in genetically modified food labeling policies. Review of International Economics 17 (3), 393–408. Gueritaine, G., Sester, M., Eber, F., Chevre, A.M., Darmency, H. (2002), Fitness of backcross six hybrids of hybrids between transgenic oilseed

194

Justus Wesseler et al.

rape (Brassica napus) and wild radish (Raphanus raphanistrum). Molecular Ecology 11, 1419–1426. Gusta, M., Smyth, S.J., Belcher, K., Phillips, P.W.B., Castle, D. (2011), Economic benefits of genetically-modified herbicide-tolerant canola for producers. AgBioForum 14 (1), 1–13. Gyamfi, S., Pfeifer, U., Stierschneider, M., Sessitsch, A. (2002), Effects of transgenic glufosinate-tolerant oilseed rape (Brassica napus) and the associated herbicide application on eubacterial and Pseudomonas communities in the rhizosphere. FEMS Microbiology Ecology 41 (3), 181–190. Herring, R. (2009), Persistent narratives: why is the ‘‘failure of Bt cotton in India’’ story still with us?. AgBioForum 12 (1), 14–22. Hossain, F., Pray, C.E., Lu, Y., Huang, J., Fan, C., Hu, R. (2004), Genetically modified cotton and farmers’ health in China. International Journal of Occupational Environmental Health 10, 296–303. Huang, J., Hu, R., Rozelle, S., Pray, C. (2005), Insect-resistance GM rice in farmers’ fields: assessing productivity and health effects in China. Science 308, 688–690. Hurley, T., Secchi, S., Bacock, B.A., Hellmich, R.L. (2004a), Managing the risk of European corn borer resistance to Bt corn. Environmental and Resource Economics 22 (4), 537–558. Hurley, T., Mitchell, P.D., Rice, M.E. (2004b), Risk and the value of BT corn. American Journal of Agricultural Economics 86 (2), 345–358. Jenczewski, E., Ronfort, J., Chevre, A.M. (2003), Crop-to-wild gene flow, introgression and possible fitness effects of transgenes. Environmental Biosafety Research 2, 9–24. Jongsma, M.A., Gould, F., Legros, M., Yang, L., van Loon, J.J.A., Dicke, M. (2010), Insect oviposition behavior affects the evolution of adaptation to Bt crops: consequences for refuge policies. Evolutionary Ecology 24 (5), 1017–1030. Kalaitzandonakes, N. (Ed.) (2003). Economic and Environmental Impacts of Agbiotech: A Global Perspective. Kluwer-Plenum Academic Publishers, New York. Karlen, D.L. (1995), Tillage effects on soil fertility. Farming for a Better Environment. Soil and Water Conservation Society, Ankeny, IA. Karlen, D.L., Sharpley, A.N. (1994), Management strategies for sustainable soil fertility. In: Hatfield, J.L., Karlen, D.L. (Eds.), Sustainable Agricultural Systems. Lewis Publishing, Boca Raton, FL. Kealey, T. (1998), Why science is endogenous: a debate with Paul Davis (and Ben Martin, Paul Romer, Chris Freeman, Luc Soete and Keith Pavitt). Research Policy 26 (7–8), 897–923. Kendall, H.W., Beachy, R., Eisner, T., Gould, F., Herdt, R., Raven, P.H., Schell, J.S., Swaminathan, M.S. (1997), Bioengineering of crops. Environmental and Socially Sustainable Development Studies and Monograph Series 23. The World Bank, Washington, DC.

The Environmental Benefits and Costs of GM Crops

195

Kleter, G., Kuiper, H. (2005), Assessing the environmental impact of changes in pesticide use on transgenic crops. In: Wesseler, J. (Ed.), Environmental Costs and Benefits of Transgenic Crops. Wageningen UR Frontis Series (vol. 7), Springer, Dordrecht, pp. 33–44. Kleter, G.A., Bhula, R., Bodnaruk, K., Carazo, E., Felsot, A.S., Harris, C.A., Katayama, A., Kuiper, H.A., Racke, K.D., Rubin, B., Shevah, Y., Stephenson, G.R., Tanaka, K., Unsworth, J., Wauchope, R.D., Wong, S.S. (2007), Altered pesticide use on transgenic crops and the associated general impact from an environmental perspective. Pest Management Science 63, 1107–1115. Koga, N., Tsuruta, H., Tsuji, H., Nakano, H. (2003), Fuel consumptionderived CO2 emissions under conventional and reduced tillage cropping systems in northern Japan. Agriculture, Ecosystems and Environment 99, 213–219. Kremer, R.J., Donald, P.A., Keaster, A.J., Minor, H.C. (2000), Herbicide impact on fusarium spp. and soybean cyst nematode in glyophosphate tolerant soybean. In: Annual Meetings Abstracts. ASA, CSSA, and SSSA, Madison. 257. Krimsky, S., Wrubel, R. (1996), Agricultural Biotechnology and the Environment. University of Illinois Press, Urbana-Champaign, IL. Kuosmanen, T., Pemsl, D., Wesseler, J. (2006), Specification and estimation of production functions involving damage control inputs: a two-stage, semi-parametric approach. American Journal of Agricultural Economic 88 (2), 499–511. Lu, Y., Wu, K., Jiang, Y., Xia, B., Li, P., Feng, H., Wyckhuys, K.A.G., Guo, Y. (2010), Mirid bug outbreaks in multiple crops correlated with wide-scale adoption of Bt cotton in China. Science 328, 1151–1153. Mellon, M., Rissler, J. (2003), Environmental effects of genetically modified crops – recent experiences. Union of Concerned Scientists. Available at http://www.ucsusa.org/. Mina, U., Khan, S.A., Choudhary, A., Choudhary, R., Aggarwal, P.K. (2008), An approach for impact assessment of transgenic plants on soil ecosystem. Applied Ecology and Environmental Research 6 (3), 1–19. Molden, D. (2007), Water for Food, Water for Life: A Comprehensive Assessment of Water Management. Earthscan, London, UK. Mooney, S., Klein, K.K. (1999), Environmental concerns and risks of genetically modified crops: economic contributions to the debate. Canadian Journal of Agricultural Economics 47 (4), 437–444. Moose, S., Below, F.E. (2008), Biotechnology approaches to improving maize nitrogen use efficiency. In: Kritz, A.L., Larkins, B.A. (Eds.), Biotechnology in Agriculture and Forestry: Molecular Genetic Approaches to Maize Improvement. Springer, Berlin, pp. 65–77.

196

Justus Wesseler et al.

Morel, B., Farrow, S., Wu, F., Casman, E. (2003), Pesticide resistance, the precautionary principle, and the regulation of Bt corn: real option and rational option approaches to decision-making. In: Laxminarayan, R. (Ed.), Battling Resistance to Antibiotics and Pesticides. Resources for the Future, Washington, DC, pp. 184–213. Motavalli, P.P., Kremer, R.J., Fang, M., Means, N.E. (2004), Impact of genetically modified crops and their management on soil microbially mediated plant nutrient transformations. Journal of Environmental Quality 33 (3), 816–824. MVS Milchvermarktungs GmBH (2009), Die faire Milch. Available at http://www.die-faire-milch.de/. National Research Council (2010), The Impact of Genetically Engineered Crops on Farm Sustainability in the United States. The National Academies Press, Washington, DC. Nickson, T.E. (2005), Crop biotechnology – the state of play. In: Poppy, G.M., Wilkinson, M.J. (Eds.), Gene Flow from GM Plants. Blackwell, Oxford, UK, pp. 12–42. Oskam, A., Meesters, G., Silvis, H. (Eds.) (2010). EU Policy for Agriculture, Food and Rural Areas. Wageningen Academic Publishers, Wageningen. Paarlberg, R. (2008), Starved for Science. How Biotechnology is Being Kept Out of Africa. Harvard University Press, Cambridge. Pemsl, D., Gutierrez, A.P., Waibel, H. (2008), The economics of biotechnology under ecosystem disruption. Ecological Economics 66, 177–183. Perry, J.N., Devos, Y., Arpaia, S., Bartsch, D., Gathmann, A., Hails, R.S., Kiss, J., Lheureux, K., Manachini, B., Mestdagh, S., Neemann, G., Ortego, F., Schiemann, J., Sweet, J.B. (2010), A mathematical model of exposure of non-target Lepidoptera to Bt-maize pollen expressing Cry 1Ab within Europe. Proceedings of the Royal Society B 277, 1417–1425. Phillips, P.W.B. (2003), The economic impact of herbicide tolerant canola in Canada. In: Kalaitzandonakes, N. (Ed.), The Economic and Environmental Impacts of Agbiotech: A Global Perspective. Kluwer Academic Publishers, New York, pp. 119–140. Pilson, D., Prendeville, H.R. (2004), Ecological effects of transgenic crops and the escape of transgenes into wild populations. Annual Review of Ecology, Evolution, and Systematics 35, 149–174. Pray, C.E., Huang, J., Hu, R., Rozelle, S. (2002), Five years of Bt cotton in China – the benefits continue. The Plant Journal 31, 423–430. Pray, C.E., Nagarajan, L., Huang, J., Hu, R., Ramaswami, B. (2011), Impact of Bt cotton, the potential future benefits from biotechnology in China and India. In: Carter, C., Moschini, G., Sheldon, I. (Eds.), Genetically Modified Food and Global Welfare. Emerald, Bingley, UK. Qaim, M. (2009), The economics of genetically modified crops. Annual Review of Resource Economics 1, 3.1–3.29.

The Environmental Benefits and Costs of GM Crops

197

Qaim, M. (2011), Genetically modified crops and global food security. In: Carter, C., Moschini, G., Sheldon, I. (Eds.), Genetically Modified Food and Global Welfare. Emerald, Bingley, UK. Qaim, M., Traxler, G. (2005), Roundup ready soybeans in Argentina: farm level and aggregate welfare effects. Agricultural Economics 32, 73–86. Rao, S.C., Dao, T.H. (1996), Nitrogen placement and tillage effects on dry matter and nitrogen accumulation and redistribution in winter wheat. Agronomic Journal 88, 365–371. Rodriguez, M.B., Godeas, A., Lavado, R.S. (2008), Soil acidity changes in bulk soil and maize rhizosphere in response to nitrogen fertilization. Communications in Soil Science and Plant Analysis 39 (17&18), 2597–2607. Rousu, M., Huffman, W., Shogren, J.F., Tegene, A. (2007), Effects and value of verifiable information in a controversial market: evidence from lab auctions of genetically modified food. Economic Inquiry 45 (3), 409–432. Saxena, D., Stotzky, G. (2001), Bacillus thuringiensis (Bt) toxin released from root exudates and biomass of Bt corn has no apparent effect on earthworms, nematodes, protozoa, bacteria and fungi in soil. Soil Biology and Biochemistry 33, 1225–1230. Scatasta, S. (2005), Comment on Van de Wiel, Groot and Den Nijs: gene flow from crop to wild plants and its population-ecological consequences in the context of GM-crop biosafety, including some recent experiences from lettuce. In: Wesseler, J. (Ed.), Environmental Costs and Benefits of Transgenic Crops, Wageningen UR Frontis Series (vol. 7), Springer, Dordrecht, pp. 143–162. Scatasta, S., Wesseler, J., Hobbs, J. (2007), Differentiating the consumer benefits from labelling of GM food products. Agriculture Economics 37 (2–3), 237–242. Serecon Management Consulting Inc. (2005), Herbicide tolerant volunteer canola management in subsequent crops. Canola Council of Canada (CCC), Winnipeg, Manitoba. Available at http://www.canolacouncil. org/uploads/HT_canola_final.pdf. Service, R.F. (2007), A growing threat down on the farm. Science 316, 114–117. Sexton, S., Zilberman, D. (2011), Biotechnology and biofuel. In: Carter, C., Moschini, G., Sheldon, I. (Eds.), Genetically Modified Food and Global Welfare. Emerald, Bingley, UK. Sharma, D.C. (2010), Bt cotton failed admits Monsanto. India Today, March 6. Available at http://indiatoday.intoday.in/site/Story/86939/ BtþcottonþhasþfailedþadmitsþMonsanto.html?complete¼1. Shleifer, A. (2010), Efficient regulation. NBER Working Paper No. 15651, National Bureau of Economic Research, Cambridge, MA. Siciliano, S.D., Germida, J.J. (1999), Taxonomic diversity of bacteria associated with the roots of field-grown transgenic Brassica napus cv.

198

Justus Wesseler et al.

Quest, compared to the nontransgenic B. napus cv. Excel and B. rapa cv. Parkland. FEMS Microbiology Ecology 29, 263–272. Siciliano, S.D., Theoret, C.M., DeFreitas, J.R., Hucl, P.J., Germida, J.J. (1998), Differences in the microbial communities associated with the roots of different cultivars of canola and wheat. Canadian Journal Microbiology 44, 844–851. Smith, P., Martino, D., Cai, Z., Gwary, D., Janzen, H., Kumar, P., McCarl, B., Ogle, S., O’Mara, F., Rice, C., Scholes, B., Sirotenko, O., Howden, M., McAllister, T., Pan, G., Romanenkov, V., Schneider, U., Towprayoon, S., Wattenbach, M., Smith, J. (2008), Greenhouse gas mitigation in agriculture. Philosophical Transactions of the Royal Society B 363, 789–813. Smyth, S.J., Gusta, M., Belcher, K., Phillips, P.W.B., Castle, D. (2011a), Environmental impacts from herbicide tolerant canola production in Western Canada. Agricultural Systems 104, 403–410. Smyth, S.J., Gusta, M., Belcher, K., Phillips, P.W.B., Castle, D. (2011b), Changes in herbicide use following the adoption of HR canola in Western Canada. Weed Technoloy. In Press. Snow, A.A., Andersen, B., Jorgensen, R.B. (1999), Costs of transgenic herbicide resistance introgressed from Brassica napus into weedy B. rapa. Molecular Ecology 8, 605–615. Tenkorang, F., Lowenberg-DeBoer, J. (2008), Forecasting long-term global fertilizer demand. FAO Report. Available at ftp://ftp.fao.org/ agl/agll/docs/globalfertdemand.pdf. Tilman, D., Cassman, K.G., Matson, P.A., Naylor, R., Polasky, S. (2002), Agricultural sustainability and intensive production practices. Nature 418, 671–677. Trigo, E.J., Cap, E.J. (2003), The impact of the introduction of transgenic crops in Argentinean agriculture. AgBioForum 6, 87–94. Uchida, E., Ouchi, T., Suzuki, Y., Yoshida, T., Habe, H., Yamaguchi, I., Omori, T., Nojiri, H. (2005), Secretion of bacterial xenobioticdegrading enzymes from transgenic plants by an apoplastic expressional system: an applicability for phytoremediation. Environmental Science and Technology 39, 7671–7677. USDA (2009), U.S. fertilizer use and price. As on January 24, 2010. Available at http://www.ers.usda.gov/Data/FertilizerUse/. United States-Environmental Protection Agency, US-EPA (1986), Pesticide fact sheet: glyphosate. No. 173. EPA 540/FS-88-124, Office of Pesticide Programs, Washington DC. United Nations, UN (2008), UN millennium ecosystem assessment. Available at http://www.millenniumassessment.org/en/Index.aspx. Van de Wiel, C., Groot, M., den Nijs, H. (2005), Gene flow from crop to wild plants and its population – ecological consequences in the context of GM-crop biosafety, including some recent experiences from lettuce. In:

The Environmental Benefits and Costs of GM Crops

199

Wesseler, J. (Ed.), Environmental Costs and Benefits of Transgenic Crops, Wageningen UR Frontis Series (vol. 7), Springer, Dordrecht, pp. 97–110. Von Blottnitz, H., Rabl, A., Boiadjiev, D., Taylor, T., Arnold, S. (2006), Damage costs of nitrogen fertilizer in Europe and their internalization. Journal of Environmental Planning and Management 49 (3), 413–433. Waibel, H., Fleischer, G. (1998), Kosten und Nutzen des chemischen Pflanzenschutzes in der deutschen Landwirtschaft aus gesamtwirtschaftlicher Sicht. Vauk, Kiel. Ward, C., Flanders, A., Isengildina, O., White, F. (2002), Efficiency of alternative technologies and cultural practices for cotton in Georgia. AgBioForum 5 (1), 10–13. Wesseler, J. (2009), The Santaniello theorem of irreversible benefits. AgBioForum 12 (1), 8–13. Wesseler, J., Scatasta, S., Nillesen, E. (2007), The maximum incremental social tolerable irreversible costs (MISTICs) and other benefits and costs of introducing transgenic maize in the EU-15. Pedobiologia 51 (3), 261–269. Wheatley, D.M., Macleod, D.A., Jessop, R.S. (1995), Influence of tillage treatments on N2 fixation of soybean. Soil Biology Biochemistry 27 (4/5), 571–574. Winston, M.L. (2002), Travels in the Genetically Modified Zone. Harvard University Press, Cambridge, MA. Wolfenbarger, L.L., Naranjo, S.E., Lundgren, J.G., Bitzer, R.J., Watrud, L.S. (2008), Bt crop effects on functional guilds of non-target arthropods: a meta-analysis. PLoS ONE 3 (5), e2118. World Bank (2007). World development report 2008. Agriculture for development. The World Bank, Washington, DC. Wu, F. (2006), Mycotoxin reduction in Bt corn: potential economic, health, and regulatory impacts. Transgenic Research 15, 277–289. Wu, K., Lu, Y., Feng, H., Jiang, Y., Zhao, J. (2008), Suppressing of cotton bollworm in multiple crops in China in areas with Bt toxin-containing cotton. Science 321, 1676–1678. Zablotowicz, R.M., Reddy, K.N. (2007), Nitrogenase activity, nitrogen content, and yield responses to glyphosate in glyphosate-resistant soybean. Crop Protection 26 (3), 370–376.

CHAPTER 9

Biotechnology and Biofuel$ Steven E. Sexton and David Zilberman Agricultural and Resource Economics, UC Berkeley, 207 Giannini Hall #3310, University of California, Berkeley, CA 94720-3310, USA E-mail addresses: [email protected]; [email protected]

Abstract Purpose – To identify how agricultural biotechnology addresses the two challenges facing agriculture: to feed a world growing to 9 billion people by 2050 and to provide a liquid fuel alternative to petroleum. Design –This chapter relies on econometric modeling, a review of existing literature, and diagrammatic modeling to articulate the impact of agricultural biotechnology on food and energy markets. Findings –Agricultural biotechnology reduces the tension between food security and biofuel production. It reduces volatility in food and fuel markets and can mitigate risk to biofuel processors. Originality – The analysis is original although it relies on previous research to some extent. The analysis is compared to and contrasted with related work. Keywords: Biofuel, agricultural biotechnology, fossil fuel, food security, climate change JEL Classifications: Q16, Q42 Concerns about global climate change and scarcity of fossil fuels have sparked renewed interest in an alternative fuel technology that was first employed by humans 400,000 years ago. Although the process of converting biomass to energy has advanced considerably over the millennia, the concept of harvesting the sun’s energy through plant photosynthesis and deploying that energy through combustion is one that has been well understood for ages. As the world seeks to clean and expand its energy supply, biofuels stand as a relatively cheap, solar-powered alternative to oil. In 2008, record biofuel production was spurred by quotas and more than $11 billion in subsidies from OECD countries (GSI, 2007). Biofuels are not, however, a panacea. Under current conditions, $

The research leading to this study was supported by the EBI.

Frontiers of Economics and Globalization Volume 10 ISSN: 1574-8715 DOI: 10.1108/S1574-8715(2011)0000010014

r 2011 by Emerald Group Publishing Limited. All rights reserved

226

Steven E. Sexton and David Zilberman

their utilization reduces food security and has an uncertain effect on greenhouse gas (GHG) emissions in the short to medium term. Prompted by evidence that biofuels contributed to a food crisis in 2008 and could cause substantial increases in GHG emissions over a 100-year horizon, policymakers began to reassess the capacity for biofuels to solve the energy challenge posed by climate change and oil scarcity. In this chapter, we consider the extent to which agricultural biotechnology can minimize the downsides of biofuel production to make it more sustainable. Employing the tools of genetic engineering, agricultural biotechnology has developed seed varieties that improve agricultural productivity in four staple crops by reducing crop damage. Productivity growth is necessary to maintain per capita calorie availability as the world population grows to 9 billion people by 2050 (Alston et al., 2009; FAO, 2009; Godfray et al., 2010). Sufficient growth in crop yields would not only guarantee food security but also free land for use in production of biofuel feedstocks. Transgenic seed varieties can also reduce the carbon intensity of some biofuel feedstock production, thereby improving the carbon benefits of biofuels. The tools of genetic plant engineering can also enable second-generation biofuels by producing plants that are particularly suited for conversion to liquid fuel. Such dedicated feedstocks would not only reduce competition for food harvest but would also minimize the carbon inputs used in feedstock production and conversion. This chapter proceeds with analysis of the economic impact of biofuel production on food and energy markets. This analysis relies on conceptual modeling and previous research. Next, the economics of agricultural biotechnology is reviewed and related to biofuel production. This analysis suggests biotechnology adoption can improve the net welfare implications of biofuel production. Then, it is argued that excessive regulation has slowed the introduction of new crop technologies that could have provided considerably welfare gains, particularly in the context of high crop demand for biofuel production. A final section concludes. 1. The economics of biofuel T.W. Schultz emphasized in 1953 that maintaining cheap food must be a major objective of government food policy. Three decades after he received the Nobel prize in economics, and one decade into the 21st century, it is evident that in much of the developed world, abundant energy has become as important as abundant food, and neither is guaranteed. The world witnessed the consequences of three decades of complacency toward crop science in 2008 as food prices soared and shortages hit the developing world. As the world population grows by more than 2 billion people from 2010 to 2050, and as diets become more meat intensive, demand for food and feed will grow considerably. Demand for energy, too, will grow through the 21st century, driven by rising incomes in emerging economies such as China and India. In 2006,

Biotechnology and Biofuel

227

there were 400 cars per 1,000 people in Western Europe, the United States, Canada, and Japan. In China, there were 18 cars per 1,000 people, and in India there were 8 (The Economist, 2009). Rates of car ownership in China and India are expected to soar. Even if car ownership rates in these emerging economies do not reach those of the west, demand for fuels will grow precipitously. Sperling and Gordon (2009) predict that within 20 years, the number of cars in the world will double from 1 billion to 2 billion. Energy demand is expected to grow 104% by 2030, with China and India accounting for nearly 30% of all energy consumption (EIA, 2009). Oil is a nonrenewable resource with finite supply. Over the years, the known reserves of oil have increased as new oil discoveries outstripped consumption. But that trend has slowed and oil has become scarce. Although vast reserves remain, the cost of extracting oil from them is substantial. Experts suggest that oil extraction is at its peak or will reach its peak in the near future. Fuel supply will increasingly rely on mining tar sands and converting coal to liquid fuel, two costly and carbon-intensive processes. The combination of increasing energy demand and stagnating and expensive supply is expected to push oil prices higher. This has been the trend since the turn of the century. As oil prices rise, biofuels become more attractive. During the energy crisis of the 1970s, a biofuel industry was born in the United States. It was short-lived, however, as the subsequent decline in oil prices caused demand for biofuel to dry up. Although Brazil began developing a biofuel industry in the 1980s through government policy, demand for biofuel did not recover until about the start of the 21st century. Then oil prices began to rise again and ethanol replaced MTBE, an immiscible chemical compound, as the preferred fuel oxygenate because MTBE had contaminated water. A new wave of investment in biofuel technologies began in the United States and elsewhere around the world. Governments worried about the availability of fuel and balance of payment constraints developed policies to support the biofuel sector. Biofuels – produced from maize, sugar cane, rapeseed, soybean, and palm oil – were the beneficiaries of mandated demand and generous subsidies. In the United States, the Biofuel Act of 2007 established evolving targets for biofuel production. By 2022, 36 billion gallons of biofuel must be blended into gasoline each year. Twenty-one million gallons are mandated to come from advanced biofuels produced from cellulosic plant material. Cellulosic biofuel technologies can convert grasses, shrubs, and crop residues to liquid fuel and can generate more fuel per unit of land than conventional technologies (Heaton et al., 2008). Biofuels attracted considerable policy support because they were expected to lower the price of fuel, improve terms of trade, provide new income to the rural sector, and aid in the effort to reduce carbon emissions. However, the extent to which biofuels improve welfare among the rural poor is unclear. Some farmers benefit from higher demand for commodities, but others suffer from higher input costs. Furthermore, welfare benefits may be captured by

228

Steven E. Sexton and David Zilberman

landowners instead of farmers if increased demand for crops is capitalized into rental rates of land. Research also suggests the climate change mitigation effects of biofuels may not be as significant as once thought. In principal, biofuel is a GHG neutral technology – the same amount of carbon dioxide that is emitted in combustion is sequestered from the atmosphere in plant photosynthesis. In reality, the feedstock production and conversion processes, however, consume energy and emit GHG, diminishing the GHG emissions savings of biofuel relative to oil. Concern about the GHG effects of certain biofuels has led to introduction of regulation based on GHG contributions (Rajagopal and Zilberman, 2008) and targeted support of biofuels that offer the greatest GHG emissions savings. The biofuel industry grew considerably from 2000 to 2008, spurred by subsidies and regulation-induced demand and encouraged by low corn prices. Figure 1 shows the growth ethanol production in the United States from 1980 to 2009. In 2009, about 30% of corn produced in the United States was used for biofuel. In Europe, significant amounts of wheat and rapeseed were diverted to biofuel uses (Rajagopal and Zilberman, 2008). Current biofuel technologies are land-intensive. As Table 1 reports, the production of just 500 gallons of ethanol from corn requires one acre of land. Even though second-generation cellulosic ethanol can reduce the land intensity of biofuel 400%, scaled-up biofuel production still creates considerable new demand for land and crop output (Heaton et al., 2008). Consequently, biofuel production is associated with adverse impacts on food markets and nature preservation. As land and harvest are recruited for biofuel production, food prices climb. New demand for land induces the conversion of natural land to production with potentially disastrous consequences for biodiversity and climate change. Agricultural biotechnology presents an opportunity to mitigate these impacts and improve the sustainability of biofuels. U.S. Fuel Ethanol Production 1980-2009 12000 10000

Ethanol (millions of gallons)

8000 6000 4000 2000

19 80 19 83 19 86 19 89 19 92 19 95 19 98 20 01 20 04 20 07

0

Fig. 1.

U.S. fuel ethanol production 1980–2009. Source: Renewable Fuels Association.

229

Biotechnology and Biofuel

Table 1.

Feedstock yields and land-use implications

Crop

Harvestable Biomass (Tons/Acre)

Ethanol (Gallons/Acre)

Acres Needed for 35 Billion Gallons Ethanol

Corn grain Corn stover Corn total Prairie Switchgrass Miscanthus

4 3 7 2 6 17

500 300 800 200 600 1700

70 105 40 210 60 18

Source: Heaton et al. (2008).

2. The impacts of biotechnology on agricultural productivity For much of the 20th century, the advent of traditional plant breeding yielded a Green Revolution that more than doubled total factor productivity of staple crops. Along with other innovations such as chemical fertilizers and pesticides, and the spread of irrigation technologies, it fueled tremendous growth in food production that enabled per capita calorie availability to climb even as the world population doubled to 6 billion people (Alston et al., 2009; WHO/FAO, 2003). At the outset of the 21st century, however, yield growth in staple crops has slowed, falling below population growth rates in some cases and stalling completely in others. The traditional sources of yield gains that fueled growth since the 1940s are largely depleted, at least in developed countries. Agricultural biotechnology, however, has been documented to drive productivity growth in regions where it has been adopted. Agricultural biotechnology applies the principle tools of molecular and cellular biology to plant improvement. By 2010, the technology had largely been applied to providing crop resistance to common herbicides or genetic resistance to common pests. Herbicide-tolerant (HT) varieties permit the use of relatively cheap, broad-spectrum herbicides, like Round-Up, that are less toxic than more targeted alternatives. Insect-resistant (IR) varieties reduce the share of crops that are damaged by pests. Qaim and Zilberman (2003) used the damage control agent approach to model the impacts of transgenic pest control traits. According to the damage control function approach (Lichtenberg and Zilberman, 1986), output is the product of potential output and the fraction of the crop not damaged by pests. Pest control agents, such as pesticides and transgenic seed varieties, reduce crop damage. The impact of transgenic seeds on yield depends on their effectiveness in controlling damage and the degree of substitutability with other damage control inputs. When traits are inserted into generic seed varieties, rather than local varieties that are adapted for heterogeneous agronomic conditions, there may be yield loss due to reduction in potential yield.

230

Steven E. Sexton and David Zilberman

This effect has been termed yield lag. The yield effect of transgenic crops are likely to be higher in developing countries that tend to have severe pest problems and fewer pest control alternatives than developed counties (Qaim and Zilberman, 2003). The effects of transgenic crop adoption on crop yields and chemical use have been empirically tested with various methodologies and in various countries. As theory predicts, the literature generally finds positive yield effects and negative pesticide use effects associated with adoption of GM crops (see Qaim, 2009, for a survey). But the empirical record demonstrates that yield gains are not ubiquitous. Australia, for instance, experienced no measurable yield improvements from adoption of IR cotton (Fitt, 2003), even though India experienced yield gains of 37% (Qaim et al., 2006). Some researchers have interpreted heterogeneous effects as suggesting that conventional plant breeding offers greater potential for yield improvements than genetic engineering (Gurian-Sherman, 2009). Perhaps, because the empirical literature on biotechnology offers an unclear picture of the magnitude of GM crop yield effects, the adoption of the technology has been constrained by regulators who weigh heavily uncertain risks to human and environmental safety. Consequently, there has been greater acceptance of biotechnology in crops such as corn, soybean, and cotton that are used for feed and fiber rather than food. The commercialization of transgenic seed varieties in staple crops has been slow. A more thorough review of the evidence on biotechnology impacts on productivity growth supports the conclusion that the technology offers nontrivial benefits, particularly amid heightened demand for land because of biofuels. Sexton et al. (2009) showed that average yields of cotton and maize – two crops for which transgenic varieties have been introduced – have grown faster in the last several years than those of wheat and rice, two crops for which transgenic varieties have not been introduced. Growth rates in cotton and maize have surpassed those associated with the Green Revolution. Productivity growth in maize can likely climb higher still because most GM maize has been produced in developed countries, rather than developing countries where the gains are likely to be much larger. GM maize has been introduced in the United States, Argentina, and Brazil, but not in India, China, or African and Eastern European countries (James, 2009). Although much of the existent literature on GM crop effects can be criticized for limiting analysis to individual countries and regions and thereby limiting the robustness of estimations, Sexton and Zilberman (2011) provided a global assessment of the yield effects of first-generation biotechnology. Using data on global adoption of GM crops, crop yields and annual production, and using a fixed effects approach to control for country heterogeneity, they found that GM crop adoption significantly increased yields in cotton, maize, soybean, and rapeseed. The yield gains were greatest for cotton, at 65%. Maize yields were 45% higher with GM

Biotechnology and Biofuel

231

seed, and soybean and rapeseed yields were 25% and 12% higher, respectively. These effects are large, but as the authors noted, once differences in estimation strategies are accounted for, the estimates are not surprising. For instance, much of the previous literature used careful controls to estimate a GM seed effect, whereas Sexton and Zilberman (2011) estimated a seed adoption effect, which incorporates yield improvements from the reoptimization of farmers input use decisions given GM seed effects, as well as other changes to farmer behavior. Theory suggests, for instance, that farmers increase productive input use, such as fertilizer, as crop damage is reduced. Consistent with theory and with most earlier research, Sexton and Zilberman (2011) also found yield effects were larger in developing countries. Transgenic crops have boosted farm production not just through yield gains. They have also enabled an expansion of cropland. Sexton and Zilberman (2011) showed that adoption of transgenic varieties expands the external margin of agriculture, essentially making it profitable to farm land that is too low in quality for conventional crops. Adoption of HT crops has also been associated with adoption of double cropping – the farming practice in which a second cash crop is planted after harvest of the primary crop. In particular, HT soybeans and HT canola have been planted in rotation with wheat. This results in a ‘‘virtual’’ land expansion that is, in fact, more accurately modeled as an intensification of land use. In Argentina, adoption of HT soybean contributed to nearly a 10 million-acre increase in the area planted to soybean from 1996 to 2003 (Trigo and Cap, 2003). The gains in soybean production from new harvest area and from yield improvements are credited with meeting rapidly growing demand from China and India. Absent the output enhancements, food prices would have climbed even higher than those observed in 2008 (Sexton et al., 2009). To date, the adoption of agricultural biotechnology has been limited to a few crops (corn, soybean, cotton, rapeseed, and sugar beets) in a few countries. For more than a decade until 2010, the European Union instituted a moratorium on the approval of new genetically modified seed varieties entirely. Other countries followed Europe’s lead, so that in spite of relatively quick adoption in a handful of countries, there is vast potential for further adoption of agricultural biotechnology and associated gains in agricultural productivity.

3. The impacts of biofuel on food, fuel, and GHG emissions There is a quickly evolving literature on the impacts of biofuels on food, fuel, and GHG emissions. Using simple, conceptual models, we analyze these impacts and consider how they are affected by agricultural biotechnology adoption, which was shown in the previous section to boost farm output. First, to analyze the effects of agricultural biotechnology on the food market impacts of biofuel, consider the market for corn

232

Steven E. Sexton and David Zilberman Price of corn

C

S0 (PC)

C

S1 (PC)

A

A

PC

C

B PC

B DC(PC, PG) DF(PC) F

F

CA CB CA

Fig. 2.

Quantity of corn

Corn market equilibrium with ethanol demand and no constraint.

depicted in Figure 1. The demand for corn for food is denoted by DF ðPC Þ, where PC is the price of corn. The supply of corn is denoted by S C 0 ðPC Þ. The demand for corn for food and biofuel, which we will call total corn demand, is denoted by DC ðPC ; PG Þ; it is the sum of the demand for biofuel and the demand for food, where PG is the price of gasoline. Total demand is kinked; for a given price of gasoline, and for high corn prices, there is no demand for corn for biofuel. Although this partial analysis assumes a fixed price for gasoline, biofuel demand is expected to increase with gasoline prices, lowering the threshold corn price at which corn is demanded for biofuel. Consider first that there is no capacity to produce biofuel (or, equivalently, no demand for biofuel). In this case, total corn demand is given by DF ðPC Þ and equilibrium occurs at point B in Figure 2, with CBF units of corn produced for food at price PBC . If biofuel is introduced, then the equilibrium occurs at point A, with CA units of corn production and a A price of PA C . The quantity of corn produced for food is C F . The quantity of A A corn used for biofuel is given by C  CF . Food consumers incur losses in B this outcome because less food is available (C A F oC F ) and the price is A B higher (PC 4PC ). Losses to food consumers, however, translate into gains for gasoline consumers who benefit from the expansion of fuel supply by the amount of ethanol produced from C A  C A F units of corn. (The difference in energetic value between a gallon of ethanol and a gallon of gasoline is assumed away.)1

1 This is admittedly a simplification. In actuality, the energetic quality of ethanol is less than that of gasoline, so that one gallon of ethanol provides less work than one gallon of gasoline. Furthermore, end-use technologies, like car engines, can accommodate only a limited concentration of ethanol without modifications.

233

Biotechnology and Biofuel

Because of gains to gasoline consumers from expansion of the fuel supply, overall social welfare may improve, but welfare of food consumers unequivocally declines with the introduction of biofuels. Because biofuel demand grows with the price of gasoline, the welfare losses to food consumers is increasing in gasoline prices. The adverse food market effects of biofuels can be mitigated with the expansion of corn supply. As was demonstrated in the previous section, adoption of GM corn boosts corn output for a fixed stock of land. Suppose adoption of GM corn expands. This shifts the corn supply curve to S C 1 in Figure 2. The corn market equilibrium, then, occurs at point C. From Figure 2, it is evident that the price of corn is lower than PA C , so that food consumers benefit from a lower price. They also benefit from increased production of corn for food, as given by the intersection of price and demand for corn for food, DF ðPC Þ: It should also be evident from the figure, however, that even though food availability expands, the quantity of corn used for biofuel increases more. Now, suppose a mandate is imposed on biofuel production, as in the United States and the United Kingdom. The mandated quantity of biofuel production can be translated through the production function into a mandated quantity of corn for biofuel. When the mandate is binding, demand for corn is given by demand for food shifted out by CM units of corn, as depicted by the demand curve labeled DF(PC)þCM in Figure 3. For corn prices below PX, the mandate no longer binds and demand reverts to DC ðPC ; PG Þ. If the supply of corn is initially S C 0 in Figure 3, then the corn market equilibrium with the mandate occurs at point D, which corresponds to a corn price PD C that is higher than the price that obtains without the mandate, PA C . Total corn output is higher, but the quantity of corn produced for food, C D  C M , is smaller than it is without the Price of corn C

S0 (PC)

C

S1 (PC)

D

D

PC

A

PC

A

PX

E DC(PC, PG) DF(PC) + CMD DF(P

CM

Fig. 3.

X

CM

C)

Quantity of corn

Food market with biofuel mandate.

234

Steven E. Sexton and David Zilberman

mandate (C D  CM oCA F ). A large, binding mandate, therefore, makes food consumers worse off than they are without the mandate. In the United States and Europe, mandates are imposed in combination with subsidies, which are intended to make the mandates acceptable to the industry. The subsidy results in the government bearing some of the cost of the biofuel that is produced beyond market demand. De Gorter and Just (2009) showed that under the current targets and subsidies, U.S. biofuel policies lead to significant inefficiency in resource allocation and extra government costs. The results of their simulation analysis would have been different, however, if supply had been increased. If productivity gains were to sufficiently increase the supply of corn, for example, to S C 1 ðPC Þ in Figure 3, then the new equilibrium, point E in Figure 2, would yield a quantity demanded for corn that is greater than the mandate and a price for corn, PEC , that is lower than price under a binding mandate and limited supply (PEC oPD C ). The equilibrium at point E is efficient and improves the welfare of food consumers relative to the outcome under a binding mandate with less corn supply. Increased adoption of agricultural biotechnology, then, can overcome the problems associated with biofuel mandates by boosting corn supply. The theoretical results developed in this conceptual model are corroborated by a considerable body of research assessing the impact of biofuel on food and fuel prices. This research has generally relied on simulations based on estimated elasticities and has largely aimed to explain the role of biofuel in the dramatic rise of food prices between 2006 and 2008. The estimated impact of biofuel on food prices varies significantly depending on methodology and elasticity assumptions. For example, a study by Mitchell (2008) estimated that biofuel contributed to a 75% increase in the global food index between January 2002 and February 2008. Several other studies suggested that biofuels pushed coarse grains prices 40% higher and wheat prices 25% higher between 2007 and 2008. Studies by the Council of Economic Advisors (Lazear, 2008) and Glauber (2008) distinguished between the impact of biofuel on commodity prices versus retail food prices. Although they estimated the impact on agricultural commodity prices to be in the range of 20%–35%, the impact on retail prices was below 5% (Wiebe, 2008). The variability in the estimated food price effect of biofuel leads to a large variation in estimated welfare impacts (Cui et al., 2010). For example, assuming low gasoline price elasticities and moderate food price elasticities, Rajagopal et al. (2009) found that U.S. biofuel production in 2007 generated welfare gains to gasoline consumers that slightly dominated losses to food consumers. Under an assumption of inelastic food demand, however, they estimated losses to total social welfare of $60 billion in 2007 because food consumer losses dominated the gains to fuel consumers. This work, which incorporates a multi-market framework and models corn, soybean, and gasoline markets, does not consider the

Biotechnology and Biofuel

235

substantial social cost of biofuel subsidies (de Gorter and Just, 2009) and does not value the gains from improved energy security. Abbott et al. (2008) emphasized the difficulty in trying to quantify all factors that contributed to the recent food crisis. Contributing factors are many. For instance, Hochman et al. (2008) found that under plausible assumptions, the increase in demand for food because of economic growth had a larger impact on food prices than biofuel production. High gasoline prices also raised food prices by increasing costs of production. Carter et al. (2008) suggest that inventory management considerations explain much of the upward pressure on food prices. Others have noted that negative supply shocks and devaluation of the American dollar likely contributed to the rapid increase in food prices from 2007 to 2008 (e.g. Carter et al., 2008). Further assessment of the impact of biofuels on food prices is needed. Still, based on existing research, it is clear they had a nontrivial effect. Although the recession that began in 2008 provided relief from high food and gas prices through early 2010 because of reduced demand, global recovery is expected to restore upward pressure on prices. Allocation of additional land and harvest to biofuel production, therefore, can be expected to induce significant price increases in the future unless agricultural productivity grows with demand. Sexton and Zilberman (2011) simulate food prices in 2008 assuming biotechnology had not been used that year. Had food production not been boosted by adoption of GM crops, then lower supply would have driven corn prices 14% higher. Prices for soybean, wheat, and rapeseed would have been 21%, 9% ,and 13% higher, respectively, under assumptions of fairly elastic demand. By presenting a simple modeling framework and reviewing a literature based on empirical simulations, we have shown that biofuel production has a significant impact on food prices that transgenic seed adoption can minimize. A separate body of research investigates how food and fuel prices affect biofuel prices, the viability of the industry, and the future of the technology. To explore these linkages, we return to the conceptual model previously introduced. Assume buyers of biofuel – fuel blenders – mix biofuel with gasoline to sell to consumers. Denote the demand for biofuel by DB ðPB ; PG Þ and assume that biofuel and gasoline are substitutes in final demand, so that individual blenders can vary their mixture of the two.2 As before, demand for biofuel is increasing in the price of gasoline, PG, and declining in the price of biofuel, PG. The supply of biofuel provided by farmers (i.e. refiners who serve as middlemen), SB ðPB ; PC Þ, is increasing in the price of biofuels and decreasing in the price of the feedstock (e.g. corn).

2 They are not perfect substitutes in terms of both their physical properties and from the consumers’ perspective. Therefore, we assume that the demand for ethanol is negatively sloped for a given price of gasoline.

236

Steven E. Sexton and David Zilberman

If biofuel production is constrained neither by mandate nor by refining capacity, then equilibrium in the biofuel market occurs at point A in Figure 4, with BA units of biofuel produced at price PA B . The prices of corn and gasoline corresponding to this equilibrium are denoted by P0C and P0G , respectively. Biofuel production may be constrained from above by biofuel-refining capacity and from below by a biofuel mandate. Production capacity is denoted by BU and the mandate by BM. Let us next assume that capacity is greater than the mandate, as depicted in Figure 4, and that the mandate is binding. In this case, the price of biofuel paid to the sellers, PBB , is higher than marginal cost and higher than the free market price, PA B . This situation in represented by point B in Figure 4. A high corn price, like P1C 4P0C , causes biofuel supply to shift in to S B ðPB ; P1C Þ. Biofuel production, then, is determined by the mandate as depicted at point B in Figure 4. The marginal cost of producing the mandated quantity of biofuel – determined by the intersection of supply and the mandated quantity – is high at PBS B . This is the price refiners must receive to produce the mandated quantity of biofuel. But at BM units of biofuel production, buyers are only willing to pay a low price of PBB determined by the intersection of demand and the mandated quantity. In this case, biofuel refiners must sell to blenders at a loss. A binding mandate, therefore, can lead to bankruptcy and jeopardize the viability of the industry. Such outcomes are expected based on the analysis and simulations in Hochman et al. (2008). Biofuel subsidies, however, reduce the risk of bankruptcies. For instance, when a biofuel subsidy of S dollars per gallon is paid to blenders, their de facto price is ðPB  SÞ, which increases demand and can lead to an unconstrained Price of Biofuel

SB(PB, PC)

BS PB CS PB

0

SB(PB, PC )

CB

PB

B

PB

1

DB(PB, PG )

1 A

A

PB

0

DB(PB, PG )

BM

Fig. 4.

BA

BU

Quantity of Biofuel

Biofuel market with mandate and capacity constraint.

Biotechnology and Biofuel

237

equilibrium. The biofuel subsidy shifts the cost of overproduction of biofuel from the blenders to the government. When gasoline prices are high, for example, P1G 4P0G , biofuel demand is high, for example, DB ðPB ; P1G Þ, and may exceed refining capacity. In this case, output is determined by the capacity constraint, and prices facing buyers and sellers are determined by the intersection of the constraint and demand and supply, respectively. The price to buyers, therefore, is PCB B in . When the capacity constraint Figure 4, and the price to sellers is PCS B binds, the sellers of biofuel (i.e. the blenders) make a profit per gallon, CS given by (PCB B  PB ). Biofuel subsidies, then, may cause the market outcome to shift from market-clearing solution like point A to an outcome in which capacity constraints bind and refiners earn extra-normal profits. In other cases, it may cause a shift from one constrained solution to a more constrained solution in which prices paid by blenders are higher and profits to refiners are greater. Because the profit impact of the biofuel subsidy varies depending on the relative prices of corn and gasoline, Tyner and Taheripour (2007) proposed that the subsidy diminish when gasoline prices are high relative to corn prices. Serra et al. (2010) used time-series data to estimate nonlinear cointegrated dynamic price relationships between ethanol, corn, and oil prices. The analysis found a positive correlation between the price of ethanol and the price of corn and oil. Moreover, the analysis found that the importance of corn prices increases as the corn price increases relative to oil prices. This is consistent with the conceptual model presented here. Furthermore, Irwin and Good (2009) report transmission of instability from field crops to the biofuel sector in the United States between 2006 and 2009, when high commodity prices led to heavy financial pressure on biofuel refineries. Our conceptual framework suggests that improvements in agricultural productivity will increase the supply of biofuel for a given price of corn, reduce the price of corn, induce lower biofuel prices, and perhaps increase biofuel output. Supply enhancements from productivity gains reduce the likelihood of equilibria with binding mandate constraints (e.g. point B in Figure 3) and increase the likelihood of unconstrained equilibria (e.g. point A in Figure 3). Consequently, adoption of agricultural biotechnology reduces financial pressure on the biofuel industry, minimizing its risk, and contributing to its viability. Our analysis thus far has demonstrated that increased agricultural productivity is likely to reduce two problematic aspects of biofuel production and biofuel policies: the increase in food prices due to competition for inputs and the financial instability of the biofuel sector – partly induced by mandates – that is used to justify biofuel subsidies. Agricultural productivity gains can also improve the GHG emissions benefits of biofuels. The contribution of biofuels to GHG emissions can be divided into ‘‘direct effects’’ associated with the use of energy in production and processing of biofuels and ‘‘indirect effects’’ associated with extra GHG

238

Steven E. Sexton and David Zilberman

emissions as natural lands are converted to farming to accommodate the biofuel-induced demand for agricultural production (Fargione et al., 2008; Searchinger et al., 2008). The direct GHG effects of biofuels vary by crop and production technology. Sugarcane ethanol and sweet sorghum ethanol tend to have a lower GHG contribution than corn ethanol, for instance (Rajagopal and Zilberman, 2008). The indirect effects of biofuel production accrue because land-use change is associated with GHG emissions. To the extent that new land is brought into production of biofuels or other crops displaced by biofuels, then a GHG emissions penalty is imposed on biofuel production. Because carbon is stored in biomass, the clearing of biomass, either by burning or by decay, causes stored GHG to be emitted back into the atmosphere. There is also a GHG opportunity cost of land conversion because natural land sequesters carbon more effectively than productive land. Land conversion causes losses associated with the forgone carbon sequestration of natural land. It has been estimated that these effects, particularly the loss of carbon stored in existing biomass, cause biofuels to increase emissions in the short to medium term, although they eventually pay off the GHG emissions debt from land conversion over a period of more than 100 years (Fargione et al., 2008; Searchinger et al., 2008). The GHG emissions from land-use change can be minimized by reducing demand for land. GM crops do just that. They reduce the land intensity of agriculture, thereby lowering demand for new land in farming. Sexton and Zilberman (2011) estimated that yield enhancements from wider adoption of GM crops could have replaced or nearly replaced the share of crops diverted to biofuel uses in 2008. If the top 10 non-adopting corn producers had adopted GM corn at the same rate as the United States, then world corn output would have increased by 75 million tons in 2008; biofuels claimed 86 million tons. Vegetable oil production enhancements from GM soybean and rapeseed adoption would have been 2.3 million tons greater than the 8.6 million tons used for biofuels. Such production gains from intensification would have not only reduced food market effects of biofuel production but also minimized land-use expansion, improving the GHG balance of biofuels and lowering the risk of biodiversity loss. The widespread adoption envisioned by Sexton and Zilberman is constrained, however, by endogenous regulations that favor a precautionary approach to biotechnology. In spite of its capacity for improving food security and increasing the sustainability of biofuels, agricultural biotechnology has received at best a lukewarm reception by most government leaders. Europe imposed a de facto ban on GM crops beginning in 1998. In 2010, the European Commission granted its first approval to a GM seed variety in more than a decade. European resistance to the technology was justified on the grounds that it posed risks to food and environmental safety. In the United States, however, no fewer than three federal agencies responsible for plant and animal safety approved dozens of seed varieties after subjecting them to

Biotechnology and Biofuel

239

the same battery of tests as other new foods and plants. Still, the precautionary approach adopted in Europe proved contagious and other countries followed suit, including African countries that depend on Europe through trade and technology sharing arrangements ( Robert Paarlberg, 2008). Even countries such as India and China that aggressively adopted GM cotton varieties have been slow to accept GM crops for food and feed. The consequences of such bans are wide-ranging. They close off markets for the seeds themselves, discouraging investment in agricultural biotechnology R&D. They also close of export markets for GM crops, discouraging adoption by farmers. Even in the United States, where farmers planted the majority of cropland to GM seeds in 2008, the regulatory burden is still significant. The cost of regulatory compliance is estimated to be $15 million per seed-trait pair (Kalaitzandonakes et al., 2007). The consequences of the precautionary approach to biotechnology are nontrivial. As Collier (2008) noted regarding the European ban, ‘‘prior to 1996 grain yields in Europe tracked those in the United States. Since 1996 they have fallen by 1% to 2% per year. European grain production could be increased by around 15% were the ban lifted.’’

4. Beyond yield effects: The future of biofuel and biotechnology Heavy regulation has not just slowed adoption of existing technologies, it has also slowed the evolution of the industry. As Graff et al. (2009) showed, existing biotechnology resulted from interactions within the educational industrial complex. Much of the innovation that originated in universities was further developed and commercialized by private companies. This innovation process frequently involved significant investments by start-ups and major companies that are based on expectations about risk and reward. As the regulatory process has become more cumbersome and as the prospect for gains has declined, investment in new biotechnology innovation has slowed. Graff et al. (2009) document contraction in agricultural biotechnology product quality innovations beginning in 1999. The number of field experiments and new products continued to decline through 2009. The European ban on biotechnology that began in 1998 is likely a principle cause of the contraction, as is growing uncertainty about the technology around the world. The first generation of biotechnology produced tools capable of introducing many traits to improve agricultural efficiency. And yet, only IR and HT traits have been introduced – and only into a few crops. Scientists have identified numerous traits that improve product quality, and at least some of them can be commercialized (Graff et al., 2009). They include traits that improve the nutritional content of feed or its digestibility. These traits can reduce the amount of grain needed to feed cattle, thus reducing the demand for land for feed and releasing land for

240

Steven E. Sexton and David Zilberman

biofuel. Other traits in the pipeline may improve shelf life, reduce allergens or toxins, and improve nutritional content of foods. Drought-tolerant varieties are expected to be commercialized by 2015. They will make non-irrigated land in developing countries more productive and may lead to conversion of low-value grasslands to crop production. This would expand food production at low environmental costs and permit subsistence farmers to transition to cash crops. The yield-increasing and external margin effects of agricultural biotechnology can make biofuel production less impactful on food markets, less carbon intensive, and less vulnerable to commodity price variability. Biotechnology may also be important in developing biofuel products. The tools of molecular and cellular biology may hold the potential for engineering trees and other perennials for feedstocks. In theory, biotechnology can be used to engineer plants that produce more energetic material or are more easily converted to fuel. Genetic plant engineering is in its infancy. With a global commitment to plant science, the Gene Revolution can help develop a biofuel alternative to oil and in doing so help overcome the critical challenges of mitigating climate change and fueling and feeding a growing world.

References Abbott, P.C., Hurt, C., Tyner, W.E. (2008), What’s driving food prices? Issue Report 48495, Farm Foundation, Oak Brook, IL, July 23. Alston, J.M., Beddow, J.M., Pardey, P.G. (2009), Agricultural research, productivity, and food prices in the long run. Science 325 (5945), 1209–1210. Carter, C., Rausser, G., Smith, A. (2008), The food price boom and bust. ARE Update 12 (2), 2–4. Collier, P. (2008), The politics of hunger: how illusion and greed fan the food crisis. Foreign Affairs 87 (6), 67–79. Cui, J., Lapan, H., Moschini, G., Cooper, J. (2010), Welfare impacts of alternative biofuel and energy policies. Working Paper No. 10016, Department of Economics, Iowa State University, Ames, IA, June. de Gorter, H., Just, D.R. (2009), The welfare economics of a biofuel tax credit and the interaction effects with price contingent farm subsidies. American Journal of Agricultural Economics 91 (2), 477–488. [EIA] Energy Information Administration. International Energy Outlook 2009. Washington, DC: U.S. Department of Energy. Fargione, J., Hill, J., Tilman, D., Polasky, S., Hawthorne, P. (2008), Land clearing and the biofuel carbon debt. Science 319 (5867), 1235–1236. Fitt, G. (2003), Implementation and impacts of transgenic cottons in Australia. ICAC Recorder 21, 14–119.

Biotechnology and Biofuel

241

Food and Agriculture Organization of the United Nations (FAO). (2009), How to feed the world in 2050. Food and Agriculture Organization of the United Nations, Rome. Glauber, J. (2008), Statement of Joseph Glauber, Chief Economist before the Joint Economic Committee, U.S. Congress, May. Global Subsidies Initiative (GSI). (2007), Biofuels at what cost: government support for ethanol and biodiesel in selected OECD Countries. International Institute for Sustainable Development, Geneva, Switzerland. Godfray, H.C.J., Beddington, J.R., Crute, I.R., Haddad, L., Lawrence, D., Muir, J.F., Pretty, J., Robinson, S., Thomas, S.M., Toulmin, C. (2010), Food security: the challenge of feeding 9 billion people. Science 32 (5967), 812–818. Graff, G.D., Zilberman, D., Bennett, A.B. (2009), Correspondence: The contraction of agbiotech product quality innovation. Nature Biotechnology 27 (8), 702–704. Gurian-Sherman, D. (2009), Failure to yield: evaluating the performance of genetically engineered crops. Union of Concerned Scientists, Cambridge, MA. Heaton, E.A., Dohleman, F.G., Long, S.P. (2008), Meeting biofuel goals with less land: the potential of Miscanthus. Global Change Biology 14 (9), 2000–2014. Hochman, G., Sexton, S.E., Zilberman, D. (2008), The economics of biofuel policy and biotechnology. Journal of Agricultural and Food Industrial Organization 6 (2), Article 8. Irwin, S.H., Good, D.L. (2009), Market instability in a new era of corn, soybean, and wheat prices. Choices 24 (1), 6–11. James, C. (2009), Global status of commercialized biotech/GM crops: 2009. Brief 39–2009, International Service for the Acquisition of Agribiotech Applications, Ithaca, NY. Kalaitzandonakes, N., Alston, J.M., Bradford, K.J. (2007), Compliance costs for regulatory approval of new biotech crops. Nature Biotechnology 25, 509–511. Lazear, E.P. (2008), Testimony of the President’s Council of Economic Advisors concerning responding the global food crisis. Presented to the Senate Committee on Foreign Relations, May 14. Lichtenberg, E., Zilberman, D. (1986), The econometrics of damage control: why specification matters. American Journal of Agricultural Economics 68 (2), 262–273. Mitchell, D. (2008), A note on rising food prices. World Bank Policy Research Working Paper Series No. 4682, World Bank, Washington, D.C. Paarlberg, R. (2008), Starved for Science: How Biotechnology is Being Kept out of Africa. Harvard University Press, Cambridge, MA. Qaim, M. (2009), The economics of genetically modified crops. Annual Review of Resource Economics 1 (1), 665–694. Qaim, M., Subramanian, A., Naik, G., Zilberman, D. (2006), Adoption of Bt cotton and impact variability: insights from India. Review of Agricultural Economics 28 (1), 48–58.

242

Steven E. Sexton and David Zilberman

Qaim, M., Zilberman, D. (2003), Yield effects of genetically modified crops in developing countries. Science 299 (5608), 900–902. Rajagopal, D., Sexton, S.E., Hochman, G., Roland-Holst, D., Zilberman, D. (2009), Model estimates food-versus-biofuel trade-off. California Agriculture 63 (4), 199–201. Rajagopal, D., Zilberman, D. (2008), Environmental, economic and policy aspects of biofuels. Foundation and Trends in Microeconomics 4 (5), 353–468. Schultz, T.W. (1953), The Economic Organization of Agriculture. McGrawHill, New York. Searchinger, T., Heimlich, R., Houghton, R.A., Dong, F., Elobeid, A., Fabiosa, J., Tokgov, S., Hayes, D., Yu, T.H. (2008), Use of U.S. croplands for biofuels increases greenhouse gases through emissions from land-use change. Science 31 (5867), 1238–1239. Serra, T., Zilberman, D., Gil, J.M., Goodwin, B.K. (2010), Price transmission in the US ethanol market. In: Khanna, M., Scheffran, J., Zilberman, D. (Eds.), Handbook of Bioenergy Economics and Policy. Springer, New York, pp. 53–72. Sexton, S. and Zilberman, D. ‘‘How Agricultural Biotechnology Boosts Food Supply and Accommodates Biofuel.’’ NBER Working Paper 16699. Washington, DC: National Bureau of Economic Research, 2011. Sexton, S.E., Zilberman, D., Rajagopal, D., Hochman, G. (2009), The role of biotechnology in a sustainable biofuel future. AgBioForum 12 (1), 130–140. Sperling, D., Gordon, D. (2009), Two Billion Cars: Driving Toward Sustainability. Oxford University Press, Oxford. The Economist. (2009), Pocket world in figures 2010 Edition. London, Profile Books, Ltd. Trigo, E.J., Cap, E.J. (2003), The impact of the introduction of transgenic crops in Argentinean agriculture. AgBioForum 6 (3), 87–94. Tyner, W.E., Taheripour, F. (2007), Future biofuels policy alternatives. Paper presented at the Farm Foundation/USDA conference on Biofuels, Food, and Feed Tradeoffs, April 12–13, 2007, St. Louis, Missouri. Available at http://www.farmfoundation.info/news/articlefiles/943-tynerandtaheripourcorrected6-11-07.pdf. Wiebe, K. (2008), Biofuels: Implications for natural resources and food security in developing countries. Paper presented at Sustainable Biofuels and Human Security Conference, University of Illinois, May 12–13, Urbana-Champaign, IL. World Health Organization (WHO), Food and Agriculture Organization of the United Nations. (2003), Diet, nutrition and the prevention of chronic diseases. WHO Technical Report Series No. 916, World Health Organization, Geneva, Switzerland.

CHAPTER 10

Consumer Preferences for Genetically Modified Food Jayson L. Lusk Department of Agricultural Economics, Oklahoma State University, 411 Ag. Hall, OSU, Stillwater, OK 74078-6026, USA E-mail address: [email protected]

Abstract Purpose – Despite the existence of hundreds of studies and several review articles on consumer preferences for genetically modified (GM) food, it remains difficult to ascertain the current state of knowledge on the topic. The purpose of this chapter is to distill some of the key findings from the body of research on consumer preferences for GM food. Approach – In reviewing key pieces of literature, including two meta-analyses, the chapter identifies four key unresolved questions and includes discussions on how the questions might be resolved. Findings – The chapter identifies four questions in need of additional thought and research. The questions relate to (1) why the market for GM-free food is so small in the United States despite the large estimated willingness-to-pay premiums for GM-free food, (2) why consumers remain so uninformed about biotechnology despite their seemingly high levels of aversion, (3) why economists have generally ignored the information-content of GM food policies, and (4) why it is so difficult to determine why U.S. and European consumers have seemingly reacted so differently to GM foods. Value – This chapter should be useful to those interested in learning about the current state of knowledge on consumer preferences for GM food, and to those seeking to identify areas in need of additional research. Keywords: Biotechnology, consumer demand, genetically modified food, willingness-to-pay JEL Classifications: Q13, Q18 1. Introduction Topics surrounding the advent of genetically modified (GM) food have fascinated agricultural economists for more than a decade. Arguably, the Frontiers of Economics and Globalization Volume 10 ISSN: 1574-8715 DOI: 10.1108/S1574-8715(2011)0000010015

r 2011 by Emerald Group Publishing Limited. All rights reserved

244

Jayson L. Lusk

primary reason the topic has been of such interest and controversy is that some consumers were averse to the new technology when it appeared on the market. For example, in 1997 Hoban reported that less than half the population of many European countries was willing to eat GM food; he also showed that almost two-thirds of Swedes, Germans, and Austrians considered GM food to be a serious health hazard. Were it not for consumer concerns about GM food both domestically and abroad, the economic analysis of GM food would amount to little more than a traditional analysis of technology adoption. Accordingly, journals are now replete with findings from research on consumer aversion to GM food. For example, Dannenberg (2009) reported that there are 51 studies that have provided 114 explicit estimates of consumer willingness-to-pay (WTP) for GM food. Of course, there have been numerous additional studies published on consumer preferences for GM food that have used measures of consumer acceptance other than WTP (e.g., Hoban, 1997, Gaskell et al., 1999). Despite this fact, and despite the existence of several reviews of consumer research (e.g., Bredahl et al., 1998; Costa-Fonta et al., 2008), it is difficult to distill what is known about consumer preferences for GM food. Moreover, such work is often looked upon with a skeptical eye by applied theorists and producer and industry groups, sometimes with good reason but sometimes out of mere prejudice. Thus, on the one hand, there seems to be general consensus that information is needed on consumer preferences for GM food, and yet, on the other hand, such work is difficult to summarize and is often looked upon with distrust. As indicated, there have been a number of reviews of the research on consumer demand for GM foods, and as such, there is little need for another summary of the literature. A more constructive way forward is to convey what I have learned from the body of consumer research, with a focus on unresolved issues that arise from a careful scrutiny of the literature. Focusing on a few unresolved questions is one way to take a critical look at the existing survey and experimental evidence on consumer preferences for GM food in a way that asks what the extant findings really mean. By attempting to resolve the questions, some weaknesses in previous survey and experimental research are exposed, but the investigation also reveals insight into topics in need of additional research.

1.1. Question #1: Most studies suggest consumers are willing to pay significant premiums to avoid GM foods, so why do foods advertised as ‘‘GM-free’’ have almost no market penetration in the United States? There is ample evidence from surveys and experiments indicating that consumers are willing to pay sizable premiums to avoid GM foods. Figure 1, for example, shows the distribution of WTP premiums for

Consumer Preferences for Genetically Modified Food

245

6

Number of Studies

5 4 3 2 1 0

>110% 105% 100% 95% 90% 85% 80% 75% 70% 65% 60% 55% 50% 45% 40% 35% 30% 25% 20% 15% 10% 5% 0% −5% −10% −15% −20% <−20% WTP Premium for Non-GM Food (%)

Fig. 1. Distribution of willingness-to-pay premiums for nongenetically modified foods over genetically modified foods from 57 studies. Source: Lusk et al. (2005).

non-GM food over GM food from the 57 studies reviewed in the metaanalysis conducted by Lusk et al. (2005). Of the 57 studies included in the meta-analysis, about half were conducted with U.S. citizens, a third were conducted with Europeans, about 9% were conducted with consumers in China, Japan, or Taiwan, and the remaining studies were conducted with consumers in other parts of the world. The data reveals that consumers, worldwide, are averse to GM foods and are often willing to pay to have non-GM rather than GM food. Indeed, 82% of the studies reported positive WTP premiums for non-GM food. Of the remaining 16% of studies that reported negative WTP premiums (meaning consumers would pay a premium for the GM food), almost all correspond to studies which asked consumers to value a GM food that provided a direct benefit to the consumer such as improved nutrition or health. Across all studies, the weighted average WTP premium (weighted by the number of observations used to estimate the respective study’s valuation) for non-GM vs. GM food is about 25%. A key observation from Figure 1 is that there is a very wide dispersion across the studies included in the meta-analysis. WTP premiums for GM food ranged from as low of –67% (meaning people valued GM over non-GM) to a high of 784% (next highest was 169%). Despite the substantial variation in WTP premiums, Lusk et al. (2005) showed that as much as 89% of the variation in WTP data showed in Figure 1 can be explained by a simple econometric model. In particular, the mean WTP

246

Jayson L. Lusk

reported by a study was accurately predicted by the characteristics of the sample of consumers studied (e.g., students vs. nonstudents, United States vs. Europe), methods for eliciting WTP (e.g., hypothetical vs. real money, choice experiment vs. contingent valuation), and the characteristics of the food valued (e.g., vegetable vs. meat, first vs. second generation GM). As shown in Figure 2, the meta-analyses revealed that European consumers’ WTP premium for non-GM food was 29% higher than U.S. consumers’ WTP premium for non-GM food. The results also indicated that consumers were WTP 28, 41, and 49% more for non-GM meat products than they were for non-GM produce, non-GM processed food, and non-GM oil, respectively. Stated differently, consumers were most averse to the use of biotechnology in meat products and were least averse to the use of biotechnology in oil. Lusk et al. (2005) also found that the WTP premium for non-GM over GM was 49% higher for so-called firstgeneration products compared to second-generation products that were engineered to provide tangible benefits to consumers, such as increased nutrition. In other words, people were willing to pay much more to avoid GM food if the biotechnology did not create tangible benefits for the consumer. The meta-analysis also revealed that WTP values were significantly influenced by the methods used to elicit consumer preferences. In

1st generation technology vs. biotechnology used to improve taste, health, or nutrition

49%

28%

Food is a meat product vs. produce

49%

Food is a meat product vs. oil

41%

Food is a meat product vs. processed food Data collected using hypothetical vs. nonhypothetical methods

40%

Data collected in-person vs. by mail or phone

63%

29%

European vs. US consumers 0%

10%

20%

30%

40%

50%

60%

70%

Change in WTP Premium for Non-GM Food (%)

Fig. 2. Effect of study and production characteristics on estimated willingness-to-pay premiums for nongenetically modified foods over genetically modified foods from 57 studies. Source: Lusk et al. (2005).

Consumer Preferences for Genetically Modified Food

247

particular, Lusk et al. (2005) reported that WTP premiums for non-GM food were 40% higher when the study authors used research methods that were hypothetical (e.g., mail or phone surveys) as compared to methods that were nonhypothetical (e.g., experimental auctions), a finding which is consistent with the vast majority of studies on ‘‘hypothetical bias’’ (see List and Gallet, 2001). In addition, approaches that elicited people’s preferences in-person generated estimates of WTP premiums for non-GM food that were 63% higher than approaches that elicited preferences from people by phone or mail. This finding is consistent with the literature showing that people are more likely to respond in socially desirable ways when interviewed in person (see Leggett et al., 2003). In a testament to the vigor with which researchers have addressed the subject, there has been almost a doubling in the number of WTP estimates reported in the literature since the original meta-analysis conducted by Lusk et al. (2005). Dannenberg (2009) extended the results of Lusk et al. (2005) to include data from 51 studies reporting 114 valuations, and her results are largely consistent with those of Lusk et al. (2005). For example, Lusk et al. (2005) reported a simple unweighted average premium of 42% for non-GM foods, which can be compared with Dannenberg’s estimate of 45%. After reviewing the existing body of consumer research, both meta-analyses clearly show that consumers are willing to pay sizable premiums to avoid GM food. Indeed, Dannenberg’s results suggest that, if anything, WTP for non-GM food has been growing over time, particularly in Europe. Given these findings, one would expect that farmers and food retailers face strong economic incentives to provide consumers what they apparently say they want in the form of non-GM foods. And yet, it is rare to find foods advertised as ‘‘non-GM’’ in the United States. As Fernandez-Cornejo and Caswell (2006) noted, ‘‘In the United States, many products contain [genetically engineered] ingredients, and the demands for these products apparently have been unaffected by negative opinions about biotechnology expressed in surveys. A few specialty brands are marketed as ‘[genetically engineered] free,’ but they represent a small percentage of supermarket sales.’’ There are several possible reasons for this apparent contradiction. It could be argued that the WTP estimates obtained from surveys and experiments are biased or simply wrong. In hypothetical surveys, it is well known that people tend to indicate higher WTP amounts than what they are really willing to pay (e.g., see List and Gallet, 2001). Moreover, in faceto-face research methods, people often face social pressure to give the sort of answers they believe the researcher wants to hear (e.g., see Lusk and Norwood, 2009). Such findings suggest that the WTP premiums for non-GM food observed in previous studies are not reflective of consumers’ true underling preferences for GM but rather reflect particular features of the methods used to elicit WTP. As indicated previously, there is some truth to this argument as Lusk et al. (2005) and Dannenberg (2009) show

248

Jayson L. Lusk

that much of the variation in WTP observed across studies results from differences in the questioning methods. However, this cannot be the complete answer because even studies that use real money and real food find that many consumers are willing to pay significant premiums for nonGM food (e.g., Lusk et al., 2006; Rousu et al., 2007) as do studies where people’s answers are completely anonymous (Hu et al., 2005). Moreover, research on the external validity of consumer research methods has yielded largely positive results. For example, Chang et al., (2009) showed that the same consumer preference elicitation methods that have been used to elicit consumer WTP for GM food performed quite well in predicting the actual market share of three new products introduced in a grocery store. As another example, both Johnston (2006) and Vossler et al. (2003) have shown that answers to contingent valuation questions accurately predicted a real referendum vote. This is not to say that survey/experimental results are perfect – only that it is difficult to believe that so many studies could be so wrong about consumer WTP for GM food. Nevertheless, studies testing the external validity of WTP estimates for GM food are badly missing. However, researchers face a Catch 22 because there are virtually no scanner data or grocery sales of ‘‘GM-free’’ food that could be used to validate the survey/experimental answers. If inaccurate estimates of consumer demand are not the culprit, perhaps the answer is on the supply-side. Estimates suggest that the cost of segregating non-GM soybeans and corn at the 5% tolerance level are only about 10% of the farm-level corn and soybean prices (Lin et al., 2000; Bullock and Desquilbet, 2002). Nevertheless, the cost difference is well below the estimated premiums consumers are WTP for non-GM food (note, however, that the segregation costs increase exponentially as tolerance levels fall below 1%). The farm- and elevator-level segregations costs are also a lower bound on the costs of segregation throughout the entire marketing channel. Because organic is a more encompassing label than ‘‘GM-free,’’ the added cost of providing ‘‘GM-free’’ must logically be less than the cost of providing organic, assuming the same scale of production. The retail price premiums for organics differ widely across food items, but range from 10% to more than 100% premiums.1 One possible reason why the market for ‘‘GM-free’’ is so small is that the cost of providing ‘‘GM-free’’ is so high.2 However, this seems like an unlikely explanation given the existence of the organic food market. 1

For example, see: http://www.ers.usda.gov/Data/OrganicPrices/. It is interesting to contrast the situation in the United States and Europe with that in Japan, where there are relatively more ‘‘non-GM’’ products sold. One reason for this difference might be that Japan enforces a relatively lenient 5% tolerance level for products labeled as ‘‘GM-free,’’ making the costs of the label significantly smaller than in other parts of the world such as Europe which imposes a 1% tolerance level. 2

Consumer Preferences for Genetically Modified Food

249

A more plausible explanation for the fact that there is no market for ‘‘GM-free’’ food in the United States is that demand for this food characteristic has been subsumed under the broader umbrella of organics. Organic labeling standards require that organic foods are also non-GM. Thus, the existence of organic foods in the United States has perhaps hindered the emergence of what would, in many ways, be a redundant market for non-GM food. However, even this explanation seems unlikely because the market share for organic foods is relatively small. Retail market-share data on organic sales is a bit unreliable, but farm-level data indicates that less than 1% of U.S. farmland is devoted to organics (Dimitri and Oberholtzer, 2009). Even in the fruit and vegetable category, where organic sales are the highest, only about 5% of fruit and vegetable cropland is devoted to organic production (Dimitri and Oberholtzer, 2009). Although organic foods have experienced rapid growth in recent years, the trend appears to be slowing and perhaps even reversing (Mintel, 2008) and probably accounts for far less than 5% of all food retail sales. Thus, while the existence of the organic food market partially explains the lack of a ‘‘GM-free’’ food market, the organic market size is too small to reconcile the size of the WTP estimates with the size of the market. This is especially true when one notes that the higher costs of organic production, which are estimated at around 10 to 30% (Lohr, 2001), are less than the estimated WTP premiums for non-GM food. The reality is that consumers do not know much about the food they consume. For example, only 26% of consumers believe they have eaten a GM food (Pew, 2006). The reality is that close to 100% of U.S. consumers have eaten a GM food. In attempting to resolve the apparent contradiction between the high WTP for non-GM food observed in surveys and the low market share of non-GM products, it is important to recognize that consumer research, by asking people what they are willing to pay for nonGM food, serves to inform people about modern agricultural production practices. Thus, the WTP observed from surveys or experiments corresponds to that of informed consumers, whereas market behavior corresponds to uninformed consumers. It could be that the market share for organic and non-GM food would be much higher if consumers knew how corn and soybeans were actually produced in the United States. That is, the ‘‘average’’ consumer lacks the information that would translate the WTP observed in consumer studies into market behavior. The difference in informed and uninformed consumers may well explain much of the difference in predictions from surveys/experiments and the actual market share for non-GM food. If this is true, one might wonder why firms do not advertise and promote non-GM products. Advertising can transform an uninformed consumer into an informed one. One reason that large agribusiness firms may not aggressively advertise non-GM products is that the firms also (primarily) sell products made with GM ingredients. While advertising could increase

250

Jayson L. Lusk

the demand for non-GM foods, large agribusiness firms may be reluctant to advertise non-GM for fear of the adverse effects such advertising may have on demand for their existing product lines.

1.2. Question #2: If consumers are concerned about GM food, why are they generally uninformed and unknowledgeable about the technology? As indicated, consumers do not know much about biotechnology. For example, research by the Pew organization (2006) found that almost 60% of U.S. citizens have not ‘‘seen, read, or heard recently about genetically modified food that is sold in grocery stores.’’ Moreover, 74% of consumers indicated ‘‘very little’’ or ‘‘no’’ knowledge of government regulation of GM food. These findings led Pew (2006) to conclude, ‘‘Public knowledge and understanding of biotechnology remains relatively low.’’ Such findings have been confirmed in various ways in numerous other studies. Gaskell et al. (1999), for example, asked United States and European consumers to answer true/false questions such as, ‘‘Ordinary tomatoes do not contain genes while genetically modified tomatoes do.’’ They found that, on average, U.S. citizens only answered about 8% of the questions correctly and Europeans only answered about 30% of the questions correctly. Although the evidence suggests low levels of consumer knowledge, it also suggests relatively high levels of concern for GM food as indicated by the WTP estimates discussed in the previous section. But, if consumers are really so concerned about biotechnology, why have they not troubled themselves to become informed about the technology? To put the issue another way, consider the results from the study conducted by Noussair et al. (2002). They asked French consumers to bid in experimental auctions on different foods, one of which had a label indicating that it was made with GM ingredients. Initially, WTP was about the same for GM and non-GM foods. Only when the researchers specifically drew attention to the differing labels did consumers bid less on the GM food; WTP for the GM food fell by about 30% following the labeling announcement. As Noussair put it (2002, p. 52), ‘‘the absence of a negative effect on demand in reaction to products containing GMOs is in large measure due to the fact that customers do not notice the labeling.’’ One would think, however, that if concern for GM food is really so pronounced, that consumers would scour labels to ensure the absence of such ingredients. But, this is not how people act. Why the discrepancy between people’s supposed aversion to GM food on the one hand and their unwillingness to seek out information about the subject on the other? One answer is that time is scarce, and time spent informing one’s self about biotechnology is time not spent doing other enjoyable activities. Presumably, people will spend time acquiring

Consumer Preferences for Genetically Modified Food

251

information up to the point where the marginal benefit of information acquisition equals the marginal cost. As we have shown, however, most people have chosen to acquire very little information about biotechnology. It would seem difficult to imagine that the lack of knowledge, however, is a result of the high costs of information acquisition as any Google search will turn up countless easy-to-read documents on the subject. Rather, one might suspect that the value of information to consumers is quite low. This is, in fact, what Rousu et al. (2007) found. Even in the most optimistic scenarios they considered, the expected value of objective third-party information about GM food was only about 5 cents per bag of potato chips, bottle of vegetable oil, or bag of potatoes. As compared to people who previously had no information on GM food at all, Rousu and Lusk (2009) found that positive information about GM food had an average value of less than a penny for cookies that they were required to eat. It seems that people are not knowledgeable about GM foods because becoming so simply isn’t worth it. But, if consumer WTP for information about GM is so small, why is their WTP to avoid GM food so large? One answer is that people use other types of information to draw inferences about the safety of food. As argued by Brossard and Nisbet (2007), ‘‘Faced with many competing demands for their time and attention, most citizens lack the ability and/or the motivation to be fully informed about an issue and instead rely heavily on information short-cuts such as values and trust in combination with the interpretations of the issue most readily available from media coverage to form an opinion.’’ If consumers have trust in regulatory agencies and the firms that supply food, they need not bother themselves with learning about each new technology and food safety issue that comes along. In a survey of over 2,000 U.S. households, for example, Lusk (2008) found that only 10% of respondents disagreed with the statement that ‘‘In general, the meat and milk I buy from grocery stores is safe to eat.’’ Thus, it may be that trust in the overall safety of food creates an environment in which people do not feel the need to know more about biotechnology, despite the fact that when informed of biotechnology consumers prefer non-GM to GM food. Numerous studies have shown that trust in government and food regulatory agencies is positively correlated with acceptance of GM food (e.g., Kuznesof and Ritson, 1996; House et al., 2004). The downside to this situation is that if trust levels fall, for example in response to a food safety scare, consumers may seek out information on other food attributes perceived as risky – such as biotechnology. Many authors have argued that more public education is needed to improve consumer acceptance of biotechnology. Such arguments are typically based on the observation that people with higher levels of education (and stated subjective, knowledge of GM food) are generally more accepting of GM food. However, as discussed by House et al. (2004), subjective knowledge may not correlate well with objective knowledge

252

Jayson L. Lusk

about GM food, and thus it is unclear whether ‘‘education’’ will really have the kind of influence many authors propose. Moreover, proponents of ‘‘education’’ rarely experimentally test the effects of education but rather draw inferences from cross-section analyses where a participant’s education level is endogenously determined with a host of other factors that also affect concern for GM.

1.3. Question #3: If consumer WTP for GM food is affected by information, why do most economic models assume WTP for GM food is unaffected by policies and labels? In an experimental setting, Rousu et al. (2007) showed that consumer WTP is significantly influenced by the provision of positive, negative, and third-party information. They found that consumers who received positive information about GM food actually placed a greater value on GM-labeled food than on non-GM-labeled food for two of the three goods analyzed. However, individuals who received negative information about GM foods discounted the GM-labeled foods by about 35%. Consumers who received both positive and negative information about GM foods discounted the GM-labeled foods anywhere from 17 to 29%. Lusk et al. (2004) also showed that information on environmental benefits, health benefits, and benefits to the third world significantly decreased the amount of money consumers demanded to consume GM food. In particular, 24% of U.S. consumers switched from preferring non-GM to preferring GM after receiving information on the potential environmental benefits of GM food, and about 16% of U.K. and French consumers switched from preferring non-GM to preferring GM after receiving information on the potential of biotechnology to increase the supply of food to consumers in the third world (see Rousu and Lusk, 2009). Similar results have been found in nonexperimental studies. Kalaitzandonakes et al. (2004), for example, showed that media coverage surrounding the Starlink corn incident in the United States affected consumer demand for taco shells. These and other studies clearly show that consumer demand for GM food is significantly affected by information. If people are influenced by information about GM food, government policies might signal consumers as to the safety of GM food. For example, suppose a consumer learned that their country banned the production of GM crops. What sort of information might this action send? The federal government directly and indirectly finances a wealth of research and employs thousands of scientists. When the government reaches a decision about a policy, it is reasonable to assume that its own experts and funded research are brought to bear on the problem. Thus, individuals might interpret policy as a decision by knowledgeable experts. Anecdotal

Consumer Preferences for Genetically Modified Food

253

evidence from interest groups suggests a common perception that policy could influence beliefs about safety. For example, many agricultural groups in the United States strongly oppose mandatory GM food labels because of the contention that consumers will interpret the label as an indication that food produced through biotechnology is less safe than food produced through traditional means. Lusk and Rozan (2008) found support for such arguments using survey data from U.S. households. They found that individuals who believed the government imposed a mandatory labeling policy for GM food also believed GM food was less safe and were less willing to eat and buy GM food than consumers who either believed no policy was in place or were uncertain on the matter. Apparently, consumers interpreted the presence of a mandatory label as a signal that GM food is risky and in need of stricter regulatory oversight as compared to non-GM food. The implication of this result is important. Many theoretical studies have investigated the welfare effects of GM food adoption and labeling by assuming that consumer WTP for GM food is independent of the policy in place (for examples see, Crespi and Marette, 2003; Fulton and Giannakas, 2004; Lapan and Moschini, 2004; Lence and Hayes, 2005). In these models, the welfare effects of a country adopting GM technology, for example, depends on the size of the demand shift (e.g., the WTP premium for non-GM food) relative to the decrease in food prices that results from the more efficient production made possible via biotechnology. However, the results in Lusk and Rozan (2008) and the many studies on effects of information on WTP for GM food suggest that these conventional modeling approaches may be incorrect because they ignore the possibility that policies themselves send information about the safety of GM food. When policies shift, WTP for GM food may also shift. Consider another example. Most consumers know very little about milk production, and when asked how much they would pay for a quart of milk in real money experiments, Kanter et al., 2009 found that WTP for conventional milk depended heavily on whether they knew organic or rBST-free milk was available for sale. The mere presence of rBST-free and organic milk stigmatized conventional milk. The amount people were willing to pay for conventional milk fell $0.35–$0.50 per quart when they learned that rBST-free and organic milk were available. These results suggest that policies which require rBST labels, for example, would cause a shift in demand for conventional milk. In traditional modeling approaches, the welfare effects of labeling policies depend primarily on the costs of labeling and the extent to which consumers are willing to pay for the labeled and unlabeled food, but these results suggest that WTP itself depends on whether a label is present. It is often difficult to know a priori how particular policies might influence WTP for GM food. However, if policies affect WTP, then the results of traditional cost–benefit analyses are likely biased. Elicited

254

Jayson L. Lusk

demand for GM food prior to a policy may provide biased estimates of the welfare consequences that occur once the policy is enacted, as individuals’ safety beliefs may change when the policy is changed. Welfare analysis itself becomes more complicated as judgment calls must be made about whether the benefits and costs of a policy should be measured relative to the ex ante or ex post demand curve.

1.4. Question #4: Why there is so little consensus on the factors explaining the difference in U.S. and European preferences for GM food? One of the most well-known stylized facts from the consumer research is that U.S. consumers are more accepting of GM food as compared to European consumers (Hoban, 1997; Gaskell et al., 1999). This early survey work has been largely confirmed by economic experiments involving real food and real money (Lusk et al., 2006). The stylized fact also conforms to the ‘‘facts on the ground.’’ Europe has been relatively slow to approve GM crops for commercialization and has enforced a mandatory labeling of GM food. By contrast, the United States has approved numerous GM crops for commercialization and has no mandatory labeling policy in place. The stylized fact has long intrigued researchers, and many authors have developed pet theories to explain the differences. The differences have been attributed to differences in trust, culture, previous food safety experiences, knowledge, preferences for nature, and politics and policies, among other explanations. However, there appears to be little consensus in the literature as to why such differences exist. Moreover, empirical research on the matter has proven to be less than persuasive. For example, Lusk et al. (2006) show that a model controlling for differences in demographics, objective and subjective knowledge, trust in government, and attitudes toward the environment, new foods, food quality, and technology could only explain about one-fifth of the difference in U.S. and French consumers’ bids to refrain from eating a GM food; the same model could explain only about one-third to one-half of the difference in U.S. and U.K. consumers’ bids. One factor that is often overlooked in focusing on mean or median differences across the United States and Europe is the wide heterogeneity in consumer preferences in Europe. Gaskell et al. (1999) and Hoban (1997) show substantive differences in GM food acceptance across different European countries. Hoban (1997), for example, reported that 64% of consumers in the Netherlands said they were willing to eat a GM food, whereas only 30% of German consumers indicated they were willing to eat a GM food. Using real food, real money experiments, Lusk et al. (2006) and Noussair et al. (2004) show that many French and British are relatively unconcerned about GM foods. The point is that the stylized fact

255

Consumer Preferences for Genetically Modified Food

on U.S. and European differences in preferences for GM food, while arguable true at the mean, belies the fact that there is substantial overlap in the U.S. and European distributions of preferences. For example, consider the results in Figure 3, which come from the study of Lusk et al. (2006) who determined the minimum amount that consumers demanded to be paid to eat a GM cookie in the United States, France, and England using experimental auction methods. The figure shows that 52% of French consumers demanded more than $2 to eat the GM cookie, whereas only 9% of U.S. participants demanded an amount this high. This finding only reinforces the stylized fact that is commonly reported in the literature. What is less well recognized is the result shown on the left-hand side of Figure 3 – the nontrivial percentage of consumers who demanded little to no compensation to eat the GM food. About 37% of English consumers and 28% of French consumers demanded less than a quarter to consume the GM food. Another surprising finding is that U.S. and European consumers are strikingly similar in their views on how scientific technologies should be regulated. Gaskell et al. (2005) asked over 1,200 U.S citizens and over 25,000 Europeans to answer two questions: (1) ‘‘should decisions about technology be left to the experts or based on the views of the public?’’ and (2) ‘‘should decisions be made on the basis of scientific evidence or on moral and ethical considerations?’’ Based on answers to the two questions, they categorized individuals into one of four groups. People who answered

70.0%

66% US (CA, FL, and TX)

60.0%

Percentage of Consumers

England

40.0%

30.0%

52%

France

50.0%

37%

28% 19%

20.0%

16% 10.0%

9% 8%

7% 3%

0.0% $0.00 to $0.24

$0.25 to $0.49

10% 6%

1% 1% $0.50 to $0.74

$0.75 to $0.99

9%

6%

7%

2% 2% 1% 1% 1% 0% $1.00 to $1.24

$1.25 to $1.49

6% 0% 1% 0%

$1.50 to $1.74

$1.75 to greater than $1.99 $2.00

Willingness-to-Accept (US dollars)

Fig. 3. Distribution of consumers’ willingness-to-accept to eat a GM cookie in the United States, France, and England. Source: Lusk et al. (2006).

256

Jayson L. Lusk

‘‘experts’’ to the first question were classified as ‘‘elitists’’ (the others were deemed ‘‘populists’’) and people who answered ‘‘scientific’’ to the second question were classified as ‘‘scientific’’ (the others were deemed ‘‘moralists’’). Table 1 shows the distribution of consumer types in the United States and Europe. The distribution of types is strikingly similar across the two locations, and the majority of citizens in both locations are ‘‘scientific elitists,’’ believing that decisions about technology should be made by experts on the basis of scientific evidence. Gaskell et al. (2005) showed that these ‘‘scientific elitists’’ were the group most accepting of new technologies, such as biotechnology, nanotechnology, and stem cells, particularly compared to the ‘‘moral populists,’’ who believed that decisions about technology should be made by the public on the basis of moral and ethical considerations. Despite the similarity in views on how technology should be regulated, the ‘‘scientific elitists’’ in the United States were much more accepting of GM food than were the ‘‘scientific elitists’’ in Europe (76% approval in United States vs. 28% approval in Europe). These results indicate that differences in views about how new technological decisions should be made cannot explain the difference in the mean levels of aversion to GM food across the Atlantic. Ultimately, we may never know exactly why Europeans appear, on average, more averse to GM food than Americans. The answer may be a result of little more than protectionism on the part of Europeans, who early on adopted policies to protect their domestic agricultural production sector in light of the new technologies being adopted in the United States. Differences in levels of protectionism, coupled with the divergent signals that these policies sent consumers about the potential safety of GM foods, may well explain the observed differences in U.S. and European acceptance of GM food. However, it would be virtually impossible to prove such a conjecture as it relies on an argument about a one-time event with dynamic effects than cannot be easily replicated in an experimental or survey setting. Table 1.

Differences in consumer types based on beliefs about how technology should be regulated

Type

Should decisions about technology be left to the experts or based on the views of the public?

Should decisions be made on the basis of scientific evidence or on moral and ethical considerations?

United States (%)

Europe (%)

Scientific elitist Moral elitist Scientific populist Moral populist

Expert Expert Public Public

Scientific evidence Moral considerations Scientific evidence Moral considerations

54 22 11 14

52 22 10 15

Consumer Preferences for Genetically Modified Food

257

2. What does it all mean? Aside from the work on meat demand, consumer-oriented agricultural economists have probably devoted more energy to studying consumer preferences for GM food than any other topic in recent decades. It is important to take a step back and ask what impact this work has had on applied theorists, producers, biotechnology firms, food retailers, and policy makers. Unfortunately, the answer may be: not much. One reason relates to the findings of Kalaitzandonakes and Bijman (2003). Despite all the discussion and handwringing about the lack of consumer acceptance of GM food in Europe, the primary reason behind the de facto ban on GM foods in Europe may relate to industry structure and strategic interaction between retailers not consumer concern per se. Kalaitzandonakes and Bijman concluded (2003, p. 369), ‘‘until the incentives and strategies of key players in the food supply chain are accounted for, little progress will be made in understanding public acceptance of agrifood biotechnology.’’ Still, the literature has provided a wealth of evidence on consumer WTP for GM food. The meta-analyses by Dannenberg (2009) and Lusk et al. (2005) suggest that existing knowledge of consumer WTP for GM food can be effectively summarized by simple econometric models. The appeal of this finding is that meta-analyses do not hinge on the results of a single study; they are based cumulatively on a variety of different studies by different authors in different locations with different foods and different methods. Previous research has done an effective job identifying what consumers’ valuations are. The most fruitful research yet to be conducted lies in explaining why consumers have a particular valuation estimate, predicting how these valuations might change and determining the effects of public policies on valuations. Some of the more interesting and important studies in recent years have focused on seemingly subtle issues related to the value of information. Hu et al., 2005, for example, argued that the value of GM food labeling policies cannot be determined by simple calculations of WTP premiums for non-GM vs. GM food. Instead, the effects of labeling policies depend on the extent to which consumers’ choices change after the provision of information, given that underlying product quality (i.e., whether the product is GM or non-GM) is unlikely to change. The value of information approach taken by Rousu et al. (2007) is similar in spirit. These studies show that the estimated value of information is relatively small in comparison to the more traditionally reported WTP premiums for GM food. More work is needed to determine consumers’ value for information, and to determine how it depends on consumers’ prior beliefs about existing market choices. Likewise, a stronger conceptual foundation is needed for thinking about the merits of policies, such as GM food labels or bans, when consumers are, and will likely remain, uninformed even after policy passage.

258

Jayson L. Lusk

Other interesting studies have emerged, which have attempted to determine whether consumers believe GM food production and consumption creates externalities. Even if consumers are averse to consuming GM foods, such aversion does not necessarily imply the presence of a market failure that would justify government action. Indeed, if consumer aversion is really so great, it would seem that firms would possess ample incentive to profit from introducing and selling non-GM foods. There are those, however, who argue that the profit incentive is insufficient because GM food production (and consumption) creates externalities that are not reflected in market prices. To determine the extent to which consumers are concerned about such externalities, Carlsson et al., 2007 devised a creative set of choice questions in which people were asked to choose between a GM food product and two different types of non-GM foods: (1) a food that was non-GM only for that consumers’ choice, and (2) a food that was non-GM for all consumers as a result of a ban. By comparing the WTP premium for the first type of non-GM food, which arguably only measures the consumers’ own private value for non-GM, to the WTP premium for the second type of food, in which one consumers’ choices affect others’ choices, Carlsson et al., 2007 argued that consumers’ perceptions and value of an externality can be determined. Their empirical results showed no evidence of a perceived externality among the Swedish consumers they studied, and they concluded regulating GM foods based on the presence of externalities would be misplaced. While I have some quibbles with the particular questioning approach taken by Carlsson et al., 2007 (e.g., the choice questions were framed in a less than believable way for respondents, and it is difficult to attribute the differences in WTP for the two attributes solely to the lack of concern for the externality because consumers may also have option values), the over-arching question asked is a good one, and more work is needed on the issue. We have suggested some other ways, using nonhypothetical experiments, to determine the extent to which people believe an externality or public good exists, and it would be useful to apply these and other related methods to the topic of GM food (Lusk, et al., 2006; Lusk and Norwood, 2011). Finally, economic research on consumer aversion to GM has focused almost entirely on eliciting consumer preferences for GM food (see Lusk and Coble 2005 for an exception). However, in a world with uncertain outcomes, choices are influenced not only by preferences for differing outcomes but also beliefs about the likelihood of observing the outcomes. More research is needed on consumers’ beliefs about GM foods, beliefs about the efficacy of GM food policies, and the role of information, policies, and other factors on consumers’ beliefs. Moreover, these beliefs should be studied in a theoretically consistent manner, rather than in the ad hoc way they are typically analyzed in few economic studies that have attempted to elicit beliefs.

Consumer Preferences for Genetically Modified Food

259

Although a case could be made that we have reached the point of diminishing returns in regard to the study of consumer preferences for GM food, there remains much to learn. As I have identified, several unresolved questions remain to be answered. If for no other reason, studies on consumer preferences for GM foods will likely continue because such work investigates a provocative topic embodying a myriad of issues of interest to applied economists: issues related to individual decision-making (from decision making under risk to time preferences to behavioral economics), issues related to the role of government and the potential existence of market failures (from information asymmetries to market power to externalities), and topics related to the value and optimal pricing of new products and technologies. Acknowledgment The author would like to thank the editors for helpful suggestions on a previous version of the chapter. References Bredahl, L., Grunert, K.G., Frewer, L.J. (1998), Consumer attitudes and decision-making with regard to genetically engineered food products – a review of the literature and a presentation of models for future research. Journal of Consumer Policy 21, 251–277. Brossard, D., Nisbet, M.C. (2007), Deference to scientific authority among a low information public: understanding U.S. opinion on agricultural biotechnology. International Journal of Public Opinion Research 19, 24–52. Bullock, D.S., Desquilbet, M. (2002), The economics of non-GMO Segregation and identity preservation. Food Policy 27, 81–99. Carlsson, F., Frykblom, P., Lagervist, C.J. (2007), Consumer benefits of labels and bans on GM foods-choice experiments with Swedish consumers. American Journal of Agricultural Economics 89, 152–161. Chang, J.B., Lusk, J.L., Norwood, F.B. (2009), How closely do hypothetical surveys and laboratory experiments predict field behavior? American Journal of Agricultural Economics 91, 518–534. Costa-Fonta, M., Gila, J.M., Traill, W.B. (2008), Consumer acceptance, valuation of and attitudes towards genetically modified food: review and implications for food policy. Food Policy 33, 99–111. Crespi, J.M., Marette, S. (2003), ‘‘Does contain’’ vs. ‘‘Does Not Contain’’: does it matter which GMO label is used? European Journal of Law and Economics 16, 327–344. Dannenberg, A. (2009), The dispersion and development of consumer preferences for genetically modified food – a meta-analysis. Ecological Economics 68, 2182–2192.

260

Jayson L. Lusk

Dimitri, C., Oberholtzer, L., 2009. Marketing U.S. organic foods: recent trends from farms to consumers. U.S. Department of Agriculture, Economic Research Service, Information Bulletin No. 58, September 2009. Available at http://www.ers.usda.gov/Publications/EIB58/EIB58.pdf Fernandez-Cornejo, J., Caswell, M., 2006. The first decade of genetically engineered crops in the United States. U.S. Department of Agriculture, Economic Research Service, Economic Information Bulletin No. 11, April 2006. Available at http://www.ers.usda.gov/publications/eib11/ eib11.pdf Fulton, M., Giannakas, K. (2004), Inserting GM products into the food chain: the market and welfare effects of different labeling and regulatory regimes. American Journal of Agricultural Economics 86, 42–61. Gaskell, G., Bauer, M.W., Durant, J., Allum, N.C. (1999), Worlds apart? The reception of genetically modified foods in Europe and the U.S. Science 285, 384–387. Gaskell, G., Einsiedel, E., Hallman, W., Priest, S.H., Jackson, J., Olsthoorn, J. (2005), Social values and the governance of science. Science 310, 1908–1909. Hoban, T.J. (1997), Consumer acceptance of biotechnology: an international perspective. Nature Biotechnology 15, 232–234. House, L., Lusk, J., Jaeger, S., Traill, W.B., Moore, M., Valli, C., Morrow, B., Yee, W.M.S. (2004), Objective and subjective knowledge: impacts on consumer demand for genetically modified foods in the United States and the European Union. AgBioForum 7, 113–123. Hu, W., Veeman, M.M., Adamowicz, W.L. (2005), Labelling genetically modified food: heterogeneous consumer preferences and the value of information. Canadian Journal of Agricultural Economics 53, 83–102. Johnston, R.J. (2006), Is hypothetical bias universal? Validating contingent valuation responses using a binding public referendum. Journal of Environmental Economics and Management 52, 469–481. Kalaitzandonakes, N., Bijman, J. (2003), Who is driving biotechnology acceptance? Nature Biotechnology 21, 366–369. Kalaitzandonakes, N., Marks, L.A., Vickner, S.S. (2004), Media coverage of biotech foods and influence on consumer choice. American Journal of Agricultural Economics 86, 1238–1246. Kanter, C., Messer, K.D., Kaiser, H.M. (2009), Does production labeling stigmatize conventional milk? American Journal of Agricultural Economics 91, 1097–1109. Kuznesof, S., Ritson, C. (1996), Consumer acceptability of genetically modified foods with special reference to farmed salmon. British Food Journal 98, 39–47. Lapan, H.E., Moschini, G. (2004), Innovation and trade with endogenous market failure: the case of genetically modified products. American Journal of Agricultural Economics 86, 634.

Consumer Preferences for Genetically Modified Food

261

Leggett, C.G., Kleckner, N.S., Boyle, K.J., Duffield, J.W., Mitchell, R.C. (2003), Social desirability bias in contingent valuation surveys administered through in-person interviews. Land Economics 79, 561–575. Lence, S.H., Hayes, D.J. (2005), Genetically modified crops: their market and welfare impacts. American Journal of Agricultural Economics 87, 931–950. Lin, W., Chambers, W., Harwood, J. (2000), Biotechnology: U.S. grain handlers look ahead. Agricultural Outlook. Washington DC: U.S. Department of Agriculture, pp. 29–34. List, J.A., Gallet, C.A. (2001), What experimental protocol influence disparities between actual and hypothetical stated values? Environmental and Resource Economics 20, 241–254. Lohr, L. (2001), Factors affecting international demand and trade in organic food products. Changing Structure of Global Food Consumption and Trade, Agriculture and Trade Report WRS-01-1, U.S. Department of Agriculture, Economic Research Service. Available at www.ers.usda.gov/publications/wrs011/wrs011j.pdf/. Lusk, J.L. (2008), Consumer preferences for cloning. Final report submitted to the U.S. Department of Agriculture, Economic Research Service. October 1, 2008. Lusk, J.L., Coble, K.H. (2005), Risk perceptions, risk preference, and acceptance of risky food. American Journal of Agricultural Economics 87, 393–405. Lusk, J.L., House, L.O., Valli, C., Jaeger, S.R., Moore, M., Morrow, B., Traill, W.B. (2004), Effect of information about benefits of biotechnology on consumer acceptance of genetically modified food: evidence from experimental auctions in United States, England, and France. European Review of Agricultural Economics 31, 179–204. Lusk, J.L., Jamal, M., Kurlander, L., Roucan, M., Taulman, L. (2005), A meta analysis of genetically modified food valuation studies. Journal of Agricultural and Resource Economics 30, 28–44. Lusk, J.L., Norwood, B., Pruitt, R. (2006), Consumer demand for a ban on subtherapeutic antibiotic use in pork production. American Journal of Agricultural Economics 88, 1015–1033. Lusk, J.L., Norwood, F.B. (2009), Bridging the gap between laboratory experiments and naturally occurring markets: an inferred valuation method. Journal of Environmental Economics and Management 58, 236–250. Lusk, J.L., Norwood, F.B. (2011), Speciesism, altruism, and the economics of farm animal welfare. European Review of Agricultural Economics. forthcoming. Lusk, J.L., Rozan, A. (2008), Public policy and endogenous beliefs: the case of genetically modified food. Journal of Agricultural and Resource Economics 33, 270–289.

262

Jayson L. Lusk

Lusk, J.L., Traill, W.B., House, L.O., Valli, C., Jaeger, S.R., Moore, M., Morrow, B. (2006), Comparative advantage in demand: experimental evidence of preferences for genetically modified food in the United States and European Union. Journal of Agricultural Economics 57, 1–21. Mintel. (2008), Perish the thought-organic market set for stunted growth. Mintel Press Release, Mintel Oxygen Reports, November, 2008. Available at http://www.mintel.com/press-release/Perish-the-thought– Organic-market-set-for-stunted-growth?id¼294 Noussair, C., Robin, S., Ruffieux, B. (2002), Do consumers not care about biotech foods or do they just not read labels?. Economics Letters 75, 47–53. Noussair, C., Robin, S., Ruffieux, B. (2004), Do consumers really refuse to buy genetically modified food?. The Economic Journal 114, 102–120. Pew Initiative on Food and Biotechnology (2006), Review of public opinion research. November 16, 2006. Available at http://www.pewtrusts.org/ uploadedFiles/wwwpewtrustsorg/Public_Opinion/Food_and_Biotechn ology/2006summary.pdf. Last accessed January 7, 2010. Rousu, M.C., Huffman, W.E., Shogren, J.F., Tegene, A. (2007), Effects and value of verifiable information in a controversial market: evidence from lab auctions of genetically modified food. Economic Inquiry 45, 409–432. Rousu, M.C., Lusk, J.L. (2009), Valuing information on GM foods in a WTA market: what information is most valuable?. AgBioForum 12, 226–231. Vossler, C.A., Kerkvliet, J., Polasky, S., Gainutdinova, O. (2003), Externally validating contingent valuation: an open space survey and referendum in Corvallis, Oregon. Journal of Economic Behavior and Organization 51, 261–277.

CHAPTER 11

The Effect of GM Labeling Regime on Market Outcomes$ Elise Golan and Fred Kuchler Economic Research Service, U.S. Department of Agriculture, Food Economics Division, Washington, DC, USA E-mail addresses: [email protected]; [email protected]

Abstract Purpose – This chapter investigates the role that mandatory genetically modified (GM) labeling versus voluntary labeling has played in the split between those countries with small GM markets and those with large GM markets. Methodology/approach – Data on product introductions and other market evidence are used to examine market outcomes and identify the likely drivers of GM market bifurcation. Findings – Labeling has negligible effects on consumer choice or on GM differentiation costs and therefore does not explain the split in GM market outcomes. Other factors have driven market outcomes: namely, consumer confidence in government and the safety of the food supply, competition among manufacturers and retailers, market momentum, and most importantly, the affordability of a non-GM strategy. Ultimately, a nonGM market strategy is feasible only if consumers are willing to cover the additional costs associated with non-GM production and marketing. The two elements composing the cost/price wedge between GM and non-GM products – the cost-reducing benefits of the GM technology and the costs of differentiating non-GM products – therefore play an important role in market outcomes. In the mid-1990s, when producers, manufacturers, and retailers were determining their strategies, neither element was very large. As a result, both GM and non-GM marketing strategies were economically feasible. Practical implication – Regardless of the labeling regime, changes in the cost/price wedge between GM and non-GM products could change the mix of GM and non-GM products on the market. $

The views expressed here do not necessarily reflect those of the Economic Research Service or the U.S. Department of Agriculture.

Frontiers of Economics and Globalization Volume 10 ISSN: 1574-8715 DOI: 10.1108/S1574-8715(2011)0000010016

r 2011 by Emerald Group Publishing Limited. All rights reserved

264

Elise Golan and Fred Kuchler

Originality/value of paper – This analysis extends the literature by focusing on the impact of labeling regime on both consumer behavior and the cost/ price wedge between GM and non-GM products. Keywords: GM labeling, mandatory labeling, GM markets, non-GM markets JEL Classifications: M38—government policy and regulation, O33— Technological change, choices and consequences, diffusion processes, Q55—Technological innovation Countries around the world have split into two camps when it comes to labeling foods made with first-generation genetically modified (GM) crops. Some, including the United States and Canada, do not require GM labeling for these foods, allowing undifferentiated GM marketing. Others, including Japan, Russia, and countries in the EU, mandate GM labeling, leaving non-GM as the undifferentiated default. In the 15 or so years since GM commercialization, there has been surprisingly little product differentiation away from these initial defaults: non-GM markets are small in the United States and other nonlabeling countries and GM markets are even smaller in the EU and other GM labeling countries. It is tempting to credit at least some of this result to the influence of labeling regime. In this chapter, we present data on market differentiation and provide evidence supporting the conclusion that the effect of labeling regime on market outcomes has been small. We extend the literature by focusing on the impact of labeling on both consumer behavior and the price/cost wedge between GM/non-GM products.

1. GM product differentiation and labeling regime: the evidence Virtually all commercialized GM crops are first-generation varieties. These varieties have cost-reducing or yield-enhancing input traits and are considered by governments around the world as substantially equivalent to their conventional counterparts (i.e., they present no novel consumer attributes and are as safe for human consumption as their conventional counterparts). Labeling regulations for these varieties are at the center of the labeling debate and are the topic of this chapter. In their review of international labeling policies, Grue`re and Rao (2007) found that as of February 2007, 10 countries (Australia, China, New Zealand, Norway, Japan, Russia, Saudi Arabia, South Korea, Switzerland, and Taiwan) plus all those in the EU had enforced mandatory labeling laws for GM foods or food products containing GM ingredients. Canada, Hong Kong, South Africa, and the United States had voluntary labeling. A number of other countries had documented partially enforced mandatory

265

The Effect of GM Labeling Regime on Market Outcomes

labeling laws, nonenforced mandatory labeling laws, or plans to introduce mandatory labeling laws. Grue`re et al. (2009) have found that countries exporting to the EU and Japan are likely to mandate labeling of GM food. The split between those countries with mandatory GM labeling laws and those without is mirrored in a split in market outcomes. The United States and other nonlabeling countries have small non-GM markets and large GM markets, while the EU and other GM labeling countries have small GM markets and large non-GM markets. In the United States, for example, organic markets, which accounted for approximately 3% of food sales in 2009 (Organic Trade Association, 2009), are the primary source of non-GM products (USDA organic food standards require organic food to be non-GM). And though data from DataMonitor show 6,899 new food and nonalcoholic beverage product introductions in the United States with explicit non-GM labeling from January 2000 to December 31, 2009 (Figure 1), this was only 3.9% of the 177,407 food and nonalcoholic beverage product introduction during this time period. The annual nonGM share of product introductions ranged from 3.1 to 4.4%. In the EU and other GM labeling countries, very few products labeled as containing GM ingredients are found on grocery store shelves. Greenpeace found only 77 GM labeled products in 10 EU countries in November 2004, mostly GM soy oil or other imported items (Greenpeace, 2005). ProductScan data show that from 2000 to 2004, manufacturers introduced only two explicitly GM products in the Japanese market: one fermented soybean product (Dr. Tomi-Chan’s natto) and a glow-in-the1000 922 900 775

800 Number of SKUs

717 700

627

613

630

2001

2002

678

695

2004 2005 Years

2006

667

575

600 500 400 300 200 100 0

2000

2003

2007

2008

2009

Fig. 1. U.S. food and nonalcoholic beverage product introductions with explicit non-GM claims, January 1, 2000–December 31, 2009 (DataMonitor).

266

Elise Golan and Fred Kuchler

dark powder. No GM product introductions were registered in the ProductScan database for Europe from 2000 to 2004. DataMonitor data show only five food and beverage product introductions with explicit GM labeling from January 2000 to January 2010 for markets in Asia, North America, South America, and Europe. Media reports (such as in The Grocer, 2004; CNN Tech, 2005) also mention the introduction into the European market of a beer made with GM corn. This beer, the development of which was partially funded by a consortium of biotech companies led by Monsanto Co., was created with the objective of testing European consumer reaction to GM labeled products. A spokesperson for Monsanto Co. said that Kenth beer was designed to ‘‘make an abstract discussion [among an inner circle of scientists, politicians, and nongovernmental organizations] more concrete’’ (CBS News, 2004). The niche status of non-GM markets in the United States and other nonlabeling countries and of GM markets throughout Europe and other labeling countries is reflected in cultivated acres. In the United States, GM crops account for the majority of planted acres for soybeans, cotton, and corn. In 2009, herbicide-tolerant soybeans accounted for 91% of soybean acreage while insect-resistant (Bt) corn accounted for 63% of U.S. corn acreage (Economic Research Service, 2009). In the EU, only 1% of acreage planted to corn in 2007 was GM (Go´mez-Barbero and Rodrı´ guesCerezo, 2007; Moschini, 2008). The difference in the market outcomes between countries with and without mandatory labeling would seem to identify labeling as the catalyst. The evidence on labeling’s effect on both consumer choice and production costs suggests, however, that labeling regime had little to do with these divergent outcomes.

2. GM labeling unlikely to have large effects on consumer choice There are two seemingly contradictory truths about labels. On the one hand, they are critical to consumer choice. Without labels, consumers would not be able to differentiate at the grocery store between regular and gluten-free pasta or between hot and mild salsa. On the other hand, however, labels are generally a weak policy tool for changing food consumption behavior because they generally fail to get consumers’ attention. Time- and attention-strapped consumers filter out information that is not a priori interesting to them. Empirical evidence shows that consumers often make hasty food choices in grocery stores and do not consult food labels (Aldrich, 1999). Even First Lady Michelle Obama confessed in a speech to the Grocery Manufacturers Association (Obama, 2010) that as a busy working mom ‘‘the last thing I had time to do was to stand in a grocery store aisle squinting at ingredients that I couldn’t pronounce to figure out whether something was healthy or not.’’ And she is

The Effect of GM Labeling Regime on Market Outcomes

267

not alone. Using data from the Diet Behavior and Nutrition module of the 2005–2006 National Health and Nutrition Examination Survey, Todd and Variyam (2008) found that 61% of respondents never or only sometimes consulted nutrition labels when buying food. And, for those using nutrition labels, most sought out particular pieces of information and ignored the rest. Todd and Variyam found that use of fiber information rose in the 10 years from 1995–1996 to 2005–2006, indicating, perhaps, an aging population’s increased awareness of dietary fiber’s health benefits. Use of other nutrient information fell over the same period. Even warning labels – large and on the front of the package – are not guaranteed to attract consumers’ attention. The U.S. Food and Drug Administration’s ‘‘Review of Research Communicating Warning Information’’ (Levy, 1997) is a testament to the difficulties in designing effective food hazard labels, as is the U.S. Environmental Protection Agency’s Read the Label First campaign, which implores consumers to heed the hazard and use labels on pesticides and household cleaning products. Some research suggests that even green-faced Mr. Yuck – the U.S. poisonwarning symbol for children – is not always a reliable deterrent (Vernberg et al., 1984). These and other experiences have taught regulators around the world the hard truth of Magat and Viscusi’s observation that labeling is generally not very effective and that there are some circumstances, such as when people do not read or do not care about the information on the label, in which it may not be effective at all (Magat and Viscusi, 1992). Could it be that GM labels succeed in grabbing consumers’ attention in a way that would impress even Mr. Yuck? Not likely. Evidence suggests that consumers are just as likely to overlook GM labels as other labels. In a willingness-to-pay experiment with French consumers, Noussair et al. (2002) found that even when they gave subjects three minutes to sit and read a list of GM flagged ingredients, the label did not have any impact on the amount participants were willing to pay for the product. Subjects simply did not pay attention to the label. And, as pointed out by Noussair and his coauthors, ‘‘What is not read in the laboratory will probably not be read in the supermarket’’ (p. 51). A number of observations explain consumers disregard for labeled information. First, food labels usually include a lot of information. In 2009, for example, a single carton of eggs sold in a U.S. grocery store chain was labeled with a ‘‘cage-free’’ claim, the grocery store ‘‘quality and satisfaction money-back guarantee’’ logo, the Orthodox Union symbol of kosher certification, and a long list of nutrient claims, including ‘‘25% of the daily value of vitamin E, 185 mcg of lutein per egg, and 100 mg of omega-3 polyunsaturated fatty acids per egg.’’ Research suggests that a lot of information on a label is likely to reduce the chance that consumers will read or correctly interpret any of it. Even if consumers do consider each piece of information on a label, they may find it difficult to order the information according to importance. For example, out of 10 warnings on

268

Elise Golan and Fred Kuchler

a label, consumers may have difficulty picking out the most important. As a result, consumers may underreact to important information or overreact to less important information (Noah, 1994). When GM products are labeled, the information is just one more piece of information on already crowded labels. And, the information is typically buried in the back in the ingredient list, making it relevant for consumers who know where to look, but not an attention grabber for the average consumer. For example, only the most motivated consumers would look past the bright big lettered ‘‘marvelous mouth-melting creamy filled pillows,’’ on the package of Tiger Food Brand’s recently released MiniMallows to find the small print ‘‘genetically modified flavors’’ in the ingredient list. Noussair et al. (2002) suggest that large food producers’ early experiments with GM labeling did not lead to reductions in sales simply because consumers did not pay attention to the labels. Kalaitzandonakes et al. (2005) also found that consumers in the Netherlands did not alter their purchasing behavior in reaction to explicit GM labeling from 1997 to 2000. The evidence on general label use – and early GM labeling experiences – suggests that mandatory labeling alone cannot explain the difference in market outcomes between the EU and other GM labeling countries and the United States and other nonlabeling countries. Labeling is simply not a powerful tool for changing consumer behavior. Other factors must have played a role.

3. Labeling regime does not change the distribution of differentiation costs Another possible explanation for the divergent market outcomes could be related to the actual costs of labeling. If mandatory labeling made it relatively more expensive to differentiate and market GM foods in GM labeling countries than in nonlabeling countries, it would make sense that GM production and marketing would be lower in countries with mandatory labeling. Labeling, however, has little effect on the level or distribution of differentiation costs. The distribution of these costs is instead determined by consumer demand for product authentication. In both GM labeling and nonlabeling countries, the demand for authentication is for non-GM foods, not GM foods. This is because the GM crops currently on the market are first-generation varieties and do not have marketable quality-enhancing attributes of value to consumers, such as better taste or a superior nutritional profile. Any demand for differentiation and authentication is therefore from consumers with preferences for non-GM products. Surveys at the time of GM introduction revealed a reservoir of consumers with potential preferences for non-GM products in both GM labeling and nonlabeling countries. Market attitude and willingness-to-pay

The Effect of GM Labeling Regime on Market Outcomes

269

studies published in the early 2000 indicate that like their EU counterparts, many US consumers (around half) had reservations about buying GM foods and on average were willing to pay a positive premium for GM-free foods (for reviews of the GM consumer attitude and willingness-to-pay literature, see Lusk et al., 2005; Fernandez-Cornejo and Caswell, 2006; Lusk, 2011). Because non-GM products are the higher-valued consumer products, the costs of GM/non-GM differentiation are borne by the producers and consumers of non-GM foods. Only consumers of non-GM products demand authentication and only non-GM producers need to supply it. Just as nonorganic producers do not have to segregate, test, or certify that their products were grown with pesticides and nonorganic fertilizers, GM producers of first-generation varieties do not have to segregate, test, or certify that their products do not contain non-GM ingredients. Nonorganic consumers will not feel cheated if organic product is mixed with their nonorganic product; non-GM consumers will not feel cheated if nonGM is mixed with their GM product. Even if GM crops and foods are labeled, producers/manufacturers who wish to market non-GM products will still need to establish their own systems to provide evidence that the products they sell are not GM. The lack of evidence identifying a container of corn as GM is unlikely to provide adequate evidence of non-GM identity. Producers/manufacturers will need to develop systems to credibly differentiate non-GM from GM products. Mandatory labeling does not change the basic differentiation dynamic described in Golan and Kuchler (2002), Lence and Hayes (2005), Moschini (2008), and Grue`re et al. (2008), where an initial competitive market equilibrium in which all commodities and foods are non-GM is described by P0ðnon-GMÞ ¼ MC0ðnon-GMÞ

(1)

where P0(non-GM) and MC0(non-GM) are the initial price and marginal cost of non-GM crops and foods. Then a first-generation, cost-reducing GM variety of the commodity is introduced, such that: MCGM oP0ðnon-GMÞ

(2)

where MCGM is the marginal cost of GM crops and foods. Through competitive market forces, at least some of the lower GM production costs are eventually passed on to consumers in the form of lower prices, so that PGM oP0ðnon-GMÞ

(3)

where PGM is the price of GM crops and foods. At this point in the narrative, markets are an indistinguishable mix of GM and non-GM products. Producers/manufacturers wishing to market non-GM products will need to establish systems for ensuring that their products remain non-GM. Non-GM differentiation costs – which may

270

Elise Golan and Fred Kuchler

include the costs of buffer zones on the farm, separate storage and transportation facilities, additional cleaning expenses, and testing and certification services – raise production costs so that MCoðnonGMÞ oMCnon-GM

(4)

where MCnon-GM is the marginal cost of non-GM crops and foods in the second period. Non-GM producers/manufacturers will bear the costs of establishing differentiation systems because it is their consumers who need assurances about product identity. Eventually, these additional production costs will be passed to consumers and the non-GM market will settle at a higher price than the initial non-GM price: P0ðnon-GMÞ oPnon-GM

(5)

where Pnon-GM is the price of non-GM crops and foods in the second period, after the introduction of genetically engineered crops and foods. In the final equilibrium, two different prices emerge, one lower than before the introduction of GM food and the other higher: PGM oP0ðnon-GMÞ oPnon-GM

(6)

Mandatory GM labeling introduces an additional element into the equation: namely, the cost of the GM label itself. This addition changes the equilibrium described in Equation (6) to PGMþlabel oP0ðnon-GMÞ oPnon-GM

(7)

where PGMþlabel is the price of labeled GM crops and foods. The difference between Equations (6) and (7) is likely to be small, since there are no verification costs associated with the label and the costs of adding a word or two more to a preexisting label are small. Other than some potential design costs for changing the spacing on the ingredient list to accommodate an additional few words, such as changing ‘‘flavorings’’ to ‘‘genetically engineered flavoring,’’ the costs of additional ink, paper, or glue should be very small or zero. And, as pointed out by French and Neighbors (1991) these costs could be absorbed in the normal label-change cycle if the compliance period is sufficiently long. It is unlikely that the costs of GM labeling played any role in changing producers’/consumers’ benefit-cost analysis of GM/non-GM products. 4. Other factors tipped the balance to a non-GM strategy in labeling countries The evidence suggests that mandatory labeling is not the root cause of the difference in market outcomes between GM labeling and nonlabeling countries, at least via any direct effects on consumer choice or producer/ manufacturer costs. Something else must have tipped the balance. A

The Effect of GM Labeling Regime on Market Outcomes

271

number of factors were likely contributors, including consumer confidence in government and the safety of the food supply, affordability of a nonGM strategy; competition among manufacturers and retailers, and market momentum. Instead of being caused by labeling, these factors themselves helped trigger the establishment of different labeling regimes and market outcomes. 4.1. Consumer confidence in government and the safety of the food supply Different levels of consumer confidence in government regulators in Europe and the United States provide some of the explanation for different GM market outcomes. A lot of this difference can be traced to Europe’s experience with bovine spongiform encephalopathy (BSE). From the 1986 discovery of BSE through 1996, when a government panel of scientists posited that beef was the source of human illnesses, British health authorities maintained that British beef was safe. Other European countries’ public health authorities displayed a similar pattern of denial. By the time GM foods began to appear, EU consumers’ trust in their food regulators was low. In a 1997–1998 survey of consumer attitudes to GM foods in Europe (15 member states of the EU plus Norway and Switzerland) and the United States, Gaskell et al. (1999) found that European respondents had the highest confidence in international organizations such as the United Nations and the World Health Organization, followed by scientific committees and national public bodies. In contrast, consumers’ trust in public health authorities in the United States was, and is, relatively high. Gaskell et al. (1999) found that when survey respondents were asked ‘‘If the U.S. Department of Agriculture (USDA) or U.S. Food and Drug Administration made a public statement about the safety of biotechnology, would you have a lot, some, or no trust in the statement about biotechnology?’’ 90% of respondents had confidence in USDA while 84 percent had confidence in FDA. This high level of trust in U.S. regulators is reflected in market studies of food safety events in the United States in the early and mid-2000s. In December 2003, U.S. agencies reported that a cow in Washington State had BSE. Federal agencies repeatedly said that consumers ought not to be alarmed by the finding. Statistical analysis of consumers’ beef purchases showed that the response was limited and dissipated within two weeks (Kuchler and Tegene, 2006). In 2006, the FDA warned consumers to avoid eating fresh spinach as it could be contaminated with potentially deadly bacterium Escherichia coli O157:H7. Statistical analysis of consumers’ purchases showed that they paid close attention to the messages from FDA, gradually returning to spinach as the agency gave the all clear signal, and substituting other leafy greens for which there was no warning (Arnade et al., 2009). European consumer’s distrust of government to protect food safety in Europe created an information void that was soon filled with an anti-GM

272

Elise Golan and Fred Kuchler

campaign by advocacy groups such as Greenpeace. This campaign was successful for at least two reasons. First, Europeans had high levels of trust in these groups. When Gaskell and his colleagues (1999) asked Europeans ‘‘Which of the following sources of information do you have most confidence in to tell you the truth about genetically modified crops grown in fields?’’ the ranking was for environmental, consumer, and farming organizations (23, 16 and 16%), followed by national public bodies (4%) and industry (1%). Second, Greenpeace’s and the other consumer advocacy groups’ campaign in Europe was strengthened by the fact that their message was a negative one. In experimental auctions to assess consumer attitudes toward GM foods, Tegene et al. (2003) found that individuals placed a greater weight on negative GM information than on positive information. In 12 separate hypothetical auctions, involving 172 consumers in two midwestern cities, participants were given the opportunity to bid for and purchase three different food products – vegetable oil, tortilla chips, and potatoes – with and without GM labels. Before the bidding, they were given information packets containing statements about biotechnology from a variety of sources. Pro-GM statements were provided by a group of leading GM companies. Greenpeace provided anti-GM statements. Science-based verifiable statements were provided by a group of individuals knowledgeable about biotechnology, none of whom had a financial stake in agricultural biotechnology. The source of each statement was identified. Participants’ bids, or the amount they were willing to pay, for GM labeled and plain-labeled foods were affected by the information packets they received. Participants who received only pro-GM information bid slightly more on the GM labeled food for two of the three products. Participants who received only anti-GM information bid 35% less, on average, on the GM labeled food than on the plain-labeled food. Those who received both pro- and anti-GM information bid on average 16–29% less on the GM labeled foods than on the plain-labeled food, depending on the food product. The high stature of groups like Greenpeace in Europe could only serve to increase the impact of negative information on European consumers.

4.2. A non-GM strategy was usually affordable Regardless of the strength of public opinion, a non-GM strategy could not work unless non-GM products could be priced so as to be affordable to enough consumers to make the strategy profitable. Ultimately, the company’s bottom line sets limits on its product differentiation strategy. Producers/manufacturers have ample evidence that price is an important factor in market formation, even in the face of seemingly strong consumer preferences. For pesticides, surveys reveal high levels of

The Effect of GM Labeling Regime on Market Outcomes

273

consumer concern about the health risks of pesticide use, with results placing the share of consumers concerned about the health risks of pesticide-treated produce at up to 70% (Ott et al., 1991; Weaver et al., 1992; Buzby et al., 1995). This high level of concern would imply a greater interest in organic produce, but with a current market share estimated at about 3% (Organic Trade Association, 2009), the organic market is certainly smaller than the survey results suggest they should be. As in this case, purchase behavior often deviates from stated purchase intentions when consumers’ own money is at stake. A non-GM market strategy would only work if consumers were willing to cover the additional costs associated with non-GM production, as described by the cost/price wedge in Equation (6). If the wedge is too large, then only consumers with strong non-GM preferences would be willing to pay the higher price for non-GM products. With lower prices, even consumers with mild preferences for non-GM products might be willing to incur the additional costs. The size of the two elements composing the price wedge – the cost-reducing benefits of the GM technology and the costs of differentiating non-GM products – therefore plays an important role in market outcome. In the mid-1990s, when producers, manufacturers, and retailers were determining their marketing strategies, neither element was very large. The early cost-reducing benefits of the technology, described by PGMoPo(non-GM), were uncertain and limited. The potential benefits of first-generation corn and soybeans included (and still include) reduced chemical use, less harmful chemical use, reduced tilling, reduced labor time, less production and financial risk, and in some cases, increased yields. Early evidence on whether or not the technology was delivering these benefits was positive, although results varied by variety, region, and year (Heimlich et al., 2000a, 2000b). Falck-Zepeda et al. (2000) developed a soybean world supply/demand model to examine the surplus created by the introduction of GM soybeans (Moschini et al., 2000 investigate similar questions). Even including rather strong positive yield effects, the largest price decrease they calculated was $0.0009 per pound. The dilution of benefits via distribution among farmers (increased profits), seed producers and biotech firms (higher seed prices, technology fees, and increased profits), and manufacturers and consumers (through lower input prices and food prices) dampened the benefits to manufacturers and consumers. Early investigations into the distribution of the benefits of GM soybeans among farmers, innovators, and consumers estimated that the majority of benefits went to farmers and innovators with only between 4 and 17% going to consumers (FalckZepeda et al., 2000; Moschini et al., 2000; Price et al., 2003). The appeal of a non-GM marketing strategy for manufacturers was strengthened by their expectations of small potential benefits from GM adoption. The early differentiation costs of non-GM production were also relatively low. Golan et al. (2001) report that in 1999 in the United States,

274

Elise Golan and Fred Kuchler

non-GM soybeans cost around 2–3% more than conventional soybeans and non-GM corn cost 2–6% more than conventional corn, though only a small number of elevators offered premiums (ERS, 2000). These premiums imply low differentiation costs, assuming that they reflect upstream estimations of non-GM value. They also translate into small price differences at the consumer level, given that the cost of corn and soybean inputs is only a small share of the retail value of food. For example, Leibtag estimates that even for products heavily based on field corn, such as corn flakes and soda, higher corn prices have a small impact on retail prices (Leibtag, 2008). He calculates that a 50% increase in corn prices from the 20-year average of $2.28 per bushel would raise the price of a box of corn flakes by about 1.6 cents, or 0.5%, and the price of soda by 1.9 cents per 2-liter bottle, or 1%. At these rates of pass through, a 2- to 6-percent increase in non-GM field corn prices would have a negligible effect on household food budgets and would be unlikely to influence consumers’ food choices. These calculations suggest that the GM/non-GM price wedge was not large enough to dissuade producers/manufacturers from pursuing a nonGM strategy. The small cost/price difference meant that producers/ manufacturers could cater to non-GM preferences without running the risk of alienating price-sensitive consumers. A larger cost/price difference would have reduced the size of the market willing to pay for non-GM products and would have reduced producers’ and manufacturers’ willingness and ability to successfully supply them. This is true for markets with either mandatory or voluntary labeling regimes. For example, the cost of replacing GM rennet in cheese production is high (Fankhauser, 2009 posits that in 2009, GM rennet cost about one-tenth as much as traditionally produced rennet cost in 1990). As a result, even cheese manufacturers in countries with mandatory labeling use GM rennet. For some countries, including those in the EU, cheese produced using GM rennet is exempt from mandatory labeling, further suggesting that cost, not a priori labeling policy, is the key driver in determining market outcomes. Producers/manufacturers may also opt for a GM strategy for products or situations in which small changes in unit input prices have relatively large effects on aggregate production costs. This dynamic may be behind widespread use of GM soybean feed in the EU – and the EU’s decision not to mandate consumer-level labeling of animal products. Again, the difference between GM and non-GM costs/prices has seemed to have a larger effect on market outcome than consumer attitudes or general labeling regime.

4.3. Wholesome competition The affordability of a non-GM strategy combined with reservoirs of sensitized consumers helped set the stage for market competition over GM content. To build their brand’s reputation for wholesomeness,

The Effect of GM Labeling Regime on Market Outcomes

275

healthfulness, and social responsibility, producers, manufacturers, and retailers across whole countries – and within segments of others – began excluding GM ingredients from their products. This strategy was particularly attractive to manufacturers of products for which wholesomeness is an important attribute, which may explain why many U.S. baby food manufacturers quickly adopted a non-GM strategy (and why Greenpeace targeted them first for negative publicity). This strategy was also important for European retailers seeking to differentiate themselves with respect to safety and quality. In early 1999, Sainsbury and Marks & Spencer, major UK food retailers, and Carrefour, a major French retailer, pledged to exclude GM ingredients in their own branded products (Poulter, 1999). Once this type of competition over reputation takes off, it can influence a whole marketplace. For example, in the United States it is almost impossible to find a can of tuna without a ‘‘dolphin-friendly’’ label. Few manufacturers want to take the risk of being painted as a dolphin killer. In Europe and some other markets around the world, few manufacturers or retailers want to take the risk of being painted with the GM brush. When an attribute becomes an important differentiator in a market, it does not matter whether labeling is mandatory or voluntary. Consumers are very good at ‘‘reading between the labels’’ and making deductions about unlabeled products. Confronted with one can of tuna labeled ‘‘dolphin friendly’’ and one with no such claim, consumers would likely assume that the unlabeled tuna was caught with dolphin-endangering practices. A mandatory ‘‘dolphin unfriendly’’ label is not necessary. Likewise, in the United States, consumers seeking non-GM processed foods infer that if the package is not labeled ‘‘non-GM’’ or ‘‘organic’’ then chances are good that it contains GM ingredients. Competitive disclosure drives firms to make explicit claims for all positive aspects of a product and allows consumers to make appropriate inferences about foods without claims. This dynamic may help explain the rise of explicit non-GM labeling in countries that have mandatory GM labeling regulations. From January 1, 2000 to December 31, 2009, manufacturers in countries in Europe, including former Soviet Union countries, introduced 2,118 new products with labels that include statements attesting to the fact that the product contains no GM ingredients (sum of the height of the bars in Figure 2). Manufactures may be calculating that this labeling, though redundant in most of these countries, further differentiates the wholesomeness of their products for domestic and international customers.

4.4. Market momentum contributes to market bifurcation Once they have established a GM/non-GM marketing strategy, producers/ manufacturers have an incentive to reduce the differentiation costs

276

Elise Golan and Fred Kuchler 450 400

380

392

397

2008

2009

350

Number of SKUs

317 300

284

250 200 150

125

100

94

78

50

35 16

0

2000

2001

2002

2003

2004

2005

2006

2007

Years

Fig. 2. Europe food and nonalcoholic beverage product introductions with explicit non-GM claims, January 1, 2000–December 31, 2009 (DataMonitor).

associated with their chosen strategy. For example, producers/manufactures may maneuver to control bulk commodity supply chains, since these are likely to be more efficient and less costly than niche supply chains. For non-GM producers/manufacturers, the best way to control bulk systems and reduce differentiation costs is to eliminate the presence of GM product in the supply chain, thereby eliminating the need for establishing differentiated supply chains. The Canadian Wheat Board is likely pursuing this tactic in its current opposition to the introduction of GM wheat into world supply chains. The International Dairy Foods Association (IDFA), a trade group for dairy suppliers and manufacturers, is pursuing this strategy in its opposition to the introduction of products of cloned animals and their offspring into the U.S. food supply. Governments may join in efforts to reduce non-GM differentiation costs by establishing restrictive GM policies or by offering public services to reduce the costs of non-GM differentiation, such as no-fee certification services. Grue`re et al. (2008) argue that non-GM certification services in countries mandating GM labeling can reduce the costs of non-GM marketing, making it less expensive to pursue a non-GM strategy. By increasing or decreasing the GM/non-GM cost wedge, these industry and government efforts ultimately serve to reinforce prevailing market

The Effect of GM Labeling Regime on Market Outcomes

277

conditions and increase the possibility of a pure GM or non-GM market outcome.

5. Conclusion: labeling regime does not determine market outcome A review of the evidence suggests that GM labeling laws do not explain the stark contrast in market outcomes between countries with and without mandatory GM labeling. Labeling has negligible effects on consumer choice or on differentiation costs. Instead, other factors, including consumer confidence in government and the safety of the food supply, competition among manufacturers and retailers, market momentum, and most importantly, the affordability of a non-GM strategy, helped tip the balance to predominantly non-GM markets in GM labeling countries and toward predominantly GM markets in nonlabeling countries. Ultimately, a non-GM market strategy is feasible only if consumers are willing to cover the additional costs associated with non-GM production and marketing. The two elements composing the cost/price wedge between GM and non-GM products – the cost-reducing benefits of the GM technology and the costs of differentiating non-GM products – therefore play an important role in market outcomes. In the mid-1990s, when producers, manufacturers, and retailers were determining their marketing strategies, neither element was very large. As a result, both GM and nonGM marketing strategies were economically feasible. Over time, a change in the cost/price wedge between GM and non-GM products could change the mix of GM and non-GM products on the market even if labeling laws remain unchanged. For example, in the United States, the marketing strategy for non-rbST dairy products has evolved over time (rbST, or recombinant bovine somatotropin, is a GM version of bST, a natural growth hormone that stimulates milk production in cows). When milk produced with rbST was first introduced into the U.S. food supply, survey results indicated that over 50% of consumers thought that rbST posed a serious health threat or were unsure if it did (Preston et al., 1991). Despite consumer sentiments and dire predictions of market collapse, few manufacturers chose to market differentiated non-rbST milk, and sales remained steady (Blisard et al., 1996) or fell slightly for retail sales of fluid milk and cheese but grew substantially for butter and frozen desserts (Kaiser, 1997). Recently, however, numerous manufacturers and retailers, including Safeway, Wal-Mart, and Starbucks have begun to market non-rbST milk. Though unclear what triggered this change in strategy, it is fairly clear that it was not due to a change in labeling laws or a preference-changing safety event. Regardless of what triggered it, however, the feasibility of the strategy was aided by the fact that only a small share of cows are treated with rbST in the United States (about 17% in 2007; APHIS, 2007) and

278

Elise Golan and Fred Kuchler

that differentiation verification is lax. As a result, the differentiation costs associated with a non-rbST strategy are not prohibitive. As with rbST, changes in the cost/price wedge between GM and non-GM products could convince producers, manufacturers, or retailers to change their marketing strategy. Whether because of changes in the benefits of a GM technology or in the costs of differentiation, GM/non-GM marketing patterns could converge or even reverse themselves, erasing the seemingly permanent split between GM labeling and nonlabeling countries.

References Aldrich, L. (1999), Consumer Use of Information: Implications for Food Policy. U.S. Department of Agriculture, Economic Research Service, Agricultural Handbook Number 715. Retrieved from http://www. ers.usda.gov/Publications/AH715/. Animal Plant Health Inspection Service, U.S. Department of Agriculture. (2007), National animal health monitoring system, dairy, Part 1: reference of dairy cattle health and management Practices in the United States. Retrieved from http://www.aphis.usda.gov/animal_health/ nahms/dairy/downloads/dairy07/Dairy07_dr_PartI.pdf. Arnade, C., Calvin, L., Kuchler, F. (2009), Consumer response to a food safety shock: the 2006 foodborne illness outbreak of E. coli 0157:H7 linked to spinach. Review of Agricultural Economics 31 (4), 734–750. Blisard, N., J. Blaylock, J., Smallwood, D. (1996), Fluid milk and cheese advertising. U.S. Department of Agriculture, Economic Research Service, Staff Report No. AGES-9601. Buzby, J., Ready, R., Skees, J. (1995), Contingent valuation in food policy analysis: a case study of a pesticide-residue risk reduction. Journal of Agricultural and Applied Economics 27 (2), 613–625. CBS News. (2004), Europeans balk at biotech beer. Retrieved from http:// www.cbsnews.com/stories/2004/07/08/tech/main628305.shtml. CNN Tech. (2005), Family-run brewery serves up GM beer. Retrieved from http://www.cnn.com/2005/TECH/07/15/gm.foods/index.html? iref¼allsearch. Economic Research Service, U.S. Department of Agriculture. (2000), Biotech corn and soybeans: changing markets and the government’s role. Issues Brief, U.S. Department of Agriculture, Economic Research Service. Economic Research Service, U.S. Department of Agriculture. (2009), Data sets: adoption of genetically engineered crops in the U.S. U.S. Dept. of Agriculture, Economic Research Service. Retrieved from http:// www.ers.usda.gov/data/biotechcrops/. Falck-Zepeda, J., Traxler, G., Nelson, R. (2000), Rent creation and distribution from biotechnology innovations: the case of Bt Cotton and Herbicide-Tolerant Soybeans in 1997. Agribusiness 16, 21–32.

The Effect of GM Labeling Regime on Market Outcomes

279

Fankhauser, D. (2009), Rennet for making cheese. Retrieved from http:// biology.clc.uc.edu/Fankhauser/cheese/rennet/rennet.html. Fernandez-Cornejo, J., Caswell, M. (2006), The first decade of genetically engineered crops in the United States. U.S. Department of Agriculture, Economic Research Service, Economic Information Bulletin No. (EIB11). Retrieved from http://www.ers.usda.gov/publications/eib11/. French, M., Neighbors, D. (1991), A model of firm costs of compliance with food labeling regulations. In: Caswell, J. (Ed.), Economics of Food Safety. Elsevier, New York, pp. 229–325. Gaskell, G., Bauer, M., Durant, J., Allum, N. (1999), Worlds apart? The reception of genetically modified foods in E rope and the U.S.. Science 285, 384–389. Golan, E., Kuchler, F. (2002), Labeling GMO’s: implications for consumer welfare and trade. In: Krissoff, B., Bohman, M. (Eds.), Global Food Trade and Consumer Demand for Quality. Kluwer Academic Press, New York. Golan, E., Kuchler, F. Mitchell, L. Greene, Jessup, A. (2001), Economics of food labeling. U.S. Department of Agriculture, Economic Research Service, Agricultural Economics Report No. (AER793). Retrieved from http://www.ers.usda.gov/Publications/AER793/. Go´mez-Barbero, M., Rodrı´ gues-Cerezo, E. (2007), GM crops in EU agriculture. A case study for the Biotechnology for Europe Study. The European Commission. Retrieved from http://bio4eu.jrc.ec.europa.eu/ documents.html. Greenpeace. (2005), EU markets report. Retrieved from http://www.greenpeace.org/raw/content/international/press/reports/european-marketsreport-2005.pdf. Grue`re, G.P., Carter, C., Farzin, Y. (2008), What labeling policy for consumer choice? The case of genetically modified food in Canada and Europe. Canadian Journal of Economics 41 (4), 1472–1497. Grue`re, G.P., Carter, C., Farzin, Y. (2009), Explaining international differences in genetically modified food labeling policies. Review of International Economics 17 (3), 393–408. Grue`re, G.P., Rao, S.R. (2007), A review of international labeling policies of genetically modified food to evaluate India’s proposed rule. AgBioForum 10 (1), 51–64. Heimlich, R.E., Fernandez-Cornejo, J., McBride, W., Klotz-Ingram, C., Jans, S., Brooks, N. (2000a), Adoption of genetically engineered seed in U.S. agriculture: implications for pesticide use. In: Fairbairn, C., Scoles, G., McHughen, A. (Eds.), Proceedings of the 6th International Symposium on the Biosafety of Genetically Modified Organisms. University Extension Press, University of Saskatchewan, Saskatoon, Canada, pp. 56–63. Heimlich, R.E., Fernandez-Cornejo, J., McBride, W., Klotz-Ingram, C., Jans, S., Brooks, N. (2000b), Genetically engineered crops: has

280

Elise Golan and Fred Kuchler

adoption reduced pesticide use? U.S. Department of Agriculture, Economic Research Service, Agricultural Outlook AGO-273, 13–17. Retrieved from http://www.ers.usda.gov/publications/agoutlook/ aug2000/ao273toc.pdf. Kaiser, H. (1997), Impact of national generic dairy advertising on dairy markets, 1984-96. NICPRE 97-03, R.B. 97-10. Retrieved from http:// commodity.aem.cornell.edu/nicpre/bulletins/rb9710.pdf. Kalaitzandonakes, N., Marks, L., Vickner, S. (2005), Sentiments and acts towards genetically modified foods. International Journal of Biotechnology 7, 161–177. Kuchler, F., Tegene, A. (2006), Did BSE announcements reduce beef purchases? U.S. Department of Agriculture, Economic Research Service, Economics Research Report No. (ERR-34). Retrieved from http://www.ers.usda.gov/publications/ERR34. Leibtag, E. (2008), Corn prices near record high, but what about food costs? Amber Waves, U.S. Department of Agriculture, Economics Research Service. Retrieved from http://www.ers.usda.gov/AmberWaves/February08/Features/CornPrices.htm. Lence, S., Hayes, D. (2005), Genetically modified crops: their market and welfare impact. American Journal of Agricultural Economics 87, 931– 950. Levy, A. (1997), Review of Research Communicating Warning Information. Consumer Studies Branch, Office of Scientific Analysis and Support, Center for Food Safety and Applied Nutrition, Food and Drug Administration. Lusk, J. (2011), Consumer preferences for genetically modified food. In: Carter, C., Moschini, G.C., Sheldon, I. (Eds.), Genetically modified food and global welfare. Emerald, Bingley, UK. Lusk, J., Jamal, M., Kurlander, L., Roucan, M., Taulman, L. (2005), A meta analysis of genetically modified food valuation studies. Journal of Agricultural and Resource Economics 30, 28–44. Magat, W., Viscusi, W.K. (1992), Informational Approaches to Regulation. MIT Press, Cambridge, MA. Moschini, G. (2008), Biotechnology and the development of food markets: retrospective and prospects. European Review of Agricultural Economics 35 (3), 331–355. Moschini, G., Lapan, H., Sobolevsky, A. (2000), Roundup ReadyTM soybeans and welfare effects in the soybean complex. Agribusiness 16, 33–55. Noah, L. (1994), The imperative to warn: disentangling the ‘‘right to know’’ from the ‘‘need to know’’ about consumer product hazards. Yale Journal on Regulation 11 (2), 293–400. Noussair, C., Robinb, S., Ruffieux, B. (2002), Do consumers not care about biotech foods or do they just not read the labels? Economics Letters 75, 47–53.

The Effect of GM Labeling Regime on Market Outcomes

281

Obama, M. (2010), Speech at the Grocery Manufacturers Association, March 16, 2010. Retrieved from http://www.whitehouse.gov/thepress-office/remarks-first-lady-a-grocery-manufacturers-associationconference. Organic Trade Association (2009), The Organic Trade Association’s 2009 Organic Industry Survey. Organic Trade Association, Greenfield, MA. Ott, S., Huang, C., Misra, S. (1991), Consumers’ perceptions of risks from pesticide residues and demand for certification of residue-free produce. In: Caswell, J. (Ed.), Economics of Food Safety. Elsevier, New York. Poulter, S. (1999), Sainsbury’s to axe GM own brands. The Daily Mail, March 17. Retrieved from http://www.thefreelibrary.com/Sainsbury’sþtoþaxeþGMþownþbrands.-a0109752477. Preston, W., McGuirk, A., Jones, G. (1991), Consumer reaction to the introduction of Bovine Somatotropin. In: Caswell, J. (Ed.), Economics of Food Safety. Elsevier, New York, pp. 229–325. Price, G., Lin, W., Falck-Zepeda, J.B., Fernandez-Cornejo, J. (2003), The size and distribution of market benefits from adopting agricultural biotechnology. U.S. Department of Agriculture, Economic Research Service, Technical Bulletin No. 1906. Retrieved from http://www.ers. usda.gov/publications/tb1906/tb1906.pdf. Tegene, A., Huffman, W. Rousu, M., Shogren, J. (2003), The effects of information on consumer demand for biotech foods: evidence from experimental auctions. U.S. Department of Agriculture, Economic Research Service, Technical Bulletin No. 1906. Retrieved from http:// www.ers.usda.gov/Publications/TB1903/. The Grocer. (2004), GM beer eyes UK. The Grocer, February 14. Retrieved from http://www.thegrocer.co.uk/articles.aspx?page ¼ articles&ID¼91453. Todd, J., Variyam, J. (2008), The decline in consumer use of food nutrition labels, 1995-2006. U.S. Department of Agriculture, Economic Research Service, Economics Research Report No. (ERR-63). Retrieved from http://www.ers.usda.gov/Publications/ERR63/. Vernberg, K., Culver-Dickinson, P., Spyker, D. (1984), The deterrent effect of poison-warning stickers. American Journal of Diseases of Children 138, 1018–1020, doi:PMID 6496418. Weaver, R., Evans, F., Luloff, A. (1992), Pesticide use in tomato production: consumer concerns and willingness-to-pay. Agribusiness 8 (2), 131–142.

CHAPTER 12

International Trade and Welfare Effects of Biotechnology Innovations: GM Food Crops in Bangladesh, India, Indonesia, and the Philippines Guillaume P. Grue`rea, Antoine Boue¨tb,c and Simon Meveld a

Environment and Production Technology Division, International Food Policy Research Institute, Washington DC, USA E-mail address: [email protected] b Markets, Trade and Institutions Division, International Food Policy Research Institute, Washington DC, USA E-mail address: [email protected] c Laboratoire d’Analyse et de Recherche en Economie et Finances Internationales, Universite´ Montesquieu Bordeaux IV, Bordeaux, France d Formerly at the World Bank, Washington DC, USA E-mail address: [email protected]

Abstract Purpose – The chapter examines the international welfare effects of biotech crop adoption, based on a transversal literature review and a case study of the introduction of genetically modified (GM) food crops in Bangladesh, India, Indonesia, and the Philippines. Methodology/approach – The analysis is based on (a) a review of lessons from the applied economic literature and (b) simulations using an improved multimarket, multicountry, computable general equilibrium (CGE) model, calibrated with productivity hypotheses formulated with local scientists in the four Asian countries. Findings – Results from the analysis show that, in the absence of traderelated regulations, GM crop adoption generates economic gains for adopting countries and importing non-adopters, that domestic regulations at adopters and especially non-adopters can reduce these gains, and that import regulations in other countries can also affect gains for exporting adopters. The case study illustrates these conclusions, but it also shows that net importers will mostly benefit from adoption in their terms of trade, and that segregation of non-GM crops for export markets can be beneficial if it is not too costly. Research limitations/implications – The use of a CGE model allows for accounting for cross-sectoral effects, and for regulations affecting bilateral Frontiers of Economics and Globalization Volume 10 ISSN: 1574-8715 DOI: 10.1108/S1574-8715(2011)0000010017

r 2011 by Emerald Group Publishing Limited. All rights reserved

284

Guillaume P. Grue`re et al.

trade flows, but it also has a number of limitations. The model used here, like the ones used in the other papers in the literature, is static, based on an aggregated representation of the global economy (GTAP database), and assumes perfect competition. This means that the absolute results of each scenario may not perfectly represent the actual welfare effects engendered by the adoption of biotech crops. Still, what matters here is the comparison of the relative welfare effects across countries and scenarios. The simulations are also done ex-ante, so, even if the model here was calibrated with country-based data, the results do depend on hypothetical assumptions about the performance of the selected technologies. Originality/value of the paper – The chapter aims to illustrate the welfare effects generated by GM crops for adopters, non-adopters, in a segmented and regulated international market. Unlike other papers, the review section provides key transversal lessons from the literature, accounting for results from both partial equilibrium and CGE model studies. The empirical application focuses on four populous Asian countries that have been largely left out of the literature. The model used in the simulation presents a number of improvement from the CGE literature on GM crops, including partial adoption, factor-biased productivity shock in each adopting country, GM labeling regulations modeled as trade filters, and the inclusion of costly non-GM segregation as observed in the international market. Keywords: Genetically modified food, international trade, Asia JEL Classifications: Q16, Q17, Q18

1. Introduction In a globalized economy, the overall welfare generated by the use of genetically modified (GM) crops does not simply consist in the sum of their effects in adopting countries. Given that the four main GM crops (canola, corn, cotton, and soybeans) are also largely traded agricultural commodities, their wide adoption by large exporting nations inevitably implies that they have had direct effects on foreign consumers and competitors. Furthermore, the successful adoption of GM crops in certain countries has generated spillover effects on others, encouraging them to develop, test, and sometimes adopt the same technologies. It also encouraged the informal introduction and/or copy of GM seeds in a number of non-adopting countries. While these effects are similar to those observed for other productivityenhancing agricultural technologies, the marketing standards and regulations applied to GM crops and the products derived thereof (except GM cotton) introduce some particularities. Domestic biosafety regulations of GM food and feed crops and the products derived thereof, coupled with

International Trade and Welfare Effects of Biotechnology Innovations

285

contrasting consumers’ and buyers’ preferences, have shaped the international market for GM products. The adoption of productivity increasing technology by a large exporter would be expected to reduce price at importers, lower export prices for exporting competitors, and potentially lower returns to production domestically in an open economy. But the same effects will not necessarily apply to an international market where some countries authorize importing none or only a limited number of authorized GM products and some of them require processing or tracking systems of GM grains before importation. To make it simple, GM crops introduction in a regulated and differentiated market setting have induced three types of international economic responses. First, they have resulted in supply shifts, most significantly in emerging economies, resulting in global price decline with gains for consumers/buyers and potential loss especially for non-adopting competitors (Elbehri and Macdonald, 2004; Frisvold et al., 2006). Second, this supply shift effect is amplified by the presence of import regulations, resulting in trade diversion, depressed price for GM products (Vigani et al., 2010), and lower gains for large adopters. Third, the differentiated demand of certain GM-averse or GM-sensitive countries has created specific niche markets for identity preserved non-GM grain1 and/or increased demand for existing substitutes (organic and conventional), which may have benefited producers and traders of these substitutes and changed the range of products available to consumers (Carter and Grue`re, 2006). The main policy question is to evaluate the economic, trade, and welfare implications of combining these responses for existing and future potential GM crops to observe which of these effects has been and is bound to be dominant. In their discourse on trade-related regulations, GM proponents have often focused their arguments on the first effect, noting that import regulations would result in welfare losses in regulating countries, given the safety of GM products, and that separating GM from non-GM would be prohibitively costly. On the other hand, anti-GM pressure groups tend to talk about the second effect, saying that regulations and/or import bans are necessary, and that countries will lose exports if they adopt GM crops because of the lack of demand and import regulations. This pattern has also been replicated at the country level. Countries embracing GM crops have often ignored export risks in their decision making, letting developers assess market effects themselves. For instance, the United States has not taken into consideration market risks engendered by import regulations and zero percent tolerant levels for unapproved GM crops in major importers in their regulatory framework. Companies had to take care of market potential and market risks on their

1 For instance, see the institutionalized market for non-GM soybean in Japan, which is quoted in the Tokyo Grain Exchange.

286

Guillaume P. Grue`re et al.

own, just like for any other product. The presence of potential production externalities (generated by the risk of mixing supplies at every level of the chain) required companies to self-regulate or face potential liability charges. While such stewardship has worked well in the case of soybeans for the past 15 years, it has not been the case for corn or rice, as evident with the reputed GM/non-GM commingling cases observed for StarLink corn, Bt 11 corn, and LL 601 rice, all of which created significant international market disruption with long-term losses for GM and nonGM producers (Ledford, 2007; Carter and Smith, 2007). In contrast, GM-sensitive countries have taken a strict precautionary approach to market risks, sometimes even considering potential but unproven export losses as sufficient to avoid testing or importing any GM technology. Several authors have shown that the fear of export loss has been a major driver in the reluctance to use GM technology in developing countries (Paarlberg, 2002; Grue`re and Sengupta, 2009). Yet, some of these fears have been irrational (Paarlberg, 2006; Smyth et al., 2006), and were found to be based on a combination of risk averse, poorly informed decision makers, with narrow or misleading views on trade potential and the feasibility of GM/non-GM segregation (Grue`re and Sengupta, 2009).2 The purpose of this chapter is to provide an economic assessment of the three above-listed effects and their implication on global welfare and the future management of market risks, striking a balance between laissezfaire and over-precautionary approach. The following section provides a rapid review of the GM and trade empirical literature. We then present an empirical application using an improved computable general equilibrium (CGE) modeling approach analysis focusing on GM rice and wheat adoption in India, Bangladesh, Indonesia, and the Philippines. The case study is used to assess the impacts of large importers’ regulations on the potential benefits of adopting particular GM crops in the four countries (i.e., opposing the first and second effects) and evaluating the opportunity cost of GM/non-GM segregation for these crops under the external constraints previously defined (third effect).

2. Lessons from past studies A significant number of articles have been published on the global trade effects of GM crop adoption. In their review of the applied economic literature on GM crops in developing countries during the first decade (1998–2007), Smale et al. (2009) found 26 peer-reviewed papers or articles in this category, but more have been published since then. Focusing on 2 In fact, virtually all large GM food or feed-producing countries (United States, Canada, Argentina, Brazil, India, and South Africa) produce alternative non-GM crops, and even organic crops for domestic and/or international markets.

287

International Trade and Welfare Effects of Biotechnology Innovations

Table 1. Crop modeling framework

Simulations by crops and modeling framework from reviewed studies Maize, canola, and soybeana

CGE 11 Partial equilibrium Total 11

Maize, canola, soybean,a rice, and wheat 4

4

Soybean Rice Golden Wheat Cottonb Other rice

1 2

4 1

3

3

5

3

1

6 2

2

1

8

2

List of reference used: Partial equilibrium studies: Annou et al. (2005); Berwald et al. (2006); Cabanilla et al. (2005); Frisvold et al.(2006); Kaye-Blake et al. (2008); Langyintuo and Lowenberg-DeBoer (2006); Moschini et al. (2000); Sobolevsky et al. (2005). CGE studies: Abdalla et al. (2003); Anderson and Jackson (2003); Anderson and Jackson (2005a); Anderson and Jackson (2005b); Anderson et al. (2004); Anderson and Valenzuela (2007); Anderson et al. (2008); Anderson and Yao (2003); Boue¨t and Grue`re (2011); Elbehri and Macdonald (2004); Grue`re et al. (2009b); Hareau et al. (2005); Huang et al. (2004); Jensen et al. (2009); Nielsen and Anderson (2001); Nielsen et al. (2001, 2003); Stone et al. (2002); Van Meijl and van Tongeren (2004). a Most CGE simulations use the GTAP categories coarse grain and oilseeds without differentiating the effects by crop. b Most studies use the plant-based fiber category of GTAP instead of only cotton.

methods, Smale et al. (2009) divide these papers into three subcategories: descriptive studies, studies using partial equilibrium models, and those using CGE models.3 Descriptive studies focused on past, present, or future market access issues with GM crop adoption (e.g., Paarlberg, 2006; Wu, 2006). In contrast, simulations in partial or general equilibrium frameworks are used to empirically determine the economic effects of GM crop adoption in the presence or absence of domestic or foreign regulations in various regions. In this section, we do not provide a complete review of the literature; instead, we identify some of the transversal results transpiring from the empirical economic literature. All of the papers we review provide measurement of the welfare effects under specified GM crop adoption and regulation scenarios in selected regions. Table 1 presents the distribution of results by crop and modeling framework of the 27 distinguishable peerreviewed papers used in this review (listed under the table). We find three main welfare results that seem to be consistent across studies. While we contend that these lessons are robust enough to provide 3 Since then, a new category could be added with the publication of the first ex-post studies using statistical and econometric tools to study the trade impact of GM regulation (Vigani et al., 2010) and/or trade as a motivation for GM regulation (Grue`re et al., 2009a).

288

Guillaume P. Grue`re et al.

useful conclusions, we also acknowledge that they are based on ex-ante simulation studies, and therefore rely on specific – often uncertain – assumptions that may not fully or accurately capture the effects of GM crops in trade. They also depend on the modeling assumptions and scenarios. The first general lesson is that in the absence of GM-specific trade regulations, adopting a GM crop is generally beneficial and non-adopting importers will gain while non-adopting competitors may lose. The gains of adopting are consistently found for adopters in CGE studies. For instance, while the amplitude of gains vary, Abdalla et al. (2003), Nielsen et al. (2001), and Anderson and Jackson (2005a) consistently find welfare gains for all adopters of GM coarse grains and oilseed. Van Meijl and van Tongeren (2004) report positive gains for Europe if it adopts these crops. Grue`re et al. (2009b) and Huang et al. (2004) find that India and China largely benefit from GM rice adoption. Anderson et al. (2008) and Elbehri and Macdonald (2004) conclude that sub-Saharan African countries will derive significant gains if they adopt GM cotton. These conclusions are confirmed by the results of partial equilibrium studies. Using an international model of rice, Annou et al. (2005) show that the adoption of drought-resistant GM rice would provide significant benefits globally, and especially benefit early investors and importing countries. Cabanilla et al. (2005) estimate that Bt cotton will result in significant benefits for West Africa compared to non-adoption, Frisvold et al. (2006) also find significant adoption gains for the same technology in the United States and China. Sobolevsky et al. (2005) uniquely model the effect of intellectual property rights in their economic model of the soybeans market and show that adopters in the United States would gain in the absence of trade regulation. Berwald et al. (2006) also find that the opportunity cost of rejecting GM wheat in the United States and Canada when others adopt it is significant. But these adoption gains do not always translate into benefits for all. While Annou et al. (2005) report positive results from GM rice adoption, their model results reveal that producers may lose from the resulting price decrease. Kaye-Blake et al. (2008), using a multicommodity partial equilibrium trade model, show that producer returns will vary largely depending on national and foreign adoption levels as well as consumer acceptance, but can be positive or negative. Even at the aggregate level, GM crop adoption can result in welfare losses, for instance, in the case of GM cotton in Argentina (Boue¨t and Grue`re, 2011) and the United States (Anderson and Valenzuela, 2007), rice in sub-Saharan Africa (Hareau et al., 2005), or soybeans in the Americas (Moschini et al., 2000), most often because of competitive pressure related to the presence of other adopters following the famous case of ‘‘Immiserizing Growth,’’ first illustrated by Bhagwati (1958). Not unexpectedly, the same price depressing effect reduces welfare for competing non-adopters as best illustrated in the case of cotton (Frisvold

International Trade and Welfare Effects of Biotechnology Innovations

289

et al., 2006); India and sub-Saharan Africa systematically lose when they do not adopt GM cotton (e.g., Anderson and Yao, 2003; Elbehri and Macdonald, 2004; Cabanilla et al., 2005; Grue`re and Boue¨t, 2011). Similarly, results from the spatial trade model simulations of the cowpea market by Langyintuo and Lowenberg-DeBoer (2006) show that nonadopting producers are bound to lose from GM cowpea adoption in neighboring countries within West Africa. In contrast, noncompetitors and importers derive gains from the adoption of productivity enhancing technologies in the absence of regulation. China, Europe, and Japan, who tend to be either self-sufficient or importers of cereals, do derive benefit from GM crops adoption in other countries (Nielsen and Anderson, 2001; Anderson and Jackson, 2005b). The second general lesson is that accounting only for their market effects, the introduction of GM-specific trade regulations reduces welfare gains especially for non-adopters. The caveat should be emphasized here; none of the models incorporates nonmarket utility gains for consumers in importing countries. Not only are biosafety regulations designed to address potential safety concerns associated with GM food, but their avoidance or strict regulatory oversight may also be found beneficial by GM-averse consumers. If consumers in Europe, Japan, or Korea overwhelmingly express a positive willingness to avoid GM products, market losses may not necessarily translate into net welfare losses. Naturally this caveat also applies to the previous result – in the absence of regulation; importing GM products in these countries may not necessarily translate into benefits for all consumers. That said, the market effects tend to be relatively consistent across studies, regardless of the modeling framework. The introduction of import bans or labeling regulations in non-adopting Europe, Japan, or Korea results in reduced welfare gains or net welfare losses (Stone et al., 2002; Abdalla et al., 2003; Anderson et al., 2004; Van Meijl and van Tongeren, 2004; Anderson and Jackson, 2005b; Berwald et al., 2006; Grue`re et al., 2009b). In the case of rice, Grue`re et al. (2009b) find positive market effects with domestic regulations in Japan, Korea, or Europe but only in the case of specific labeling regulation scenarios. Part of it may due to protectionist producer gains, as found in the case of coarse grains in Europe (Anderson and Jackson, 2003). Lastly, results from the literature show that importers’ regulations can reduce gains for exporting adopters. Perhaps the most common experiment with trade simulations is the introduction of import regulations, especially import bans, and the measurement of its effect on exporting GMcrop-producing countries. Most often the regulating countries are Western Europe (EU) and Japan (sometimes with Korea). Results from simulations suggest that trade regulations abroad can significantly reduce gains for adopting countries, but mostly in the case of coarse grains, oilseeds, and cereals in Argentina and the United States (Abdalla et al.,

290

Guillaume P. Grue`re et al.

2003; Van Meijl and van Tongeren, 2004; Anderson and Jackson, 2005b). The effects are almost null for India or China (Huang et al., 2004; Grue`re et al., 2009b), and negligible, if not positive for sub-Saharan African countries (Nielsen and Anderson, 2001; Anderson and Jackson, 2005a). The same studies show that adopting GM crops in these countries will result in welfare gains that largely exceed any potential loss of market in Europe. With costless segregation, markets can adjust and developing countries are bound to gain from GM crop adoption (Nielsen et al., 2001).4 These results however may not apply to all adopting region, as adopting GM crops can result in market losses if the main export target is banning GM imports (e.g., for Australia and New Zealand, see Anderson and Jackson, 2005b).5 Overall, these three lessons provide useful insights on the international welfare effect of GM crop adoption, but they also leave many ambiguities and do not provide a complete picture of the complexity of market scenarios. Among others, few studies focus on the case of GM food crops or on the effect of GM crop adoption in net importers, or provide insights into the effect of segregation on welfare. As noted in Smale et al. (2009), many studies also used ad hoc productivity shocks to calibrate their models. The purpose of the following section is to provide a specific application, with importers and exporters, in the case of GM wheat and rice, two major food crops in less well covered countries based on expertsolicited data.

3. The case of GM food crops in Bangladesh, India, Indonesia and the Philippines In this application, we build on previous literature using CGE models by improving the representation of trade policies and refining the assumptions on the productivity effects of biotechnology in at least two respects.6 First, we account for the ‘‘trade filter’’ effect of GM food labeling policies in sensitive countries, which allows certain marketable products to be imported, but not others.7 Second, we include the option of costly segregation of non-GM crops for export, and we model GM crop adoption as a factor-biased productivity shock based on expert-based disaggregated data on agricultural constraints. The main assumptions, modeling framework, and policy scenarios are explained below. 4 In contrast, costly segregation can benefit non-adopting exporters and consumers in Western countries (Sobolevsky et al., 2005). 5 Furthermore with segregation and labeling, small open economies will not always gain from adopting GM crops (Plastina and Giannakas, 2007). 6 See Smale et al. (2009) for details. 7 For more on international GM food labeling regulations, see Grue`re and Rao (2007).

International Trade and Welfare Effects of Biotechnology Innovations

291

3.1. Productivity shock The productivity shocks associated with the use of GM rice and wheat are derived from primary and secondary data. We conducted a series of consultations and focus group meetings with scientific, agricultural, and regulatory experts in India and Bangladesh in July 2005 and Indonesia and the Philippines in September 2005 on the potential effects of biotechnology improvements to resist to biotic and abiotic stresses eliciting subjective estimates. The discussions in our meetings helped us to decide to select four types of relevant traits: insect resistance, virus resistance, drought resistance, and salt tolerance applied specifically to rice and wheat in each country. In parallel, we obtained existing studies of GM technology, productivity constraints, and technology potential publicly available for these and other countries (e.g., Evenson et al., 1996). The productivity effect of each GM crop/trait combination is modeled based on its effect on yields, use of chemical inputs (mainly pesticides), and its assumed effect on labor at the water basin level in India and national level for the three other countries.8 Combining expert estimates on constraints and productivity potential and secondary data on yield constraints, we derived expected yield effects in rainfed versus irrigated land in the 10 water basins of India and the other 3 countries.9 For the case of drought- and salt-tolerant crops, we also estimated the share of affected areas in each subregion in order to account for the fact that not all land is affected by drought or soil salinity constraints. To do so, we used categorical indicators of drought and salinity constraints by areas of production, type of land, and water basin based on a satellite imagery and agricultural study developed by the spatial team of IFPRI.10 By filtering these indicators with production area in each spatial unit, we obtained share of affected areas in each subregion. We used exogenous adoption rates based on collected expert data, and adjusted them according to the type of land (rainfed versus irrigated) and subregion.11 We also corrected the production share of each Indian region by a proportional factor linked to historical data of the adoption of high yielding varieties of each crop. The adoption rate in each subregion was then multiplied by each yield and area factors to obtain a total expected

8 We did not have the proper data to include the cost of seeds as a third factor. Yet we can justify the exclusion of seed premiums by using exogenous partial adoption rates. 9 The detailed aggregation and calibration procedures are explained in Grue`re et al. (2007). 10 The detailed mapping methodology, using an entropy approach to spatial disaggregation is explained in detail in You and Wood (2006). 11 In particular, we assume that producers in irrigated areas tend to have a better access to new technologies, but at the same time, rainfed producers may benefit more from certain technologies.

292

Guillaume P. Grue`re et al.

Table 2. Aggregate relative productivity factors and adoption rates of the composite GM crops for the countries of studies used in the simulation model Crop

Country

% Yield effects

% Input effects Chemicals

Rice

Wheat

Bangladesh India Indonesia Philippines Bangladesh India

5.12 8.38 3.44 4.39 5.63 24.6

% Adoption

Labor

29.72 14.59 14.50 60 0 6.02

5.2 3.78 8.79 7.82 0 4.5

49.2 71.74 63.92 60.34 14.75 24.43

Source: Authors’ computations.

Table 3. Crop

Relative productivity effects and initial adoption rates assumed for other countries Country

% Yield effects

% Input effects Chemicals

Rice Wheat

China China Argentina

7.03 7 7

65 0 0

% Adoption initial

Labor 9.1 7.7 7.7

80 50 50

Sources: Authors’ assumptions based on Huang et al. (2004).

yield effect of the technology.12 To translate these data into usable inputs into the model, we computed the aggregated relative productive effects of the composite GM crops, as shown in Table 2. In our consultation meetings, we found that local experts in the four Asian countries agreed that GM food would only be introduced in their country if China adopted it first. We therefore assume that China will be a technology leader for rice, but because of the lack of data, we only let China adopt Bt rice. For wheat, we hypothesize that China and Argentina use herbicide-resistant varieties with identical effects.13 Our assumptions are shown in Table 3 with relative yields, input effects, and adoption rates. 12 The details of the procedure can be found in Grue´re et al. (2007). The assumptions derived from this process for the four countries of study in terms of yield effects and input effects may be requested from the authors. 13 Some of the productivity assumptions shown in Table 7 (e.g., wheat) are not completely comparable to the ones shown for our countries of study in Table 6 simply because the traits are not the same, and because the relative productivity effects shown in Table 6 represent composite (multitrait) GM varieties rather than simple varieties.

International Trade and Welfare Effects of Biotechnology Innovations

293

3.2. Trade modeling and scenarios A modified version of the MIRAGE CGE model (Bchir et al., 2002)14 was used to simulate a range of scenarios on the productivity effect, trade restrictions, and segregation options. This model is based on the GTAP 6.1 database, which represents the world as of 2001. For this application, we divide the economy into 21 regions, including GMcrop-producing countries, sensitive importing countries, and other important countries, and 19 sectors, including the relevant production sectors, as well as the chemical sector.15 The MIRAGE model includes a representation of trade policies and trade preferential agreements (using MacMap-HS6 2001 data). We first modified the MIRAGE model by dividing the two production sectors (wheat and rice) into GM and non-GM substitutes for all GMadopting countries. Second, the model was changed to allow for the use of specific productivity shocks only on GM products in each GM sector for each adopting countries. The model was also modified to allow for the ban of GM and/or non-GM imports in selected countries only from GMproducing nations going toward the final consumption to reflect the current effects of labeling policies (Carter and Grue`re 2006). Lastly, the model permits the introduction of a segregation cost for non-GM going from GMcrop-adopting countries to sensitive importing ones. To calibrate the model, we used the assumed parameters provided above regarding the productivity shocks and the proposed initial adoption rates.16 However, because of the relative aggregated level of the GTAP database, we reduced the productivity shock proportionally to account for the share of pesticide costs into the aggregated GTAP chemical sector for each GM crop concerned in each country.17 After this data adjustment, under each set of scenarios, the model was calibrated to incorporate the assumed productivity shocks in all selected GMcrop-adopting nations. We then ran the model only once to simulate a comparative static shock. It should be noted that, like others, we use a perfect competition representation of the economy for simplification.18 We define two distinct sets of scenarios. The first set, entitled RICE, represents the case of the adoption of GM rice in the four countries of 14 The MIRAGE model was developed at the Centre d’Etudes Prospectives et d’Informations Internationales (CEPII) in Paris. The version used in this paper is based on developments of the model done at the International Food Policy Research Institute (IFPRI). More information is available at http://www.mirage-model.eu/miragewiki/. 15 The sector disaggregation is available upon request. 16 The calibration process is explained in detail in the appendix of Grue`re et al. (2007). 17 We use a two-step approach, first deriving the share of fertilizer in chemical use from FAOSTAT 2001, and second by using general data on the share of insecticides in total pesticide use at the continental level (Yudelman et al., 1998). 18 Further refinements of our simulations could include dynamics and imperfect competition modeling.

294

Guillaume P. Grue`re et al.

study and in China. The second set is named WHEAT, and presents the introduction of GM wheat in Bangladesh, India, China, and Argentina. Each set of scenarios comprises of eight individual scenarios, as shown in Table 4. Scenario 0 is run as a benchmark without GM production. Scenario 1 simulates a productivity shock associated with the adoption of GM crops and no trade restriction. Scenario 2a represents the short-run effect of the adoption of new GM varieties, namely the import ban of GM and non-GM crops from the adopting countries in sensitive countries. Scenario 2b represents current trade restrictions on GM imports in sensitive countries. Current marketing regulations, private standards, and consumer reactions in these countries act as a ‘‘trade filter.’’ Products to be used for final consumption are not purchased, but products for intermediate consumption (such as animal feed or food ingredient) can enter the market in sensitive countries because the corresponding final products are not necessarily subject to labeling requirements (e.g., meat in the EU, soy oil in Japan, for more on labeling, see Grue`re and Rao, 2007). Lastly, scenarios 3a-i, 3a-ii, 3b-i and 3b-ii allow for the segregation of nonGM products in GMcrop-adopting countries to export to sensitive importing countries. The four scenarios are proposed to study the implication of segregation costs under trade ban or trade filter. 3a-i is run with costless segregation of non-GM but a ban of GM toward both final and intermediate consumption; 3a-ii is the same scenario with the addition of a 5% basic segregation cost.19 Similarly, scenario 3b-i represents the case of a trade filter in sensitive countries but costless segregation of non-GM toward the final consumption; 3b-ii adds a 5% segregation cost.

3.3. Simulation results We present the results in terms of production, trade, and welfare effects, defined as the equivalent variation (or real income) between each scenario and the base (0) for each set. Both absolute values in 2001 US $ million per year and percentage of total real income are shown for each region in each scenario. 3.3.1. GM rice adoption Changes in production and trade for rice are presented in Table 5. Changes in welfare effects for RICE are presented in Table 6. In this case, 19 We choose to impose a 5% cost for two reasons: first, it corresponds to a median value in the literature on segregation cost (where estimates vary from a few percent to 10–15%), second because it corresponds to the premium reported on the market for non-GM products. For instance, maize traders in South Africa reported in June 2007 that identity preserved nonGM maize was sold for a 5% price premium compared to GM maize (Grue`re and Sengupta, 2010).

Scenario number and title

0. Base 1. Productivity shock 2a. Import ban, no segregation 2b. Import filter, no segregation 3a-i. Import ban, costless segregation 3a-ii. Import ban, 5% segregation costs 3b-i. Import filter, costless segregation 3b-ii. Import filter, 5% cost segregation a

Definition of scenarios for GM adopting countries

Productivity shock on GM crops

No Yes Yes Yes Yes Yes Yes Yes

Ban toward the intermediate consumption in sensitive countriesa of

Ban toward the final consumption in sensitive countriesa of

Non-GM

GM

Non-GM

GM

No No Yes No No No No No

No No Yes No Yes Yes No No

No No Yes Yes No No No No

No No Yes Yes Yes Yes Yes Yes

The sensitive countries are the European Union, Rest of Europe, Japan, South Korea, and Australia–New Zealand.

Segregation of non-GM product exported toward sensitive countriesa No No No No Yes Yes Yes Yes

International Trade and Welfare Effects of Biotechnology Innovations

Table 4.

295

296

Table 5.

Guillaume P. Grue`re et al.

Percentage changes in production, export and import volumes for selected set RICE scenarios in GM adopting countries

SET RICE

Scenario

Production

1 2a 2b 3b-ii 1 2a 2b 3b-ii 1 2a 2b 3b-ii

Exports

Imports

China 19.7 19.5 17.9 19.5 18.3 5.6 18.0 15.6 45.9 46.1 46.7 46.1

Bangladesh 7.5 7.5 7.5 7.5 9.8 5.1 31.4 3.0 27.0 27.1 27.4 27.1

India 19.8 19.7 16.4 19.6 15.3 7.9 6.0 13.5 49.4 49.7 50.4 49.6

Indonesia 20.4 20.2 19.3 20.2 150.7 100.5 8.0 130.4 –65.2 –65.4 –65.8 –65.3

Philippines 19.1 18.9 17.8 18.9 84.5 46.4 24.1 67.6 56.2 56.4 57.0 56.3

Source: Authors’ derivations.

five countries adopt GM rice: China, India, Bangladesh, the Philippines, and Indonesia. The global welfare gains with the adoption of GM rice range from $9.9 billion to $11.5 billion per year a total exceeding previous estimates because of the intense adoption in major rice-producing country. Second, trade restrictions in the form of an import ban in sensitive countries reduce global gains by 14%. Segregation at a 5% cost reduces global gains by about 6%. Interestingly, segregation at 5% costs also results in lower global gains than a trade filter, which indicates that, provided rice is accepted in intermediate consumption, segregation would increase global welfare gains if it does not cost too much. The major welfare gains occur in the five adopting countries. First, China, with a relatively large adoption rate, gains over $4.6 billion per year (or 0.6% of total real income). This total is slightly larger than former studies, including the maximum effect found by Huang et al. (2004) because we do not explicitly reduce the gains from GM crops due to the price of seeds.20 In the GTAP database, Chinese rice imports are just slightly inferior to its exports, making China a small net exporter. This may explain why an embargo on rice slightly reduces its gains, a trade filter reduces it a little less, and the welfare gains with costless segregation are close to the first scenario. But at the same time, China obtains slightly higher gains with costly segregation in sensitive countries than under other scenarios. 20 Therefore, the gains presented here include the returns to the developers and adopting producers together.

Table 6.

Change in welfare effects ($million per year) under each scenario with GM rice adoption in selected Asian countries

Region

2a. Import ban, no segregation

2b. Import filter, no segregation

3a-i. Import ban, costless segregation

3a-ii. Import ban, 5% segregation cost

3b-i. Import filter, costless segregation

3b-ii. Import filter, 5% segregation cost

$ million

$ million

$ million

$ million

$ million

$ million

$ million

5.5 4640.5 529.7 191.1 8.5 1106.8 638.8 452.6 3258.8 10.3 104.3 6.1 28.1 1.8 0.8 350.2 38.6 105.5 74.1 13.4 1.6 11533.9 (0.05)

Source: Authors’ results from simulations.

4.3 4617.6 292.5 159.5 1 1102.3 637.6 452.8 3241.4 10.9 101.6 5.2 33.1 1.5 0.4 61.0 14.8 106.0 73.5 13.2 1.6 9891.5 (0.04)

6.1 4632 211.2 165.2 5.8 1105.1 638.5 452.7 3252.8 10.5 104.9 5.9 29.4 1.8 0.7 194.8 27.3 106.5 74.4 13.4 1.6 11012.6 (0.05)

4.9 4627.7 94 22 6.4 1105.5 638.2 452.7 3252.4 10.5 103.2 5.7 30.2 1.6 0.6 165 26.2 105.4 73.7 13.3 1.6 10713.5 (0.04)

0.4 4640.5 131.6 182.7 0.2 1106.4 638.0 452.8 3250.8 16.3 105.3 8.7 36.1 10.2 30.1 58.7 7.8 105.3 74.7 12.5 1.5 10044.2 (0.04)

5.8 4636 357.9 177.6 7.4 1106.2 638.6 452.7 3256.3 10.4 104.6 6 28.7 1.8 0.8 276.5 32.6 105.9 74.2 13.4 1.6 11263.4 (0.05)

0.5 4649.7 153.2 10.0 1.3 1107.3 638.5 452.8 3255.5 16.2 107 9.1 34.4 10.3 30.3 67.1 14.7 105.9 75.3 12.6 1.6 10648 (0.04)

297

Australia and New Zealand China Japan South Korea Rest of Asia Indonesia Philippines Bangladesh India Canada United States Mexico Rest of Latin America Argentina Brazil European Union Rest of Europe North Africa and Middle East Rest of sub-Saharan Africa South Africa Tanzania and Uganda World % of total

1. Productivity shock

International Trade and Welfare Effects of Biotechnology Innovations

GM rice adopted in bold regions

298

Guillaume P. Grue`re et al.

India also obtains large positive gains from adoption, exceeding $3.2 billion, or 0.87% of total welfare, a total close to the maximum found in the literature. But India is a net exporter of rice and therefore it gains less with trade restrictions and more with segregation. The net reduction in welfare gains with a complete ban of rice in sensitive countries only amounts to $17 million – a result consistent with the literature. This small change can be explained by the fact that India does not export as much to Europe (about 16% according to the original GTAP database) as it does to other regions, like North Africa and Middle East (47%) and other African countries (19%), and that trade diversion occurs with selective bans. Bilateral trade flows show that, under scenario 2a, Indian rice is slightly diverted from Europe to South and North America, and African countries. In total, under the trade ban, India produces 16% more than without GM rice, which is less than under other scenarios; it reduces total rice exports by 6%, but still reduces its total rice imports by over 50%. Under scenario 2b, India’s welfare is reduced by a much smaller amount. Segregation allows to increases welfare by a few million dollars even at 5% costs. Bangladesh obtains the largest relative gains with the adoption of GM rice, with an additional 1.2% gains in total welfare per year, which is equivalent to over $450 million per year. Rice production increases by 7.5% under all scenarios, which allows the country to reduce rice imports by over 27%. The trade ban results in a significant relative reduction in exports, but this loss is limited as the absolute value of its rice exports is small ($1 million in the GTAP database) compared to its imports ($74 million). As a net importer, Bangladesh is slightly better under the most restrictive trade scenarios because they are associated with relatively lower import prices. Indonesia also obtains very significant gains from GM rice adoption, exceeding $1.1 billion per year, or 1.1% of total welfare. Indonesia is also a large net importer of rice. With GM rice, Indonesia increases its production by 20% and reduces its imports by 66%. A total trade ban in the short run has small effects on Indonesia’s welfare with a reduction of $4 million. Indonesia is slightly better off with a trade filter, and with segregation, but the differences are really small relative to the total gains. The Philippines increases its welfare by 1% (or about $640 million) annually by adopting GM rice. Originally a net importer, the introduction of GM rice results in an increase in production by about 17–19% under all scenarios, and reduces imports by more than half. The changes across scenarios are really small resulting from production and import differences with price changes. The relative loss of sensitive importers from a total ban on rice coming from GMcrop-adopting countries explains the almost entirety of the difference in global welfare across scenarios. In the group, Japan loses the

299

International Trade and Welfare Effects of Biotechnology Innovations

most from a total ban and gains the most from a trade filter and costless segregation. Japan’s welfare rapidly declines with an increase in the segregation costs. Europe loses a small relative amount under the total ban but still gains under all other scenarios as a net importer of rice. 3.3.2. GM wheat adoption Tables 7 and 8 show the results in terms of production and trade and welfare obtained with the WHEAT set, in which China, Argentina, India, and Bangladesh adopt GM wheat. The global gains are much less than with GM rice adoption, ranging between $1.6 and $2.3 billion annually. It is important to note that although Argentina, India, and China are relatively large producers of wheat, other countries of North America, Europe, or Oceania dominate the global wheat market. Global real income decreases only minimally with trade restrictions (2a or 2b) compared to the simple productivity shock (1), but reduces more significantly with the introduction of costs of segregation (3a-ii and 3b-ii). For comparison with previous scenarios, trade restrictions reduce welfare gains by 1.3%, while costly segregation reduces welfare gains by up to 30%. Most of the relative losses with costly segregation occur in sensitive countries, mostly Japan and South Korea, who have to pay more for imports of wheat. These same importers only incur relatively small reductions in welfare gains with trade bans because they are able to source their imported wheat from other countries, notably North America and Australia. China increases its welfare and its wheat production, respectively, by 0.09% (or $690 million) and 9% with adoption of GM wheat, and reduces its wheat imports by over 40%. Trade restrictions do not affect its welfare Table 7.

Percentage changes in production, export, and import volumes for selected set WHEAT scenarios in GM adopting countries

SET WHEAT

Scenario

Production

1 2a 2b 3b-ii 1 2a 2b 3b-ii 1 2a 2b 3b-ii

Exports

Imports

Source: Authors’ derivations.

China 9.1 9.1 8.3 9.0 44.7 44.1 63.8 10.7 43.6 43.6 44.1 43.7

Bangladesh 4.3 4.3 4.4 4.3 36.8 36.8 37.2 36.9 0.6 0.6 0.6 0.6

India 16.8 16.8 14.5 16.2 50.5 50.3 35.2 45.6 42.7 42.7 43.3 42.9

Argentina 30.8 30.8 30.3 30.7 5.8 5.8 6.1 5.9 36.9 36.9 37.1 37.0

GM wheat adopted in bold regions

Region

1. Productivity shock

2a. Import ban, no segregation

2b. Import filter, no segregation

$ million

$ million

$ million

21.2 687.8 57.9 15.4 14.8 4.3 9.1 10.4 945.2 28.4 14.3 2.5 10.7 215.4 76.9 75.7 10.8 132.9 16.9 9.2 0.6 2261.3 (0.009)

Source: Authors’ results from simulations.

16 684 51.5 2.5 13.8 4 8.9 10.6 940.5 26.3 20.2 2 10.4 212.7 77.3 73.4 9 131.6 16.7 9.2 0.5 2231.8 (0.009)

21.2 687.8 57.9 15.3 14.8 4.3 9.1 10.4 945.2 28.4 14.3 2.5 10.7 215.2 77 75.6 10.6 133 17 9.2 0.6 2260.8 (0.009)

3a-i. Import ban, 3a-ii. Import ban, costless 5% segregation segregation cost $ million 20. 686.5 56.54 10.9 14.5 4.2 9.1 10.4 944.7 27.9 15.7 2.4 10.7 214.2 77.1 74.9 10.2 132.6 16.9 9.2 0.6 2253.3

$ million 14.4 698.9 175.3 196.8 21 5.3 8.8 10.5 942.6 22 18.1 5.4 16.6 205.8 47.5 146.4 8.4 132.5 17.8 8.4 0.6 1576.5

3b-i. Import filter, costless segregation

3b-ii. Import filter, 5% segregation cost

$ million

$ million

21.2 687.8 57.9 15.4 14.8 4.3 9.1 10.4 945.2 28.4 14.3 2.5 10.7 215.3 77 75.6 10.7 132.9 16.9 9.2 0.6 2261.1

15.7 700.3 173.8 192.1 21.3 5.4 8.9 10.5 943.3 22.6 16.5 5.5 16.6 207 47.3 145.7 7.8 132.9 17.9 8.4 0.6 1584.7 (0.006)

Guillaume P. Grue`re et al.

Australia and New Zealand China Japan South Korea Rest of Asia Indonesia Philippines Bangladesh India Canada United States Mexico Rest of Latin America Argentina Brazil European Union Rest of Europe North Africa and Middle East Rest of Sub-Saharan Africa South Africa Tanzania and Uganda World

Change in welfare effects ($ million per year and %) under each scenario with GM wheat adoption in selected Asian countries

300

Table 8.

International Trade and Welfare Effects of Biotechnology Innovations

301

gains significantly; however, adding a cost of segregation does increase the total welfare gains of China because it results in a small increase in exports toward other countries. China exports about $42 million and imports 10 times more wheat. The costly segregation scenario divides the market into GM (or mixed) and pure non-GM, and the non-GM export price to sensitive countries goes up significantly, while the GM price is slightly reduced down, which could explain the observed gain. Bangladesh is also a large importer of wheat, and only adopts GM wheat at a partial scale in this set of scenarios, for a small overall production, which is reflected by the small gains. Overall, Bangladesh produces 4% less wheat, imports a little less wheat and exports less wheat. Thus, under our assumptions, Bangladesh absorbs less wheat overall, and is not able to compete with India and the other GM wheat producers. India is the main winner from GM wheat adoption, with gains over $940 million per year (or 0.25% of total real income). As a net exporter India gains more under scenarios 1 and 3a/3b and less with complete trade restriction under scenario 2a or 2b. Once again, the loss with a complete ban in sensitive countries is negligible ($5 million) compared to the gains with the adoption of GM wheat. Costless segregation does not make much difference with the trade filter scenario, which can be understood by the fact that virtually all wheat is used in the intermediate consumption not the final consumption of importing-sensitive countries. Costly segregation reduces slightly the gains but still allows India to be better off than under a complete ban. Argentina follows the same pattern as India with a smaller absolute gain with GM wheat.

3.4. Discussion Overall, results from this case study do confirm the three lessons of the literature, but provide more detailed analysis of the different types of conflicting effects. First, our simulations show that the adoption of GM food crops can be translated into significant economic gains in the large majority of regions (consistently with the first lesson of our literature review), but this is true in the presence or absence of trade restrictions in certain sensitive countries. Only a few regions experience net losses with the adoption of GM crops, and these losses occur mostly in importers with domestic regulations (second lesson). Second, our simulations show that trade regulations can affect the gains from GM crops (third lesson), but that this effect is relatively small compared to the gains with the adoption of GM crops. Table 9 shows the relative change in gains out of the total gain from GM crops under the most restrictive scenarios of each set. Even if globally, the gains are reduced by up to 30% overall, we find that the gain reduction is closer to 1% in most cases for our countries of study and China, and does not

302

Guillaume P. Grue`re et al.

Table 9. Relative effect of trade restriction on total gains from GM crop adoption for selected countries in different sets of scenarios Set Scenarios compared China Bangladesh India Indonesia Philippines World

Rice 1 vs. 2a

Wheat 1 vs. 2a

0.5% 0.0% 0.5% 0.4% 0.2% 14.2%

0.6% 1.8% 0.5% 6.7% 2.5% 1.3%

Source: Simulation results.

exceed 6.7% of total gains. Interestingly, in certain cases, trade restriction even results in relative increase in gains for certain net importers or nonadopters. These results also provide light on the welfare implications of market differentiation. Results show that the use of segregation for non-GM crops can help offset some of the relative losses due to trade restrictions. Differences between the trade scenario and the hypothetical case with costless segregation provide benchmark values for the opportunity cost of segregation, defined as the most a country could spend on segregation to avoid losing compared to trade restrictions with no segregation. Estimates of these opportunity costs are reported for selected countries in Tables 10 and 11. Exporting GMcrop-producing countries, like India, have a positive but relatively limited opportunity cost of segregation. Most of the global benefits of segregation would occur in importing sensitive countries rather than exporting GMcrop-producing countries. Thus, traders in sensitive countries will likely have a larger incentive to set up segregation systems in GMcrop-adopting countries than the exporters in these latter countries themselves. More generally, we show that segregation is not a silver bullet; it all depends on cost. In some cases, GM crop adopting countries will gain from segregation even with a 5% cost (e.g., India for rice), while in other cases, these countries will only gain if the cost of segregation is lower (e.g., India for wheat). Emerging countries that are already able to supply highquality agricultural products to sensitive importing countries should be able to take advantage of this option, particularly if the sunk cost is partially assumed by importers. Smaller developing countries may have less incentive and support to set up segregation systems, except if it is driven domestically by a strong niche market for non-GM products. Lastly, with the examples of Bangladesh, Indonesia, or the Philippines, we see that large importers will not become net exporters with partial

International Trade and Welfare Effects of Biotechnology Innovations

Table 10.

303

Opportunity cost ($ million per year) of segregation of non-GM rice for adopting and sensitive countries

Type of country

Country

Segregation of non-GM Segregation of non-GM rice rice for final for final and intermediate consumption only consumption

GM producers China India Indonesia Bangladesh Philippines

3.1 3.6 1.9 0.06 0.16 4.9

10.09 10.92 3.2 0.12 0.64 24.73

Australia–NZ 0.28 Japan 113.67 South Korea 12.47 EU 81.64 Rest of 5.37 Europe 213.43

0.64 198.53 137.55 226.05 11.36

250.78

821.98

Total GM producers Sensitive countries

Total sensitive countries WORLD Global

572.85

Source: Authors’ derivations.

adoption of GM food crops. Their opportunity cost of segregation is negligible and even negative in some cases, when segregation results in slight increases in import prices relative to no segregation. But thanks to the increase in production associated with GM crops, they can dramatically reduce their imports of agricultural commodities to feed their large and growing population. For these countries, the effect of trade restriction is limited to the changes in prices. They can be slightly better-off overall under the most restrictive trade policies because the prices of GM and especially non-GM products decrease under these scenarios compared to no trade restrictions. But these relative differences are minimal and most often negligible compared to the overall gains with the adoption of GM crops.

4. Conclusion The introduction of transgenic crops is perceived as a relative success by some, as revealed by its reported global adoption by millions of farmers, but it is perceived as a relative failure by others, in part because of its limitations to a few countries, crops, and traits, and because of consumer

304

Table 11.

Guillaume P. Grue`re et al.

Opportunity cost ($ million per year) of segregation of non-GM wheat for adopting and sensitive countries

Type of country

Country

Segregation of non-GM Segregation of non-GM wheat for final wheat for final and consumption only intermediate consumption

China India Bangladesh Argentina

0.01 0.04 0 0.08 0.13

2.46 4.14 0.17 1.44 7.87

Australia–NZ Japan South Korea EU Rest of Europe

0 0.01 0.01 0.07 0.14

3.99 4.96 13.42 1.56 1.09

0.23

17.04

0.33

21.46

GM producers

Total GM producers Sensitive countries

Total sensitive countries WORLD Global Source: Authors’ derivations.

concerns in a number of countries. One of the reasons for the limitation of transgenic or GM crops to certain traits and countries is related to market sensitivity, international trade risks, and the fear of export losses. In this chapter, we provide a rapid review of the literature on the international welfare effects GM crops and presented a specific case study of GM rice and wheat introduction in four countries of Asia. We find that, in the absence of regulations, GM crop adoption is generally beneficial for adopting countries and importing non-adopters, that domestic regulations at adopters and especially non-adopters can reduce these gains, and that import regulations in other countries can also affect gains for exporting adopters. The case study illustrates these conclusions, but it also shows that net importers will mostly benefit from adoption in their terms of trade, and that segregation of non-GM crops for export markets can be beneficial if it is not too costly. Thus, our analysis shows that the combination of productivity increase, import regulations, trade diversion, and market differentiation observed in the international markets for GM products will generally advantage those countries that are the most able to adapt to the changing environment – adopting GM when competitors do, segregating when there is a niche market, and not strictly restricting their own domestic market regardless of adoption. But other countries, that do not adopt GM crops nor regulate

International Trade and Welfare Effects of Biotechnology Innovations

305

them, will also benefit from technology adoption either directly or indirectly. One implication of our findings is that market risks, while potentially present for a new adopter of GM crops, are not insurmountable and should be assessed on a case-by-case basis. Neither a complete laissez-faire nor a complete ban will be sustainable in a market where an increasing number of GM crops circulate between countries with diverse regulatory requirements. To reduce the likelihood of trade disruption, countries should be encouraged to develop simple market risk assessment frameworks and ask developers to include them in their dossiers. In case where export risks are significant, a more thorough analysis would have to be conducted, potentially resulting in the setting up of management schemes to avoid large trade disruption. More generally, countries should be encouraged to continue their discussions and efforts toward a progressive harmonization of standards and practices, such as the tolerance levels for unapproved GM or for non-GM products. Acknowledgments The authors wish to thank Rowena Valmonte-Santos, Liang You, Cynthia Rossi, Imdadul Hoque, Purvi Mehta-Bhatt, Sutrisno, and Rey Ebora for their help and contributions at various stages of the first phase of the project on productivity modeling. They would also like to thank all the participants to the meetings in the four countries. This research was supported by the Program for Biosafety Systems and the South Asia Biosafety Program, two programs managed by the International Food Policy Research Institute and funded by the United States Agency for International Development. References Abdalla, A., Berry, P., Connell, P., Tran, Q. T., Buetre, B. (2003). Agricultural biotechnology: potential for use in developing countries. ABARE eReport 03-17. Australian Bureau of Agricultural and Resource Economics (ABARE), Canberra. Anderson, K., Jackson, L.-A. (2003), Why are US and EU policies toward GMOs so different? AgBioForum 6 (3), 95–100. Anderson, K., Jackson, L.-A. (2005a), Some implications of GM food technology policies for sub-Saharan Africa. Journal of African Economies 14 (3), 385–410. Anderson, K., Jackson, L.-A. (2005b), GM crop technology and trade restraints: economic implications for Australia and New Zealand. The Australian Journal of Agricultural and Resource Economics 49 (3), 263–281.

306

Guillaume P. Grue`re et al.

Anderson, K., Valenzuela, E. (2007), The World Trade Organisation’s Doha Cotton Initiative: a tale of two issues. The World Economy 30 (8), 1281–1304. Anderson, K., Jackson, L.-A., Nielsen, C.P. (2004), Genetically Modified Rice Adoption: Implications for Welfare and Poverty Alleviation. World Bank Policy Research Working Paper, Washington, DC. Anderson, K., Valenzuela, E., Jackson, L.-A. (2008), Recent and prospective adoption of genetically modified cotton: a global computable general equilibrium analysis of economic impacts. Economic Development and Cultural Change 56 (2), 265–296. Anderson, K., Yao, S. (2003), China, GMOs and World Trade in agricultural and textile products. Pacific Economic Review 8 (2), 157–169. Annou, M., Fuller, F., Wailes, E. (2005), Innovation dissemination and the market impacts of drought-tolerant, genetically modified rice. International Journal of Biotechnology 7 (1/2/3), 113–127. Bchir, M. H., Decreux, Y., Gue`rin, J-L., Jean, S. (2002), MIRAGE, a computable general equilibrium for trade policy analysis. Working Paper No. 2002-17. Centre d’Etudes Prospectives et d’Informations Internationales (CEPII), Paris, France. Berwald, D., Carter, C.A., Grue`re, G.P. (2006), Rejecting new technology: the case of genetically modified wheat. American Journal of Agricultural Economics 88 (May), 432–447. Bhagwati, J.N. (1958), Immiserizing growth: a geometrical note. Review of Economic Studies 25 (3), 201–205. Boue¨t, A., Grue`re, G. P. (2011), Refining estimates of the opportunity cost of non-adoption of Bt cotton: an application to seven countries in subSaharan Africa. Forthcoming in Applied Economic Perspectives and Policy. Cabanilla, L.S., Abdoulaye, T., Sanders, J.H. (2005), Economic cost of non-adoption of Bt-cotton in West Africa: with special reference to Mali. International Journal of Biotechnology 7 (1/2/3), 46. Carter, C.A., Grue`re, G.P. (2006), International approval and labeling regulations of genetically modified food in major trading countries. In: Just, R.E., Zilberman, D., Alston, J. (Eds.), Regulating Agricultural Biotechnology: Economics and Policy. Springer, New York. Carter, C.A., Smith, A.D. (2007), Estimating the market effect of a food scare: the case of genetically modified StarLink corn. Review of Economics and Statistics 89 (3), 522–533. Elbehri, A., Macdonald, S. (2004), Estimating the impact of transgenic Bt cotton on west and central Africa: a general equilibrium approach. World Development 32 (12), 2049–2064. Evenson, R.E., Herdt, R.W., Hossain, M. (1996), Rice Research in Asia: Progress and Priorities. CAB International, Wallington, UK. Frisvold, G., Reeves, J.M., Tronstad, R. (2006), Bt Cotton adoption in the United States and China: international trade and welfare effects. AgBioForum 9 (2), 69–78.

International Trade and Welfare Effects of Biotechnology Innovations

307

Gruere, G.P., Carter, C.A., Farzin, Y.H. (2009a), Explaining international differences in genetically modified food labeling regulations. Review of International Economics 17 (3), 393–408. Grue`re, G.P., Mevel, S., Boue¨t, A. (2009b), Balancing productivity and trade objectives in a competing environment: should India commercialize GM rice with or without China? Agricultural Economics 40 (4), 459–475. Grue`re, G.P., Rao, S.R. (2007), A review of international labeling policies to evaluate India’s proposed rules. AgBioForum 10 (1), 51–64. Grue`re, G., Sengupta, D. (2009), The effects of GM-free private standards on biosafety policymaking in developing countries. Food Policy 34 (5), 399–406. Grue`re, G.P., Sengupta, D. (2010), Reviewing South Africa’s marketing and trade related policies for genetically modified products. Development Southern Africa 27 (3), 333–352. Hareau, G., Norton, G.W., Mills, B.F., Peterson, E. (2005), Potential benefits of transgenic rice in Asia: a general equilibrium analysis. Quarterly Journal Of International Agriculture 44 (3). Huang, J., Hu, R., van Meijl, H., van Tongeren, F. (2004), Biotechnology boosts to crop productivity in China: trade and welfare implications. Journal of Development Economics 75 (1), 27–54. Jensen, T. J., Jensen, G. J., Gylling, M. (2009). Adoption of GM food crop varieties in the European Union. Paper presented the 13th Annual Conference on Global Economic Analysis, Penang, Malaysia in 2010. Available at https://www.gtap.agecon.purdue.edu/resources/download/ 4765.pdf Kaye-Blake, W.H., Saunders, C.M., Cagatay, S. (2008), Genetic modification technology and producer returns: the impacts of productivity, preferences, and technology uptake. Review of Agricultural Economics 30 (4), 692–710. Langyintuo, A.S., Lowenberg-DeBoer, J. (2006), Potential regional trade implications of adopting Bt Cowpea in west and central Africa. AgBioForum 9 (2), 111–120. Ledford, H. (2007), Out of bounds. Nature 445, 132–133, 01/11/2007. Moschini, G., Lapan, H., Sobolevsky, A. (2000), Roundup Readys Soybeans and welfare effects in the Soybean Complex. Agribusiness-An International Journal 16 (1), 33–55. Nielsen, C., Anderson, K. (2001), Global market effects of alternative European responses to genetically modified organisms. Review of world economics 137 (2), 320–346. Nielsen, C.P., Robinson, S., Thierfelder, K. (2001), Genetic engineering and trade: panacea or dilemma for developing countries. World Development 29 (8), 1307–1324. Nielsen, C., Thierfelder, K., Robinson, S. (2003), Consumer preferences and trade in genetically modified foods. Journal of Policy Modeling 25 (8), 777–794.

308

Guillaume P. Grue`re et al.

Paarlberg, R.L. (2002), The real threat to GM crops in poor countries: consumer and policy resistance in rich countries. Food Policy 27 (3), 247–250. Paarlberg, R.L. (2006), Are genetically modified (GM) crops a commercial risk for Africa? International Journal of Technology and Globalisation 2 (1/2), 81–92. Plastina, A., Giannakas, K. (2007), Market and welfare effects of GMO introduction in small open economies. AgBioForum 10 (2), 104–123. Smale, M., Zambrano, P., Grue`re, G., Falck-Zepeda, J., Matushke, I., Horna, D., Nagarajan, L., Yrramareddy, I., Jones, H. (2009), Measuring the Economic Impacts of Transgenic Crops in Developing Agriculture During the First Decade. IFPRI Food Policy Review 10. International Food Policy Research Institute (IFPRI), Washington, DC. Smyth, S., Kerr, W.A., Davey, K.A. (2006), Closing markets to biotechnology: does it pose an economic risk if markets are globalised? International Journal of Technology and Globalisation 2 (3/4), 377–389. Sobolevsky, A., Moschini, G., Lapan, H. (2005), Genetically modified crops and product differentiation: trade and welfare effects in the Soybean Complex. American Journal of Agricultural Economics 87 (3), 621–644. Stone, S., Matysek, A., Dolling, A. (2002), Modelling possible impacts of GM crops on Australian trade. papers.ssrn.com. Canberra, Australia. Van Meijl, H., van Tongeren, F. (2004), International diffusion of gains from biotechnology and the European Union’s Common Agricultural Policy. Agricultural Economics 31 (2-3), 307–316. Vigani, M., Raimondi, V., Olper, A. (2010), GMO regulations, international trade and the imperialism of standards. LICOS Discussion Paper 255/2009. LICOS Centre for Institutions and Economic Performance, Katholieke Universiteit Leuven, Leuven, Belgium. Retrieved from http://www.econ.kuleuven.be/licos/DP/DP2010/DP255.pdf. Wu, F. (2006), Mycotoxin reduction in Bt corn: potential economic, health and regulatory impacts. Transgenic Research 15 (3), 277–289. You, L., Wood, S. (2006), An entropy approach to spatial disaggregation of agricultural production. Agricultural Systems 90, 329–347. Yudelman, M., Ratta, A., Nygaard, D. (1998), Pest management and food production: looking to the future. 2020 Food Agriculture and the Environment Discussion Paper No. 25, International Food Policy Research Institute, Washington DC.

CHAPTER 13

Global Welfare and Trade-Related Regulations of GM Food: Biosafety, Markets, and Politics Guillaume P. Grue`re Environment and Production Technology Division, International Food Policy Research Institute, Washington, DC 20006-1002, USA E-mail address: [email protected]

Abstract Purpose – The chapter provides a comprehensive review of trade-related regulations of genetically modified (GM) food, identifies their main effects, and analyzes the main motivations behind their support. Methodology/approach – The analysis is substantiated by (a) results from the literature on GM food regulations and (b) comparative statics results from a simplified three-country partial equilibrium welfare and political economic model. Findings – The analysis shows that in a non-GM producing country, traderelated regulations will benefit producers, but not necessarily consumers. Producers’ support is found to be instrumental to push for a ban, for information requirements on shipments, or for mandatory labeling of GM food products. Outside pressure groups will play the role of swing voters in cases where consumers and producers do not agree. Research limitations/implications – The analytical model is based on simplifying assumptions on the groups and market effects of each regulation. Future research is needed to empirically validate some of the main results. Originality/value of the chapter – The goal of the chapter is to inform economic and policy researchers on the effects of GM food trade-related regulations. The chapter provides an updated comprehensive overview of the key trade regulations of GM food. It uses a unique model to derive the main welfare effects of GM food regulations. By comparing the effects of GM food regulations in different types of countries for different pressure groups, the findings provide new insights in this area.

Keywords: Genetically modified food, trade, biosafety policy JEL Classifications: Q18, Q17, D78, Q56 Frontiers of Economics and Globalization Volume 10 ISSN: 1574-8715 DOI: 10.1108/S1574-8715(2011)0000010018

r 2011 by Emerald Group Publishing Limited. All rights reserved

310

Guillaume P. Grue`re

1. Introduction During the last 15 years, genetically modified (GM) crops have been produced, consumed, and traded in an increasing number of countries. Still, only a few GM crops have been commercialized, and even fewer have been produced at a large scale – soybeans, cotton, maize, and canola still represent the almost entirety of the global GM crop area (James, 2010). An increasing number of GM events – unique crop/trait combinations – have been commercialized for each of these crops, but few of the same traits have been used to improve other crops. This specialization coupled with the continued growth in adoption has resulted in an increasing share of these four crops being GM. Today, most of soybeans, half of maize, about half of global cotton production and an increasing share of canola are likely GM.1 Because these four crops are major agricultural commodities, GM products have been largely traded internationally,2 but the presence of diverging importers’ preference and heterogeneous trade-related regulations has resulted in a double segmentation of the international market (Grue`re and Sengupta, 2010). First, geographical differences in national, regional, and international regulations, in association with GM-free private standards, have contributed to the separation of markets for products that contain or may contain GM from their conventional counterparts purely based on non-GM materials (with the purity level depending on each product, regulation, or standard). Second, each GM event has been approved for production or import only in a limited set of countries, whereas many other countries have not adopted any regulations on the use of GM products. This double division (geographic and event-based) has required an increased sophistication of food marketing systems worldwide. In particular, traceability and identity preservation (IP) systems have been developed to separate and track specific GM and/or non-GM products from the field to their end users. Several examples illustrate this complexity. Brazil exports large volume of GM soybeans to China, which does not allow the cultivation of GM soybeans, and therefore segregates its GM imports from non-GM domestic soybeans. South Africa produces GM and non-GM maize, imports GM maize from Argentina but not from the United States, and exports milled GM maize and non-GM maize grains to specific Southern African countries, but not others (Grue`re and 1 For instance, the production of corn and soybeans in GM-adopting countries represented 56% and 84% of the global production in 2005, respectively (own calculations based on FAOSTAT), and probably much more in 2010. 2 GM soybean adopting nations covered over 94% of total soybean export volume and value in 2005, while GM corn adopting nations represented 80% of total volume and 77% of total value of corn the same year (estimates based on UN Comtrade).

Global Welfare and Trade Regulations of GM Food

311

Sengupta, 2010). The United States exports large quantities of GM maize to countries like Mexico and the Philippines, but also non-GM soybean products to Japan and non-GM maize to South Korea. While the marketing system has adapted relatively quickly, it has also faced new challenges. On the seed side, some GM crops approved for particular purposes have moved to other countries or marketing channels without approval, creating trade tensions and regulatory reactions.3 Moreover, several GM varieties that were being tested but not approved for use in any country have been unintentionally found in domestic and foreign markets. For example, two types of unapproved GM rice that had been field tested in China and the United States were detected there and in the European Union (EU) – creating a series of import bans despite limited risks involved (Ledford, 2007). In this setting, trade-related regulations of GM food are key factors in understanding the dynamic of adoption of GM crops and the contribution of these technologies to global welfare. They affect exporters, importers, but can also have spillover effects on the regulatory and adoption decisions of countries that do not produce nor consume GM products (e.g., Paarlberg, 2008). While officially designed for public goals, these regulations may also serve less laudable political goals, and have unwanted market effects (either domestically or internationally). The purpose of this chapter is to present an overview of current and upcoming trade-related regulations of GM food, and to analyze their effects on trade, consumers, and producers in a comprehensive manner. More specifically, we review the effects of import approval authorization, the possible impact of documentation requirements for GM commodities, and the effects of GM food labeling and GM-free private standards. Each type of regulation is analyzed based on available evidence from the literature, a qualitative assessment of key political actors and their role, and an analytical model assessing their main trade and welfare effects and the possible rationale for their adoption. Past literature on GM food and international trade has mostly focused on three issues: (a) the choice of import regulations, (b) the export side or the economic effects of trade-related regulations, and (c) the issue of global competition and GM technology adoption (e.g., Smale et al., 2009). More specifically, most published papers used simulation models to evaluate the potential effects of GM crop adoption in various regions under different regulatory and competition scenarios.4 Other ex-ante analyses focus on the optimal choice of regulations (e.g., Lapan and Moschini, 2007; Plastina and Giannakas, 2007), or on positive evaluations of introducing new trade-related regulations (e.g., Grue`re and Rosegrant, 2008). Fewer studies

3 4

StarLink corn, for example, was intended for animal feed, not for human consumption. See Smale et al. (2009) for a list of recent papers that include developing countries.

312

Guillaume P. Grue`re

provide ex-post analyses, perhaps because of insufficient time series data available. Among those, a few papers analyze the observed trade effects of GM adoption in the presence of existing regulations (Purcell and Kalaitzandonakes, 2004; Smyth et al., 2006; Vigani et al., 2009), while others try to explain the current pattern of regulations (Anderson and Jackson, 2003; Lapan and Moschini, 2004; Graff et al. 2009; Grue`re et al., 2009). This chapter aims to add to the very last category and provide a primer on trade-related regulations of GM food. While referring to the results of other studies, the chapter presents a positive analysis of current and upcoming trade-related regulations. Unlike previous contributions, the goal is to assess trade-related regulations in a comprehensive manner and provide the reader with a political economic outlook of the growing complexity of the GM trade regulatory world. A simplified economic model is used to derive the qualitative price, and welfare and political economic effects of the key identified regulations. In what follows, we define GM food as raw and processed products derived from GM crops and used for food and/or animal feed.5 To our knowledge, there are no trade-related regulations on nonfood or nonfeed products from GM crops (e.g., cotton fibers derived from GM cotton),6 so we do not consider those. The chapter is organized as follows. After defining our model, we first look at import approval, then consider traded shipment requirements, and market regulations and standards in the importing country. We conclude on the current and future challenges of adapting to an increasingly complex trading environment.

2. The basic model The use of a model in this chapter serves the purpose of complementing the argumentation, but also allows explicitly linking the international market and welfare effects resulting from these regulations to the national political motivations behind their adoption. To do so, three modeling layers have to be accounted for; the international market effect of each regulation, the domestic welfare effects in the regulating country, and the political economic implications – derived partly from the welfare effects it may have. The challenge in building such a model is to ensure that it provides a relatively reasonable picture of the effects of those regulations on key political groups in different types of countries (producing GM or not, net importers, or exporters), while providing tractable, consistent, and 5 These products represented an average trade value of $42 billion per year for 2000–2004 (Grue`re, 2006). 6 Cotton is however subject to GM-free private standards like organic or fair trade labeling.

Global Welfare and Trade Regulations of GM Food

313

comparable results. The only way to do that is to subject the model to a number of assumptions that will be presented in the following paragraphs. The selected analytical framework is a three-country partial equilibrium model aimed to represent the present situation for a GM food or feed crop. Let us assume that there are no trade regulations apart from GM product regulations. Because most of the current GM crop producers are also exporters,7 GM exporters are separated from two types of non-adopting importers. Let country A be producing GM and non-GM variants of the product and exporting both to the world – representing North or South American countries. More specifically, A produces two GM events, a past event g1 and a new event g2 of the same product. Based on export market requirements, A can also produce a conventional non-GM version n which can be separated from the GM marketing channel using a costly IP system at an exogenous marketing cost cs.8 Country B produces n but is also a net importer of products from A and has specific regulations for GM crops. This represents a number of OECD countries that have not planted GM crops but largely imported certain types of GM commodities (Japan, South Korea, and most countries in Europe). Country C, representing a number of developing nations, is another importer of products from A without specific regulations for GM or non-GM. The welfare analysis focuses on specific identified groups in each country. In A, we focus on exporting producers that are a priori the most affected by trade regulations. In B, we assess the effects of new regulations on consumers and non-GM producers. To capture consumer heterogeneity, we differentiate three groups of B consumers:9 non-GM consumers, indifferent consumers who disregard the presence of GM or non-GM in their purchasing choices, and switching consumers (following the denomination in Bansal and Grue`re, 2010) who will change their purchasing decision according to what they see on the label. This characterization is a simplification of the proposed double consumer differentiation made by Berwald et al. (2006) and Grue`re et al. (2009). The models developed in these papers include heterogeneous GM aversion and the distinction between distrusting/precautionary consumers who know and do not necessarily trust GM regulations, and trusting/tolerant consumers who base their perceptions and utility on the presence of regulations and labels. Here, 7

See Footnote 2. The most important limitation here is the fact that we do not explicitly link the segregation costs to the market outcome (share of market being covered by GM or non-GM). Lapan and Moschini (2004) provide a comprehensive review of why such link does matter. Here costs of segregation are only embedded into prices. This sets our analysis in a relatively short-term to medium term horizon. 9 While this distinction differs from the usual uniform taste distribution (used in Fulton and Giannakas, 2004 and many others), weighing the three groups allows the inclusion of potential difference in tastes and allows for the asymmetrical response of consumers depending on the labeling policy (see section below). 8

314

Guillaume P. Grue`re

non-GM consumers represent the group of GM averse distrusting/ precautionary consumers, switching consumers represent the GM averse tolerant/trusting consumers, and indifferent consumers represent the indifferent consumers in each of these previous group. By assumption, these groups are exogenous, which sets this analysis in a relatively short- or medium-term horizon (assuming stable preferences). In country C, we evaluate the effects of B regulations on producers and consumers altogether. Regulations in country B are the results of political decisions, which indirectly depend on the welfare effects derived by the main pressure groups. We use the proportional voting model by Persson and Tabellini (2000) framework, as adapted by Grue`re et al. (2009) on GM food labeling as a basis of analysis. Three main groups are represented: agricultural producers AP, consumers/voters CV (which includes the three aforementioned subgroups), and other interested parties OP that will represent both anti-GM groups (including NGOs) and pro-GM groups (including biotech companies).10 While producers and consumers’ positions are directly determined by economic welfare, welfare in the other group depends on the balance of forces on the ground – if the anti-GM forces11 dominate they will push for a non-GM agenda, and conversely. We assume that each group determines his position according to a vote equivalent. Most democratic countries need to submit a new regulation to a parliament or a council, and obtain a majority vote. For instance, if the whole population is against a new measure, but some pressure groups are effectively campaigning in favor of it, the result may be measured in the number of votes at the parliamentary level. The intensity of the message delivered by each group depends on the proportion of members approving it. The adoption decision by policymakers will be made according to the weighted sum of support of each group. In each group l, an individual voting member k prefers regulating (R) to not regulating (NR) if and only if W l ðRÞ  W l ðNRÞ þ skl , where W l ð:Þ is the welfare function associated with the regulation, and skl is an idiosyncratic parameter representing voter k’s preference for or against the regulation. For simplicity, we assume that skl is drawn from a uniform distribution, on a range centered on zero, that is, skl ! U½1=ð2jl Þ; 1=ð2jl Þ with jlW0 representing the homogeneity of each group. In each group, the indifferent member (or swing voter) is defined by the parameter sl so that sl ¼ W 1 ðRÞ-W l ðNRÞ. Any voter k such that sklosl prefers regulating, and conversely any individual voter k such that sklWsl opposes regulating. 10 We could separate the biotech lobbies from the environmental NGOs, but given that they tend to have completely opposite objective, the balance is what ultimately matters. This assumption implies that potential users and voters may count before any other group, but these groups may be able to influence voters and users. 11 Note that in addition with vocal anti-GM NGOs, the chemical industry may also play a role in pushing decisions toward non-GM (Graff et al., 2009).

Global Welfare and Trade Regulations of GM Food

315

We compute the probability of voting or the P proportion of vote equivalents for a particular regulation R as VðRÞ ¼ gl Fðsl Þ, where F(.) is the cumulative density function (c.d.f.) of s and gl is the political weight of group l. Substituting the c.d.f of the uniform distribution, this gives:    X 1 1 X l l l l l l ¼ (1) þ gjs VðRÞ ¼ gj s þ 2jl 2 To obtain a majority of vote for the regulation, the second P l equivalents term has to be positive: g jl sl  0. Thus, three factors affect the outcome: the weight given to each group in the final decision (gl), the degree of homogeneity in each group (jl),12 and the welfare change associated with the new policy for each group (sl). At the end, the policymaker will balance the will of its constituents V(R) with his own assessment of safety-related matters, which is represented by the relative difference in perceived safety DS with the regulation compared to the status quo DS ¼ (S(R)-S(NR))/S(NR). The utility score of the decision maker for a particular regulation is assumed to be the following: U DM ðRÞ ¼ DS þ ð1  DSÞVðRÞ

(2)

Because DS and V(R) are set to be in [0,1], this function is also within [0,1], and it is set up to ensure that perceived safety effects dominate constituents’ preference. This ‘‘total utility score’’ will lead to an adoption decision for any score greater or equal to 1/2. While the weight and concentration parameters depend on the country and group, the welfare effects sl will be computed in the respective groups (producers AP, consumers CV) in each country. For simplification, we assume a linear supply and demand in the three countries with the following functional forms for product i in country j: pji ¼ cji Qji on the supply side and pji ¼ aj Qji þ bj þ Ci ðlÞ on the demand side. Unless specified for particular scenarios, we assume that 8j 2 fA; B; Cg aj o0; bj 40; 0ocg2 ocg1 ocjn . Ci(l) represents the utility shifter for the quality component of the product as perceived by consumers of country B. It depends on the type of consumers and labeling regulations as defined in Table 1 for the three groups of consumers.13 We assume that non-GM consumers will try to avoid GM products and switching consumers will only buy non-GM with mandatory labeling.14 12 A consequence of the simplifying grouping of anti- and pro-GM interests in the OP group is that homogeneity may be low in general, but this could be compensated by the weight of the group in political decisions. 13 This schedule follows the setting of Berwald et al. (2006) and Grue`re et al. (2009) for equivalent groups. 14 Exogenous consumers’ preference is a simplification that needs to be accounted for in the results; changing the share of each group may provide a means to more flexibility. More sophisticated models could also be used to capture changes in variants especially in the long run –(e.g., see Grue`re et al., 2008).

316

Table 1.

Guillaume P. Grue`re

Product consumed and quality function Wk (k) for the three consumer groups in country B

Type of consumers No labelling Voluntary labelling Mandatory labeling GM-free private standards

Non-GM

Switching

Indifferent

Mixed/l Non-GM/l Non-GM/l Non-GM/l

Mixed/0 GM/0 Non-GM/l Non-GM/0

Mixed/0 GM/0 GM/0 Non-GM/0

Prices are obtained by setting excess supply equal to excess demand in the world market for each product. Producer welfare in all countries is defined by profits Pi ¼ pi Qi  Ci ðQi ÞQi , where pi and Qi are the prices and quantities, and C i ðQÞ ¼ ci Qi =2 is the unit cost function for variant i. Consumer welfare is equal to the Marshallian consumer surplus (CS) in countries A and C. In B, we define it as: ! X at CS t  TðRÞ ð1  DSÞ (3) U B ðRÞ ¼ t

where T(R) is the tax imposed with each regulation and at is the share of consumers of group t (non-GM noted NG, switching S, indifferent I). The model is solved by (a) determining the price(s) at equilibrium, (b) computing the welfare effects, (c) assessing the majority of vote equivalents in B, and (d) determining the decision maker’s utility score.

3. Regulating imports: import approval and low level presence The first and most important regulation is the authorization to import a specific GM event in a given country. There are two distinct issues: importing seeds or planting materials for experiment or crop production, and importing GM products or processed products derived thereof. Since we focus on GM commodity trade, we will only treat of the approval for products intended for processing, food, or feed (often noted FFPs). Not all countries have implemented import regulations for GM products, but those that have typically require an application with food or feed safety data from the developer for each new GM product (Grue`re, 2006). The product then goes through an evaluation by a scientific body giving an opinion, but the final approval decision depends on policymakers. Although the exact modalities differ across countries, the food safety data requirements tend to be relatively similar, following OECD and Codex Alimentarius principles, even if the assessment may be more or less strict. Carter and Grue`re (2006) provide a review of the approval processes in Japan, Australia and New Zealand, and the EU. They note

Global Welfare and Trade Regulations of GM Food

317

that the observed differences are mainly at the policy level, where different layers of political representations are involved. While most developed countries have functional import regulatory systems, other countries can be divided into those that have not enforced their regulations, those with no regulations, and those with official bans of GM products (Grue`re, 2006). Very few countries have official bans of GM products, and they are mostly in Southern Africa (Grue`re and Sengupta, 2010). Many developing countries have no regulations, but inted to introduce biosasfety regulations in the future, following guidance under the Cartagena Protocol on Biosafety (CPB) to which they are members. Among others, the CPB allows rejecting imports even in the absence of proven risk, but this bears the risk of being in violation of countries’ obligations under the World Trade Organization (WTO) (Winham, 2003). Even among developed countries, several regulations have been slow to become functional, and still face political challenges. A key example is that of the EU, which applied a moratorium on new GM events from 1998 to 2003. This moratorium was the object of a WTO dispute launched by Canada, Argentina, and the United States in 2003. The 2006 ruling by the WTO Dispute Settlement Body did not provide any qualified opinion on scientific matters, but did find that the EU violated the Sanitary and Phytosanitary (SPS) Agreement (WTO, 2006) and requested the EU members to change their regulations and stop the moratorium. If the EU revised its regulatory system to a certain extent, several EU members still have to comply with this decision.15 The main issue of an importing regulating country is to ensure the safety of incoming products. Therefore, the goal is to reject all unapproved material at the border, regardless of quantities, but this proves to be challenging if not unrealistic in a globalized economy. In principle, a country would prefer to have a 0% tolerance for non-approved GM events. However, because GM products are largely traded commodities in a bulk marketing system, accidentally mixing frequently occurs, and is often unavoidable, resulting in low level presence of different types of GM events in commodity shipments.16 The multiplication of approval of GM events in large exporting countries associated with lengthy approval processes at importers has resulted in an increasing pressure on regulating importers to lift their 0% tolerance. This phenomenon, called asynchronous approval of GM events, has created a number of incidents where the importing countries had to reject large shipments of grains or oilseeds because of minute traces of unapproved GM material at the port (Grue`re, 2009; Stein and Rodrı´ guez-Cerezo, 2009). 15 Since then, Canada has concluded an agreement with the EU to settle the dispute. The United States has also been discussing a similar agreement with the EU (ITCSD, 2009). 16 In fact, the accidental presence of non-approved material is a well-known reality – for instance, the presence of non-grain organic material (animal and plant waste) is commonly accepted in traded shipments at the tolerance level.

318

Guillaume P. Grue`re

As a consequence, importers had to substitute for other non-GM products at a higher cost, sometimes paying a significant premium to avoid traces of unapproved GM events. Economically, while approval regulations will be beneficial to all consumers if they do increase the safety of imported food, they bear two types of costs. First, obtaining approval in each country of import is a fixed cost for biotech developing companies in the exporting country. While these companies had to obtain approval for production in the country, each importer has slightly different testing and documentation requirements that add to the overall cost of marketing a product. Second, if the importing country’s approval decision happens after the commercialization in the exporting country (assuming a 0% tolerance level), the importers face a temporary tariff equivalent, in the sense that they have to procure a substitute for the original mixed GM commodity at a higher price until approval is granted. This tariff equivalent also applies to commodities transported in the same fashion – for example, soybean or wheat shipments can be found to contain trace levels of unapproved maize. This temporary barrier is eliminated when the import authorization is granted, but it can also remain in place indefinitely if it is used as pretext for regulatory reform. South Africa provides a notable example (Grue`re and Sengupta, 2010). In 2003, under pressure by domestic maize producers, the GM import authorization procedure (called commodity clearance) was eliminated, and since then imports have been subject to the regular procedure for a new GM crop planting approval. Thus, no GM product can be imported unless it shows no environmental risk and a significant agronomic advantage for farmers. This decision, which was highly contested by the animal feed industry, has resulted in the blocking of U.S. imports of maize that may contain a GM maize event resistant to corn rootworm, because this GM event does not provide any agronomic advantage in South Africa. The use of zero percent tolerance levels by major importing countries has also created a disincentive to the development and use of publically developed GM crops especially in developing countries. The fact that a minimum trace of unapproved GM event in countries of the EU or Japan can trigger bans or penalties creates new challenges for public biotech research organizations. To avoid this situation, public sector developers would need to obtain approval in all major countries, a significant additional cost they may not always be able to bear. In some cases, these organizations can also face a rejection by domestic authorities if potential export risks are considered in an approval decision. South Africa recently rejected the commercialization of the first publically developed GM potato variety intended solely for small-scale nonexporters, mostly because of the fear of export loss of the commercial potato industry. Past studies have tried to track the effect of GM regulations on trade it to be (Cadot et al., 2003; Disdier and Fontagne´, 2010; Parcell and Kalaitzandonakes, 2004; Smyth et al., 2006). These studies generally did

319

Global Welfare and Trade Regulations of GM Food

not separate the effects by type of regulation, and it is therefore difficult to assess the effect of import approvals per se. Nonetheless, while focusing on transatlantic exchanges, they tend to find that trade was not affected significantly by GM policies in the recent past, even if trade diversion may have occurred because of GM regulations. Still, Disdier and Fontagne´ (2010) do find that the EU policies condemned by the WTO panel had a negative effects on the exports of potentially GM products from Argentina, Canada and the United States. Furthermore, cases where a domestically unapproved GM event did enter the global commodity chain, like StarLink corn, or the more recent Liberty Link rice 601, did have significant market effects (e.g., Carter and Smith, 2007). These cases generated series of bans of commodities in many countries at a huge cost for domestic producers. To assess the potential welfare implications of these regulations, we compare five scenarios, defined with price notations in Table 2. Each scenario represents a step in the approval procedure of new GM events in country A, with different regulatory response in country B. It should be noted that scenarios 0 (base), 2, and 4 (GM events approved in B) present a world market with only one price and one type of good. In contrast, scenarios 1 (import ban) and 3 (B approves g1 not g2) result in product differentiation and consequently GM or non-GM segregation in country A. While this phenomenon is modeled as a simple cost, segregation is not a trivial issue; it can be defined as special type of externality of GM production on conventional production (Lapan and Moschini, 2004) due to the verification costs that arise for GM-free products. If segregation costs dominate the productivity effects, the use of GM can be welfare reducing (Lapan and Moschini, 2004; Moschini et al., 2005). The goal is not to assess the general welfare of GM adoption here but, the interference of regulations and consumer preferences do favor segregation and may therefore affect the outcome under GM adoption. Here, we assume that A will produce non-GM

Table 2.

Scenarios with import approval and price notations

Country Scenario 0: No regulation Scenario 1: B bans g1 Scenario 2: B approves g1 Scenario 3: A approves g2 Scenario 4: g2 approved in B

A

B

C

Exports g1 Price: pW 0 Exports g1 and n Prices: pg and pn Exports g1 Price: pW 2 Exports g1 and g2 Prices: pg1 and pg2 Exports g2 Price: pW 4

Imports g1 Price: pW 0 Imports n at a cost cs Price: pn Imports g1 Price: pW 2 Imports g1 at a cost cs Price: pg1 Imports g2 Price: pW 4

Imports g1 Price: pW 0 Imports g1 Price: pg Imports g1 Price: pW 2 Imports g2 Price: pg2 Imports g2 Price: pW 4

320

Guillaume P. Grue`re

Table 3. Comparison of scenarios in terms of welfare in the three countries Producer welfare A B

C

Due to segregation costs, 1 and 3 less beneficial, 4 is better than 2 and 0. Most likely ranking: 1W3W2Z0W4.

Better off under 0 or 2, compared to 4, 1 is better than 3.

Consumer welfare Best outcome under scenarios 3 or 4. Overall, 4 better than 2, which is better than 0. 3, is the worst market outcome, 1 is best for non-GM consumers. Best market outcome under 3 or 4, but it depends on additional risk with g2.

under scenario 1, which will only happen with a sufficiently small cost of segregation, to reflect current market realities (e.g., the United States or Argentina do export non-GM soybeans to Europe or Japan). If it is prohibitive, the price of non-GM is the autarky price in B. Setting up demands and supplies, we obtain the equilibrium prices (shown in Table A1 in appendix) under each scenario. Comparative statics show that, under our assumptions, the mixed GM/non-GM price under scenarios W W 0 and 2 exceeds the one under scenario 4 ðpW 0 ¼ p2 4p4 Þ, simply because the second GM event is more cost-advantageaous than the first. We also find that the price of event 2 is lower than the price of event 1 when sold alone ðpg2  pg Þ, and the price of event 1 is lower than the price of non-GM under scenario 2 ðpg1  pn Þ. Under most likely conditions,17 the price of event 1 is lower than that of 2 and the price of GM is lower than that of non-GM when sold together in B: pg  pn and pg2  pg1 . The comparisons between the world prices and pg or pg2 are more ambiguous and depend on demand and supply parameters. If the supply effect dominates, the price of GM in scenarios 1 or 3 will be higher than the price without trade restriction (scenarios 0, 2, and 4). If the drop in demand dominates, the price of GM will be lower than world prices with the same technology. Nonetheless, this partial comparison can help rank scenarios for welfare as shown in Table 3. While approval regulations can help guarantee the safety of imported GM food products for consumers (scenario 2 is better than 0), delayed approval processes (scenario 3) can be expensive and detrimental to consumers, particularly those who are not adamantly opposed to GM. Consumers in C may gain from such situation (scenario 3), if the exporter’s product approval has been sufficiently rigorous and the product does not represent additional risks. On the supply side, B producers naturally benefit from stricter regulations on imports, while producers in A may lose if the segmentation under (1 or 3) is costly, and C producers can win or lose depending on structural parameters. 17 Especially – but not only – if the share of non-GM consumers is important, that they are strongly opposed to GM and/or if the cost of segregation is nontrivial.

321

Global Welfare and Trade Regulations of GM Food

P

In vote equivalents, in country B, a sufficient condition for gl jl sl  0 is that all groups benefit from the regulations, that is, sl  0. Producers will have positive welfare gains compared to the status quo (scenario 0) in scenarios 1, 2, and 3. Of these, 1 and 3 may lead to positive or negative welfare effects for consumers,18 while 2 is positive for both consumers and producers. This means that, assuming the nongovernmental group OP is not strictly opposed to an approval regulation,19 V(R)W1/2. Overall, given these results, if the decision maker believes that an import approval procedure increases (or does not decrease) consumer safety (DSW0), he will pass a regulation as U DM ðRÞ  DS þ ð1  DS=2Þ41=2. A ban (scenario 1) may occur if the decision maker is convinced that any GM product is risky (even despite contradicting scientific evidence), if producers have a strong political weight (gAP),20 and/or if non-GM consumers largely dominate the population ðaNG  1Þ with a very high willingness to avoid GM (l), ceteris paribus. A nontariff barrier (scenario 3) can appear if import volumes are low, and if the local producers have a strong representation, a higher voice than consumers (gAPWgCV), and are potentially supported by anti-GM groups. 4. Information requirements on shipments under the Biosafety Protocol The CPB entered into force in September 2003 with the goal of setting up a harmonized framework of risk assessment, risk management, and information sharing on the transboundary movements of living modified organisms (LMOs). Among the key measures of the Protocol, there are specific rules for LMOs intended for direct uses as food, feed, or processing (noted LMOFFPs), which are essentially unprocessed GM commodities (Grue`re and Rosegrant, 2008). These products represent more than half of total import values of the four main GM commodities (Grue`re, 2006).21 In particular, Article 18.2.a of the Protocol requires that each traded shipment of LMO-FFPs be labeled as ‘‘may contain’’ LMO-FFPs not intended for release in the environment, though it also noted that a more 18 In their model with endogenous GM adoption, labeling, and costly segregation, that could represent a more long-run horizon, Lapan and Moschini (2004) further show that consumers in an importing countries can gain from trade restrictive policies because it may result in lower prices for GM-free goods. 19 This is unlikely given that B does not produce GM, that anti-GM will push for stricter regulations and that biotech companies do believe that a seal of approval is useful (Carter and Grue`re, 2006), and that it may help them avoid competition. 20 For instance, in Zimbabwe, where consumers have no voice, domestic producer lobbies have been strong supporters of a GM maize import ban, despite expected disastrous consequences (Gomo, 2010). In Zambia, the president did ban imports in 2002 largely to support domestic producers’ interest (Grue`re and Sengupta, 2009a). 21 Approximately 51% of soybeans and 88% of maize import value comes from unprocessed commodities (Grue`re, 2006).

322

Guillaume P. Grue`re

specific rule on information requirements should be determined at a later date. At a March 2006 meeting in Brazil, Protocol members agreed to adopt a two-option rule consisting of a more stringent option and the less stringent one that had previously been in effect. Under the stringent option, shipments containing LMO-FFPs identified through means such as IP systems would be labeled as ‘‘does contain’’ LMO-FFPs and would include a precise list of all GM events present in each shipment. Shipments containing LMO-FFPs that are not well identified would follow previous practice and would be labeled as ‘‘may contain’’ LMO-FFPs. At the same time, a complete list of GM events commercialized in the exporting country would be available to importers via the Biosafety Clearing House (BCH), an internet database. At the same meeting, Protocol members also agreed that the two-option rule would be reconsidered, with the possibility of making the stringent ‘‘does contain’’ option mandatory for all countries (Grue`re and Rosegrant, 2008).22 While the benefits of this proposed regulatory change are not clear, its implementation would generate significant new costs (e.g., Grue`re and Rosegrant, 2008). More specifically, under the ‘‘does contain’’ rule, countries that export GM would have to test each shipment to verify the accuracy of GM-event identification. Even if the GM-producing countries export a non-GM commodity, they would still have to conduct additional tests in order to make sure the quantity of GM crops in the shipment was lower than the potential threshold levels set up by importers. Importing CPB member countries would also need to pay for the IP system or to conduct tests to confirm the validity of shipment statements in order to ensure enforcement of these requirements. Previous studies have analyzed the likely economic implications of adopting the ‘‘does contain’’ rule in different countries, such as Argentina (Direccion Nacional de Mercados Agroalimentarios, 2004), the United States (Kalaitzandonakes, 2004), or Australia (Foster and Galeano, 2006), reporting that the costs of such change would be potentially significant. Using a multiregion computable general equilibrium model, Huang et al. (2008) show that it would affect the prices of maize and soybeans, increasing world prices overall. While their results show that the cost of implementation would be large globally, but not really significant for China (their focus country), they note that smaller developing countries would likely pay a higher price. Grue`re and Rosegrant (2008) assess the potential implementation costs of Article 18.2.a on all member countries of the Asia Pacific Economic Cooperation and provide a range of cost estimates for exporters and importers, noting the disproportional cost for less developing countries that have been supportive of this measure. They also show that it would

22 During the October 2010 5th Meeting of Parties in Nagoya, Japan, it was decided that the measure would be postponed to the 7th meeting of Parties to leave more time for observation.

Global Welfare and Trade Regulations of GM Food

323

effectively constitute a new entry cost for GM adoption and for Protocol membership in this region. As a follow-up, Bouye¨t et al. (2010) use a spatial equilibrium model for eighty countries in the maize market and show that the measure would increase prices and also result in trade diversion at the detriment of large exporters and importers. Economically, the benefits are difficult to measure, simply because they are not obvious, unless one counts the use a GM event list as justification for a ban, for the benefit of a policymaker’s popularity. As long as there are at least two GM events being exported, the measure will act as a selective tariff that depends on testing requirement and the degree of enforcement by importers. This technical barrier will likely apply to all products, whether GM or not, coming from GM-producing countries, but unlike scenario 3, it will be a lasting and limited tariff, not a temporary potentially prohibitive one. Furthermore, it will apply to many more countries because many developing countries without regulations are CPB members. The proposed regulation is modeled as an additional transport cost for GM and non-GM for A to B and C, assuming that B and C are CPB members. We use an alternative to scenario 4 as the basis of analysis. In scenario 4, A only produces and exports g2 to A and B, but here we assume that g1 is still present in the commodity chain (as a residual presence or due to stocks) when the measure enters into force. All shipments have to be tested to know whether there is g1 or not. Assuming a per unit cost t, we obtain a new equilibrium price pW 5 (see Table A1). Naturally, we find that this price is higher than without the regulation W pW 5 4p4 . For B, it will slightly increase the cost of imports and raise domestic prices, for C it will include new costs for consumers.23 A exporters will all lose. The rent will go to testing companies present in A, B, and C. The situation is slightly different if C is not member: the price of the good may fall in C to the benefit of consumers and at the detriment of producers. The observed support among CPB members for this new regulation is singular, as it is essentially unrelated to safety (Grue`re and Rosegrant, 2008) and therefore may be driven by anti-GM sentiments. Assuming that the decision maker does not consider it a safety-related his decision P matter, l l s . Knowing that function simplifies to U DM ðRÞ ¼ VðRÞ ¼ 1=2 þ gl jP consumers will lose, B’s decision maker will support it gl jl sl  0 only if producers vocally push for it (large gAP jAP sAP ) with support from the more vocal anti-GM lobbies (that tend to dominate the group OP in Protocol meetings24) and minimum opposition from consumers

23 Those results have been verified with an empirical model in the case maize, see Boue¨t et al. (2010). 24 The biotech industry coalition tends to keep a low profile in meetings of the Protocol, with scarce declarations, while NGOs do not hesitate to stand up.

324

Guillaume P. Grue`re

(e.g., assuming t is relatively small or that they are uninformed about this technical measure). At the international level, country C may or may not support the decision, depending on the voice of consumers. A lot of developing countries (especially in Africa) have let their environmental representatives support this measure as an additional restriction to what they see as potentially hazardous products, that need to be labeled as such, without accounting for unaware consumers and the effect it would have on import prices (Boue¨t et al., 2010).

5. Marketing regulations and standards Three types of marketing policies have trade-related implications: traceability, GM labeling, and GM-free private standards. This section will only briefly discuss the EU traceability requirements, to mainly focus on labeling policies and private standards, and their role on affecting demand, prices, and market access. Traceability is an important regulatory requirement for GM traded products, but it is only applied in the EU since 2004. The main differences between traceability and other marketing policies are its focus on safety and the fact that its implementation cost focuses solely on GM products and suppliers. Labeling policies are not designed as safety measures (Grue`re and Rao, 2007). Moreover, in most cases, the cost of labeling will be borne by non-GM providers, because either exporters want to signal their non-GM status (voluntary labeling/private standard) or because they want to avoid a GM label (mandatory labeling) in markets adverse to GM. On the one hand, this GM target is justified by the fact that traceability is designed to help conduct recall in cases of food safety crises. On the other hand, this regulatory requirement adds a new cost of entry to the EU for GM products only, which may be beneficial to domestic non-GM producers that represent a very large majority of producers in the EU. Essentially, traceability can be modeled as a safety measure with a significant fixed cost (setting up a tracking system) plus some variable costs (labor, tests, documentation, archives) that are applied solely to GM imports. While industrialized or large-scale exporters may be able to comply with it without much difficulty, traceability may prevent others to export GM products to the EU, as observed in other markets. In contrast, many countries have adopted GM food labeling policies. The economic implications of GM labeling have been largely discussed in the literature; this section will only focus on labeling and trade. While a large number of labeling policies have been adopted over the world (Grue`re and Rao, 2007), there is a clear dichotomy in labeling approaches between voluntary and mandatory labeling (e.g., Runge and Jackson, 2000).

Global Welfare and Trade Regulations of GM Food

325

Mandatory labeling requires food companies to display the presence of certain GM ingredients (or ingredients derived from GM crops) on food products over a specific threshold level. Each regulation has a different coverage and different exemptions. While developed countries have relatively well-enforced regulations, virtually all developing countries still have to enforce their regulations (Grue`re and Rao, 2007). Unlike other consumer regulations, mandatory labeling directly applies to decisions for food companies, not consumers, and therefore does not guarantee consumer choice (Grue`re et al., 2008). Still, the companies’ decision to use GM ingredients is indirectly related to consumer preference. The cost structure of labeling and its effect on demand by food companies (and ultimately consumers) are primordial in its ultimate market effect. Together with the degree of enforcement, these factors largely predict what the effect of labeling will be in a particular market (Bansal and Grue`re, 2010). Consequently, the trade effect of labeling also depends on these three factors. So far, the evidence has shown that companies have been keen to avoid labeling their products as GM in virtually all countries with enforced mandatory labeling regulation. China is the only confirmed exception where GM labeling resulted in all targeted products being labeled as GM (with a limited demand shift see Lin et al., 2008). While there is evidence that trade considerations play a role in explaining labeling regulations (Grue`re et al., 2009; Vigani et al., 2009), introducing a mandatory labeling policy does not guarantee market access. Having a domestic labeling policy in the exporting country may be reassuring for importers, assuming the labeling is actually enforced. But labeling domestic consumer goods will generally not be sufficient to obtain market access to a labeling country. For instance, introducing GM labeling in India would not help basmati rice exporters to keep their nonGM market in Europe, while it would bear significant costs to the entire rice sector (Bansal and Grue`re, 2010). Regardless of the domestic labeling policy, these exporters will need to ensure that their rice is non-GM. A more convincing explanation relates to trade agreements and political influence; in a bipolar regulatory world, memberships to free trade agreements seem to be increasingly used to support a particular position on GM regulations.25 The case of voluntary labeling is fundamentally different in that it is a bottom-up approach. Companies may or may not decide to label their products as GM or non-GM; nonlabeling is an option. In a perfectly competitive market with no political distortion, voluntary labeling may 25 Certain countries in Asia were reportedly encouraged to adopt mandatory labeling of GM food as a precursor to discussing free trade agreements (FTAs) with the EU (personal conversation with national Codex representatives on labeling, February 2009), while others (like Malaysia or South Korea) had to balance their interest for FTAs with the United States with their labeling regulations (Merrett, 2007; Ahn 2008).

326

Guillaume P. Grue`re

lead to exactly the same outcome as mandatory labeling for consumers (Bansal and Ramaswami, 2010). But its outcome will be more directly linked to consumer demand than mandatory labeling (Grue`re et al., 2008). Currently, voluntary labeling schemes apply to non-GM products in market in the presence of consumers willing to pay to avoid GM products. A large share of the non-GM products available in retail market is in fact organic before being non-GM. Still, non-GM, nonorganic grains are traded internationally. For instance, the Japanese Tokyo Grain Exchange does have a quote for non-GM soybeans that are sold for a significant price premium.26 In certain countries like Japan or South Korea, non-GM claims have long been displayed on numerous products, despite the existence of a mandatory labeling policy, as publicity and potentially to support stricter standards than those defined in the official mandatory labeling policy.27 The growing market presence of non-GM labels in these countries demonstrates the low informative value of mandatory labeling of GM food. Non-GM labels would appear without mandatory labeling in these countries, but it would not prevent other GM products to appear on the retail shelves (Grue`re et al., 2008). Voluntary labeling claims are generally supported by GM-free private standards that are set up by food companies, traders, or retailers (Knight et al., 2008). But while these standards can use labeling as a selling argument, not all GM-free private standards are associated with non-GM labels. A number of large food companies use blanket GM-free private standards (often on all products), without using it as a selling point. For instance, General Mills and Frito-Lay apply non-GM policies without non-GM claims in the United States (Grue`re et al., 2008). McDonalds’ decision to reject GM potatoes has resulted in the abandonment of Bt potato in North America without any non-GM promotional campaign (Grue`re and Sengupta, 2009a). The rationale for such standard may be a combination of brand and reputation insurance at a low cost. While set up privately and therefore not official ‘‘regulations,’’ these standards increasingly decide what product gains market access. Before mandatory labeling was fully enforced in the EU, several retailers companies had already set up GM-free policies (Bernauer, 2003; Grue`re and Sengupta, 2009a). With market concentration, these standards rapidly started to affect the type of GM products sold in countries of Europe, Australia, New Zealand, Japan, or South Korea, in the favor of animal feed.28 There is also evidence that GM-free private standards at importers 26 Grue`re (2009) shows that the average premium for non-GM soybean sold in this market was about 24% in 2008. 27 The same trend is occurring in certain Europe an countries where non-GM claims have recently been allowed. 28 Labeling may have generalized this phenomenon in these markets to filter only products exempted from regulatory requirements.

327

Global Welfare and Trade Regulations of GM Food

Table 4. Country Scenario 6: Voluntary labeling in B Scenario 7: Mandatory labeling in B Scenario 8: GM-free private standards in B

Scenarios with marketing standards A

B

C

Exports g2 and n

aNG consume non-GM

Imports g2

Exports g2 and n

aNGþaS consume non-GM

Imports g2

Exports g2 and n

All consume non-GM

Imports g2

have played a significant role in biosafety policymaking in developing countries (Grue`re and Sengupta, 2009a). The model is used to assess the welfare effects and rationale for these different approaches. Three scenarios are considered, as shown in Table 4. Country B’s demand differs according to the group of consumers and the standards (following Table 1). We assume that mandatory labeling results in a small share of GM products, not a corner solution with only non-GM, but scenario 8 represents a case with a corner solution that represents this specific case. Segregation costs differ according to the specific policy, with voluntary labeling bearing an additional marketing cost K (following Grue`re et al., 2008). Lastly, we assume that A will export non-GM to B in the three scenarios, which will only occur if it is at least as profitable as GM for these producers.29 The derived equilibrium prices are shown in Table A2 in appendix. While the price of GM is always inferior to the price of non-GM, the comparison across scenario is more ambiguous (see Appendix for details). The price of non-GM will be the highest under the private standard policy if the demand shift effect dominates. Voluntary labeling may also generate a higher non-GM price than mandatory labeling if the cost of labeling is large and the proportion of switching consumers is relatively small (two realistic assumptions). The price of GM will be the lowest under the private standard scenario, but the ranking of the two others will depend on the demand and supply parameters. If the demand shift dominates, voluntary labeling will generate a higher GM price than mandatory labeling. Comparing these prices with the case of scenario 4, under a nondifferentiated market, we find that under plausible parameters, prices of non-GM will be higher, while the price of GM may be lower or higher than the unified price. To sum up, we find that the most likely rankings will be PS ML W ML VL PS  pVL pW g and p4  pn  pn  pn . Using these qualitative 4  pg  pg 4 4 rankings, we compare the welfare effects of each policy in Table 5.

29 More flexible models could address this issue; here, we focus on the current market situation with GM/non-GM coexistence in the key exporting countries.

328

Guillaume P. Grue`re

Table 5. Country

Welfare effects under the three scenarios

Producer welfare

Consumer welfare

A

If non-GM is a market niche, scenario 6 is the best outcome and may be better than 4.

B

Unambiguous ranking: 8Z6Z7W4

C

Scenarios 6 and 7 are better than 8 and 4.

Scenario 8 will be better than scenarios 6 and 7. The comparison between 8 and 4 is ambiguous. aNG, aI, and l are critical in the ranking: marketing regulations are better if aNG is sufficiently large; if T(R) is small compared to l, 7 will be preferred. In contrast, if aI is large, the ranking becomes: 4Z6Z7Z8. Same ranking as consumers in A.

Voluntary labeling appears to be potentially welfare-enhancing regulations for producers of A and C, and has a low effect on these countries’ consumers, which may explain why it is rare to see voluntary labeling measures in regulating countries that do not produce GM. In contrast, country B’s producers and consumers (if largely in favor of non-GM, and assuming low tax rates) will prefer mandatory labeling, which will reduce consumer welfare in A and C private Scenario 8 (standards) is the best option for producers in B and potentially for consumers in A and C. Interestingly, the results show that the main gains of mandatory labeling (with or without a corner solution)30 are obtained by the domestic producers in any labeling country that does not produce GM. In effect, whether labeling results in a demand shift toward non-GM or if it only creates additional costs to import GM products (that are in fact mixed GM/non-GM), domestic producers cannot lose and may gain substantial rents. This protectionist effect is unambiguous and may be visible in the push for mandatory labeling in non-GM food producing countries. For example, in India, the push for a strict mandatory labeling policy in India in 2006 was vocally supported by domestic producers of vegetable oils that wanted to limit imports and increase the domestic price of their products at a time of low world prices (Bansal and Grue`re, 2010). At the decision-making level, assuming labeling is considered a nonsafety P related measure (as argued in the EU and elsewhere),31 U DM ðRÞ ¼ 1=2 þ gl jl sl and strict mandatory labeling in the EU and other

30 As noted above, scenario 8 is equivalent to mandatory labeling reaching a corner solution with no GM product sold in B. 31 Grue`re and Rao (2007).

Global Welfare and Trade Regulations of GM Food

329

countries can be explained by a consensus between domestic producers ðsAP 40Þ, market dominating non-GM consumers (that would drive sCV to be large and positive), and anti-GM organizations (large gOP jOP sOP ) that have dominated the public debate and do seem to appear stronger than pro-GM organization.32 In importing countries with less vocal antiGM civil society (sOP close to zero or positive) and/or consumer concerns (small gCV ), domestic producers may be sufficient to the introduction of mandatory labeling, as seen in some transition in developing countries. GM-free standards are in principle driven by consumer aversion to GM, but the results suggest that non-GM consumers may not gain as much as domestic producers from such policy. Given that these standards are set up by food companies rather than governments, we cannot apply the same decision-making model. But it is interesting to see that food companies may raise the cost of their non-GM inputs by setting up these standards. Anti-GM and ‘‘green’’ organizations probably play a key role in pushing companies to go GM-free (especially via their targeted internet campaigns against GM products). Until recently, avoiding GM products in food items was not excessively costly because of the small share of potentially GM ingredients in food products, and the wide availability of non-GM substitutes. But recent developments in the international food markets have demonstrated that prices matter in these decisions. While they were avoiding GM maize, the spike in maize prices in 2007–2008 did result in certain Korean and Japanese maize processing companies to switch back to GM maize. The recent declarations by executives in European food companies, on the potential role of GM in food production, seem to indicate a possible change of direction in the future (Grue`re and Sengupta, 2009b). 6. Conclusion: the challenges of an increasingly complex trading environment This chapter provided a synthesis of current trade-related regulations of GM food, from import authorization to market access and commercialization. A simplified welfare and political economic model was developed to assess the welfare effects of each type of regulation and analyze their possible adoption. The results of the analysis show that in a non-GM producing country, trade-related regulations will benefit producers, but not necessarily consumers. An import authorization that increases safety without raising costs significantly (i.e., with measures to avoid asynchronous approvals) will benefit consumers and producers. The effect of labeling regulations on consumers is ambiguous and depends on the shares of GM averse and of indifferent consumers and on the willingness to pay 32

This is consistent with Grue`re et al. (2009).

330

Guillaume P. Grue`re

of the former to avoid GM products. A developing country with low consumer aversion and/or consumer willingness to pay will be better off with voluntary labeling than mandatory labeling. Lastly, proposed information requirements under the CPB do not provide any market benefit to consumers, regardless of their characteristics. Based on this welfare analysis, we assessed the political support needed for an importing country to introduce each regulation. While consumers may play a role in supporting an import approval process, provided it does not result in a ban, producers’ support is likely to be instrumental to push for a ban, for information requirements on shipments, or for mandatory labeling of GM food products. Outside pressure groups (pro- or anti-GM) will play the role of swing voters in cases where consumers and producers do not agree, such as on mandatory labeling, information requirements, or potentially to support GM-free private standards. While these conclusions are derived from a simplified model, with a relatively inflexible structure reflecting a relatively short-term horizon, they tend to be consistent with the few positive political economic analyses in the literature. Anderson and Jackson (2003) and Graff et al. (2009) found that European protectionism may play a role in their regulations. Lapan and Moschini (2004) showed that trade regulations of GM as in Europe may not only protect consumers but also producers. Grue`re et al. (2009), although focusing mostly on export considerations, connected labeling to trade-related reasons rather than others. Still, more work is needed in this area to remove the limitations of the presented analysis and empirically validate some of the main results. While the model illustrates the fact that trade-related regulations are likely to affect third countries via price transmission effect, it does not capture the extent of their actual spillover effects. These regulations have played a critical role in limiting the market to few GM crops in few countries. In particular, the fear of export losses to countries with GM regulations has been a key constraint in the use of public-driven GM crops particularly in developing countries (Gruere, 2006; Paarlberg, 2008). Market access considerations have also resulted in decisions by biotech companies to shelve new GM crops, like wheat or rice. More generally, export-related considerations are now progressively entering decision making in an increasing number of countries. The multiplication of import rejections due to the presence of unapproved GM events in shipments is pushing exporting countries toward more caution. On the one hand, in view of the huge losses incurred to non-adopting producers in past cases of accidental commingling (e.g., LL601 rice), considering export risk before commercial release appears to be a justified and welcome decision. On the other hand, the management of market risk needs to be done rationally in a case-by-case basis. Precautionary measures based on hypothetical or unproven risks are bound to be detrimental to consumers and producers (e.g. potatoes in South Africa).

331

Global Welfare and Trade Regulations of GM Food

There has also been a gradual shift in power between exporters and importers in the last few years. For most of the last decade, exporters had to adapt to market requirements on GM food and had no word to say. But the increasingly rapid pipeline of new GM products has made regulations challenging for importers, especially those with zero percent tolerance levels. In these countries, buyers are starting to call for changes of regulation, toward more practical and less precautionary measures, something that had never happened before. Ultimately, the future global welfare effects of GM crops will depend on the evolution of trade-related regulations. The most important challenge of regulators will remain to ensure that new GM food products are safe for consumers in any importing country. The second challenge will be to manage export risks associated with new GM events in an increasingly complex international regulatory system. Harmonization in risk assessment and management procedures would greatly facilitate tackling these two challenges. On the other hand, the use of non-safety related information regulations, including labeling, will likely continue to be the object of trade tensions. Whether countries can address these three issues successfully will depend on their capacity to integrate biosafety, markets, and political considerations in a constructive manner. Appendix

Table A1. Scenario

Equilibrium prices under the five scenarios without marketing regulations Countries A and C

Country B A

0

C

A

C

B

B

B

b =a b =a b =a þaNG l=a pW 0 ¼ 1=cg þ1=cB þ1=cC 1=aA 1=aB 1=aC n

1

1 pg ¼

n

bA =aA bC =aC A C 1=cg1 þ1=cC n 1=a 1=a

2

pn ¼ A

B

A

B

C

C

bB =aB aNG l=aB þcS =cA n B B 1=cA n þ1=cn 1=a

B

b =a b =a b =a þaNG l=a pW 2 ¼ 1=cg þ1=cB þ1=cC 1=aA 1=aB 1=aC n

1

n

3 A

A

C

C

=a b =a pg2 ¼ 1=cg b þ1=cC 1=aA 1=aC 2

4

pg1 ¼

n

A

C

A

C

B

B

bB =aB þaNG l=aB þcS =cg1 1=cg1 þ1=cBn 1=aB

B

b =a b =a b =a þaNG l=a pW 4 ¼ 1=cg þ1=cB þ1=cC 1=aA 1=aB 1=aC 2

5 pW 5 ¼

n

n

bA =aA bC =aC þbB =aB þaNG l=aB þt=cg2 A B C 1=cg2 þ1=cBn þ1=cC n 1=a 1=a 1=a

332

Guillaume P. Grue`re

Table A2.

Equilibrium prices in the three scenarios with marketing regulations

Scenario 6. Voluntary labeling

GM (sold in A, B, and C) A

B

C

B

B

B

A

aNG ðb þlÞ=a þcs þK=cn b =a b =a ðaI þaS Þ b =a VL pVL g ¼ 1=cg þ1=cC 1=aA aI þaS =aB 1=aC pn ¼ 1=cA þ1=cB aNG =aB n

2

7. Mandatory labeling

C

A

Non-GM (sold in B)

A

n

C

A

C

E

B

=a b =a aI ðb =a Þ ¼ 1=cb pML C A B C g g þ1=c 1=a aI =a 1=a 2

n

bA =aA bC =aC 8. GM-free private standard pPS ¼ g 1=cg þ1=cC 1=aA 1=aC 2

n

n

pML ¼ n

ðaNG þaS ÞðbB þlÞ=aB þcs =cA n B B 1=cA n þ1=cn ðaNG þaS Þ=a

pPS n ¼

bB =aB aNG l=aB þðcs þKÞ=cA n B B 1=cA n þ1=cn 1=a

Comparison of prices used in the text.  The price of non-GM will be the highest under the private standard policy if the demand shift effect dominates:     1 l 1 1 B VL PS pn  pn 2aNG A ðcs þ KÞ  B  ðb  lÞ A þ B cn a cn cn

 Voluntary labeling generates a higher non-GM price than mandatory labeling if the cost of labeling is large and the proportion of switching consumers relatively small:    aS B ðaNG ðbB þ lÞ=aB Þ þ ððcs þ KÞ=cA K ML VL nÞ  A pn  pn 2 B b   A B B a ð1=cn Þ þ ð1=cn Þ þ ðaNG =a Þ cn

 Voluntary labeling will generate a higher GM price than mandatory labeling if the demand shift dominates:     1 bA 1 bC 1 1 ML VL 1  þ C p g  pg 2 A B  1  C B a b a c c b g2 n

References Ahn, C. (2008), Food safety on the butcher’s block. Foreign Policy in Focus. April 18, 2008. Available at http://www.fpif.org/articles/ food_safety_on_the_butchers_block. Anderson, K., Jackson, L.A. (2003), Why are US and EU policies toward GMOs so different?. AgBioForum 6 (3), 95–100, Available at http:// www.agbioforum.org/v6n3/v6n3a02-jackson.pdf.

Global Welfare and Trade Regulations of GM Food

333

Bansal, S., Grue`re, G. (2010), Labeling genetically modified food in India: economic consequences in four marketing channels. IFPRI Discussion Paper No. 946. International Food Policy Research Institute, Washington, DC. Available at http://www.ifpri.org/sites/default/files/ publications/ifpridp00946.pdf. Bansal, S., Ramaswami, B. (2010), Label for GM foods: what can they do? Economic and Political Weekly 45 (26&27), 167–173, Available at http:// epw.in/epw/uploads/articles/14919.pdf. Bernauer, T. (2003), Genes, Trade and Regulation. Princeton University Press, Princeton, NJ. Berwald, D., Carter, C.A., Grue`re, G.P. (2006), Rejecting new technology: the case of genetically modified wheat. American Journal of Agricultural Economics 88 (2), 432–447. Boue¨t, A., Grue`re, G., Leroy, L. (2010), From ‘‘May Contain’’ to ‘‘Does Contain’’: The price and trade effects of strict information requirements for GM maize under the Cartagena Protocol on Biosafety. Selected paper presented at the 2010 meeting of the Agricultural & Applied Economics Association in Denver, CO. Available at http://ageconsearch.umn.edu/bitstream/61533/2/AAEApaperv3.pdf. Cadot, O., Suwa-Eisenman, S., Trac- a, D. (2003), OGM et relations commerciales transatlantiques. Cahiers d’e´conomie et sociologie rurales 68–69, 104–148. Carter, C.A., Grue`re, G.P. (2006), International approval and labeling regulations of genetically modified food in major trading countries. In: Just, R.E., Zilberman, D., Alston, J. (Eds.), Regulating Agricultural Biotechnology: Economics and Policy. Springer, New York. Carter, C.A., Smith, A. (2007), Estimating the market effect of a food scare: the case of genetically modified StarLink Corn. Review of Economics and Statistics 89 (3), 522–533. Direccion Nacional de Mercados Agroalimentarios. (2004), Contexto y Opciones Para La Exportacion Segregada de Maiz y Soja OVM y No OVM en Condiciones de Bioseguridad, Conforme al Protocolo de Cartagena. Proyecto FAO/SAGPYA TCP/ARG 2903. UN Food and Agriculture Organization, Rome. Disdier, A.-C., Fontagne´, L. (2010), Trade impact of European measures on GMOs condemned by the WTO panel. Review of World Economics 146 (3), 495–514. Foster, M., Galeano, D. (2006), Biosafety Protocol: implications of the documentation regime. eReport 06.2. Australian Bureau of Agricultural and Resource Economics (ABARE), Canberra, Australia. Fulton, M., Giannakas, K. (2004), Inserting GM products into the food chain: the market and welfare effects of different labeling and regulatory regimes. American Journal of Agricultural Economics 86, 42–60. Gomo, T. (2010), Minister motivated by self-interest. The Zimbabwe Times, January 31. Available at http://www.thezimbabwetimes.com/ ?p ¼27115.

334

Guillaume P. Grue`re

Graff, G.D., Hochman, G., Zilberman, D. (2009), The political economy of agricultural biotechnology policies. AgBioForum 12 (1), 34–46, Available at http://www.agbioforum.org/v12n1/v12n1a04-graff.htm. Grue`re, G.P. (2006), An analysis of trade related regulations of genetically modified food and their effects on developing countries. EPT Discussion Paper 147. International Food Policy Research Institute, Washington, DC. Available at http://www.ifpri.org/sites/default/files/ publications/eptdp147.pdf. Grue`re, G.P. (2009), Asynchronous approvals of GM products, price inflation, and the codex annex: what low level presence policy for APEC countries? Selected paper presented at the 2009 symposium of the International Agricultural Trade Research Consortium. Available at http://iatrc.software.umn.edu/activities/symposia/2009Seattle/seattle-Gruere.pdf. Grue`re, G.P., Carter, C.A., Farzin, Y.H. (2008), What labelling policy for consumer choice? The case of genetically modified food in Canada and Europe. Canadian Journal of Economics 41 (4), 1472–1497. Grue`re, G.P., Carter, C.A., Farzin, Y.H. (2009), Explaining international differences in genetically modified food labeling regulations. Review of International Economics 17 (3), 393–408. Grue`re, G.P., Rao, S.R. (2007), A review of international labeling policies of genetically modified food to evaluate India’s proposed rule. AgBioForum 10 (1), 51–64. Available at http://www.agbioforum.org/ v10n1/v10n1a06-gruere.htm. Grue`re, G.P., Rosegrant, M.W. (2008), Assessing the implementation effects of the Biosafety Protocol’s stringent information requirements in countries of the Asia Pacific Economic Cooperation. Review of Agricultural Economics 30 (2), 214–232. Grue`re, G., Sengupta, D. (2009a), The effects of GM-free private standards on biosafety policymaking in developing countries. Food Policy 34 (5), 399–406. Grue`re, G., Sengupta, D. (2009b). Biosafety and perceived commercial risks: the role of GM free private standards. IFPRI Discussion Paper 00847. International Food Policy Research Institute, Washington, DC. Available at http://www.ifpri.org/sites/default/files/publications/ ifpridp00847.pdf. Grue`re, G.P., Sengupta, D. (2010), An analysis of South Africa’s marketing and trade related policies for genetically modified products. Development Southern Africa 27 (3), 333–352. Huang, J., Zhang, D., Yang, J., Rozelle, S., Kalaitzandonakes, N. (2008), Will the Biosafety Protocol hinder or protect the developing world: learning from China’s experience. Food Policy 33 (1), 1–12. International Center for Trade and Sustainable Development (ITCSD). (2009), Canada and EU resolve trade dispute on GMOs. Bridges Weekly Trade News 13 (27). Available at http://ictsd.org/i/news/ bridgesweekly/51287/.

Global Welfare and Trade Regulations of GM Food

335

James, C. (2010), Global status of commercialized biotech/GM crops: 2009. ISAAA Brief No. 41-2009. International Service for the Acquisition of Agri-biotech Applications, Ithaqa, NY. Kalaitzandonakes, N. (2004), The potential impacts of the Biosafety Protocol on agricultural commodity trade. IPC Technology Issue Brief. International Food & Agricultural Trade Policy Council, Washington, DC. Available at http://www.agritrade.org/Publications/IBs/Techy/ BSP.pdf. Knight, J.G., Holdsworth, D., Mather, D.W. (2008), GM food and neophobia: connecting with the gate keepers of consumer choice. Journal of the Science of Food and Agriculture 88, 739–744. Lapan, H.E., Moschini, G. (2004), Innovation and trade with endogenous market failure: the case of genetically modified products. American Journal of Agricultural Economics 86 (3), 634–648. Lapan, H.E., Moschini, G. (2007), Grading, minimum quality standards, and the labeling of genetically modified products. American Journal of Agricultural Economics 89 (3), 769–783. Ledford, H. (2007), Out of bounds. Nature 445, 132–133, Available at www.nature.com/news/2007/070108/full/445132a.html. Lin, W., Tuan, F., Dai, Y., Zhong, F., Chen, X. (2008), Does biotech labeling affect consumers’ purchasing decisions? A case study of vegetable oils in Nanjing, China. AgBioForum 11 (2), 123–133, Available at http://www.agbioforum.org/v11n2/v11n2a06-lin.htm. Merrett, N. (2007), Malaysian trade talks stall on GM labelling. Foodnavigator.Com April 24. Available at http://www.foodnavigator.com/news/ ng.asp?n ¼ 76002-gm-malaysia-labelling. Moschini, G., Bulut, H., Cembalo, L. (2005), On the segregation of genetically modified, conventional and organic products in European agriculture: a multi-market equilibrium analysis. Journal of Agricultural Economics 56 (3), 347–372. Paarlberg, R. (2008), Starved for Science. How Biotechnology is Being Kept Out of Africa. Harvard University Press, Cambridge, MA. Parcell, J., Kalaitzandonakes, N. (2004), Do agricultural commodity prices respond to bans against bioengineered crops?. Canadian Journal of Agricultural Economics 52 (2), 201–209. Persson, T., Tabellini, G. (2000), Political Economics. Explaining Economic Policy. MIT Press, Cambridge, MA. Plastina, A., Giannakas, K. (2007), Market and welfare effects of GMO introduction in small open economies. AgBioForum 10 (2), 104–123, Available at http://www.agbioforum.org/v10n2/v10n2a05-giannakas.htm. Runge, C.F., Jackson, L.A. (2000), Labeling, trade and genetically modified organisms: a proposed solution. Journal of World Trade 33 (6), 111–122. Smale, M., Zambrano, P., Grue`re, G., Falck-Zepeda, J., Matushke, I., Horna, D., Nagarajan, L., Yrramareddy, I., Jones, H. (2009),

336

Guillaume P. Grue`re

Measuring the economic impacts of transgenic crops in developing agriculture during the first decade. IFPRI Food Policy Review 10. International Food Policy Research Institute, Washington, DC. Doi: 10.2499/0896295117FPRev10. Smyth, S., Kerr, W.A., Davey, K.A. (2006), Closing markets to biotechnology: does it pose an economic risk if markets are globalised? International Journal of Technology and Globalisation 2 (3/4), 377–389. Stein, A.J., Rodrı´ guez-Cerezo, E. (2009), The global pipeline of new GM crops: implications of asynchronous approval for international trade. JRC Technical Report EUR 23486 EN, European Communities, Luxemburg. Available at http://ipts.jrc.ec.europa.eu/publications/pub.cfm?id ¼ 2420. Vigani, M., Raimondi, V., Olper, A. (2009), GMO regulations, international trade and the imperialism of standards. LICOS Discussion Paper 255/2009. LICOS Centre for Institutions and Economic Performance, Katholieke Universiteit Leuven, Leuven, Belgium. Available at http:// www.econ.kuleuven.be/licos/DP/DP2010/DP255.pdf. Winham, G.R. (2003), International regime conflict in trade and environment: the Biosafety Protocol and the WTO. World Trade Review 2 (2), 131–155. World Trade Organization (WTO). (2006), European communities – measures affecting the approval and marketing of biotech products, Reports of the Panel, WTO Docs WT/DS291/R, WT/DS292/R, WT/ DS293/R, September 29. Available at http://www.wto.org/english/ news_e/news06_e/291r_e.htm.

CHAPTER 14

Innovation, Risk, Precaution, and the Regulation of GM Crops Alan RandallT Agricultural and Resource Economics, The University of Sydney, A20, Sydney 2006, NSW, Australia; Agricultural, Environmental, and Development Economics, The Ohio State University, Columbus, OH 43210-1067, USA E-mail address: [email protected]

Abstract Purpose – New genetically modified (GM) crops are novel but risky interventions, offering a variety of potential benefits but also the possibility of serious unintended consequences. I address the regulatory framework for GM crops, seeking protection from disproportionate risks without unduly stifling innovation. Approach – Conditions that may justify precautionary interventions are identified, and an idealized regulatory protocol (screening, pre-release testing, and post-release surveillance, STS) is developed to provide protection, encourage research and learning, and focus-in quickly on the cases that pose serious threats of harm. This protocol is adapted to the case of GM crops, and compared with current regulatory practice in the United States, the EU, and Canada, as well as international agreements exemplified by the Cartagena Protocol on Biosafety. Two real-world cases are considered, Starlinks corn and Roundup-Readys canola, and some speculations are offered as to how the stylized protocol might have handled them. Findings for policy – Pre-release, US regulatory practice is more fragmented and incomplete than the stylized protocol; EU practice is more systematic T Alan Randall is a professor of agricultural and resource economics at the University of Sydney and a professor emeritus of agricultural, environmental, and development economics at The Ohio State University, Columbus, OH 43210-1067. This chapter is part of a larger project that includes a review and commentary on the precautionary principle literature (Randall, 2009) and a book that explores the role of precaution in an integrated risk management framework (Randall, 2011). A complementary relationship has developed among the three works. In this chapter, the discussion of the precautionary principle as it applies to proposed innovations is greatly informed by the aforementioned review article and book. The stylized case study of proposed GM crop innovations (Section 5) was developed with this chapter in mind, and was subsequently adapted for use also in the book. Helpful suggestions from the editors of this volume are much appreciated.

Frontiers of Economics and Globalization Volume 10 ISSN: 1574-8715 DOI: 10.1108/S1574-8715(2011)0000010019

r 2011 by Emerald Group Publishing Limited. All rights reserved

338

Alan Randall

and streamlined, but some critics perceive over-regulation; and Canadian regulatory practice is more consistent with the protocol. Only the EU performs systematic post-release surveillance. International agreements have various weaknesses, beginning with fragmentation: for example, food safety and biosafety are regulated separately. Implications for further research – Embracing the STS framework opens a broad new avenue of research about to how the mix of pre-release testing and post-release surveillance might be streamlined to provide adequate protection while reducing further the costs and delays entailed. Keywords: Genetically modified crops, risk assessment, risk management, precaution, regulation JEL Classifications: Qi, Q2, D8

1. Introduction Genetically modified (GM) crops offer a variety of potential benefits, including higher yields, resistance to diseases and pests, specific characteristics desirable to consumers, and characteristics that reduce production costs. However, decisions about introducing new GM varieties must be made in the context of possible unintended consequences including threats to food safety and human health, potential gene flow and other threats to biodiversity, and development of resistance in targeted weeds, pests, and diseases. This chapter considers new GM crop varieties as cases in the broader category of novel but risky interventions. The regulatory options often are characterized in rather extreme form – we could release the new variety with minimal prior scrutiny (rolling the dice, so to speak) and address any consequent harm later; or we could impose severe ex ante restrictions on release in the event that threats of plausible harm could not be dismissed. Neither strategy seems ideal. There have been more than a few innovations that turned out very badly. On the contrary, an overly cautious approach to novel interventions in general could be very costly in terms of innovation benefits foregone. But perhaps we do not have to choose between these extremes. The overarching objective of this chapter is to explore the possibility of protocols that provide substantial protection from potential harm without stifling innovation in general. Following a brief summary of the argument for precaution in the case of proposed innovations, I specify the form of a coherent PP framework that explicitly relates threat, evidence, and remedy. Proposed novel interventions provide opportunities for pre-screening and iterative pre-release testing with sequential decision points, and I explore these possibilities and their potential for reconciling precaution and innovation. A stylized

Innovation, Risk, Precaution, and the Regulation of GM Crops

339

regulatory protocol for GM crops is presented in some detail, showing the potential roles of pre-screening, pre-release testing, and post-release surveillance. As it turns out, US regulatory practice incorporates some (but not nearly all) of the pre-release steps identified in the stylized protocol, but leaves post-release surveillance to relatively informal processes. EU practice is more systematic and streamlined, but some critics perceive over-regulation. Perhaps Canadian pre-release regulatory practice most nearly approaches the spirit of the suggested protocol, but post-release surveillance is informal. International agreements exhibit several weaknesses, starting with fragmented approaches (e.g., food safety and biodiversity are addressed by different agreements) rather than an integrated framework. To illustrate the workings of the stylized protocol I consider two real-world cases, Starlinks corn and Roundup-Readys canola, speculating as to how the protocol might have handled them. I conclude with further discussion of pre-release testing, and how the mix of pre-release testing and post-release surveillance might be improved to provide adequate precaution while reducing the costs and delays entailed. To the extent that the argument against precaution tends to focus on costs and delays (rather than objections to precaution per se), it makes sense to attend seriously in bringing those costs and delays within acceptable bounds.

2. Innovation, risk, precaution – framing the issue Faced with extraordinary risk, uncertainty, and/or gross ignorance about future consequences, we can stay the course and attend later to harmful outcomes in the cases where they occur, or we can try to get ahead of the game by taking precautions before the potential for harm has been established. We can think of downsides to both approaches. When things go badly, the costs of damage, clean-up, and remediation can be enormous, as we see in the cases of asbestos and polychlorinated biphenyls (PCBs), to give just two examples, which were widely dispersed throughout the economy and environment before harm was determined. Invasive exotic species provide another example, in this case one in which agriculture has often been implicated. If we prohibited every innovation that introduced a plausible threat of harm, it seems likely that we would bear a substantial cost in terms of beneficial innovations foregone. Cast in these terms, the dilemma is clear – a continuing sequence of occasional disasters on the one hand versus systematic repression of innovation on the other. But perhaps the choice does not have to be so grim. If precaution can be focused on the cases that present extraordinary risk, and can be implemented iteratively to encourage learning and reassessment and to provide sequential decision points, then precaution would be less intrusive and costly while still providing substantial protection from harm. The word ‘‘precaution’’ signals the two key concepts involved: caution suggests a desire to avoid harm, and especially catastrophic harm; and ‘‘pre’’

340

Alan Randall

means before. A universal feature of precaution (as opposed to caution) is that measures are invoked before we are really sure about the potential for harm. However, the case of proposed innovations offers an additional kind of before – measures can be invoked before the innovation is released irrevocably.1 With proposed innovations, there is opportunity for prescreening and testing in ways that limit public risk (in vitro, in confinement, with animal subjects before testing with human subjects, etc.) before decisions are made about release. These possibilities expand the scope of inquiry into precaution and the precautionary principle. In addition to the standard question – what (if any) conditions might justify precaution (Hughes, 2006) – a set of questions emerges about ways to implement precaution so as to minimize its cost, intrusiveness, and potentially chilling effect on innovation. 2.1. The precautionary principle Raffensperger and Tichner (1999) attribute the origin of the current precautionary movement to Germany’s emerging environmental movement of the 1970s, which gave a prominent role to the ‘‘Vorsorgeprinzip’’ (the direct translation is foresight principle). Beginning in the 1980s, international conferences, agreements, and treaties endorsed precautionary measures (the Montreal Protocol on ozone-depleting substances, 1987; and the Framework Convention on Climate Change, 1992), the precautionary approach (the Rio Declaration on Environment and Development, 1992; the Cartagena Protocol on Biosafety, 2000; and the Stockholm Convention on Persistent Organic Pollutants. 2001), and the precautionary principle (the Third North Sea Conference, 1990; the [Maastricht] Treaty on European Union, 1992; and the UNESCO World Commission on the Ethics of Scientific Knowledge and Technology, 2005).2 There are many definitions of the precautionary principle, PP, in the literature (Cooney, 2004), but most of them can be grouped into three broad categories, on a weaker–stronger scale. The threads common to all three categories are the focus on uncertain consequences and precaution, 1 Colloquial appeals to precaution include ‘‘First, do no harm’’ and ‘‘Be careful not to drive over the cliff,’’ which capture the two kinds of cases where precaution might be thought to apply. The first invokes the idea of active intervention that might go awry (as, for example, when an innovation is released with unintended adverse consequences), whereas the second warns that unpredictable and disastrous systemic change may be induced by overstressing familiar systems (e.g., by exploitation and/or pollution) in the course of business as usual. The second case is not addressed further in this chapter. 2 Some commentators argue that choice of noun matters (Peel, 2004) – the precautionary approach signals more flexibility than the precautionary principle, especially in regard to social, economic, and political caveats. Furthermore, US negotiators, perhaps fearful of protectionism in international trade as well as excessive litigation at home, have insisted on precautionary ‘‘measures’’ or ‘‘approaches’’, rather than ‘‘principle’’ in multilateral environmental agreements (Shaw and Schwartz, 2005).

Innovation, Risk, Precaution, and the Regulation of GM Crops

341

which together imply caution to forestall uncertain future harm. Below are examples of each category, abbreviated, and paraphrased in every case, with emphasis added to highlight key differences:  Uncertainty about harmful consequences does not justify failure to take precautionary action (Bergen Declaration, 1990).  Plausible but uncertain harm justifies precautionary intervention (UNESCO, 2005).  Uncertain harm requires intervention, and the burden of proof is assigned to the proponent of the proposed risky action (Wingspread Statement, 1998).3 Later in this section, I address several critiques objecting to the persistence of multiple PP formulations. In Section 3, I suggest a PP framework that focuses precaution on disproportionate threats, and relates the stringency of the remedy to the magnitude of the threat and the strength of the evidence. 2.2. Applications to proposed innovations The PP, while it remains controversial, has been adopted by governments, multinational organizations, and various global conventions. It has potential application to medicine, pharmacy, workplace safety, and public health and safety. Government entities including the European Union and Canada have committed to the PP as a guiding principle (European Commission, 2000; Canadian Perspective y , 2001). The United States has been more circumspect about the PP (Wiener and Rogers, 2002). Pharmaceutical products are tightly regulated by the Food and Drug Administration, which requires evidence of safety and effectiveness before approving drugs for general release. However, the United States does not deny patentability for potentially harmful medical technologies, as some other countries have done (Kolitch, 2006). A precautionary approach applies also to approval of pesticides, which are valued precisely because of their toxicity to target species. For most other aspects of protecting the environment and public health and safety, the US approach typically is to wait until there is evidence of damage and then set a regulatory standard.4 Frequently the standard provides a margin of safety – which suggests an element of caution, but not precaution. 3 Note that the Wingspread Statement, while widely circulated and influential, has no official status, being the product of a working group of prominent environmentalists, mostly from nongovernment organizations. 4 Jasanoff (2000), Garrity (2004) and others have argued that US laws and regulations permit a stronger precautionary stance (e.g., routine pre-release testing of whole categories of newly developed substances). However, there seems to be some regulatory reluctance to implement the precautions that are permitted.

342

Alan Randall

2.3. PP controversies The PP is controversial and has attracted a considerable variety of critiques. In order to hone in on those that are most relevant to our inquiry, we need to apply some filters. Critiques addressed here have survived two filters: they are taken seriously in the scholarly literature, and they are directly relevant to precaution in the context of innovation in general and the GM foods case in particular. 2.3.1. The legalistic critique PP is ill-defined, confusing, and applied inconsistently. Sandin (1999) has observed that there are about 20 different formulations of the PP in circulation, so how (he asks) can it be taken seriously in real-world policy? Lofstedt et al. (2002) note that the PP conflicts with other important values that good law and policy would surely respect, and it fails to provide clear instructions for resolving these conflicts. Similarly, discussion of the PP in the context of international trade seems mostly to proceed in legalistic terms, seeking enforceable interpretations, and exhibiting discomfort with competing definitions and inconsistencies, either internal or with other widely honored values (Peterson, 2006). M. Peterson (2007) swung for the home run, claiming that the PP lacks normative content. Specifically, it cannot guide action and decision because it is not structured and used as a stand-alone rule that tells us what to do and what not to do for each possible input of information. Each of these critiques in its own way demands too much. The PP is positioned as a principle, so we must ask what reasonably can be expected of a principle. A principle captures a serious moral intuition, but enunciating a principle neither claims nor establishes its lexical priority over other principles. In complicated exercises in law and policy, important principles will come into conflict and resolution requires a weighing of the principles involved, the values at stake, and the facts of the case. The bottom line is that the critics are asking too much of PP as a principle. Nevertheless, having enunciated PP as a principle, there remains serious work to do, designing decision frameworks that honor the PP within the context of competing, complementary, and conflicting principles and values. 2.3.2. The scholastic critique PP is meaningless, self-defeating, and logically incoherent. Sunstein (2005) argued that the PP is meaningless and self-defeating: it would forbid any and all risky alternatives, yet even the no-action alternative involves risk. Viewed charitably as well-meant hyperbole, Sunstein’s complaint highlights the need to take risk–risk trade-offs seriously and suggests that a serious PP should be specified carefully so as to direct it toward threats that are in some sense extraordinary.

Innovation, Risk, Precaution, and the Regulation of GM Crops

343

Both Sunstein and M. Peterson (2006) attacked the PP by contrasting it to the standard economic-rationality approach to decision-making and risk management, labeled ordinary risk management, ORM, by Randall (2009).5 Peterson proposed a proof that the PP is logically incoherent. What he actually showed is that a decision-maker applying rationality axioms and facing ideal conditions familiar to economists would choose ORM rather than a precautionary restraint. This is reasoning of impregnable circularity, since ORM is exactly what one gets when one derives optimal decisions from economic rationality assumptions and ideal conditions. The take-home lesson is that the scope for a meaningful PP is likely to be concentrated in the class of cases where conditions ideal for ORM are absent, e.g. unlikely but potentially catastrophic threats, and cases where the outcome set is poorly specified or unbounded.

2.3.3. The ‘‘scaredy-cat’’ critique PP takes irrational fears seriously. Sunstein (2005) offers several variants of this critique, including some we have seen already: PP is an injunction to avoid all risk; PP empowers the ignorant and fearful, whereas risk management decisions would be better left to professionals; and PP can be invoked by unfounded panic. Basically, PP stifles innovation and venturesome spirit,6 and it does so because it elevates the influence of the fearful and ignorant over that of rational risk managers. We have addressed the first part of this complex proposition already: it makes sense to focus the PP on threats that are extraordinary. There is legitimate controversy about the second part. Do ordinary people really get risk so badly wrong (Willis, 2001)? Do we miss some important subtleties in what ordinary people tell us about controversial issues (Kahan et al., 2006)? The bottom line seems to be that serious PP proposals must pay attention to the relationships between scientific evidence and precautionary intervention, and to design and implementation of procedures for informing ordinary people and consulting them in decision-making.7

5 ORM is committed to a weighing of the costs and benefits of threat avoidance or mitigation. Although ORM sanctions a number of approaches in assessing benefits and costs, weighing is at the heart of them all. In this context, expected utility approaches open the door for explicit risk aversion; the extended benefit cost analysis approaches expand the menu of ways to weigh, by including option value and passive use value; and the real options approach brings environmental risk under a general theory of asset valuation under risk. 6 There is a variant of this argument directed explicitly to fears that science would be stifled (Harris and Holm, 1999, 2002; Foster et al., 2000). 7 Note that something is missing here. Because it seems that the jury is still out, I do not offer a bottom line on the question of whether ordinary people given credible information do or do not deal decently well with risk issues. But I tilt toward the proposition that they do.

344

Alan Randall

2.3.4. The trade critique PP serves as cover for protective trade barriers. The PP has become an issue in international trade – witness United States objections to European Union restrictions on importing GM food commodities8 and beef raised with artificial hormones9 – and the US and Australian skepticism about the PP has been motivated at least in part by concern that it may serve as an excuse for trade barriers (D. Peterson, 2006). In unsubtle application, PP-based objections to international food aid may invite certain disaster for the very poor as the ‘‘acceptable’’ cost of avoiding a more speculative risk to future trade (Turvey and Mojduszka, 2005). This critique is not so fundamental, it seems to me. In the trade context, there is always the fear that scoundrels will invoke available pretexts for protectionism. In that regard, PP is hardly unique – similar fears have been expressed regarding, for example, multifunctional agriculture and payments for ecosystems services. Clearly, it is inappropriate to camouflage protectionist measures in a cloak of precautionary virtue. But trade is merely an instrument for welfare, and as such its virtues are not unlimited. There are legitimate differences among countries about the appropriateness of certain inputs into production of food products, and it is hard to argue that trade imperatives should over-ride importing country concerns. A case can be made for the compromise of requiring that such imports be admitted but with labels informing consumers about the use of production inputs and methods the importing country considers questionable.

2.3.5. The critiques in context Summarizing and critiquing some of the major critiques of the PP prepares us for what follows, in two ways. The critiques themselves have been questioned, a process that has blunted some of them and placed the remaining challenges in perspective. The concerns that remain should be taken seriously, in the sense that a serious PP framework must deal with them in constructive ways.

8

The World Trade Organization found that the EU had maintained a de facto moratorium on GM imports through 2003 by failure to complete procedures without undue delay, and had acted inconsistently with its obligations in that safeguard measures were not based on risk assessments as specified in the Sanitary and Phytosanitary (SPS) Agreement and hence could be presumed to be maintained without sufficient scientific evidence (WTO Dispute Settlement DS291, 2006; Matthee and Vermersch, 2000). 9 In the EC-Hormones case, the WTO Appellate Body found (1998) that the European Community ban on imports of beef from cattle treated with certain hormones violated the SPS agreement because it reflected a higher level of protection and was not justified by a risk assessment; in five of six cases, there was no rational relationship between the EC measure and the scientific evidence submitted; and that invoking the ‘‘precautionary principle’’ did not override a country’s obligations under the SPS agreement.

Innovation, Risk, Precaution, and the Regulation of GM Crops

345

3. Innovation, risk management, and scope for precaution Innovation, especially innovation of a genuinely novel kind, is a venture into the unknown. Unintended consequences are possible for two distinct reasons: we may be unaware of all the pertinent properties of the innovation, and we may be unable to predict the response of the system into which it is introduced. Yet the unintended nature of the consequences does not absolve us of responsibility – we could have been more cautious, trying to learn more about likely outcomes before releasing the innovation irrevocably. 3.1. Weaknesses of ORM The standard approach to risk management (which I label ordinary risk management, ORM) combines a reductive model of risk assessment with a rational utilitarian decision framework. 3.1.1. Assessment of risk ORM risk assessment models tend to be overly reductive, whereas complex systems thinking is more alert to unpredictability, gross ignorance,10 unknown unknowns,11 and the possibility of surprises.12 For establishing harm, ORM accepts the norms of science (which are designed to minimize the acceptance of false empirical generalizations into the body of scientific knowledge), whereas a case can be made for an approach that is more evenhanded in its stance toward diagnostic risks (balancing the desires to avoid false positives and false negatives regarding harm and threat thereof) and open to adjusting the evidence requirement to the severity of the threat. 3.1.2. The decision framework ORM is founded on the paradigm case of well-specified games of chance, where the ex ante outcome set is fully defined in terms of magnitude and probability of each possible outcome. It struggles to make sense of uncertainties and is tongue-tied in the face of gross ignorance. ORM, being founded on the law of large numbers – why else would statistical expectations loom so large in ORM? – is more comfortable with insurable 10 Gross ignorance refers to situations where the ex ante outcome set is unbounded (Henry and Henry, 2002). 11 Donald Rumsfeld gets the credit for this term: ‘‘but there are also unknown unknowns, the ones we don’t know we don’t know’’ (US Department of Defense briefing, February 12, 2002). 12 Surprises are defined as outcomes that were not included in the ex ante outcome set (as opposed to common usage, where outcomes that were assigned low probability ex ante are called surprises).

346

Alan Randall

risk, whereas some of the big worries of our age (e.g., climate change) raise the possibility of systemic risk. Deviations from the base-case, where the law of large numbers rules and the decision rule is based on expected value, have been developed to introduce risk aversion; but there is always debate about when their use is appropriate. Being utilitarian at the core, ORM has unconvincing responses to the challenge of low-probability, highdamage threats (witness the St. Petersburg paradox). In practice, ORM combines a ‘‘safe until proven harmful’’ stance to risk assessment with a reluctance to get into serious pre-release testing, whereas PP thinking is more open to prescribing pre-screening and pre-testing procedures for proposed innovations – and perhaps for whole classes of cases where the possible risks are high enough, or the exposed systems sensitive enough, to pose a serious threat – and to prohibit a specific innovation that presents a disproportionate threat. 3.1.3. Scope for precaution It makes sense, I think, to accept ORM as a well-tested and proven framework for the management of ordinary risks. It offers less convincing solutions to risks that go beyond the ordinary – for example, problems where systems are complex and unpredictable, we are not guaranteed repeated trials (rendering the law of large numbers irrelevant), and there is a credible threat of disproportionate harm. Nevertheless, the limitations of ORM do not alone make the case for PP – a coherent PP must steer clear of those criticisms that survived scrutiny in Section 2. In particular, it should be focused on threats that are in some sense extraordinary, e.g. credible threats of disproportionate harm and cases where the ex ante outcome set is ill-defined; and it should pay rigorous attention to the relationships between knowledge, harm, and action (Hughes, 2006).

3.2. A coherent PP framework The general conceptual form of the PP suggested here relates the knowledge concept (evidence), the harm concept (threat, defined as chance of harm),13 and the action concept (remedy) as follows: If there is evidence stronger than E that an activity raises a threat more serious than T, we should invoke a remedy more potent than R. 13 A note on language: In common usage, risk jumbles the concepts of harm and chance, sometimes referring to one, sometimes to the other, and sometimes to the interaction of both. Through this chapter, I use risk, risky, risk-taking, etc., in the colloquial manner; and I follow standard practice by using risk colloquially in terms such as risk assessment and risk management. However, I make every effort to use our ‘‘chance of harm’’ concept, threat, consistently and rigorously.

Innovation, Risk, Precaution, and the Regulation of GM Crops

347

A PP of this form would focus on the key issues: what sorts of threats might invoke precaution, what sorts of evidence might justify a precautionary intervention, what sorts of remedies might be appropriate, and how might these key elements, E, T, and R, interact? Careful consideration of these questions suggests a tentative working definition of the precautionary principle, in an ETR framework (Randall, 2009, 2011): Credible scientific evidence of plausible threat of disproportionate (and often asymmetric) harm calls for avoidance and remediation measures beyond those recommended by ordinary risk management. The call to action is triggered by scientifically credible evidence of a disproportionate threat, such that the loss from the worst case outcome (even if unlikely) is disproportionately large relative to the gain from the most likely outcomes; if the threat is also asymmetric, the very bad outcome is more likely than it would be if outcomes where normally distributed.14 Remedies indicated are not restricted to those that would pass a benefit–cost filter even if substantial risk aversion is built-in. This precautionary principle addresses plausible but uncertain harm, and it may call for approaches that consciously depart from ORM. It is simply not limited in figuring out how much risk aversion to front-load into a benefit–cost analysis.

4. Looking more closely at remedies Among the categories of potential PP applications, proposed novel interventions offer the most complete array of pre- and post-release remedies (Box 1) – once an innovation has been released irrevocably, it is already too late for pre-release testing; and pre-release testing is inapplicable to the ‘‘novel outcomes from overstressing systems in the course of business as usual’’ category. For proposed innovations, flexible, iterative pre-release testing protocols can be designed to generate information useful in evaluating the threat and potential remedies even as they provide protection. In this process, empty threats are identified and the intervention allowed to proceed, manageable risks are identified and the intervention allowed to proceed perhaps subject to continuing regulatory oversight, and in the relatively few cases where unacceptable threats are confirmed the proposed intervention can be abandoned pre-release. Examples that fall into the proposed novel interventions category include new drugs, purposeful releases of nonindigenous species, genetically modified organisms (GMOs), and synthetic organic compounds. Yet, as we have seen, pre-release testing is routine for only some of these interventions, and only in 14 For example, Weitzman (2009) argues that the consensus among reputable climate models suggests a ‘‘fat-tailed’’ probability distribution generating a nontrivial probability (around 7%) of catastrophic temperature rise, clearly a threat that is both disproportionate and asymmetric.

348

Alan Randall

Box 1. Stylized pre- and post-release strategies for threats from novel interventions Introduction of a novel organism (a GMO, or a nonindigenous species) is proposed. Potential benefits are substantial, but there is a credible threat of serious harm. Pre-release (continue/terminate decision at each step) 1. Pre-screening for potential harm (chemistry, biology, genetics, ecology, etc.) 2. Laboratory testing in vitro

Test in isolation for potential harm

3. Laboratory testing with whole organisms 4. Release protocol with iteratively less restrictive quarantine – proceed to next step only if harm is below a threshold of concern  Step 1  Step 2  Etc. 5. Release for general use (conditional on what, if any, restrictions?)

Post-release 1. Post-release surveillance 2. If signs of harm  Preliminary evaluation (chemistry, biology, genetics, ecology, etc.)  Laboratory testing in vitro  Laboratory testing with whole organisms  Testing in confined environment with intensive monitoring 3. If it is established that the harm was caused by the introduced organism  Prohibit, restrict, and/or regulate further use of the organism  Mitigate, adapt, and/or remediate damage

Determine pathways to harm in environment

Innovation, Risk, Precaution, and the Regulation of GM Crops

349

some jurisdictions. If the opportunity is missed and suspicions of harmful impact arise, available remedies are limited to the post-release category. There is a utilitarian tradition that values maintaining flexibility while we wait to learn more (Arrow and Fisher, 1974; Henry, 1974; Pindyck, 2007), and Gollier and Treich (2003) write of precautionary measures as managing the wait for better scientific information. Pre-release testing is not just about waiting for better information, but actively designing and guiding processes for learning more about the threat and methods of managing it. The basic idea is to stick a toe in the water, i.e. find ways to gather information without risking very much. We could learn more about the proposed novel intervention, perhaps in the laboratory; model the system to identify possible reactions to the intervention and to identify indicators that might provide early warning of harm; perhaps even conduct limited trial releases under controlled conditions that allow revocability (pulling the plug on the experiment, and containing and eliminating any further harm); all before exposing the system to the risks entailed in general release. 4.1. Screening, pre-release testing, and post-release surveillance, STS Consider STS as a stylized representation of the kind of research-intensive release process we have in mind. STS is feasible if the effects of an innovation can, at least in the early stages of the process when uncertainty about chance of harm is greatest, be confined in space and time.15 Beginning with screening, and testing under laboratory conditions, the innovation goes through a stepwise pre-release testing process with monitoring, study, reassessment of the threat and the remedy, and adjustments in remedy as warranted by emerging evidence, at every step – a process that may be iterated perhaps many times (Figure 1). Each step ends with a decision selecting one of the three options: proceed to general release, continue to the next step in the testing process, or terminate testing and prohibit general release. The termination decision is made only when the evidence suggests unacceptable risks in the next step. For risks thought acceptable, research should focus also on risk management, mitigation, and adaptation. Should evidence of unacceptable risk fail to arise during screening and pre-release testing, general release may be undertaken with much more assurance than if we had simply rolled the dice at the outset. In STS, release would be followed by a program of post-release follow-up research to check for unexpected harmful consequences – possible but less likely now, 15 Containment during pre-release testing is an essential component of STS protocols. Effective containment can be expensive, but so can failure to contain. During pre-release testing, the US rice supply became contaminated with LibertyLinks, a GM variety designed for resistance to glufosinate herbicide. In a series of court cases the developer, Bayer AG, has been ruled responsible for consequent economic damages (In Re Genetically Modified Rice Litigation, 06-md-1811, US District Court, Eastern District of Missouri, 2010).

350

Alan Randall Screen

Prohibit

PRT 1

Prohibit

PRT 2

Prohibit

PRT 3

Release

Prohibit

Release Release Release

Post-release surveillance

Fig. 1. Stylized screening, pre-release testing, post-release surveillance (STS) process – assume three steps of pre-release testing, PRT. Source: Randall (2011, p. 220).

because extensive pre-release testing should have revealed and eliminated the more likely harmful possibilities. Post-release research should focus also on methods of remediating any damage discovered.

4.2. Innovation and the PP If all goes well, step-by-step through the STS process, general release is not prohibited indefinitely but delayed just long enough to enable screening and any pre-release testing indicated; precautionary accommodations may turn out to be temporary; and levels of self-protection may be adjusted as more is learned. Precaution is directed toward those innovations that present unusual threats, and is therefore a lesser impediment to innovation in general. In the polar case where we can never learn anything about the chances of a specific innovation being disastrous until the die is cast, and we cannot contain the disaster or mitigate it should it occur, precaution can be exercised only with respect to the whole program of innovation, a much less promising circumstance (Box 2). More often, we can expect to be able to identify more and less hazardous classes of innovations or, better yet, learn some things about particular innovations before their general release. Ideally, from a precautionary standpoint, we may be able to implement fully the STS procedure. The iterative and sequential remedies provided by pre- and post-release testing protocols offer some reassurance that the PP can be implemented in ways that

Innovation, Risk, Precaution, and the Regulation of GM Crops

351

Box 2. Venturing into the Unknown – Innovation and the PP Given that exploring the unknown has upside possibilities, critics have worried that PP would stifle innovation and growth. Despite the infrequent innovation that results in truly disastrous harm, and the many innovations that yield modest net losses and are soon abandoned, assume that the expected value of a long sequence of innovations is positive. There may even be occasional upside surprises such as the laser, which turned out to have beneficial uses not imagined at the time of its discovery. Provided we can save enough from the gains that we can bear the infrequent disasters, we would want to keep on innovating. But the disaster may happen on any trial and, whenever it happens, it will be a huge set-back. If it happens early in the sequence, recovery will be difficult, perhaps impossible. Would the PP prohibit innovations? Consider four cases: 1. Suppose we can never learn anything about the chances of a specific innovation being disastrous until the die is cast, and we cannot contain the disaster or mitigate it when it occurs. Then, our PP deliberations must focus on the program of innovation. T is large, and E is clear (we know that some innovations are disproportionately harmful, and ex ante we know nothing specific about the threat in any particular trial). It is reasonable that admissible R would include prohibition (but prohibition will not necessarily be the chosen R). 2. Suppose, instead, that we know some things ex ante. Then we can sort prospective innovations into ex ante more and less hazardous classes, and confine extreme remedies to cases that are ex ante more than ordinarily prone to generate serious threats. 3. Suppose we can learn some things about each innovation before implementation, perhaps through laboratory analyses, and modeling. Then, we can sort the proposed innovations with more confidence, and quarantine those that E suggests are more likely to be disastrous. 4. Suppose we are not limited to binary remedies (implement/ prohibit). Perhaps we can implement each innovation stepwise, relaxing the degree of revocability following each step that reveals no unacceptable risks. Then, if E generated by this STS procedure signals alarm early in this process, implementation can be aborted and disaster contained. Case 1, where all innovations must be treated as equally likely to turn out disastrously and prohibiting the whole program of innovation is an admissible remedy, seems unlikely. Case 2, where

352

Alan Randall

all we can do is sort potential innovations into ex ante more and less hazardous classes, is likely to be rare and good policy and practice would aim to make it more so. In cases 3 and 4, the PP serves the learning function envisioned, and provides (it can be argued) needed protections as we venture into the unknown – and these protections are directed not against innovation generically but against particular innovations identified as especially risky.

protect against disproportionate threats without systematically stifling the process of innovation. Ideally, they allow PP interventions for a broad class of ex ante risky innovations, with particular cases moving rapidly through the process toward general release while the precautionary focus hones in on a few cases where testing tends to confirm a serious threat. Perhaps more precaution at the outset results in lower costs and less unnecessary precaution in the end. 5. Genetically modified crops as a case of innovation An agricultural biotechnology research program of global dimensions, in which the corporate role has come to dwarf that of universities and research institutes (Huffman, 2011), has demonstrated the capacity to produce a continuing stream of potential commercial innovations. Human discomfort with innovation may apply to particular GMO innovations viewed as one-shot events, and to the program of biotechnological innovation that some perceive to have the potential to rapidly accelerate the pace of change, perhaps with inadequate scrutiny. In addition to risks of harm to human health and the environment, concerns have been raised about impacts on the structure of the agricultural industry and associated socio-economic effects, and a threat to public trust generated by reluctance of industry and some governments to label GM foods as such (FAO, 2000). Below, a stylized protocol pre-release screening and testing, and postrelease surveillance is applied to the health and environmental threats from a stream of innovations in GM crops. 5.1. A stylized release protocol for genetically modified crops GMO risk is a particular example of risk from technical innovation. GM crops offer a variety of potential benefits (Qaim, 2011; Raney and Matuschke, 2011; Wesseler et al., 2011), including higher yields, resistance to diseases and pests, specific characteristics desirable to consumers (e.g., nutritional enhancements, and perhaps disease-fighting properties), and characteristics that reduce production costs (e.g., resistance to weed-

Innovation, Risk, Precaution, and the Regulation of GM Crops

353

control chemicals). However, specific threats have been raised. Among them, food safety and other effects on human health, gene flow, and other effects on biodiversity (e.g., toxicity to nontarget species), and development of resistance in targeted weeds, pests, and diseases (superbugs and superweeds) are addressed here. Food safety and human health 1. The case where we are unable to learn anything about the chances of a specific innovation being disastrous before its general release (i.e., case 1, Box 2), is quite unlikely. 2. By systematic pre-screening, we can learn some things ex ante. For example, by chemical assessment, analysis of pathways, etc., we should be able to predict cases that increase the likelihood of toxicity and food allergies.16 3. Testing in the laboratory should distinguish ex ante some of the cases where release would be risky. 4. STS applies quite readily – analysis, modeling, laboratory testing, limited and controlled trials with animals and eventually human volunteers, and eventually general release with follow-up review and assessment. Even for GM foods judged safe, individual exposure does not have to be involuntary – remedies may include labeling that provides individual consumers the opportunity to choose. Biodiversity 1. Case 1 (Box 2) is quite unlikely. 2. Pre-screening may be fruitful – we know some important things ex ante. For example,  We can distinguish open-pollinated species, where risk of gene flow (spread to the non-GM population, and to related species) is higher, from others.  We know something about susceptible species and varieties and their roles in the host environment, and we can model the impact of the proposed intervention on host ecosystems. 3. If the results of pre-screening are inconclusive, STS may apply quite readily – ecosystems modeling, limited and controlled trials, and eventually general release with follow-up review and assessment. 4. If it is decided to permit commercial use of GM species with higher risk of gene flow, the choice is between quarantine and general release. Quarantine may be implemented in commercial operations by emphasizing containment. In the case of general release, the emphasis should be on post-release review and assessment and dealing with the damage, if any, after it has occurred. Quarantine arrangements in 16 Cellini et al., (2004) provide a helpful review of the potential for unintended consequences from genetic modification as well as from conventional plant breeding, and review methods of screening and testing for potential adverse effects on human health.

354

Alan Randall

commercial production, being expensive, would be concentrated on high-value crops and varieties. This implies that, for low-value crops, the options may be limited to prohibition and general release with mitigation, adaptation, and remediation of any damage. Resistance 1. We know ex ante that evolutionary processes favor mutations exhibiting resistance, and that evolution proceeds more rapidly among simpler kinds of organisms. However, the natural selection that drives evolution requires fairly broad exposure to the resistant crop. This consideration suggests only limited opportunity to learn anything before general release about the chances of a specific innovation being especially risky in this regard. 2. Nevertheless, we have at least a little ex ante knowledge that would be useful in pre-screening. For example, we can make some inferences about the likelihood of resistance developing in this case by examining the experience with similar prior cases. Also, factors relevant to gene flow (the biodiversity discussion, above) may be relevant also to development of resistance. 3. There seems to be only a limited role for laboratory testing. However, modeling may help distinguish ex ante the cases where general release is likely to stimulate development of resistance. 4. If it is decided to permit commercial use of GM species with higher risk that resistance will develop, the choice is between quarantine and general release. Quarantine may be implemented in commercial operations by emphasizing containment, and is plausible in the case of high-value crops. In the case of field crops, commercialization implies general release, and the emphasis must be on post-release review and assessment and dealing with the damage, if any, after it has occurred.

5.2. Where might PP prohibitions apply? The pre- and post-release testing protocols outlined in Section 4 and sketched here for proposed GM crop innovations are designed to protect society from disproportionate threats, and do so at acceptable cost by iteratively focusing more intense scrutiny on cases where the threat of harm cannot be dismissed. This approach is sensitive to the costs and opportunity costs of precaution, and to the benefits lost when innovation is burdened with unnecessary costs, delays, and prohibitions. Nevertheless, the possibility remains that a serious PP would, when other remedies have been exhausted, prohibit some innovations that pose disproportionate threats. In the case of GM crops, prohibitions are most likely to apply in the STS domain, where results in the early stages suggest aborting the process before general release. The argument for PP prohibitions always requires careful assessment of the evidence, threat, and remedy conditions in the

Innovation, Risk, Precaution, and the Regulation of GM Crops

355

particular case. However, we can identify some kinds of cases where resort to PP prohibitions is more plausible. Food safety and human health  Laboratory testing demonstrates toxicity or allergenic properties.  Limited and controlled trials demonstrate toxicity or allergenic properties.  Post-release review and assessment demonstrates toxicity or allergenic properties. Biodiversity  High-risk GMOs may be restricted to operations capable of effective containment.  General release may be prohibited if effective containment cannot be demonstrated, or is economically infeasible (e.g., low-value field crops). Resistance  High-risk GMOs may be restricted to operations capable of effective containment.  General release may be prohibited if effective containment cannot be demonstrated, or is infeasible economically (e.g., for low-value field crops).17

5.3. Current regulatory practice applied to GM crops Many countries review and/or regulate GM crops and/or foods, incorporating some of the procedures sketched in the stylized protocol, above. The United States made two key decisions early in the game: to regulate GM products rather than GM technology; and to rely on existing health, safety, and environmental laws rather than developing a new regulatory framework designed for GM products. As a result, three federal agencies may become involved when GM crop cases touch on their regulatory domain (http://www.agbios.com/static/cscontent/REGUSACAN-USAEN_printer.html). The Food and Drug Administration maintains a voluntary consultation process in which it reviews tests (typically theoretical and in vitro – animal studies are rare) conducted by developers of GM foods and, if satisfied (the substantive standard is that they pose no greater risk than non-GM foods), issues a letter stating that they have no further questions. This modest level of scrutiny is initiated at the request of product developers eager to offer assurances to the buying public. 17 In practice, US regulatory authorities have been reluctant to prohibit practices that threaten to advance development of resistance. The routine administration of antibiotics in the feed of healthy livestock in confined animal feeding operations continues in the United States, despite fears that it will contribute to development of antibiotic resistant human pathogens.

356

Alan Randall

The USDA Animal and Plant Health Inspection Service regulates new GM varieties as it would exotic varieties and species proposed for introduction, i.e. it is concerned mostly with harmful characteristics and invasive potential. Developers apply for nonregulated status, and APHIS reviews risk factors (i.e. conducts a limited screening) and solicits public comments before making a determination.18 GM crops with pesticidal properties must be registered, and the Environmental Protection Agency conducts a risk assessment (a kind of screening that may involve reviewing tests conducted by developers) and invites public comment before making the registration decision. Post-release, none of these agencies conducts serious surveillance, but all three agencies have the authority to remove GM products from the market if new, valid data demonstrate serious threats to consumers or the environment. At that stage, for some organisms, elimination may be very costly and incompletely effective. The Canadian regulatory process resembles that of the United States in some ways, e.g. the commitment to regulating GM products rather than processes, and the involvement of three separate agencies (Canadian Food Inspection Agency, Health Canada, and Environment Canada). Yet the prerelease risk assessment process is more structured than that in the United States, and it appears to be more rigorous, especially when it comes to assessing environmental risks. Post-release processes are similar to those in the United States: no systematic post-release surveillance, but regulators may remove GMOs from the market if presented with new, valid evidence of harm. The European Union regulatory framework, being designed specifically for GMOs, is more comprehensive and consistent. All foods and feeds containing nontrivial proportions of GMOs or GMO derivatives are regulated via a one-stop authorization procedure (http://europa.eu/ legislation_summaries/agriculture/food/l21154_en.htm). A risk assessment is conducted, and the authorization criterion is that the product must be determined, by tests using the most advanced knowledge and technology available, to be as safe to humans, animals, and the environment as their conventionally derived counterparts. Consumers, farmers, and businesses must be given the freedom to either use or to reject authorized products made from GMOs, a goal that leads to requirements for GMO labeling, traceability, and co-existence (GM plants must be grown and handled so as to prevent uncontrolled mixing with conventional products). None of these three regulatory frameworks (the United States, Canada, and the EU) mirrors precisely the stylized process sketched earlier. The EU framework is precautionary in several dimensions, and it is more attentive to coexistence and freedom to choose than are the other frameworks 18 Recent court rulings (Monsanto v. Geertson Seed Farms, 561 U. S. Supreme Court, 2010, concerning Roundup Readys alfalfa; and Center for Food Safety v. Vilsack, U. S. District Court C 08-00484 JSW, 2010, concerning Roundup Readys sugar beets) have questioned APHIS’ diligence in assessing environmental risks from open-pollinated GM varieties.

Innovation, Risk, Precaution, and the Regulation of GM Crops

357

discussed.19 Post-release environmental monitoring is required for all GMOs. The EU framework requires the same kind of scrutiny for all applicants, whereas the STS process uses a sequence of filters aimed at focusing quickly on the more problematic cases, and could for that reason be thought less sensitive to minimizing impediments to innovation.20 The Canadian process has the virtue, from my perspective, of taking a relatively comprehensive approach to environmental risks (gene flow, etc), and requiring pre-release tests first in contained environments and then in confined field trials, as proposed in Randall (2011). However, it has been criticized for failure to learn from experience – the argument is that, although experience has tended to confirm the benign nature of GM varieties, the regulatory burden on new GM varieties has not been adjusted downward (Smyth and McHughen, 2007). The US regulatory process is less complete in several dimensions than STS. Precaution is exhibited, in that there are ex ante risk assessments involving screening for risk factors (and, in the case of the FDA and EPA, review of the rather limited pre-release tests conducted by developers), but it looks like a cautious sort of precaution – tilting more toward minimizing costs and impediments to innovation, a critic might infer, than toward protection from harmful consequences. In the United States and Canada, post-release surveillance is informal – provided by NGOs, academic researchers, and independent observers – rather than systematic, with the inevitable result that claims of harm are debated in the media, the regulatory structure, and perhaps the courts before remedial action is taken. Nevertheless, Fedoroff et al. (2010) argue that US pre-release screening and testing procedures amount not to too little regulation but too much. One of their points is that corporate biotechnology shows little interest in raising crop production in the lowest income countries, many of which are likely to be severely impacted by global warming, leaving that market segment to government and university programs that cannot afford the regulatory burden (see also Smyth and McHughen, 2007).21 19 Coexistence and freedom to choose are lightning rods for some critics of the EU approach, who suspect inappropriate barriers to trade. A counter-argument might be that European preferences, so long as they are imposed on European as well as foreign suppliers, should be respected. 20 Nevertheless, the EU process pays some attention to streamlining the process, by establishing a one-stop authorization procedure and committing to timely decisions. 21 I would debate Federoff et al. on two points. First, The STS-based process sketched above is not so onerous, I would claim, in that many GM crops are likely to be released following prescreening, or pre-screening plus limited pre-release testing. Furthermore, as we gain experience with GM crops it is likely that pre-screening will become more reliable, which would speed more varieties toward general release. Second, while Federoff et al. note that only a few GM crops have been released, I would argue that this can hardly be blamed on the rather modest regulatory controls in place. It is true that many of the GM varieties in development have yet to make it to market, but I suspect that the reason often has more to do with public concerns expressed in a variety of informal ways than with excessive formal regulation.

358

Alan Randall

5.4. International agreements: the case of the Cartagena Protocol on Biosafety Trade in GM crops is subject to various internationally agreed standards and protocols, including the Codex Alimentarius addressing food safety. The authority of these institutions can be complicated – for example, the Codex is voluntary, but it is recognized by the World Trade Organization as an international reference standard for the resolution of disputes involving food safety – and jurisdiction is usually limited to countries that are signatories to some broader international agreement. Having compared and contrasted existing regulatory practice in the United States, the EU, and Canada with the stylized STS protocol, we now consider existing international agreements, focusing on the case of the Cartagena Protocol on Biosafety (CPB) adopted by more than 130 countries to address threats to biodiversity. It establishes a Biosafety Clearing House for information, and rules applying to trade in living modified organisms (LMOs), i.e. those that can grow and/or reproduce. An advance informed agreement (AIA) procedure for LMOs intended to be introduced into the environment (e.g., seeds for planting, fish for release, etc.) requires exporters to seek consent before the initial shipment. The Protocol is explicitly precautionary in principle (Preamble, and Article 1), and in practice it allows importing countries to apply a precautionary approach in the AIA process, taking action to avoid potential adverse effects even in the absence of scientific consensus (Articles 10.6 and 11.8, and Annex III). However, other provisions of the Protocol soften its precautionary stance: expedited entry is granted for food and feed commodities intended for direct use and accompanied by documentation that they may contain LMOs and are not intended for intentional introduction into the environment, and for LMOs in transit and/or containment. The Protocol does not address food safety concerns (see Codex Alimentarius, earlier), and it requires neither segregation of LMO commodities nor consumer product labeling. The stylized STS process would, arguably, endorse the CPB’s precautionary stance to LMOs intended for release into the environment. Provisions for importing country consent respect more than national preferences and attitudes – they also respect importing country knowledge. The effects of such introductions depend at least as much on the attributes of the receptor ecosystems (and the economies and societies that have evolved along with those ecosystems) as on the attributes of the LMO itself, and importing countries may well have unique insights on the vulnerability of their ecosystems. Compared to CPB, the STS process would take systematic post-release surveillance more seriously, and would pay more systematic attention to the categories excepted from the AIA process (foods and feeds intended for direct use, and LMOs in transit and/or in containment). In all of these

Innovation, Risk, Precaution, and the Regulation of GM Crops

359

cases, there is a need to ensure that benign announced intentions (direct use, transit, and/or containment) do indeed materialize as stated; and the STS protocol would be more attentive than the CPB to the detailed remedies that should be implemented to ensure compliance with stated intentions. In the case of foods and feeds intended for direct use, the absence of segregation requirements risks contamination of the non-GMO supply. One might argue that contamination issues extend far beyond threats to biodiversity, and we can hardly expect a protocol on biosafety to attend seriously to contamination concerns. I am not entirely convinced that this justifies the CPB’s unconcern with segregation but, either way, this example highlights a major difference between the stylized STS protocol and the patchwork of international institutions and agreements addressing GMOs. The STS protocol is inherently more integrative, ideally addressing the broad suite of threats and their interactions rather than compartmentalizing the threats and assigning them to different institutions with different missions and jurisdictions.

5.5. Applying the STS-based protocol: two case examples Here we consider two actual cases, contrasting the regulatory process and outcome in the United States with how they might have been handled had the stylized STS protocol (above) been in place.22 Starlinks corn, a Bt (Bacillus thuringiensis) corn, produces a protein that kills corn borer larvae. It was authorized in the United States by EPA (under its mandate to regulate GMOs with pesticidal properties) for livestock feed. However, it was not approved for human use because of concerns that the protein involved had some attributes of an allergenic protein. It never obtained EU approval, which involved the United States in trade issues with the EU and food aid issues with some African countries that trade with the EU. Eventually Starlink corn was found in corn destined for human consumption. Allergic reactions were reported in a few cases, but the Centers for Disease Control were unable to confirm the association with Starlink. The FDA issued guidelines for segregating Starlink from non-Bt corn, and eventually Starlink corn virtually disappeared from the US corn supply. The FDA guidelines were withdrawn and superseded by less stringent EPA guidelines. The Starlink episode was a public relations disaster for the biotechnology industry, and its costs to US corn growers may have exceeded $100 million. 22 Being aware of the hazards inherent in this process – perhaps the least of which is comparing real life, warts and all, with an idealized hypothetical alternative – I will attempt to compensate with a good deal of circumspection about how the stylized process would have worked.

360

Alan Randall

The stylized protocol would have screened for human health and food safety, biodiversity, and resistance effects. The protocol may have addressed concerns about resistance and biodiversity by requiring a series of test in contained environments, confined field tests, etc. But, because the protocol is open to adjusting the stringency of pre-release testing as more is learned about Bt crops as a category, it is also possible that concerns about resistance and biodiversity may have been overridden on grounds that Bt corn presents risks no greater than those of other Bt crops, or risks no greater than naturally occurring Bt, etc. However, it should be noted that these kinds of judgments are inherently quantitative, yet there is little quantitative evidence that can be brought to bear on these questions. The interesting issue concerns human health and safety – given that screening alerted authorities to the possibility of allergenicity, the protocol would call for rigorous testing. If testing had confirmed allergenic properties, it is likely that Starlink corn would have been withdrawn from the market much earlier in the process. If testing had rejected allergenicity convincingly, the controversy would likely have been contained. Yet, pre-release testing was curtailed. Perhaps approval for livestock feed only seemed an easier way out. It is likely that the protocol would have rejected the ‘‘livestock feed only’’ solution as an unworkable remedy, because the infrastructure for handling corn from farm to consumer was inadequate to ensure segregation of Starlink from non-Bt corn. The developer, surely, would have responded either by pursuing a rigorous program of testing for allergic reactions in human consumers, or by withdrawing Starlink corn. Either response would have saved both the developer and corn producers considerable harm relative to the way things actually played out. STS calls for systematic post-release surveillance. In contrast, claims of allergic reactions to Starlink corn came from the grassroots and gained public currency before they could be vetted by respected and responsible institutions (in this case, the Centers for Disease Control). Roundup Readys canola has been approved in many countries. In the United States, the key approval came from EPA under its authority to regulate pesticides. A major consideration in the approval decision seemed to be that other RR crop varieties had been approved already. The STS-based protocol would have screened for human health and food safety, biodiversity, and resistance effects. It is likely that resistance would have been the most compelling of the threats. Screening would have identified canola as an open-pollinated species with many wild weedy relatives (which are obvious filters in STS screening for possible gene-flow and resistance), and called for rigorous scrutiny. We cannot be sure whether Roundup Ready canola would or would not have been approved under the stylized protocol. The protocol would have been attentive to evidence that even Roundup-resistant weeds are susceptible to other

Innovation, Risk, Precaution, and the Regulation of GM Crops

361

readily-available herbicides (which was a clinching argument in several countries that approved Roundup Ready canola), but it may also have attended to claims that accelerated resistance to Roundup would surely have increased costs for farmers (and also for suburbanites, who use substantial quantities of the stuff).

5.6. A digression: where does market risk fit in? Some of the examples in this chapter touch on market risk, in addition to the threats to human health, biodiversity, and resistance that are the focus of this chapter. The documented harm from the Starlinks corn and LibertyLinks rice (see footnote 16) episodes was economic – the costs of eliminating the GM varieties from the respective supply chains and the interim loss of market opportunities – and claims of threat to human health have not yet been affirmed. So the question arises: what role should considerations of market risk play in the proposed integrated risk management approach? Unlike the threats that are my primary focus, market risk is a secondorder risk. The threat to market opportunities posed by GM crops arises because consumers and the public worry, or are thought to worry, about the first-order threats, e.g. to human health and biodiversity. A risk management framework addressed to consumer and public wellbeing is unlikely to identify market risk as a key concern. Furthermore, a risk management framework that includes a role for precaution is unlikely to identify market risk as a disproportionate threat, because the worst-case aggregate harm from market impacts is unlikely to be disproportionately greater than the expected gains. Nevertheless, market risk is a serious concern for the crop production, marketing, and trade sectors, for subsectors vulnerable to GM contamination (e.g., organic products), and for the biotechnology sector. It makes sense that risk management strategies for those sectors pay close attention to consumer and public concerns in their intended markets. To the extent that public policy is concerned with the wellbeing of domestic producers, market risk is a relevant policy issue. Furthermore, it is appropriate that courts be attentive to economic damages resulting from contamination of supply chains by GM varieties. Given that market risk is a second-order risk derived from consumer and public concerns about first-order risks, attending to the first-order risks is likely to reduce market risk. The discussion of how the stylized STS framework would have handled Starlinks corn (above) emphasized that it would have been attentive to resolving the issue of allergenicity one way or the other. In this, the STS protocol is motivated by concern with human health and food safety. Yet, market risk would also have been reduced had the food safety issue been resolved convincingly.

362

Alan Randall

6. Lessons for precaution and innovation Cost of remedy considerations has a legitimate role, along with costs of damage, in crafting remedies for credible threats of harm. It is easy to find claims or instances of daunting costs on both sides – the costs of PP prohibitions that stifled innovation and reduced economic growth could readily amount to hundreds of billions of dollars, and the costs of postrelease remediation of PCBs in the United States (just one country’s experience with one troublesome innovation) have already exceeded $50 billion with no end in sight. Cost-effectiveness is always a consideration in the choice of remedies, and is a source of motivation for developing and refining iterative and sequential remedies that encourage learning and reassessment, and provide sequential decision points, making precaution less intrusive and costly while still providing substantial protection from harm.

6.1. Is routine pre-release testing too costly? There is a widely circulated utilitarian claim that routine pre-release testing of innovations is too costly, and that the costs that really matter are the hidden ones: the benefits forgone when innovation is delayed. The FDA has become the poster child for this argument, as economists highlight the costs in lives lost and afflictions suffered due to delays in the release of promising new drugs. Interestingly, there is very little systematic empirical evidence in support of this claim – apparently, this is one of those intuitions that economists think is too self-evident to require empirical support. Yet the FDA operates as it does in response to clear political intent – more rigorous pre-release testing of new drugs was mandated following the thalidomide disaster in the early 1960s.23 This divergence between the views of the public and many economists hinges on two issues, First, the claim that FDA’s caution actually puts lives at risk by delaying release of new treatments rests on an expected value criterion: the claim is that the expected value of aggregate benefits would increase if we accepted more adverse effects from new drugs as the price for gaining earlier access to their benefits. Yet it can be argued that people expect more from approved drugs and medical treatments than that they, on average, do at least as much good as harm; and that people’s willingness to submit to prescribed treatments hinges on their trust that the 23 The FDA actually saved the US from the thalidomide disaster that befell many other countries, but that good fortune was attributed to the persistent skepticism of one FDA scientist, i.e. to the happenstance of having ‘‘the right person in the right place at the right time’’. Subsequent legislation was motivated by a desire to institutionalize a stronger precautionary stance in FDA.

Innovation, Risk, Precaution, and the Regulation of GM Crops

363

pharmaceutical industry and its regulators have eliminated the unacceptable risks in the recommended treatment. Second, in the absence of firm empirical evidence, there is a lot of uncertainty about the magnitude of the benefits lost due to excessive precaution. Whatever the magnitude of lost benefits, the kinds of iterative and sequential remedies discussed earlier provide ample scope for fine-tuning – e.g. for reducing lost benefits by increasing the proportion of truly desperate patients included in prerelease tests – to reduce the costs of precaution and still provide substantial protection from harm.

6.2. Could the mix of pre- and post-release remedies be improved? FDA’s approval procedure for new drugs follows laboratory and animal testing with three phases of testing with human subjects, each phase involving more subjects and conditions closer to actual clinical practice than its predecessor. If the drug is approved, FDA does post-market surveillance, analyzing outcomes from use among the general population. The approval process may take around eight years, long enough to motivate complaints that innovation is stifled and its benefits delayed, but short enough that it may fail to identify long-term harmful effects. Even with this rather exhaustive pre-approval testing, mistakes occur.24 Manski (2009) asks whether FDA has settled on the optimal combination of pre-approval and post-market research. He proposes a shift in the mix – new drugs would reach the user population more quickly but follow-up surveillance of users would be more systematic. Quicker release would increase risk for all early adopters of new treatments, whereas the benefits would be concentrated among those with serious afflictions. So, it would make sense to accelerate availability not across the board, but for new treatments for serious afflictions, and to direct early post-market use toward patients with serious cases of the indicated condition.25 Compared with current practice, the drug approval process 24

Consider the case of Vioxx, a Merck product prescribed as an anti-inflammatory following approval in 1999. Post-market surveillance revealed increased incidence of strokes and heart disease among users, and Merck recalled Vioxx in 2004 and set aside almost $5 billion to settle legal claims. It has been argued that the phase-3 testing of Vioxx was inadequate (despite FDA approval) because the subjects were not representative of the population likely to use it post-approval (Manski, 2009). 25 Recent developments regarding thalidomide illustrate some elements of the Manski proposal. Thalidomide was prescribed originally for nausea and morning sickness, rather mundane afflictions, and the resulting serious birth defects were judged to be disproportionately harmful. However, new applications for much more serious afflictions – where the patients’ prospects are grim in the absence of treatment – have been approved, subject to rigorous safeguards to minimize the chance of harm. Perhaps FDA’s response to this situation reveals one of the America’s few fundamentally precautionary institutions exhibiting some sensitivity to the economic-utilitarian critique.

364

Alan Randall

would be more continuous, giving the patients with greatest need earlier access to potentially beneficial drugs while encouraging more systematic long-term studies of users.

7. Concluding comments The standard approach to risk management, here labeled ordinary risk management (ORM), is well-adapted in managing well-specified risks involving predictable systems. The PP has been proposed to address uncertain threats in systems that are not well-understood, but it has attracted a lot of controversy. Here I conclude that there is scope for a PP addressed to disproportionate threats, but that a coherent PP must be constructed to steer clear of some valid concerns raised by the critics. The case of proposed innovations (including GM crops and food products) is especially interesting, in that it raises the potential for disproportionate threats and it allows for the most complete menu of remedies, both preand post-release. If precaution is a blunt instrument, prohibiting innovations en masse based on uncertain threats of harm, then clearly it comes at a nontrivial cost – it can systematically impede innovation at substantial cost to human progress. But, if precautionary remedies can be implemented in iterative sequenced decision processes that take full advantage of pre-release screening and testing opportunities, we can think in terms of a winnowing process that cuts a relatively wide swath at the outset but focuses quite quickly on a much smaller set of genuine threat cases, thereby saving time, direct costs, and benefits foregone due to delayed innovation, while affording more protection from disproportionate threats than provided by current procedures. The stylized remedy processes sketched in this chapter start to bridge the gap from principle to application in a way that provides more precaution at the outset, but imposes less unnecessary and intrusive precaution in the majority of cases where threats can readily be dismissed or managed. Finally, a recent proposal by Manski (2009) makes the case that there are potential gains to be made by optimizing the mix of pre- and postrelease remedies. In the particular case of new drugs for serious afflictions, he argues for expediting pre-release testing so long as post-release surveillance is ramped-up. The analogy between new pharmaceutical drugs and GMOs is quite incomplete: food safety, gene flow, and resistance threats associated with GMOs are more in the nature of public ‘‘bads,’’ in that individual exposure is involuntary (although labeling requirements may mitigate this concern in the case of food safety). Nevertheless, the idea of seeking the right mix of pre- and post-release remedies makes sense for the GM crops case, too. It opens a substantial research agenda that I hope is taken seriously.

Innovation, Risk, Precaution, and the Regulation of GM Crops

365

References Arrow, K., Fisher, A. (1974), Environmental preservation, uncertainty, and irreversibility. Quarterly Journal of Economics 55, 313–319. Bergen Ministerial Declaration on Sustainable Development in the ECE Region (1990), UN Doc. A/CONF.151/PC/10; 1 Yearbook on International Environmental Law 429, 4312. Canadian Perspective on the Precautionary Approach/Principle – Proposed Guiding Principles. (2001), www.pco-bcp.gc.ca/raoicssrdc/ default.asp?Language¼E&Page¼Precaution&Sub¼Booklet#3. Cellini, F., Chesson, A., Colquhoun, I., Constable, A., Davies, H., Engel, K., Gatehouse, A., Karenlampi, S., Kok, E., Leguay, J-J., Lehesranta, S., Noteborn, H., Pedersen, J., Smith, M. (2004), Unintended effects and their detection in genetically modified crops. Food and Chemical Toxicology 42, 1089–1125. Cooney, R. (2004), The precautionary principle in biodiversity conservation and natural resource management – an issues paper for policy-makers, researchers, and practitioners. IUCN Policy and Global Change Series No. 2. http://www.pprinciple.net/publications/PrecautionaryPrincipleissuespaper.pdf European Commission (2000), Communication on the Precautionary Principle. Brussels, February 2, 28pp. FAO (Food and Agriculture Organization of the United Nations) (2000), FAO Statement on Biotechnology. http://www.fao.org/biotech Fedoroff, N., Battisti, D., Beachy, R., Cooper, P., Fischhoff, D., Hodges, C., Knauf, V., Lobell, D., Mazur, B., Molden, D., Reynolds, M., Ronald, P., Rosegrant, M., Sanchez, P., Vonshak, A., Zhu, J. (2010), Radically rethinking agriculture for the 21st century. Science 327 (5967), 833–834. Foster, K., Vecchia, P., Repacholi, M. (2000), Risk management: science and the precautionary principle. Science 288 (5468), 979–981. Garrity, T. (2004), MTBE: a precautionary tale. Harvard Environmental Law Review 28, 281–342. Gollier, C., Treich, N. (2003), Decision-making under scientific uncertainty: the economics of the precautionary principle. Journal of Risk and Uncertainty 27, 77–103. Harris, J., Holm, S. (1999), Precautionary principle stifles discovery. Nature 400, 398. Harris, J., Holm, S. (2002), Extending human life span and the precautionary principle. Journal of Medicine and Philosophy 27, 355–368. Henry, C. (1974), Option values in the economics of irreplaceable assets. Review of Economic Studies 41 (Symposium), 89–104. Henry, C., Henry, M. (2002), Formalization and applications of the precautionary principle. Universite´ Catholique de Louvain, Institut de Recherches Economiques et Sociales (IRES) Discussion Paper 2002009.

366

Alan Randall

Huffman, W. (2011), Contributions of public and private R&D to biotechnology innovation. In: Carter, C., Moschini, G., Sheldon, I. (Eds.), Genetically Modified Food and Global Welfare. Emerald, Bingley, UK. Hughes, J. (2006), How not to criticize the precautionary principle. Journal of Medicine and Philosophy 31, 447–464. Jasanoff, S. (2000), Commentary: between risk assessment and precaution – reassessing the future of GM crops. Journal of Risk Research 3 (3), 277–282. Kahan, D., Slovic, P., Braman, D., Gastil, J. (2006), Fear of democracy: A cultural evaluation of Sunstein on risk. Harvard Law Review 119, 1071–1109. Kolitch, S. (2006), The environmental and public health impacts of US patent law: making the case for incorporating a precautionary principle. Environmental Law 36, 221–256. Lofstedt, R., Fischhof, B., Fischhof, I. (2002), Precautionary principles: general definitions and specific applications to genetically modified organisms. Journal of Policy Analysis and Management 21 (3), 381–407. Manski, C. (2009), Adaptive partial drug approval: a health policy proposal. The Economists’ Voice 6 (4), Article 9. http://www.bepress.com/ev/vol6/ iss4/art9 Matthee, M., Vermersch, D. (2000), Are the precautionary principle and the international trade of genetically modified organisms reconcilable? Journal of Agricultural and Environmental Ethics 12 (1), 59–70. Peel, J. (2004), Precaution – a matter of principle, approach, or process? Melbourne Journal of International Law 5 (2), 483–501. Peterson, D. (2006), Precaution: principles and practice in Australian environmental and natural resource management. Australian Journal of Agricultural and Resource Economics 50, 469–489. Peterson, M. (2006), The precautionary principle is incoherent. Risk Analysis 26, 595–601. Peterson, M. (2007), Should the precautionary principle guide our actions or our beliefs? Journal of Medical Ethics 33, 5–10. Pindyck, R. (2007), Uncertainty in environmental economics. Review of Environmental Economics and Policy 1, 45–65. Qaim, M. (2011), Genetically modified crops and global food security. In: Carter, C., Moschini, G., Sheldon, I. (Eds.), Genetically Modified Food and Global Welfare. Emerald, Bingley, UK. Raffensperger, C., Tichner, J. (1999), Protecting Public Health and the Environment: Implementing the Precautionary Principle. Island Press, Washington, DC. Randall, A. (2009), We already have risk management – do we really need the precautionary principle? International Review of Environmental and Resource Economics 3 (1), 39–74. Randall, A. (2011), Risk and Precaution. Cambridge University Press, Cambridge, UK.

Innovation, Risk, Precaution, and the Regulation of GM Crops

367

Raney, T., Matuschke, I. (2011), Current and potential farm-level impacts of genetically modified crops in developing countries. In: Carter, C., Moschini, G., Sheldon, I. (Eds.), Genetically Modified Food and Global Welfare. Emerald, Bingley, UK. Sandin, P. (1999), Dimensions of the precautionary principle. Human and Ecology Risk Assessment 5, 889–907. Shaw, S., Schwartz, R. (2005), Trading Precaution: The Precautionary Principle and the WTO. United Nations University, Institute for Advanced Studies, Tokyo. Smyth, S., McHughen, A. (2007), Regulating innovative crop technologies in Canada: the case of regulating genetically modified crops. Plant Biotechnology Journal 6, 213–225. Sunstein, C. (2005), Laws of Fear: Beyond the Precautionary Principle. Cambridge University Press, Cambridge, UK. Turvey, C., Mojduszka, E. (2005), The precautionary principle and the law of unintended consequences. Food Policy 30, 145–161. UNESCO (2005), The Precautionary Principle, World Commission on the Ethics of Scientific Knowledge and Technology. UNESCO, Paris. Weitzman, M. (2009), On modeling and interpreting the economics of catastrophic climate change. Review of Economics and Statistics 91 (1), 1–19. Wesseler, J., Scatasta, S., Fall, E. H. (2011), The environmental benefits and costs of GM crops. In: Carter, C., Moschini, G., Sheldon, I. (Eds.), Genetically Modified Food and Global Welfare. Emerald, Bingley, UK. Wiener, J., Rogers, M. (2002), Comparing precaution in the United States and Europe. Journal of Risk Research 5 (4), 317–349. Willis, R. (2001), Lighting the leap in the dark. http://www.forumforthefuture.org/greenfutures/articles/60288 Wingspread Statement on the Precautionary Principle (1998), http:// www.gdrc.org/u-gov/precaution-3.html. WTO Dispute Settlement DS291, (2006). http://www.wto.org/english/ tratop_e/dispu_e/cases_e/ds291_e.htm.

SUBJECT INDEX

Page numbers followed by f and t indicate figures and tables, respectively. ABA. See Abscisic acid (ABA) Abiotic stresses, 68 Abscisic acid (ABA), 11 ABSP II. See Agricultural Biotechnology Support Project II (ABSP II) Acreage planted GM corn, proportions of, 157 Ad hoc productivity, 290 Aflatoxins, 13 AFLPs. See Amplified fragment length polymorphisms (AFLPs) Africa sweet potato in, 14 AgrEvo, 133 Agricultural biotechnology, 226, 228, 229, 240 adoption of, 231 economics of, 226 Agricultural biotechnology patents, issue rate of, 116, 117f Agricultural Biotechnology Support Project II (ABSP II), 107 Agricultural chemical companies, 132 Agricultural commodities, 310 Agricultural crops. See also Plants cisgenic, 9 development in, DNA-based techniques and, 2–3, 3f genetically modified. See Genetically modified (GM) crops genetics and, 2–4 herbicide-tolerant crops, 36–38 losses in, due to pests, 33, 33f transgenic, 7–8 Agricultural productivity biotechnology, impacts of, 229–231 GHG emissions, 237 increasing, 179 Agricultural research, public expenditures on, 141, 143t Agrobacterium tumefaciens, 64 gene transfer with, 123 Agro-biological dynamics, 203

Agro-climatic conditions, 159 Agro-ecological dynamics, 203 Agroeste Sementis, 135 Allele, 4 Aly Paticipacoes, 135 Amaranthus palmeri, 181 Amplified fragmentlength polymorphisms (AFLPs), 4 Animal and Plant Inspection Service (APHIS), 123–125, 124f, 125f Animals biopharm, 9–10 transgenic, 9–10 Antithrombin, 10 APHIS. See Animal and Plant Inspection Service (APHIS) APHIS reviews risk factors, 356 Argentina HT soybean adoption, 37, 67 transgenic crops in, 8 Asgrow Seed Company, 135 Asia GM crops in, 57–63 ASSOCHAM, 100 Astra Zeneca, 137 ATryn, 10 Avery, Oswald, 119t, 120 Bacillus thuringiensis (Bt) crop varieties, advantages for, 122 genes, 116, 122 Bacillus thuringiensis (BT) corn, 359 category, 360 risks, 360 Bacillus thuringiensis (Bt) gene, 8 cotton, in Asia, 58–63 crops, 38–40, 39t Cry1Ac/Cry1Ab, 91 eggplant in India, 42–43, 105–108, 106t household income effects of, 41f white maize, 64 yellow maize, 64

370

Subject Index

Bangladesh adoption of GM rice, 298 GM food crops, 290 importer of wheat, 301 regulatory experts, 291 Bargaining capabilities, 217 Bayer CropScience, 133–134 with Chinese Academic of Agricultural Sciences, 143 on R&D, 140 Beatle, George, 119t, 120 Beta carotene, 13 Bikaneri Narma, 91 Bill & Melinda Gates Foundation, 12, 13 Binary variable for double-stacking (Kij), 156 coefficients of, 164 Biocentury Transgene Technology Company (BTCC), 91 Biocide Index, 66 Biofuels, 15–16, 225, 226, 227 carbon benefits of, 226 demand for, 235 corn, 232 economics of, 226–229 food market, 233 future of, 239–240 GHG neutral technology, 228 industry, 228 market with mandate and capacity constraint, 236 production, 236, 240 risk of bankruptcies, 236 subsidies, 236–237 sustainability of, 228 Biopharm animals, 9–10 Bioremediation, 16 Biosafety Bill for GM crops, 127 Biosafety Clearing House (BCH), 322 Biosafety policy, 327 Biosafety policymaking, 327 Biosafety regulations, domestic of GM food, 284 Biotech crops in 2009, by major producing country, 126t Biotech event, 133 Biotechnology APHIS in, 123 future of, 239–240 impacts of on agricultural productivity, 229–231 use of, 246

Biotechnology, agricultural, 1–19 adoption of, 88–92 diversity analysis, 4, 5–6 genetics, 2–4 investments in, 85–92 molecular marker, 4–5, 6–7 Biotechnology Regulatory Services, of APHIS, 123 Biotech seed companies, 129 Biotic stresses, 68 Bovine spongiform encephalopathy (BSE), 271 Boyer, Herbert, 120 Brazil biofuels, 227 HT soybean, adoption in, 127, 128f transgenic crops in, 8 Breeder’s rights, 121 Breeding conventional, 3–4 Bt. See Bacillus thuringiensis (Bt) BTCC. See Biocentury Transgene Technology Company (BTCC) Bt cotton, 176. See also Bacillus thuringiensis (Bt) CAAS. See Chinese Academy of Agricultural Sciences (CAAS) Canada economic loss, 203 HT canola, adoption in, 127, 128f, 129 precautionary principle, government, 341 regulatory model, 202 transgenic crops in, 8 Canadian Food Inspection Agency (CFIA), 204 Canadian regulatory process, 356 Canadian Wheat Board, 276 Canola HT, adoption in Canada, 127, 128f, 129 Cargill, 135 Cartagena Protocol on Biosafety (CPB), 317, 358 Cartagena Protocol on Biosafety to the Convention on Biological Conservation, 174 Cellulosic biofuel technologies, 227 Center for Chinese Agricultural Policy, 85 Chemical companies agricultural, 132 Ciba, 138 DuPont, 137 Monsanto, 134

Subject Index Pioneer Hi-Bred Corn Company, 118, 134–135, 135, 137 in seed industry, 132 China adoption gains, 288 adoption of Bt rice, 292 adoption of GM crops in, 88–89, 89f, 89t Bt cotton in, 40, 88–89, 89f, 89t, 128f, 129 commercialization of, 62 ecological impact of, 63, 84, 93–96, 94f, 95f health effects, 96–97 GMO crops in pipeline in, 103t, 104–105 GM rice in, 104–105 investments, in agricultural biotechnology, 85–86, 85f in public agricultural research, 142–143 transgenic crops in, 8 unapproved GM rice, 202 Chinese Academy of Agricultural Sciences (CAAS), 91, 104, 139 Chromosomes, 4 Ciba, 138 CIMMYT. See International Maize and Wheat Improvement Center (CIMMYT) Cisgenic products, 5, 9 Climate change, 56, 225 Coexistence regulations, 202–205 economic implications of, 206 economic point of, 208 economic problem of, 203 implications of, 215–216 liability rights, 209–214 property rights systems, equivalence of, 216–219 regional value of, 208–209 Cohen, Stanley, 119t, 120 Cohen–Boyer gene splicing technique, 119t, 120 Commercialization of biopharm animals, 10 of Bt cotton, in China, 62 of genetically modified crops, 36–42 Commodity clearance, 318 Computable general equilibrium (CGE) modeling, 286 Consumer-oriented agricultural economists, 257 Consumers about GM food, 244 biofuel, 235 biotechnology, 250

371

confidence, in government regulators, 271 controversy, 244 demand for GM foods, 244, 252 differences, about technology, 256 distribution of, 255 distrusting/precautionary, 313 economic research on, 258 in Europe, 272, 289 food, 232, 233, 234 French, 255 gasoline, 232 German, 254 in importing country, 331 in Japan, 289 in Korea, 289 market risk, 361 in Netherlands, 254 pesticide, health risks of, 273 preferences for GM food, 258 price differences, 274 trade filter, 294 trusting/tolerant, 313 type of, 316 welfare effects, 328 willingness-to-pay for, 244, 248, 252 Conventional breeding, 3–4 in plants, 5 Conventional markets, concentrations, 165 Corn for GM trait development, 122 GM varieties adoption, in US, 127, 127f patenting activity on varieties for, 125 public/private research, SYs in, 139, 140t, 142t Corn Belt, 167 Corn demand, 232 Corn market equilibrium with ethanol demand, 232 Corn seed prices, 155 in U.S. Corn Belt, 152 Cost-effectiveness, 362 Cost-minimizing behavior non-GM farmer, 212 Cost-reducing benefits of technology, 273 Cotton Bacillus thuringiensis in, 122–123 Bt, 39, 40 adoption in China, 128f, 129 adoption in India, 89–92, 90f, 92t, 128f, 129 impact of, 97–104, 99f, 99t, 102t in Asia, 58–63, 60t

372

Subject Index

cultivation, and farmer suicides in India, 61–62 farm-level impacts of, 59–63, 60t, 65–66 impact in China, 92–97, 94f, 95f net revenue from, 95, 95t production in South Africa, 65 for GM trait development, 122 GM varieties adoption, in US, 126, 127, 127f public/private research, SYs in, 139, 140t, 142t stacking of HT and IR traits in, 129 Cournot game, 154 Crick, Francis, 120 Crop adoption, GM, 287, 290 cotton, 288 countries definition of, 295 RICE scenarios, 296 WHEAT scenarios, 299 global trade effects of, 286 Crop biotechnology, 122 private and public R&D in, 138–144 private sector investment in, 143 public sector investment in, 143 Crop reporting districts (CRD) level, 157 Crops. See Agricultural crops Cross-GHHIs to affect prices, 158 Cross-market impacts, 166 Cry1Ac/Cry1Ab Bt gene, 91 DALYs. See Disability-adjusted life years (DALYs) Damage compensation claiming for, 220 Darwin, Charles, 118, 119t, 120 DBT. See Department of Biotechnology (DBT) DeKalb Genetics Corporation, 135 Delta & Pine Land, 135 Denghai Seeds, 86 Deoxyribonucleic acid (DNA), structure in genetic engineering, 117, 120 recombinant, discovery of, 120 Department of Biotechnology (DBT), 86, 87 Department of Science and Technology (DST), 87, 88 De Ruier, 135 Developed countries IPRs in, 121 Diamond v. Chakrabarty, 121

Dietary fiber’s health benefits, 267 Disability-adjusted life years (DALYs), 45–47 Distribution effects, poverty and, 40–42 Diversity analysis, 4, 5–6 DMRKYNETIC, data collection, 155 DNA-based techniques impact of, 19 schematic of crop improvement, 2–3, 3f DNA-marker technologies, 4–5 DNA sequencing, 4, 17 Dow AgroSciences, 134, 137, 140 Drought tolerance, in plants, 10–12 Drought-resistant GM rice adoption of, 288 Drought-Tolerant Maize for Africa (DTMA), 73 Drug approval process, 363 DST. See Department of Science and Technology (DST) DTMA. See Drought-Tolerant Maize for Africa (DTMA) DuPont, 137 Economic analysis, of GM food, 244 Economic damages contamination of supply chains, 361 cross-pollination, 206 Economic growth demand for food, 235 Economic implications adopting corn cost estimates, 322 Economic loss US and Canadian, 203 Economic model of soybeans market, 288 Economy insect-resistant GM crops, 38–40 Eggplant, Bt, 8, 105–108, 106t ‘‘863’’ program, 85 Energy demand, 227 Environmental benefits of GM crops, 175–177 Environmental Impact Quotient (EIQ), 66, 180–181 Environmental Protection Agency (EPA), 132, 204 Environmental safety issues of GM crops, 184–186 EPA. See Environmental Protection Agency (EPA) Escherichia coli, 271

Subject Index Ethanol, 15, 229 production, 228 European consumers, distrust of government, 271 European corn borer (ECB), 157 European Food Safety Agency (EFSA), 204 European Food Safety Authority (EFSA), 185 European Union (EU), 202 genetically modified seed, approval of, 231 GM field trials in, 124–125 GM food commodities, 344 precautionary principle, government, 341 regulatory model, 202 European Union regulatory framework, 356 Ex ante restrictions, 338 Exporting GMcrop-producing countries, 302 Exports of cotton, 100–101 FAO. See Food and Agriculture Organization (FAO) Farmer positive transaction costs liable for, 214 not liable for, 214 prohibitive transaction costs liable for, 212–213 not liable for, 212 property rights, distribution of liable for damages, 215–216 not liable for damages, 215 zero transaction costs liable for, 213–214 not liable for, 213 Farmer Field School (FFS), 96 Farmers planting seeds with rootworm protection, 155 Farm-level impacts of Bt cotton, 59–63, 60t, 65–66 FDA. See Food and Drug Administration (FDA) Feed production, increasing efficiency in, 176 Fermentable sugars lignocellulosic biomass to, conversion from, 15–16 Fertilizer use and GM crops, 182–184 FFS. See Farmer Field School (FFS) 4-Firm concentration ratios (CR4), 149 concentration ratios of, 150

373

Flavr Savr tomato, 7 FLD. See Front-Line Demonstrations (FLD) Flooding, 6 Food adoption, 253 availability, GM corps and, 32–34 biofuel, impacts of, 231–239 security. See Food security Food and Agriculture Organization (FAO), 30, 31f, 68–69 food balance sheet data, 32 Food and Drug Administration (FDA), 9 Food and Drug Authority (FDA), 204, 341, 355 Food consumers, 232 Food market, with biofuel mandate, 233 Food safety data, 316 Food security, 226 genetically modified crops and, 29–49 Fossil fuels, 225 Foundation seed companies, 118 Free-market system, 202 Frey, K.J., 139 Front-Line Demonstrations (FLD), 98 Fuel availability of, 227 biofuel, impacts of, 231–239 demand and supply, 227 Funding, for molecular biology research, 142 Garrod, Archibald, 119t, 120 Gasoline, 232 markets, 234 prices, 237 GEAC. See Genetic Engineering Approval Committee (GEAC) Generalized Herfindahl–Hirschman indices (GHHI), 151 Genes Bacillus thuringiensis, 8 defined, 118 transfer of, 9 Genetically modified (GM) crops adoption rates of, 292 alfalfa, 205 in Asia, 57–63 coexistence regulations, 202–205 economic implications of, 206 implications of, 215–216 liability rights, 209–214 property rights systems, equivalence of, 216–219 regional value of, 208–209

374

Subject Index

commercialization of, 36–42, 264 commodities, 311, 313 cookie, 255 cotton, 288 crop effects, 230 crop producers, 313 digression, 361 drought-tolerant varieties, 44 effects of, 44 environmental benefits of, 175–177 environmental safety issues of, 184–186 ex-ante regulations, 207 ex-post liability rules, 207 farmer liable for damages, 215–216 not liable for damages, 215 fertilizer use and, 182–184 and food availability, 32–34 and food security, 29–49 free-market system, for farmer with zero transaction costs, 202 future impacts, 42–47 generic model, assessing coexistence, 206–208 innovation, risk, precaution, 339–344 insect-resistant, 38–42 institutional/policy issues, 47–49 international agreements, 358–359 in Latin America, 66–68 nutritionally enhanced, 44–47 and nutritional value, 35–36 pesticide use effects of, 180–182 pest-resistant, 43–44, 43t in pipeline in China, 103t, 104–105 planting of, 202 precaution/innovation, lessons for, 339 pre- and post-release remedies, 363–364 pre-release testing, 362–363 product differentiation and labeling regime, 264–266 consumer choice, effects on, 266–268 differentiation costs, distribution of, 268–270 productivity shocks rice, use of, 291 regulatory practice, 355–357 research pipeline, 68–75, 71t–73t rice, 74–75, 104–105 simulation of rice adoption, 294–299 wheat adoption, 299–301 in South Africa, 63–66

trade modeling, 293–294 transaction cost, types of, 209 positive, 214 prohibitive, 212–213 zero, 213–214 use of, 284 wheat, 75 yield effects of, 178–180 Genetically modified (GM) farming coexistence value (vc) of, 208 Genetically modified (GM) food consumer preferences for, 259 consumers concern, 250–252 exporters and importers, 330, 331 knowledge of consumer WTP for, 257 labeling policies, 257, 264 effect of, 290 labels/bans, 257 mandatory labeling laws, 264 policies, 258 production and consumption, 258 trade-related regulations of, 311 U.S. and European preferences for, 254–256 Genetically modified (GM) hybrids market share of, 159 Genetically modified (GM) labeling laws, 265 Genetically modified (GM) markets, 264 Genetically modified (GM) technology adoption rate of, 212 costs for, 215 profitability of, 160 Genetically modified (GM) wheat, 288, 294 Genetically modified (GM) wheat adoption, 299–301 in Asian countries, 300 Genetically modified organisms (GMOs), 347. See also Transgenics innovations, 352 pre- and post-release strategies, 348 precautionary principle prohibitions, 354 biodiversity, 355 food safety and human health, 355 resistance, 355 technical innovation, risk, 352 biodiversity, 353–354 food safety and human health, 353 resistance, 354 Genetic Engineering Approval Committee (GEAC), 8, 43, 90 Genetic modification (GM) crops

Subject Index scientific discoveries, foundation for, 118–120 varieties, adoption of, 125–129 recombinant DNA in, discovery, 120 traits for major crops, 122–125 pricing and benefit distribution from, 129–131, 130f Genetics, 2–4 marker-assisted analysis, 5 Genetics, public sector research expenditures on, 141, 142, 143t Gene-transfer techniques, 5 Genomic selection, 5, 16–17 Germany’s, environmental movement, 340 GHG emission effects, tillage and, 186–188 Global climate change, 225 Global trade effects crop modeling framework, 287 Glyphosate, 64, 123 resistance, 181 GM crop-adopting countries, 293, 298 non- GM products in, 294 GM crops. See Genetically modified (GM) crops GM crops, environmental benefits and costs of, 177 environmental safety issues of GM crops, 184–186 fertilizer use and GM crops, 182–184 pesticide use effects of GM crops, 180–182 tillage and GHG emission effects, 186–188 yield effects of GM crops, 178–180 GM farmer. See also Genetically modified (GM) crops compensation payments, 218 incentives for, 213 liable, 212 for damages, 215 farm-level coexistence values, 211 not liable farm-level coexistence values, 210 property rights for, 202, 206, 214, 220 GM/non-GM cost wedge, 276 GM/non-GM marketing strategy, 275–277 GM/non-GM price wedge, 274 GM organism (GMO), 204 Golden Rice, 13–14, 35, 44, 45–47 Greenhouse gas (GHG) emissions, 226 biofuel, impacts of, 231–239 Greenpeace, 272 Green Revolution, 32, 230

375

HarvestPlus initiative, 14 Health effects, adoption of Bt cotton in China, 96–97 Henan, 93 Herbicide tolerance, 8 Herbicide tolerance 1 (HT1), 158 Herbicide-tolerant (HT) crops, 36–38, 69, 229 Herbicide tolerant (HT) genes, 116 Herbicide-tolerant (HT) maize, 64–65 Herfindahl–Hirschman index (HHI), 151 Herfindahl indices (Hii), 164 Hi-Bred Corn Company, 135 High erucic acid rapeseed (HEAR), 205 Holden Foundation Seeds, 118 HR canola cultivation, 179 HT crops. See Herbicide-tolerant (HT) crops HT genes. See Herbicide tolerant (HT) genes HT maize. See Herbicide-tolerant (HT) maize Huazhong Agricultural University, 139 Hybrid corn seeds, genetically modified price-dependent demand, 154 spatial pricing, 149 biotech and seed firms, 152 data, 155–157 econometric results, 162–165 estimation, 157–162 implications for, 165–166 market structure, role of, 153 model, 153–155 oligopoly structure of, 151 U.S. agricultural biotechnology seed markets, 150 Hybrid seeds corn, 117, 149 oligopoly structure of, 151 ICAR. See Indian Council of Agricultural Research (ICAR) Identity preservation (IP) systems, 310 IIATA. See International Institute of Tropical Agriculture (IIATA) Importers’ regulations, 289 Import regulatory systems, 317 Inbred lines, 118 Income agricultural, 34 distribution, poverty and, 40–42 increased farm, in India, 99–101

376

Subject Index

India adoption of GM cotton in, 89–92, 90f, 92t agricultural biotech R&D investments in, 86, 87–88 Bt cotton adoption, 40, 42, 128f, 129 Bt eggplant in, 42–43 cotton exports, 100–101 exporter of rice, 298 farmer suicides in, 61–62 GM crops research in, 105–108, 106t GM food crops, 290 GM wheat adoption, 301 regulatory experts, 291 transgenic crops in, 8 yield gains, 230 Indian Council of Agricultural Research (ICAR), 87 Indonesia adoption of GM rice, 298 GM food crops, 290 Innovations, 341 Insecticide, usage of, 96 reduction in, 97–98, 99f, 99t Insect-resistant (Bt) corn, 266 Insect-resistant GM crops, 38–42, 69–70 Insect-resistant (IR) crops, 229 Insect-resistant traited hybrids, 153 Integrated pest management (IPM), 96, 98 Intellectual property (IP), protection of, 116 Intellectual property rights (IPR), 40 in plants, 120–122 International Center of Genetic Engineering and Biotechnology, 87 International Convention for Protection of New Varieties of Plants, 121 International Dairy Foods Association (IDFA), 276 International Food Policy Research Institute, 67 International Institute of Tropical Agriculture (IIATA), 73 International labeling policies, 264 International Maize and Wheat Improvement Center (CIMMYT), 73 International Potato Center, 14 International Union for the Protection of New Varieties of Plants (UPOV), 121 Investments in agricultural biotechnology, 85–92 agricultural biotech R&D in India, 86, 87–88 in Chinese agricultural biotech, 85–86

IP. See Intellectual property (IP) IPM. See Integrated pest management (IPM) IPR. See Intellectual property rights (IPR) Jacob Hartz Seeds, 134 JK AgriGenetics Ltd, 91 Labeling policies, 324 Labeling regulations, 264 Latin America GM crops in, 66–68 Lerner indices, 155, 166 simulated effects of, 167 Life science companies, 133 Lignocellulosic biomass, 15 Livestock genetic engineering, 10 Living modified organisms (LMOs), 358 advance informed agreement (AIA) of, 358 trades in, 358 transboundary movements of, 321 Maharashtra Hybrid Seed Company, 91 Mahyco, 91, 97 Mahyco Monsanto Biotech (MMB), 91, 101 Maize Bt, 39, 40 as major crop in South Africa, 64 GM, 230 streak virus and MAS, 7 Makhatini Flats, farmers in, 65–66 Marker-assisted selection (MAS), 3, 5, 6, 12, 14, 16, 18 maize streak virus and, 7 Market concentration own/cross, 166 simulated effects of, 168 Market effects, 289 Market risks, 361 economic assessment of, 286 Market strategy non-rbST dairy products, 277 Marshallian consumer surplus, 316 MAS. See Marker-assisted selection (MAS) MDG. See Millennium development goal (MDG) Mendel, Gregory, 118, 119t, 120 Metahelix, 91 Micronutrient deficiencies, 35–36 Microorganisms, 16 Microsatellites, 4 Miescher, Johnann, 119t, 120 Millennium development goal (MDG), 29

Subject Index MIRAGE CGE model, 293 MMB. See Mahyco Monsanto Biotech (MMB) Molecular biology research expenditures on, 141, 143t Molecular marker, 4–5, 6–7 MON810, 133 Monopolies and Restrictive Trade Practices Commission (MRTPC), 101 Monsanto, 123, 133, 134–135 Bt traits for cotton varieties from, 129 China with, 142 merger and acquisition tree of, 136f on R&D, 140 RR trait, Pioneer and, 137 Morgan, Thomas, 119t, 120 MRTPC. See Monopolies and Restrictive Trade Practices Commission (MRTPC) Mycogen Corporaton, 137 Nath Seeds Ltd, 91 National Center for Plant Genome Research (NCPGR), 87 National Crime Records Bureau, 61 Natural resource degradation of, 56 NavBharat Company, 90 NB-151, Bt cotton hybrid, 90 NCPGR. See National Center for Plant Genome Research (NCPGR) ‘‘973’’ program, 85 Non-Bt corn, 360 Nongenetically modified foods willingness-to-pay (WTP) premiums for distribution of, 245 study/production characteristics, effect of, 246 Non-GM claims Europe food and nonalcoholic beverage product, 276 Non-GM crops. See also Non-GM farmers cost of, 269, 270 to Europe, 203 for export markets, 283 planting of, 202 value of, 212 Non-GM differentiation costs, 269 Non-GM farmers, 202 benefits for, 207 compensation payment to, 219 cost-minimizing behavior, 212 incentives for, 216 property rights, 206, 208 US/Canadian, economic loss for, 203

377

Non-GM labeling U.S. food and nonalcoholic beverage product, 265 Non-GM markets, 264 strategies of, 277 Non-GM producers/manufacturers, 270 Non-GM product, cost/price wedge, 277 Non-GM rice opportunity cost, 303 Non-GM strategy, in labeling countries, 270 affordable to consumers, 272–274 food supply, safety of consumer confidence, 271–272 government regulators consumer confidence, 271–272 market momentum, 275–277 market outcome, 277–278 wholesome competition, 274–275 Non-GM wheat opportunity cost, 304 Non-irrigated land, 240 Northrup King, 138 Novartis, 133, 137–138 Nutrition value, GM crops and, 35–36 Oil extraction, 227 Optimum Quality Grains LLC, 137 Ordinary risk management (ORM) ex ante outcome, 345 risk assessment, 345 risk management, 343 safe until proven harmful, 346 weaknesses of decision framework, 345–346 precaution scope, 346 risk assessment, 345 Organic food market, 248, 249 Organic milk, 253 Origin Seed, 85 Orphan crops, 68 Paraguay transgenic crops in, 8 Paris Convention for Protection of Industrial Property (1883), 120–121 Pesticide use effects of GM crops, 180–182 Pest-resistant GM crops, 43–44, 43t Peterson Seed Company, 137 Pharmaceutical products, 341 Pharmaceuticals from transgenic plants, 14–15 Pharmacia Corporation, 135

378

Subject Index

Pharmacia & Upjohn, 135 Pharming, 14 Phenotype defined, 3 optimization, 11 Philippines adoption of GM rice, 298 Bt maize in, 40 Pioneer-DuPont, 134, 140 Pioneer Hi-Bred Corn Company, 118, 134–135, 135, 137 Pioneer Hi-Bred International v. Holden Foundation Seeds Inc., 118 Plant breeding research SYs, in, 139, 140t, 141t, 142t SYs in US, 139 Plant-incorporated protectant, 132 Plants. See also Agricultural crops breeders’ rights, 121 conventional breeding in, 5 drought tolerance in, 10–12 genetic modification of, 12–14 HT in, 123 IPR in, 120–122 pharmaceuticals from, 14–15 transgenics, 11 Plant Variety Protection Act (1970), 121 Plant Variety Protection Certificates (PVPC), 121 Plant Variety Protection Law, 86 Pollen, 5 Polychlorinated biphenyls (PCBs), 339 Post-release surveillance, 350 Potato industry export loss of, 318 Poverty and distribution effects, 40–42 reducing, agricultural technology and, 34 Precaution, scope for, 345 Precautionary principle (PP) coherent framework, 346–347 controversies critiques, in context, 344 legalistic critique, 342 scaredy-cat critique, 343 scholastic critique, 342–343 trade critique, 344 definitions of, 340 on disproportionate threats, 341 in ETR framework, 347 ex ante risky innovations, 352 Germany, 340 innovation and, 350–352 lessons for

pre/post-release remedies, 363–364 pre-release testing, 362–363 potential applications categories of, 347 prohibitions, 354 biodiversity, 355 food safety and human health, 355 resistance, 355 US and Australian skepticism, 344 Pricing from GM trait, 129–131, 130f in U.S. agricultural biotechnology seed markets, 149 Private sector, plant breeding research, 138–144 Pro/anti-GM, 330 Property rights, distribution of liable for damages, 215–216 not liable for damages, 215 Property rights system damage compensation, claiming for, 220 Public sector, in plant breeding research, 138–144 PVPC. See Plant Variety Protection Certificates (PVPC) QTL. See Quantitative trait loci (QTL) Quantitative trait loci (QTL), 7, 12 Quantitative traits, 7 RbST, cost/price wedge, 278 R&D China in agricultural, 142–143 expenditures on, 141, 142, 143t funding for, 142 investments for, 140–141 private and public, SYs, 139–140, 140t, 142t Research pipeline, 68–75, 71t–73t Restriction fragment-length polymorphisms (RFLPs), 4 RFLPs. See Restriction fragment-length polymorphisms (RFLPs) Ricardian rent model, 178 Rice genetically modified, 74–75, 104–105 Golden. See Golden Rice Rice adoption, GM, 288, 294–299 in Asian countries, 297 Rice production, 298 Rockefeller Foundation, 11, 13 Rogers Brother, 138 Rootworm (RW), 158 Roundup Readys (RR) trait, 137

Subject Index Roundup-resistant weeds, 360 RR trait. See Roundup Readys (RR) trait Saccharomyces cervisiae, 16 SAES. See State agricultural experiment station (SAES) Salt-tolerant crops, 291 SAM. See Social accounting matrix (SAM) Sandoz, 138 Sanitary and Phytosanitary (SPS) Agreement, 317 Scientists years (SYs), in plant breeding research, 139, 140t, 141t, 142t Screening, pre-release testing, and postrelease surveillance (STS) process, 349–350 protocol, applying, 359–361 SCS approach, 152 Seed expenditures, 1995–2009, 118 Seed industry, transformation of research and, 131–138 Seeds, price for, 161 Seminis, 135 Simple sequence repeats (SSRs), 4 Single nucleotide polymorphisms (SNPs), 4 SmartStaxtt, 129 SNPs. See Single nucleotide polymorphisms (SNPs) Social accounting matrix (SAM), 41 Socioeconomic development, 178 Socio-economic effects, 352 South Africa Bt maize hybrids in, 40 GM crops in, 63–66 transgenic crops in, 8 Soybeans for GM trait development, 122 GM varieties adoption, in US, 126–127, 127f herbicide-tolerant crops, 37, 67–68 HT, adoption in Brazil, 127, 128f patenting activity on varieties for, 125 public/private research, SYs in, 139, 140t, 142t seed company Asgrow Seed Company, 135 Jacob Hartz Seeds, 134 SSRs. See Simple sequence repeats (SSRs) Stacked traits, 129 Starlink corn, 359 withdrawn from market, 360 State agricultural experiment station (SAES), 140 Sugar beets, GM, 205

379

Sugars, fermentable lignocellulosic biomass to, conversion from, 15–16 Suicide, farmers in India, Bt cotton cultivation and, 61–62 Sun’s energy, 225 Swarna, 6, 7 Sweet potato, in Africa, 14 Syngenta, 134, 137–138, 140 SYs. See Scientists years (SYs) Tatum, Edward, 120 Technology adoption, GM, 311 TFP. See Total factor productivity (TFP) Tillage and GHG emission effects, 186–188 Total factor productivity (TFP), 100 Trade filter, 294 Trade regulatory, GM, 301, 312 basic model, 312–316 biosafety protocol information requirements, 321–324 complexity of, 312 GM crops, 301 GM food, 311 important regulation, 316–321 marketing regulations equilibrium prices, 331–332 and standards, 324–329 trading environment challenges of, 329–331 Trade restriction from GM crop, 302 Traditional breeding. See Conventional breeding Trait-based markers, 3, 5 Transaction costs profits, 217 Transaction costs, types of, 209 positive, 214 liable for, 214 not liable for, 214 prohibitive, 212–213 liable for, 212–213 not liable for, 212 zero, 213–214 liable for, 213–214 not liable for, 213 Transgenic crop adoption effects of, 230 Transgenic crops farm production, 231

380

Subject Index

Transgenics, 2, 5 animals, 9–10 plants, 11 products, 7–8 Transgenic seeds impact of, 229 Triple/quadruple-stacked hybrids, 160 Two-stage least squares (2SLS), 160 econometric results, 162 robust standard errors, 163 United Nations millennium development goal of, 29 United States (US) adoption gains, 288 for adoption of GM field crops, 126 biofuels, 227 economic loss, 203 ethanol production, 228, 229 GM food commodities, 344 GMOs, regulatory framework, 356 GM varieties adoption, 127f Patent Office, agricultural biotechnology patents by, 116, 117f PVPC in, 122 regulatory model, 202 SYs, in plant breeding research, 139 transgenic crops in, 8 unapproved GM rice, 202, 231 UPOV. See International Union for the Protection of New Varieties of Plants (UPOV) U.S. Corn Belt, map of, 156 U.S. Department of Agriculture (USDA), 204, 271 US Court of Customs and Patent Appeals, 121 USDA Animal and Plant Health Inspection Service, 356 VAD. See Vitamin A deficiency (VAD)

Vegetable oil production, 238 Veterinary Services National Center for Import-Export, of APHIS, 123 Vitamin A deficiency. See Vitamin A deficiency (VAD) sweet potato and, 14 Vitamin A deficiency (VAD), 44–47 Von-GM production differentiation costs of, 273 Vunisa Cotton, 65, 66 Water Efficient Maize for Africa (WEMA), 73 Watson, James, 119t, 120 Welfare effects changes, 297 import approval, 319 producer vs. consumer, 320 three scenarios, 328 WEMA. See Water Efficient Maize for Africa (WEMA) Wheat genetically modified, 75 Willingness-to-accept distribution of consumers, 255 Willingness-to-pay (WTP) consumers, 244–250 for GM food, 250–252, 252–254 consumer studies, 249 cost–benefit analyses, 253 GM and non-GM foods, 250 meta-analysis, 246 for nongenetically modified foods, 246, 247 premium for non-GM food, 245 World Bank, 46 World Trade Organization (WTO), 317 Yield effects of GM crops, 178–180 Yield-increasing technologies, 34–35