Growth Miracles and Growth Debacles: Exploring Root Causes

Growth Miracles and Growth Debacles

M2545 - BHATTACHARYYA PRINT.indd i

02/03/2011 10:09

To Anna, Sahana and Swapan

M2545 - BHATTACHARYYA PRINT.indd ii

02/03/2011 10:09

Growth Miracles and Growth Debacles Exploring Root Causes

Sambit Bhattacharyya Research Fellow, Department of Economics, University of Oxford, UK

Edward Elgar Cheltenham, UK • Northampton, MA, USA

M2545 - BHATTACHARYYA PRINT.indd iii

02/03/2011 10:09

© Sambit Bhattacharyya 2011 All rights reserved. No part of this publication may be reproduced, stored in a retrieval system or transmitted in any form or by any means, electronic, mechanical or photocopying, recording, or otherwise without the prior permission of the publisher. Published by Edward Elgar Publishing Limited The Lypiatts 15 Lansdown Road Cheltenham Glos GL50 2JA UK Edward Elgar Publishing, Inc. William Pratt House 9 Dewey Court Northampton Massachusetts 01060 USA

A catalogue record for this book is available from the British Library Library of Congress Control Number: 2010934015

ISBN 978 1 84844 631 1

03

Typeset by Servis Filmsetting Ltd, Stockport, Cheshire Printed and bound by MPG Books Group, UK

M2545 - BHATTACHARYYA PRINT.indd iv

02/03/2011 10:09

Contents Preface Acknowledgements 1

Introduction

PART I 2 3 4 5

7 8 9

1

HISTORY AND ECONOMIC DEVELOPMENT

The Great Divergence: an account of the growth miracles and growth debacles since AD 1000 Theories of root causes of economic progress Empirical evidence Root causes of economic progress: a unifying framework

PART II

6

vi viii

9 15 48 95

PROMOTING GROWTH IN THE CURRENT ENVIRONMENT: EVIDENCE AND POLICIES

Institutions and trade: competitors or complements in economic development Improving institutions with trade policy: myth or a possibility Which institutions matter most for economic growth? Making policy work: a road map for future growth

Data appendix References Index

119 140 147 166 179 184 199

v

M2545 - BHATTACHARYYA PRINT.indd v

02/03/2011 10:09

Preface No new light has been thrown on the reason why poor countries are poor and rich countries are rich. Paul Samuelson (1976), Illogic of Neo-Marxian Doctrine, p. 107.

Wide variations in living standards are observed across countries. Going further back in time, we notice that this variation is largely attributable to economic performance across countries since the start of the sixteenth century and not just how they performed post-Second World War. This book presents, in two parts, a detailed account of this process of divergence, the data involved and policies for the future. The first part opens with theories of institutions, geography, human capital, trade, religion and culture, and state formation and war explaining this divergence. Following which, it discusses some empirical results and illustrates the difficulties in quantifying the relationships between root causes and economic development. The novelty of the book is the ‘unifying framework’ which explains the process of development in Western Europe. This framework is also compared with growth narratives in other countries and continents namely Africa, China, India, the Americas, Russia and Australia. A narrative style is adopted throughout to create a bridge between the empirical literature and history. The ‘unifying framework’ is an attempt to merge all seemingly disparate theories of economic progress. The main message is that diseases and geography matter at an early stage of development. Institutions, however, become much more important as the economy develops. Geography and, in particular, disease epidemics are a crucial explanator of lack of development in Africa. In contrast, in China and India, the Malthusian population growth and disease cycle was broken fairly early and institutional weaknesses played a crucial role in their respective declines. In the Americas and Australia, colonial institutions were a crucial factor. In Russia, it was crippling political institutions of the nineteenth century and restrictive political and economic institutions of the Soviet Union that did the damage. The second part of the book focuses on growth promoting policies. First, it documents some macro evidence on the role of policy in yielding growth. It shows that trade can benefit nations in situations when institutions are adequately strong. Property rights and contracting institutions vi

M2545 - BHATTACHARYYA PRINT.indd vi

02/03/2011 10:09

Preface

vii

are good for growth. Market stabilizing institutions are good for growth and regulations are important only up to a certain extent. Second, it outlines growth promoting policies namely the ‘first principles of growth’. They are property rights, contracts, regulatory institutions, rule of law, macroeconomic stabilization, representative politics, human capital investments, market access and international trade. Third, it discusses the cases of India and China, two recent success stories. It shows how these countries have preserved incentives for private investments even without rigidly following the first principles. But more importantly, they have been able to create institutions which are well grounded in local traditions and culture and are also able to create appropriate incentives for investments. Fourth, the book outlines steps that could be taken to facilitate growth in situations of state failure, disease trap, poverty trap and scarcity of skilled workers.

M2545 - BHATTACHARYYA PRINT.indd vii

02/03/2011 10:09

Acknowledgements I gratefully acknowledge discussions with my colleagues and friends at different stages of development of this project. In particular, I would like to thank Jeffrey Williamson and Tim Hatton for sparing their time to discuss some aspects of this project. I would also like to acknowledge help and encouragement from Prema-chandra Athukorala. I have benefited from comments by and discussions with Akihito Asano, John Braithwaite, Steve Dowrick, Gary Magee, Mark Rogers and Jonathan Temple. I also gratefully acknowledge financial support from Jonathan and Jennifer Oppenheimer. Last, but not least, I would like to thank my wife and family for their love, affection and words of encouragement. The majority of the manuscript was written when I was a research fellow at the Arndt-Corden Division of Economics of the Australian National University.

viii

M2545 - BHATTACHARYYA PRINT.indd viii

02/03/2011 10:09

1.

Introduction

Hidaya Mohamed is a fifth grade student. She dreams of going to the Benjamin Mkapa Secondary School, one of the better government schools in her town, but no one in her family has received secondary education before. Her older brother does not go to school and her sister is in a government primary school. Her mother, a single parent, earns a living selling maandazi (traditional buns) which is not enough to finance Hidaya’s dream. ‘I don’t think I can go there,’ reflects Hidaya. ‘My mother can’t pay her contribution and I don’t know who can help me.’1 Hidaya’s family resides in Tanzania. The average income per capita in Tanzania today is $1141, which is approximately one-sixty-third of what it is for Luxembourg, the richest nation in the world.2 Tanzania is poor in spite of its rich natural resource endowments of diamonds, gold, iron ore, coal, natural gas and nickel. Tanzanian soil is suitable for coffee, cotton and clove plantations. In spite of all this, the Tanzanian development record is disappointing. A colonial history marred with slave trade reveals a very sad tale of exploitation and underdevelopment. Even after independence in 1961, things did not change much. Almost all of the small and big development initiatives of the government in independent Tanzania culminated in disastrous failures. One of the better known accounts of disaster is that of the Morogoro Shoe Factory. The Morogoro Shoe Factory was established in Tanzania with the help of the World Bank in the 1970s. It was endowed with labour, machinery and the latest shoe-making technology. But it hardly produced any shoes, only utilizing 4 per cent of its capacity, largely due to the absence of production incentives for the firm (Easterly, 2001). The plant was not well designed either. It had aluminium walls and no ventilation system, which was unsuitable for the Tanzanian conditions. After two decades of struggle, production finally stopped in 1990.3 Tanzania now stands as one of the poorest countries in the world, struggling with poverty, high infant mortality, HIV/AIDS and malaria. By 2001, Tanzania had accumulated external debt worth $6.7 billion which was cancelled under the Heavily Indebted Poor Countries (HIPC) Initiative.4 The Tanzanian economy in its current state cannot generate enough wealth to ensure a decent living standard for someone like Hidaya. Hidaya’s story is not unique. There are many more like Hidaya

1

M2545 - BHATTACHARYYA PRINT.indd 1

02/03/2011 10:09

2

Growth miracles and growth debacles

in the poor tropics and subtropics whose minimum aspirations cannot be fulfilled because their respective economies fail to generate enough wealth for them. The obvious question that follows is the most common yet crucial in the field of economic development. Why is there a 63-fold difference in income per capita between Luxembourg and Tanzania? The question is common because it has been addressed by social scientists on numerous occasions over the last 50 years. However, it is important to note that this debate is far from being over. It is perhaps fair to say that we are only starting to understand the complex process of development. It is crucial because it has the potential to improve living standards and reduce the proportion of people who are suffering from starvation, poverty and deprivation. In the past, the main focus of the literature was to discover and evaluate proximate causes (physical capital accumulation, technological progress and so forth). In recent years, however, there has been a welcome shift in the focus. Economists have stressed the importance of root causes (institutions, human capital, religion and culture, openness to trade and geography) in their quest for growth. Ever since Francois Quesnay wrote about the giant economic machine in 1763, the development literature has always been in search of the factor that propels the machine (see Banerjee, 2007, p. 125). No wonder that phrases of the likes of ‘Holy Grail of growth’ and ‘Elixir of growth’ have been used so many times in the literature. Perhaps the idea of an all encompassing law governing society is too irresistible to the researchers of the dismal science. The recent literature on the root causes of economic progress is not an exception. A lot of effort has gone into identifying the root cause. As a result the abovementioned explanations are often posed as competing explanations of economic development. Needless to say, this need not be the case. It will be surprising if the process of development, as complex as it is, turns out to be a result of a single factor. Perhaps the most likely outcome is a series of causally interlinked explanations. I make an attempt to establish some of these linkages. In the first part of this book, I provide empirical evidence on the role of root causes. I also offer a unifying framework for Western Europe – a development success story. The framework links the seemingly competing explanations of longrun development. Then the framework is put to the test by comparing and contrasting it with the historical process of development in the Americas, Africa, China, India, Russia and Australia. It appears that the framework does well in explaining the development process in these different continents. In the second part of the book, I focus on policies that could improve growth. I provide contemporary macroeconomic evidence on policies that

M2545 - BHATTACHARYYA PRINT.indd 2

02/03/2011 10:09

Introduction

3

may have worked better for growth. I then outline the first principles (property rights, contracts, rule of law, regulation, macroeconomic stability, representative politics, access to markets, human capital investments and trade) that are essential for growth. I also discuss the case of India and China, two recent success stories, and how they have applied the first principles. The book concludes with a discussion of growth policy in situations of state failure, poverty trap and scarcity of skilled labour. All along, I take into consideration the following issues. First, the literature is rich with various explanations and theories of development. However, it is difficult to find a single source which covers the literature well. I make an attempt here to fill this vacuum. Hopefully, graduate students working on this topic will find this useful. However, my attempt to cover the literature may not be exhaustive. Nevertheless, I do make a conscious effort to cover all of the major studies in history, economics and politics. Second, I make an attempt to present the material in a non-technical fashion wherever I can. However, some technical materials are presented in Chapters 3, 4, 6, 7 and 8. These models are followed by a non-technical summary of the main message. Therefore, I am assuming that this is not an insurmountable hurdle for readers outside the discipline. Furthermore, readers may also choose to skip the technical parts and focus on the summary instead. In my view, this will have very little impact on their understanding of the main message of the book. To put it simply: readers won’t miss much by skipping the technical details. Nevertheless, only time will tell how successful I am in making it within the reach of the elusive ‘intelligent layperson’. Third, for the sake of brevity, I skipped technical details in some cases. The analysis proceeds in two parts. The first part focuses on the role of history and the second part is more concerned about contemporary growth policy. Chapter 2 in Part I introduces the core empirical fact in the literature – the Great Divergence. Using simple diagrams from previously published research, I take a brief look at the growth miracles and growth debacles of the last millennium. I give special attention to the economic performance of North America, South America, Africa, the United Kingdom, Western Europe, Southern Europe, Russia, China, India and Indonesia. I also revisit the ‘reversal of fortune’ hypothesis proposed by Daron Acemoglu, Simon Johnson and James Robinson. Chapter 3 critically evaluates theory and evidence on the root causes of economic progress. It begins by distinguishing between root and proximate causes. It also shows how the idea of root causes can be augmented into a standard Solow growth model. Some of the prominent political economy models and theories of institutions are reviewed in this chapter. It is followed by a review of theory and evidence on the role of institutions, the role of religion and culture, the role of geography, the role of

M2545 - BHATTACHARYYA PRINT.indd 3

02/03/2011 10:09

4

Growth miracles and growth debacles

trade openness, the role of human capital and the role of state formation and war. Chapter 4 presents empirical evidence. It highlights some of the problems related to the empirical estimates in the root causes literature. It shows that malaria is crucial in explaining long-run development in Africa. It shows that factors such as institutions are statistically insignificant in an African sample. Chapter 5 presents the ‘unifying framework’. The framework describes the trajectory of capitalist development in Western Europe. It compares and contrasts the Western European trajectory with the trajectories of Africa, the Americas, China, India, Russia and Australia and tries to establish what deep structural factors are behind the development or lack of development of these regions. My conjectures are backed by the existing empirical evidence. Empirical evidence drawn from published research is presented with limited technical details. However, the source research papers are cited so that interested readers can follow it up if they please. In this chapter, I also try to present a case that the African process of development is different from the rest of the world. I argue that high incidence of malaria has an adverse impact on household savings which dampens growth. Part II of the book focuses on contemporary policymaking. Chapter 6 presents evidence on the role of trade in promoting development over the last two decades. It shows that trade is effective only when a country is above a certain threshold level of institutional quality. Chapter 7 shows that a country’s institutions could be improved through trade policy. It presents theoretical explanation and empirical evidence in support of such view. Chapter 8 tackles a related issue of institutional effectiveness. It empirically tests which institutions were more effective in delivering growth over the last two decades using international panel data. It finds that property rights, contract and macroeconomic institutions are important. Regulatory institutions are also important but overregulation is not good for growth. Political institutions come out to be statistically insignificant. Finally, Chapter 9 concludes with a list of policies essential for growth. I call this list the first principles. I argue why they are important for growth. I also discuss the Chinese and the Indian cases to illustrate how important it is to apply the first principles so that they are best suited for local conditions. I outline a way forward in situations of state failure, poverty trap and scarcity of skilled labour. Without doubt, collectively the literature has made immense progress towards understanding and identifying the root causes of economic progress. However, a lot of work remains to be done to establish linkages between factors. This book is an attempt to explore these linkages.

M2545 - BHATTACHARYYA PRINT.indd 4

02/03/2011 10:09

Introduction

5

NOTES 1. Hidaya’s story is quoted from an article titled ‘All these children want is a decent education’ by Sakina Zainul Datoo published in the October 2003 edition of the journal Africawoman on page 3. The online source of this article is the following: http://www.africawoman.net/newsdetails.php?NewsID5182&AuthorID553&CountryI D512&NewsTypeID56&IssueID524 2. These are figures for 2007 expressed in constant 2005 ‘international’ dollars adjusted for purchasing power parity (PPP) differences. The source is the World Development Indicators (WDI) Online, The World Bank. 3. The story of the Morogoro Shoe Factory is quoted from Easterly (2001, p.68). 4. The Heavily Indebted Poor Countries (HIPC) Initiative was established in 1996 as a joint collaboration between the World Bank and the International Monetary Fund (IMF).  The aim of the initiative is to reduce the excessive debt burdens faced by the world’s poorest nations. In 1999, the international community endorsed enhancements to the HIPC initiative allowing more countries to qualify for HIPC assistance, accelerating and deepening the delivery of relief, and strengthening the link between debt relief and poverty reduction.

M2545 - BHATTACHARYYA PRINT.indd 5

02/03/2011 10:09

M2545 - BHATTACHARYYA PRINT.indd 6

02/03/2011 10:09

PART I

History and economic development

M2545 - BHATTACHARYYA PRINT.indd 7

02/03/2011 10:09

M2545 - BHATTACHARYYA PRINT.indd 8

02/03/2011 10:09

2.

The Great Divergence: an account of the growth miracles and growth debacles since AD 1000

As I have illustrated in the introduction with the Tanzania and Luxembourg example, living standards do vary enormously across countries. Let’s take stock of the situation by subjecting this empirical fact to close examination. In Figure 2.1, I plot the trend in per capita income in seven different countries over the period 1950 to 2004 using data from the Penn World Table version 6.2. What I notice is, indeed, a significant variation in standard of living across countries. China, India and Brazil seem to be part of a completely different club when compared with the living standards in the United States, Australia and the United Kingdom. The other important point to note is that the difference across countries is persistent and there is very little change in relative positions since 1950. Therefore, it is quite 11 USA Australia UK

Log GDP per capita

10

9

Brazil China

8

India

Ghana

7

6 1950

1960

1970

1980

1990

2000

2010

Year

Figure 2.1

Evolution of per capita income in USA, Australia, UK, Brazil, China, India and Ghana over the period 1950 to 2005 9

M2545 - BHATTACHARYYA PRINT.indd 9

02/03/2011 10:09

10

Growth miracles and growth debacles USA 10

Japan

Log GDP per capita

9

China 8 Netherlands UK

India

Africa

7

6 0

200

400

600

800

1000

1200

1400

1600

1800

2000

Year

Source: Author’s plot using data from Maddison (2004).

Figure 2.2

Evolution of per capita income in USA, UK, the Netherlands, Japan, China, India and Africa over the period AD 1–2003

evident that the origin of this divergence is certainly not the post-war period of 1950 to 2004. To trace out the origin, we need to look further back. Looking further back in time, of course, is a challenge. Do we have the data to do this successfully? The answer is yes, thanks to the excellent research done by economic historians over the last four decades. The most widely used numbers are from Angus Maddison’s (2004) The World Economy: Historical Statistics. However, there is no consensus among historians regarding the validity of these numbers. Some historians argue that the values for China, Japan and other parts of Asia were comparable or even higher than those in Western Europe in the late nineteenth century. Needless to say, Maddison’s figures show the opposite. I shall come back to some of these debates very briefly later. Nevertheless, this debate is not central to my story and the majority of historians agree on the broad patterns of evolution of income across countries since AD 1. In Figure 2.2 we plot the evolution of per capita income since AD 1 in the United States, United Kingdom, the Netherlands, Japan, China, India and Africa. What we notice is quite striking. Prior to 1400, living

M2545 - BHATTACHARYYA PRINT.indd 10

02/03/2011 10:09

The Great Divergence

11

40

England Netherlands

Urbanization rate

30 Italy 20 France

10 Russia Japan India China

0 1000

1100

1200

1300

1400

1500

1600

1700

1800

1900

Year

Source:

Author’s plot using data from Acemoglu et al. (2002).

Figure 2.3

Urbanization in the world powers over the period AD 1000–1850

standards across countries were largely comparable. Countries diverged from each other from the sixteenth century onwards. This difference has been persistent and widening, particularly over the last two centuries, especially since the Industrial Revolution in the United Kingdom. The only exception is steady growth in India and China over the last two decades. But this, without doubt, is not enough to match the standards of living in the United States and United Kingdom any time soon for these countries. Indeed, the developed countries are a long way away. Therefore, the gap that we notice in Figure 2.1 is more of a continuation of the divergence that started in the sixteenth century. If we compare urbanization across major world powers over the period AD 1000 to 1850, we also notice a similar trend (see Figure 2.3). The level of urbanization across the world was very similar till the twelfth century with the exception of Italy, where the city states were more urbanized than the rest of the world. Figure 2.3 shows that the Netherlands experienced rapid urbanization during the thirteenth century and thereafter, and caught up with Italy, then the most progressive and prosperous part of Europe, during the early fifteenth century. This rapid growth in urbanization was unmatched by any other parts of the world till the mideighteenth century when Britain experienced a similar trend. The British

M2545 - BHATTACHARYYA PRINT.indd 11

02/03/2011 10:09

12

Growth miracles and growth debacles Western Europe

Urbanization rate

20

15

10 Asia Eastern Europe

5

1300

1400

1500

1600

1700

1800

1900

Year

Source: Author’s plot using data from Acemoglu et al. (2005b).

Figure 2.4

Urbanization rate in Western Europe, Eastern Europe and Asia

growth in urbanization was largely propelled by the Industrial Revolution and an expansion in Atlantic trade. In contrast, the French growth in urbanization was relatively modest. This is perhaps explained by the slow progress of industrial development in the European continent following the Industrial Revolution in Britain. Landes (2003) documents why the progress of industrial development and technological change in continental Europe was so slow. Urbanization in India, Russia, China and Japan remained largely unchanged over this millennium again suggesting that the genesis of divergence between the East and the West (or the North and the South) is somewhere rooted around the fifteenth century. This trend in urbanization rate persists even when we look at aggregate data. Figure 2.4 presents urbanization rates weighted by population at the continental level from Acemoglu et al. (2005b). The diagram shows that there is very little change in urbanization rates across continents over the period 1300 to 1500. It is only after 1500 that Western Europe surged ahead leaving Asia and Eastern Europe way behind. Eastern Europe experienced some growth after 1700, but the gap with Western Europe in fact widened. The urbanization rate in Asia declined after 1600 and became on a par with Eastern Europe in 1850. So much for urbanization. How about industrial production per capita?

M2545 - BHATTACHARYYA PRINT.indd 12

02/03/2011 10:09

Industrial production per capita, UK in 1900 =100

The Great Divergence

13

400 US

300

200

Canada Australia New Zealand

100

Brazil India

0 1750

1800

1850 Year

1900

1950

Source: Author’s plot using data from Bairoch (1982).

Figure 2.5

Industrial production per capita relative to the UK in the New World and emerging economies

Figure 2.5 compares industrial production per capita relative to Britain in the New World economies and the emerging economies starting from the mid-eighteenth century. We notice that industrial production in the United States relative to Britain increased rapidly, starting in 1830. Canada also experienced rapid growth, starting in 1880. Compared to these, the growth in industrial production in the other two Anglo-Saxon colonies, namely Australia and New Zealand, was fairly modest. They did experience growth, but only in the twentieth century. The growth in the United States was rapid and persistent and, in 1953, the United States industrial production per capita reached 3.5 times the level of Britain. In contrast, industrial production in Brazil and India experienced very little growth during this period. Finally, to summarize, in this chapter I show that there is a significant gap in living standards across countries. The genesis of this gap is not the post-war period. In fact, looking at data on earlier periods reveals that countries diverged from each other in terms of living standards around the fifteenth century. Economic historians often describe this as the Great Divergence. The question, however, is what caused this divergence. We shall plunge into a pool of answers in the following chapters. But before we do that, let’s take a first pass on the type of explanations that will follow. In an influential paper published in the Quarterly Journal of Economics,

M2545 - BHATTACHARYYA PRINT.indd 13

02/03/2011 10:09

14

Growth miracles and growth debacles

Acemoglu et al. (2002) argue that the Great Divergence can be explained via institutional1 divergence. They show that the fortunes of the nonWestern countries reversed around 1500. This is because of institutional reversal. The Western colonial model set up extractive institutions in the resource rich periphery where they never intended to settle. The reverse is true for settlement colonies where they wanted to settle. So, they set up institutions that are favourable to capitalism. This led to an institutional reversal which explains the Great Divergence. Not everyone, however, agrees with this story. The human capital theorists, the social capital theorists, the free trade group and many others, all have their own explanation. I shall provide a brief review of all of these theories in the next section. I shall also explain how this can all be related to the neoclassical growth model.

NOTE 1. Acemoglu et al. (2002) mainly focus on property rights and contracting institutions. For a more detailed definition of institutions, see Chapter 3.

M2545 - BHATTACHARYYA PRINT.indd 14

02/03/2011 10:09

3.

Theories of root causes of economic progress

One of the major predictions of the neoclassical growth model is that given the level of technology and income at a particular point in time, and a common growth rate of technology, a country accumulating more physical and human capital grows faster than a country accumulating less of the same. Empirical evidence suggests that this is only a part of the story. In a simple regression model with two factors of production, physical and human capital, and log initial income as an additional control, a large fraction of the cross-national growth variance remains unexplained. This leads to two types of reaction. First, a search for additional correlates of growth with the majority of the correlates having no or very little connection with theoretical models of optimizing behaviour. Second, a move towards more structured growth empirics with the primary objective of teasing out causality. The second approach is relatively recent and has gained a fair amount of acceptance of late. However, the question of how the literature has moved in this direction remains. To find an answer, we need to trace out the origin and the evolution of the growth regression. This is exactly what I seek to do in the following section.

3.1 THE NEOCLASSICAL GROWTH MODEL Let’s kick off with an illustration of the standard neoclassical growth model and show how the growth regression emerges from it. An earlier version of this model was independently developed by Robert Solow (1956) and Trevor Swan (1956) which only focused on physical capital accumulation. This version was developed by Mankiw et al. (1992) who augment the original version with human capital to accommodate the ideas of knowledge-based growth proposed and formalized by Paul Romer in the 1980s. The model is as follows. The model considers an economy at a particular point in time t with the following constant returns to scale production technology producing income Y: 15

M2545 - BHATTACHARYYA PRINT.indd 15

02/03/2011 10:09

16

Growth miracles and growth debacles

Yt 5 F (Kt, Ht, Et) 5 Kat H bt E 1t 2a 2b

(3.1)

where Kt and Ht are the stock of physical and human capital, respectively, Et 5 AtLt is effective labour growing at the aggregate rate of exogenous population growth n plus technological progress g. Dividing both sides by Et and suppressing the time subscript t, one can rewrite the production function in per capita terms as follows y 5 k ahb

(3.1a)

where k and h are stocks of physical and human capital per efficiency unit of labour, respectively. It is assumed that a fixed proportion sk and sh of the per capita output y are invested in physical and human capital, respectively. The evolution of the economy is determined by the following equations of motion. Both physical and human capital depreciate at an exogenous rate d. # k 5 sky 2 (n 1 g 1 d) k (3.2) # h 5 sh y 2 (n 1 g 1 d) h (3.3) # # At steady state k 5 h 5 0 and the long-run steady-state equilibrium values of per capita physical capital, human capital and output are given by k* 5 a

1/1 2a 2b 1/1 2a 2b s1k 2bsbh sak s1h 2a , h* 5 c , and b d n1g1d n1g1d

y* 5 c

1/1 2a 2b sak sbh d a 1b (n 1 g 1 d)

(3.4)

The unique steady-state value is ensured by the Cobb-Douglas technology, which guarantees that the Inada conditions are satisfied. By linearizing around the steady state using the first order Taylor expansion and integrating, one can derive the following expression for country i: yˆiT 5 g1 1 g2 ln yit 1 g3XiT

(3.5)

where yˆiT ; (1/T) ln [ yiT /yit ] and the vector XiT includes the investment rates in physical and human capital, rate of depreciation of capital and the rate of growth of population. It follows from the model that the lower the level of initial output of a country, the higher will be its growth rate for a given value of XiT. In other words, all other factors remaining unchanged,

M2545 - BHATTACHARYYA PRINT.indd 16

02/03/2011 10:09

Theories of root causes of economic progress

17

poor countries grow faster than the rich countries. Therefore, the sign of g2 is negative. This is known as ‘conditional convergence’ in the literature. Therefore, the theory predicts that a part of the variation in crossnational income can be explained by the variation in investment rates in physical and human capital. In other words, the central problem in economic development is to understand the process of factor accumulation. This leaves us with an empirical question to answer. Is the prediction made by the theoretical model supported by the data? This is what motivated the expansion of the now voluminous empirical growth literature, which I discuss next.

3.2 GROWTH, FACTOR ACCUMULATION AND THE MEASURE OF OUR IGNORANCE One can test the steady-state properties of the neoclassical growth model by estimating Equation 3.5 using regression techniques. In order to do that Equation 3.5 is rewritten as follows: yˆiT 5 g1 1 g2 ln yit 1 g3XiT 1 eiT

(3.5a)

where eiT is the unobserved residual. Many empirical growth studies estimate this model using a variety of techniques. Mankiw et al. (1992), King and Levine (1994), Caselli et al. (1996), Hall and Jones (1999) and Easterly and Levine (2001) are among the most cited ones. The final conclusion, however, is loud and clear. The residual rather than factor accumulation accounts for most of the growth differences across nations. This leads to two types of research strategy in the literature. The first strategy is to increase the number of regressors to account for the omitted factors which explains the variance. The inevitable fallout of this line of research is a never ending list of regressors. This is useful if the regressors are derived from models of optimizing behaviour indicating the possibility of a strong causal relationship. Unfortunately, this is not the case. The majority of the regressors are chosen on an ad hoc basis. As a result one can identify correlates, but not causal relationships. The list of correlates includes physical capital investment, human capital accumulation, low income inequality, being located further from the tropics, fewer tropical diseases, low fertility, less government spending, trade policy openness, political freedom, British legal origin, rule of law, foreign direct investment, foreign aid, ethno-linguistic fractionalization and so forth. The list is growing over time and it is far from being exhaustive. The second strategy is to look at the deep structural features of the

M2545 - BHATTACHARYYA PRINT.indd 17

02/03/2011 10:09

18

Growth miracles and growth debacles

economy. The idea is that the deep structural features influence proximate causes (factor accumulation) over the long horizon and shape the longrun economic performance of the economy. To put it more formally, the ‘proximate causes’ are functions of ‘deep determinants’. Following is the relationship in mathematical terms: XiT ; F (ZiT)

(3.6)

where ZiT is a vector of deep determinants. Sachs and Warner (1997a) and Bhattacharyya (2004) use this structure in their models. The ‘deep structural determinants’ are trade, institutions, geography, religion and culture and knowledge. The listing of knowledge and human capital as a deep determinant, however, is debatable. Rodrik et al. (2004) and Acemoglu et al. (2001) identify human capital as a proximate determinant. Glaeser et al. (2004) in the empirical literature and Mokyr (1990) in the economic history literature identify knowledge as a deep determinant. Given the weight of evidence in some of the recent research, it is perhaps appropriate to classify it as a deep determinant. Nevertheless, it is still open to scrutiny. Therefore, one can rewrite (3.5a) as a reduced form regression: yˆiT 5 a ln yit 1 gZiT 1 eiT

(3.7)

A standard way of estimating this equation is to regress Zit on exogenous historical variables that may have influenced Zit in the past and use the predicted value of Zit from this regression to estimate (3.7). It may appear that this two-step procedure is a handy way to circumvent endogeneity problems; however, it has its own problems. The two-step procedure applied to levels regression suffers from identification problems and multicollinearity issues. Bhattacharyya (2009a) illustrates this problem in a cross-section model with institutions and human capital. Dollar and Kraay (2003) show this with institutions and trade. I shall discuss this more in the following chapters. Nevertheless, this allows researchers to estimate the direct effect of deep structural determinants on current growth. The advantage of the structured growth empiric is that it is closer to theory and it gives us useful insights into the establishment of causality. This is perhaps one of the major reasons behind the majority of the recently published research taking up this approach. However, further work needs to be done to identify the chain of causation and to support it with a general equilibrium model.

M2545 - BHATTACHARYYA PRINT.indd 18

02/03/2011 10:09

Theories of root causes of economic progress

19

3.3 ‘ROOT’ AND ‘PROXIMATE’ CAUSES OF DEVELOPMENT: WHAT ARE THE DIFFERENCES? A major criticism of the ‘root causes’ literature is that its distinction of root and proximate causes is somewhat arbitrary. Furthermore, the justification for putting institutions and diseases into the root causes basket is also unclear. Before we proceed any further with the theories of root causes, it is perhaps useful to clarify these definitional aspects right now. A quick review indicates that the literature has used at least three different ways to justify this distinction. First, root causes are more fundamental than proximate causes. In other words, they cause long-run economic progress or decline by influencing the proximate causes (see Acemoglu et al., 2001; and Rodrik et al., 2004). Second, root causes are durable, whereas proximate causes vary over time (Glaeser et al., 2004). Third, it is difficult to influence root causes through direct policy intervention, whereas one can do the same with proximate causes relatively easily (Glaeser et al., 2004). One can certainly question the merits of making this distinction, the justification provided by the literature and the validity of labelling institutions and diseases as root causes. Some of the major ones are as follows. First, is the distinction between proximate and root causes meaningful and the same in all contexts, especially when there is virtually nothing except geography that is truly fixed?1 The answer to this is yes, as the distinction between proximate and root causes relies more on causal linkages between factors than time invariance of a particular factor. The problem with relying too heavily on time invariance is that no factor is truly fixed when a big enough time frame is considered. In contrast, a time frame looking at the chain of causation may provide some useful information about which factor is more fundamental. For the purpose of this book, I tend to take the view that institutions and diseases are the root causes, as they create incentives or disincentives for investments in physical and human capital, which promote development. Although one should not write off the possibility of causation running in the opposite direction, recent evidence shows that the possibility of such occurrence is negligible.2 Second, are the root causes more durable than proximate causes as it is often the case that they change rapidly due to revolution or technological breakthrough in medicine? Data from the Polity IV project shows that economic and political institutions are indeed more durable compared to investments in physical or human or financial capital. Some empirical research, however, is guilty of using expropriation risk from the Political Risk Services, a private company which assesses the risk of expropriation of foreign investment in different countries, as a measure

M2545 - BHATTACHARYYA PRINT.indd 19

02/03/2011 10:09

20

Growth miracles and growth debacles

of institutions. This measure is not durable as it assesses institutional outcome rather than institutions themselves, and it is also a measure based on perception (Glaeser et al., 2004). But, overall, there is evidence both in pre-independence historical data (Acemoglu et al., 2005b) and postindependence data from Polity IV that institutions persist. The changes that we observe are due to exogenous shocks consistent with the critical juncture view. This, however, may not be the case with diseases. But it is perhaps fair, at least for the purpose of this book, to classify diseases as a root cause when we are looking at the whole stretch of development history, including the period of primitive hunter-gatherer societies. After all, back then, diseases were very much a result of geography, rather than lack of intervention of medical science. Third, there is evidence that policy influences institutional quality and disease environment over time. Therefore, what is the use of labelling institutions and diseases as root causes when it can also be influenced by good policy like many other proximate causes? This is perhaps the most difficult case to argue against. Recent research on policy and institutional development show that good policy does influence institutions in the short run (Thelen, 2000). However, the benefits of good policy never get translated into long-run economic development unless they are locked into the institutional structure. Examples of short-lived policy-driven growth with poor institutions are not that difficult to find. Argentina in 1870s to 1920s, Czarist Russia in the decade leading up to World War I, Colombia during 1900 to 1940s, Côte d’Ivoire in the first two decades after independence – all experienced short-lived growth when the benefits of good policies were not locked into institutions (Robinson, 2006). Therefore, what is critical for institutional change is to lock in either the benefits of good policy or the positives of an exogenous shock at a critical historical juncture. For my purpose, however, it still makes sense to label institutions as a root cause, since policies are short lived compared to institutions in the time frame that I am looking at. I can present the same argument in the case of diseases, since I start from a period when the Malthusian cycle was operational and there is no recorded evidence of public health intervention that far back in time.3 Nevertheless, we shall come back to some of these details in Chapter 4, when I present the unifying framework to explain development history.

3.4 THEORIES OF DEEP STRUCTURAL DETERMINANTS In this section, I review the theories and empirics of deep structural determinants of economic progress. This provides a background for the

M2545 - BHATTACHARYYA PRINT.indd 20

02/03/2011 10:09

Theories of root causes of economic progress

21

unifying structure to be presented in Chapter 4. Theories of state, institutions, religion and culture, geography, trade and human capital and knowledge are reviewed. I include human capital and knowledge in this list in order to review the theories that strongly argue knowledge is a deep determinant of long-term economic development.

3.5 THE INSTITUTIONS THEORY Little else is required to carry a state to the highest degree of opulence from the lowest barbarism, but peace, easy taxes and tolerable administration of justice; all the rest being brought about by the natural course of things – Adam Smith (cited in Jones, 1981, p. 253)

This is what Adam Smith had to say about the role of institutions in economic development. Even though many scholars have referred to the role of institutions time and again in their work, institution theory perhaps is most strongly associated with Nobel laureate Douglass North. According to North (1994), ‘institutions are humanly devised constraints that structure human interaction’. He adds that they are made up of formal and informal constraints and they depend on the enforcement characteristic of these constraints. All these taken together form the incentive structure of the society. North (1989) argues that the long-run economic performance of a society is shaped by its incentive structure. In this study, he looks at the effects of the United States Constitution, the British common law and the centralized enforcement mechanism of Spain on economic performance and argues that institutional differences do have economic consequences. The recent empirical literature on long-run growth endorses this view (Hall and Jones, 1999; Acemoglu et al., 2001; Rodrik et al., 2004). This perhaps explains why in the recent past the institutional view has received so much attention in the growth literature. Empirical studies on growth use the instrumental variable approach in order to estimate the relationship. This is essential in order to take account of the possibility of reverse causality. In other words, as better institutions influence growth positively, it is also possible that faster growing economies can afford to develop better institutions. Knack and Keefer (1995) and Hall and Jones (1999) are among the first to use institutions as one of the explanators of economic progress. They argue that Western Europeans are historically associated with high quality institutions characterized by protected property rights and efficient enforcement of contracts. Western Europeans also migrated and settled in large numbers in similar temperate climates. When they migrated, they

M2545 - BHATTACHARYYA PRINT.indd 21

02/03/2011 10:09

22

Growth miracles and growth debacles

carried these values and institutions with them. Therefore, it can be argued that the more temperate the climate the better is the institutional quality. This allows Hall and Jones (1999) to use distance from the equator as an instrument for institutional quality. They also use the fraction of the population speaking English and the fraction of population speaking other Western European languages as alternative instruments for institutions. A high proportion of the population speaking English and other Western European languages is indicative of a significant presence of Western Europeans in that country and hence better institutional quality. Their study reports a strong positive effect of institutions on income per capita. Acemoglu et al. (2001) put forward a strong argument in favour of the proposition that institutional quality is the fundamental determinant of economic development. They use European settler mortality as an instrument for institutional quality. They argue that Europeans resorted to different styles of colonization depending on the feasibility of settlement. In a tropical environment, the settlers had to deal with the killers malaria and yellow fever and, hence, a high mortality rate. This prevented colonizers from settling in a tropical environment and resource extraction became the most important, if not the only, activity in tropical colonies. In order to support these activities, the colonizers in the tropics and the subtropics erected institutions which were extractive in nature. On the other hand, in temperate conditions European settlers felt more at home and decided to settle. In these places, they erected institutions characterized by strong protection of property rights and efficient enforcement of contracts. These institutions created by the colonizers have persisted in the colonies even after independence. Therefore, they look at the extent of European settler mortality in the past in order to use it as an instrument of the current quality of institutions. In a subsequent paper, Acemoglu et al. (2002) argue that the settlement decision of European settlers also depended on the rate of urbanization prior to 1500, the start of the period of colonization, and hence it can be used as an alternative instrument for institutions. According to this argument, a high density of population in the colonies ruled out any possibility of settlement for the settlers and they used the natives to erect extractive institutions without bothering about the overall protection of private property in the society. On the other hand, a low rate of urbanization and low population density in the colonies encouraged the settlers to settle in large numbers and build institutions to protect property rights. They use the urbanization rate in 1500, measured by the then density of population, as an instrument for institutions in this study. In both cases they report strong effects of institutions on development. Rodrik et al. (2004) also report strong effects of institutions on economic development. In a cross-country regression with the level of income

M2545 - BHATTACHARYYA PRINT.indd 22

02/03/2011 10:09

Theories of root causes of economic progress

23

per capita as the dependent variable, they control for institutions, trade openness and geography, and institutions comes out to be the only variable having a statistically significant relationship with income. In a similar exercise, Easterly and Levine (2003) report no direct effect of geography and policy on economic performance if institutional quality is present as a control. They find that the only channel through which geography affects income is institutional quality. They use geographic endowment measures as their geography variable. Many more, including the abovementioned empirical studies, identify institutions as one of the important correlates of growth. This is perhaps convincing enough to say that institutions matter. However, the question that still remains is how institutions affect economic performance. I now discuss three important papers that deal with this issue. One of the early works on this issue is by Engerman and Sokoloff (2001). In their study, they trace the origins of the contrasting growth experiences of the North and South Americas. They argue that the settlement colonies of the North and South Americas differed significantly in terms of factor endowments creating incongruous initial conditions. The initial conditions have influenced the growth path of these two continents that ensued via economic and political institutions. The incongruity in the initial conditions thus explains the variability in their growth performance. Engerman and Sokoloff (2001) write, a hemispheric perspective across the range of European colonies in the New World indicates that although there were many influences, the factor endowment or initial conditions more generally had profound and enduring impacts on the evolution of economic institutions, on the structure of the colonial economies, and ultimately on their long-run paths of institutional and economic development. While all began with an abundance of land and other resources relative to labour, at least after the initial depopulation, other aspects of their factor endowments varied – contributing to extreme differences in the distributions of land holdings, wealth, and of political power.

They argue that the climatic conditions of the North were favourable to mixed farming of grains and livestock which exhibited limited economies of scale in production. This encouraged the development of small family-size farms and a relatively homogeneous population in terms of the distribution of wealth and political power. The result was better institutions favouring broader access to economic opportunities, more extensive domestic markets and better overall growth. The South and the Caribbean, in contrast, were endowed with vast mining resources and climates and soil conditions conducive to commercial crops such as sugar, tobacco and cotton. To exploit these resources, which exhibit large

M2545 - BHATTACHARYYA PRINT.indd 23

02/03/2011 10:09

24

Growth miracles and growth debacles

economies of scale, it was efficient on the part of the colonizers to set up large mining firms and plantations run with cheap slave labour. This led to the establishment of a society where a large section of the population remained impoverished with no political or economic rights, whereas a small section, the political and economic elite, took control of all the wealth. The high level of inequality in the distribution of wealth and political power created institutions which are extractive, exploitative and extremely unjust. These institutions persisted over time causing persistent inequality and lack of growth. The highly concentrated landholdings and extreme inequality in Mexico, Colombia and Peru during the colonial period are supportive of their story. In a recent paper, Acemoglu et al. (2005a) formalize some of the ideas of Engerman and Sokoloff (2001) using a schematic model. The following is a discussion of that model. Acemoglu et al. (2005a) argue that a nation’s economic and political institutions are broadly endogenous. They are, at least in part, determined by the society or a segment of it. They distinguish between two components of political power – de jure and de facto. They define de jure political power as the power that originates from formal political institutions in the society, for example, Monarchy, Democracy, Autocracy or Dictatorship and so forth. De facto political power, on the other hand, originates from informal political institutions which often challenge the formal ones, for example, economic and social interest groups and so forth. They assert that the existing political institutions at the current period determine the distribution of de jure and de facto political power in the society. The strength of de facto political power also depends upon the distribution of resources in that period. This shapes economic institutions of that period and political institutions of the future. Current economic institutions influence current economic performance and the future distribution of resources. Future distribution of resources and future political institutions in turn influence subsequent distribution of political power, economic institutions and economic performance. The following schematic diagram summarizes the link.

{

PIt 1 dj PPt & df PPt DRt 1 df PPt

} { 1

EIt 1 Yt & DRt11 PIt11

}

(3.8)

where PI, DR, djPP, dfPP, EI and Y are political institutions, distribution of resources, de jure political power, de facto political power, economic institutions and economic performance respectively. In another related paper Acemoglu et al. (2005b) explore the origin of capitalist institutions in early modern Europe which can be cited as

M2545 - BHATTACHARYYA PRINT.indd 24

02/03/2011 10:09

Theories of root causes of economic progress

25

further evidence for the hypothesis that economic and political institutions influence economic performance and vice versa. They describe the interactive role of trade and institutional development in stimulating economic growth in certain Western European countries from the sixteenth century onwards. They assert that Atlantic trade and colonial activity enriched and strengthened commercial interest groups both economically and politically in these countries (most notably Britain and the Netherlands), which then demanded and obtained institutional reforms to protect their property rights, enabling them to trade and invest more, triggering a circular and cumulative pattern of economic growth. However, this was not the case with some of the early starters in the colonization process (notably Portugal and Spain). Even though they managed to expand their trading and ransom-seeking activities rapidly by exploiting their early mover advantage, their growth was not sustainable. A major part of the explanation is the traders failing to secure their rights from the monarch. The monarch and the clergy in Portugal and Spain remained firmly in control of the activities of the trading companies, never allowing them to accumulate de facto political power to challenge the existing institutions. As a result, the incentive to undertake investments in physical and human capital eroded fast and the growth was short lived. Bhattacharyya et al. (2009) test this hypothesis for the period 1980 to 2004 using international data. They find that there is a threshold level of institutional quality below which trade is not effective for economic development. However, countries above this institutional threshold are able to reap benefits of trade to the fullest extent. They present estimates of the threshold level of institutional quality (property rights institutions, in particular) in their paper. A further extension of the political power theory is presented by Acemoglu and Robinson (2000). Their argument is akin to the Marxist theory of the capitalists’ and the workers’ class struggle. Marx argued that capital accumulation and the associated decline of the surplus of capitalists would intensify exploitation of the workers which would lead to intense class struggle between the capitalists and the workers eventually terminating the class society. Acemoglu and Robinson (2000) accept the basic premise of this theory and argue that social and political reforms in Western Europe during the nineteenth century were an outcome of deliberate concessions made by the elite, designed to avert political instability, expropriation and possibly a revolution. As a result of these concessions inequality reduced in Western society, creating the necessary precondition for the development of property rights institutions protecting private property, which translated into faster economic growth in the periods that followed. However, this theory is not free from its critics. Galor and Moav

M2545 - BHATTACHARYYA PRINT.indd 25

02/03/2011 10:09

26

Growth miracles and growth debacles

(2006) come up with an alternative model to explain the termination of the class society. They argue that it was not class struggle but human capital and knowledge that brought about the demise. I discuss the findings of this study in detail in Section 3.9. 3.5.1

How About Institutions in Africa?

A great deal of research in recent years suggests that institutions are the key for economic development. Institutions shape the incentive structure for factor accumulation and technology development which is critically important for a country’s economic progress. Therefore, success or failure in development can be traced back to success or failure in institution building. I have reviewed the relevant literature that deals with this view above. This view, however, is also relevant for Africa (Easterly and Levine, 1997; Herbst, 2000; Acemoglu et al., 2001; Acemoglu et al., 2002). The institutions school argues that the weak institutions in Africa largely explain her state of underdevelopment. Easterly and Levine (1997) use several indicators of public policy which can also be categorized as institutional measures even though they were not explicit about this in their study. They argue that the lack of growth in Sub-saharan Africa is associated with poor public policy, such as low schooling, political instability, underdeveloped financial systems, high government deficits and insufficient infrastructure. They link the choice of poor public policies on the part of the African governments to Africa’s high ethnic fragmentation. In other words, they argue that high ethnic fragmentation increases polarization and the degree of social conflict in terms of policy choice in African societies. This increases the cost of conflict resolution. The government often fails to internalize these costs and selects socially suboptimal policies. This is perhaps the first systematic attempt to explore African institutional weaknesses and the reason behind it in our profession. Even though Easterly and Levine (1997) were the first to apply this idea in Africa, it is older than this. Alesina and Tabellini (1989) and Alesina and Drazen (1991) theoretically modelled it almost a decade before their article was published. Herbst (2000) also seeks an explanation for institutional weakness in Africa. Robinson (2002) provides an excellent review of Herbst’s book. Herbst’s work predominantly focuses on the following questions: why are states are so ineffective in Africa? Why are African states often plagued by chaos and lawlessness? He comes up with an interesting answer.4 He argues that in Africa, labour was scarce and not land, which is reflected by the low population density figures across the continent. Therefore, the pre-colonial states in the continent did not fight over land but over people.

M2545 - BHATTACHARYYA PRINT.indd 26

02/03/2011 10:09

Theories of root causes of economic progress

27

This explains why property rights over people were so well defined in the form of slavery, whereas property rights over land were not, as many of the lands were held communally. This implies that it was not necessary for the pre-colonial states to defend a well-defined territory, as there was no fierce competition over land. In other words, there was less need to invest in defence and maintain bureaucracies. Therefore, tax collection in these states was poor and often absent. This allowed African states to survive without having to engage in institution building (tax collection, investing in defence, maintaining bureaucracies and provide rule of law) which made them very fragile. The trend of almost no external threat coupled with low population density persisted during the colonial period. Therefore the colonizers had little incentive to develop institutions either. After independence, the situation did not change. The United Nations decision to enforce the colonial boundaries as national boundaries and Cold War politics reinforced this trend. Hence, what we observe now are the weak institutions of contemporary Africa. Herbst’s argument is an extension of the work done by Max Weber on the development of European states. Even though Herbst’s thesis has received kudos for its intuitive appeal, it is still not free from criticism. Robinson (2002) criticizes Herbst for not emphasizing the impact of colonialism on institutions and state formation in Africa. He also adds that the relationship between population density and institution building becomes unclear, particularly in the case of Africa when one takes into account the ‘reversal of fortune’ effect and the impact of colonialism (Acemoglu et al., 2001; Acemoglu et al., 2002). The following paragraph deals with this theory in more detail. Acemoglu et al. (2002) observe that the relatively prosperous and densely populated areas of the world outside Europe prior to 1500 are relatively poorer now, whereas the less prosperous and sparsely populated regions outside Europe before 1500 are relatively prosperous at present. This empirical fact is termed by Acemoglu et al. (2002) as the ‘reversal of fortune’. They argue that this is because of the nature of the colonial policy pursued by the European colonizers. In colonies with high population density prior to 1500, the European colonizers faced relatively more resistance from the natives, which ruled out the possibility of settlement and, hence, they erected extractive institutions without bothering about the overall protection of private property. On the other hand, in colonies with low population density prior to 1500 the European colonizers faced less resistance from the natives and decided to settle and erect institutions with better protection of private property. This argument may not be true in the case of Africa. The African continent has always been a place with low population density. In spite of the low population density, the

M2545 - BHATTACHARYYA PRINT.indd 27

02/03/2011 10:09

28

Growth miracles and growth debacles

colonizers decided not to settle in Africa and built ‘institutions of plunder’ instead.5 The answer to this apparent paradox perhaps lies in the argument of Acemoglu et al. (2001). Acemoglu et al. (2001) argue that one of the key factors that influenced the decision of not settling in Africa by the European colonizers was the continent’s disease environment, which is well reflected by the settler mortality figures from the colonial records. They use settler mortality data mostly constructed from the historian Philip Curtin’s work on the disease environment of the European colonies to establish a statistically significant inverse relationship between the disease environment and institutional quality. As Robinson (2002) puts it, these factors are not taken into account in Herbst’s story. 3.5.2

The Effects of the Institutions of Slavery and Slave Trade in Africa

The theories of the impact of slave trade on Africa’s current development can be classified into two broad categories: the depopulation view and the sociopolitical view. The strong proponents of the depopulation view are the historians Joseph Inikori and Patrick Manning. Gemery and Hogendorn (1979) also add an alternative dimension to the view, basing their argument on trade theory. Inikori (1992) argues that the slave trade out of Africa during the period 1450 to 1870 resulted in massive depopulation of the continent. The population throughout this period remained too low to trigger division of labour, growth in internal trade, specialization and diversification, transformation of the technology and organization of production in the manufacturing sector. As a result, manufacturing in pre-colonial Africa could not develop beyond the handicraft stage. In the agricultural sector, the effects were more immediate. The low ratio of population to cultivable land encouraged dispersed settlements throughout the continent, particularly in Sub-saharan Africa. The population moved towards an extensive rather than intensive form of agriculture, which made subsistence and local self-sufficiency predominant. This had a dampening effect on technology adoption and production organization in the long run. African agriculture largely remained primitive and undercommercialized during this period. Fage (1978) and Lovejoy (1982) strongly disagree with Inikori’s view. They argue that Inikori’s claim of an absolute decline in the African population due to the slave trade is unfounded. Fage (1978) also adds that the trade had a minimal impact demographically. However, Manning (1981, 1982) and Thornton (1980) show that the demographic effects are significant no matter what the absolute totals are. Manning (1981, 1982) adds another dimension to the depopulation view.

M2545 - BHATTACHARYYA PRINT.indd 28

02/03/2011 10:09

Theories of root causes of economic progress

29

He shows that the African continent, faced with an increasing demand for slaves from the New World, reacted by increasing the supply of slaves. This increase in availability of slaves also raised the African demand for slaves. Africans preferred female slaves, whereas young male slaves were exported across the Atlantic to work in the plantations. This engendered huge imbalance in African sex ratios, which had a long-term impact on the continent’s demographic structure. According to his estimates, the continent’s population was held in check during the eighteenth and the nineteenth century by the slave trade, restricting economic progress. Finally, Gemery and Hogendorn (1979) add that the mass removal of the working-age population from the continent caused an implosion of the African production possibility frontier as the lost labour input affected virtually all production choices. This resulted in unambiguous reduction in the welfare of the continent. The secular decline in welfare continued over more than two centuries, plunging the continent into economic backwardness. The major contributors to the sociopolitical view are Inikori (1977, 1992) Manning (1981, 1982) and Miller (1988). Inikori (1992) argues that a vast majority of slaves that were exported were free individuals captured by force. The capture took a number of forms, notably kidnapping, raids organized by the state, warfare, pawning, via the judicial procedures, tributes and so forth. Firearms were imported from the Europeans in exchange for slaves, particularly during the period 1750 to 1807, which were used for capturing more slaves (Inikori, 1977). Rodney (1966) and Meillassoux (1976) show that an increase in the Atlantic slave trade led to more being captured and expansion of the African slave trade, often by violence. Lovejoy (1994) shows that warfare, raiding and kidnapping were the means of enslavement for more than three-quarters of the slaves captured during 1805 to 1850 from central Sudan. The judicial process became a tool for enslaving people within the community (Klein, 2001). Klein (2001) observes that judicial penalties in the form of compensation, exile or beatings were converted to enslavement. Inikori (2000) argues that the increase in the trade for captives institutionalized banditry and corruption for more than three hundred years in the continent, which retarded socioeconomic development. Curtin (1975) partially disapproves of Inikori’s view. He argues that it is improper to view conflicts that arose from political causes in Africa in the same light as those that originated from purely economic motives, such as the slave trade. Inikori (1992), however, questions the usefulness of this dichotomy as these relationships were far more complex. According to his view, ‘attempts to establish any form of simplistic relationship are therefore misleading’ (Inikori, 1992, p. 26).

M2545 - BHATTACHARYYA PRINT.indd 29

02/03/2011 10:09

30

Growth miracles and growth debacles

Manning’s work focuses on Dahomey, which is roughly the area around the Bight of Benin, during the period 1640 to 1860 (Manning, 1981, 1982). He observes that the immense increase in slave prices with a price elasticity of supply of 1.5 created the incentive for capturing more slaves. Institutions were set up during 1640 to 1670 in this area, which were conducive to capturing slaves. The institutions included warfare, raiding, kidnapping, judicial procedures and tributes. Manning (1982) writes, ‘These structures further reinforced a willingness to tolerate or justify the enslavement of one’s enemies or even one’s own’ (p. 9). In the case of Dahomey, the state became an active participant in the collection and delivery of slaves, which is evidence favouring Inikori’s argument. Miller (1988) gives an account of the deadly nature of the process of capture in Angola. He writes that warfare and violence stimulated the capture of slaves in West Central Africa, often capturing slaves in exhausted, shaken or physically wounded condition. Populations were raided consistently by stronger neighbours, and harassed and driven out from their homes and land. Gemery and Hogendorn (1979) add that the culture of raiding and warfare created a distinct minority class in African society who became far more powerful than the rest of the population, both in economic as well as political terms. The slave traders of Africa started enjoying European goods, currency and guns in return for slaves. They made significant gains from trade, at least in the short run, even though everyone was unambiguously worse off in the long run due to the implosion of the production possibility frontier (Gemery and Hogendorn, 1979). The extreme inequality of wealth distribution tilted the existing institutions in favour of the slave traders, which created the foundation for further inequality in the future. This is consistent with the theory of institutions, inequality and growth proposed by Engerman and Sokoloff (1994) and Acemoglu et al. (2005a) with respect to the New World. These institutions persisted throughout the pre-colonial period, further intensifying the problem. After colonization, the colonial powers did not interfere with the existing extractive institutions of the natives. They were reinforced by them instead (Robinson, 2002). Persistence of weak institutions had a negative impact on the long-run economic development of the continent. Nunn (2008) also provides a similar account. He shows that the slave trade prevented state development, encouraged ethnic fractionalization and weakened legal institutions. Through these channels, the slave trade continues to affect current economic development in Africa. However, Bhattacharyya (2009b) shows that Nunn’s evidence is not robust to the inclusion of malaria and other geographic variables.

M2545 - BHATTACHARYYA PRINT.indd 30

02/03/2011 10:09

Theories of root causes of economic progress

31

3.6 THE RELIGION AND CULTURE THEORY Respect the altar of every belief. Spanish America, limited to Catholicism to the exclusion of any other religion, resembles a solitary and silent convent of nuns .  .  . To exclude different religions in South America is to exclude the English, the Germans, the Swiss, the North Americans, which is to say the very people this continent most needs. To bring them without their religion is to bring them without the agent that makes them what they are. Juan Bautista Alberdi (cited in Landes, 2000, p. 5)

Juan Bautista Alberdi, a distinguished Argentinian of the nineteenth century, believed that religion and culture play the most significant role in shaping the economic destiny of a nation. This is what he wrote in 1852, half a century before Max Weber’s remarkable work The Protestant Ethic and the Spirit of Capitalism. In this section, I discuss some of the notable theories that link religion, culture and economic performance. I classify the religion and culture theory into five broad categories, as follows: the Protestant work ethic view, the religious intolerance view, the collective belief view, the middle class virtues view and the trust view. The Protestant work ethic view is due to Max Weber and it goes back to the early twentieth century when he published his thesis entitled The Protestant Ethic and the Spirit of Capitalism. Weber (1930) argues that the epoch-making Protestant Reformation changed societal outlook and sowed the seeds of modern capitalism. Protestantism’s emphasis on industry, thrift and frugality, along with its moral approval of risk taking and financial self-development, created a social environment conducive to investment and the accumulation of private capital. This became an inspirational force in transforming traditional societies into modern capitalist ones. Catholicism, on the other hand, considered trade and accumulating riches to be sinful, an attitude which was quite inimical to economic progress. Samuelsson (1993) points out that Weber’s theory hardly applies to Lutheranism, which retained the traditional attitudes of Catholicism towards trade and commerce at least at the time of its inception. But it undoubtedly applies to Calvinism, which supports trade and commerce. One of the central themes of Calvinism is predestination. According to this belief, some individuals are chosen and saved by God while the others are not. This process of selection by God is predestined and unalterable. Calvin, in his discourses, did not mention who had been chosen. Therefore, the only practical way for the followers of Calvin to breed hope of attaining salvation is by performing intense worldly activities. Weber (1930, p. 69), also cited in Acemoglu et al. (2005a, p. 15), states that, ‘however useless good works might be as a means of attaining salvation,

M2545 - BHATTACHARYYA PRINT.indd 31

02/03/2011 10:09

32

Growth miracles and growth debacles

they are indispensable as a sign of election. They are the technical means, not of purchasing salvation, but of getting rid of the fear of damnation.’ Calvin’s doctrine encourages economic activity, but condemns enjoying the fruits of such activity by means of leisure. The followers of Calvin consider morality to be a virtue not because it can bring salvation but because it helps in acquiring moral credit. All these characteristics of Calvinism are essential preconditions for economic progress and this is how, according to Weber (1930), Calvinism influences economic performance. Given the provocative nature of Weber’s thesis, it is not hard to understand why it fails to achieve universal acceptance. In an influential study on religion and growth, Tawney (1926) rejects Weber’s view. He argues that the English economy took off in the sixteenth century when the religious influence in society was replaced by secular attitudes. The religious intolerance view is due to Landes (1998). This adds another dimension to the theory that links religion and economic performance. In his book The Wealth and Poverty of Nations, he argues that the profit motive of the entrepreneurial class coupled with invention and productivity gain are core to achieving material progress. Therefore, it is crucial for religious beliefs and practices within the society to encourage hard work and not inhibit rational and scientific thought and profit making. He uses the example of eighteenth century Britain to illustrate the effect of Protestant work ethics and values on economic performance. According to his thesis, Britain’s numerous scientific successes after the Industrial Revolution were made possible due to the supportive and tolerant role played by the Church. In contrast, the orthodoxy of the Catholic Church in Portugal and the culture of intolerance diffused by the Catholic Church in Spain and Italy strangulated scientific thinking and halted progress even though they were the first to make significant progress in cartography and ocean navigation. The infamous ‘Roman Inquisition’ in sixteenth century Europe is a testimony to the religious intolerance and orthodoxy of the Catholic Church which virtually stalled the European renaissance. To summarize, Landes (1998) maintains that the culture of intolerance in Catholicism and Islam limits the potential of a country to grow. Protestantism, on the other hand, facilitates economic growth by encouraging hard work and showing a more tolerant attitude towards scientific research. The collective belief view is due to Avner Grief. Grief (1994) holds that different cultures and religions generate different sets of beliefs about how people behave and people’s behaviour has a direct role in shaping economic performance. According to his argument, this can generate different growth equilibria across the globe even with the same set of institutions. These beliefs can also influence institutional quality in the long run and affect economic performance indirectly.

M2545 - BHATTACHARYYA PRINT.indd 32

02/03/2011 10:09

Theories of root causes of economic progress

33

The ‘middle class virtues’ view is due to Gregory Clark (2007). Clark’s main proposition is that the lack of growth in the preindustrial world is due to the lack of ‘middle class’ virtues of thrift and hard work. He argues that industrial revolution took place in England because the middle class virtues spread down the social scale in the sixteenth and seventeenth centuries. The poor experienced relatively high death rates during this time. Therefore, they did not have enough surviving children. In contrast, the infant survival rate amongst the rich was higher. As a result, many of the rich offspring were forced down the social scale carrying middle class virtues with them. This led to a change in work ethic of the entire society and rapid economic progress. This, however, did not take place in Asia, Africa and Latin America. The infant survival rates amongst the rich and the poor were largely similar in these societies. Therefore, the spread of middle class virtues was at best slack. This, in Clark’s view, explains the difference between the rich and the poor countries. Clark’s view, however, is criticized by many including Allen (2008) for its lack of evidence. Even the evidence provided by Clark appears to be unreliable (Allen, 2008). Finally, the trust view is due to Putnam (1993). He holds that the Catholic tradition emphasizes vertical bonds with the Church which tend to undermine horizontal bonds with fellow citizens. This reduces the level of trust in society and increases the transaction costs of economic exchange. Therefore, Catholicism has a negative effect on institutions and hence a negative effect on growth. Trust also has a direct impact on creditworthiness. Lack of trust in society creates a creditor unfriendly environment that limits economic progress. The creditor unfriendly environment is also due to the anti-usury culture prominent in the Catholic tradition (Stulz and Williamson, 2001). In a related thesis, Kuran (1997) argues that a large divergence in individual and social preferences may lead to a decline in social capital. He formally defines the phenomenon of divergence in individual and social preferences as preference falsification. To illustrate this further, one could think of a situation where social preference is moulded by an authoritarian state and is different from individual preferences. In such a situation, individuals would never reveal their true individual preference in public, which is different from the social preference put forward by the state, due to the fear of a backlash. Kuran describes this phenomenon as preference falsification. In his book, he describes both positive and negative implications of preference falsification. One of the negative implications could be a decline in social capital and trust which may have negative effects on long-term development.

M2545 - BHATTACHARYYA PRINT.indd 33

02/03/2011 10:09

34

Growth miracles and growth debacles

3.7 THE GEOGRAPHY THEORY The striking differences between the long-term histories of peoples of the different continents have been due not to innate differences in the peoples themselves but to differences in their environments .  .  . if the populations of Aboriginal Australia and Eurasia could have been interchanged during the Late Pleistocene, the original Aboriginal Australians would now be the ones occupying most of the Americas and Australia, as well as Eurasia, while the original Aboriginal Eurasians would be the ones now reduced to downtrodden population fragments in Australia. (Jared Diamond (1997), Guns, Germs and Steel, p. 405.)

In this section, I discuss the theories that link geography with development. I classify the existing theories of geography into five different categories. They are as follows: the climate view, the agriculture view, the market proximity view, the disease view and the sophisticated geography view. The climate view holds that the population in the tropics is not industrious enough largely due to the energy-sapping heat. Natural availability of food in abundance also makes tropical people idle. This has a direct and negative effect on human productivity and hence economic growth (Montesquieu, 1748). In a recent study, Parker (2000) supports Montesquieu’s argument. According to his thesis, an individual’s desire to maximize utility is dependent on motivation, homeostasis and neural, autonomic and hormonal adjustments. These physiological factors are governed by the hypothalamus. The activity of the hypothalamus is heavily dependent on thermodynamics. In hot conditions, the hypothalamus secretes hormones which negatively affect motivation and enterprise whereas, in cold climates, individuals are naturally hard working. These tendencies affect the steady-state level of income in these two regions. The average steady-state income in cold climates is naturally higher than the average steady state in hot climates. Hence, climate explains two-thirds of the per capita income differences between the tropics and temperate regions. The agriculture view is due to Gallup and Sachs (2000). This view maintains that high relative humidity and high night-time temperature in the tropics cause high plant respiration and slow down plant growth. They argue that the deficiency in plant growth in the tropics is also related to the lack of nutrients in tropical soil. Humid tropical soils (alfisols, oxisols and ultisols) are typically low in nutrients and organic matter. This limits plant growth and also causes soil erosion and acidification. In addition, the lack of frost allows a greater number of pests to survive and breed. These factors have a debilitating impact on agricultural productivity and inhibit economic progress. In addition to this, an alternative and sophisticated

M2545 - BHATTACHARYYA PRINT.indd 34

02/03/2011 10:09

Theories of root causes of economic progress

35

agriculture view exists due to Diamond (1997). According to this view, prior to the period of colonization, Europeans used technology which was specifically designed to suit temperate conditions. In the late eighteenth and early nineteenth centuries, when the Europeans embarked on a colonization drive, they introduced these technologies in the colonies. They worked well in colonies with temperate conditions, but failed to deliver the same goods in the tropical environment. This explains the low productivity of tropical agriculture and hence slow economic progress. The market proximity view is due to Sachs and Warner (1995b), Sachs and Warner (1997b) and Gallup et al. (1998). There is no disagreement within the economics community that trade and commerce generate wealth and prosperity (Smith, 1776 [1976]). One of the important preconditions for trade is easy access to major markets. According to the market proximity view, unfavourable geographic location characterized by no or limited access to ports or ocean-navigable waterways, being landlocked, is a major impediment to trade and commerce. Access to a port or major markets in this situation often involves crossing international boundaries, which makes the cost of transportation relatively high and limits international trade. Absence of international trade in these economies confines all commercial activities to small internal markets. This causes an inefficient division of labour and underdevelopment. If one looks at inland Africa, which is also one of the poorest areas in the world, it is quite evident that most of the countries of this region are landlocked. This prevents these countries from effectively participating in international trade because transport costs are too high (Sachs and Warner, 1997b). The disease view is due to Bloom and Sachs (1998), Gallup et al. (1998) and Gallup and Sachs (2001). According to this view, infectious malaria has a debilitating effect on human productivity and directly affects economic performance. Gallup and Sachs (2001) point out that the countries with intensive malaria grow 1.3 percentage points slower per person per year than countries without malaria and a 10 percentage point reduction in malaria might result in a 0.3 percentage point increase in annual per capita income growth. Bloom and Sachs (1998) also claim that the high incidence of malaria in sub-Saharan Africa reduces the annual growth rate by 1.3 percentage points a year. In other words, eradication of malaria in 1950 would have resulted in a doubling of current per capita income. Sachs (2003a) and Carstensen and Gundlach (2006) in empirical studies report strong and negative effects of malaria on economic progress, even after controlling for institutions and openness. Finally, the sophisticated geography view is due to Jared Diamond (1997). Diamond, in his book entitled Guns, Germs and Steel asks the question: why did history unfold so differently on different continents? He

M2545 - BHATTACHARYYA PRINT.indd 35

02/03/2011 10:09

36

Growth miracles and growth debacles

argues that geography and biogeography moulded the contrasting fates of human lives in different continents. In summary, his hypothesis is as follows. He argues that geography has endowed mankind with different sources of food and livestock. The Eurasian climate, especially southwest Asia which he calls the ‘Fertile Crescent’6 was best suited for the growth of the maximum number of edible wild grains and large mammals. Early hunter-gatherers living in this region domesticated these wild grains and adopted a sedentary agriculture-based lifestyle. They domesticated the large mammals for meat, milk and muscle power which they could use in farming. The use of large mammals in agriculture immensely improved farm productivity yielding more and more food surplus. This technology and knowledge spread all across Eurasia along the same latitude. Close contact with the large mammals also led to frequent outbreaks of epidemic diseases among the human population in this region. This helped Eurasians to develop immunity to many of these diseases over the long run. The gift of food surplus and a sedentary lifestyle allowed them to invest more time in the development of guns, steel swords, ocean-going ships and so forth. The societal structure became hierarchical and far more complex than that of the hunter-gatherers. This, however, was not the case in other continents. In places like New Guinea humans were left with very limited choices of food. No wild grains or large mammals were available for domestication. So humans in New Guinea remained hunter-gatherers. In the Americas, corn was the major grain that was domesticated. Other grains were not available. The continent also lacked large mammals for domestication. The only available option was the llama, which is weak and yields less meat and milk than cows, goats or sheep. The indigenous American population were also not familiar with horses and they lacked immunity to the fatal Eurasian diseases. As a result the Europeans faced little or no resistance from the indigenous American population in their colonial conquest. The Aztecs, the Incas and the Mayas were fighting an unequal battle with the Europeans powered with guns, steel swords, horses and germs. More than two-thirds of the population were wiped out by a smallpox epidemic when they first came into contact with the Europeans. Riding on the power of guns, germs and steel, the Europeans colonized most of the known world, getting access to a large pool of resources, which helped them to develop an advanced industrialized society. The Americas, Australia and New Zealand gained from European migration and the migration of European technology along with them. The rest of the world remained largely impoverished and the gap widened over time with the development of more and more advanced technology in the West.

M2545 - BHATTACHARYYA PRINT.indd 36

02/03/2011 10:09

Theories of root causes of economic progress

3.7.1

37

Diseases, War and Urbanization in Europe

A related but somewhat different theory of geography is from Voigtländer and Voth (2008). They pose the question of why Europe was prosperous relative to the rest of the world during the 1700s. They argue that even though pre-modern Europe was largely Malthusian, the reason behind the divergence in income between Europe and the rest of the world can be explained by three factors, namely, diseases, war and urbanization. War, urbanization and trade-driven disease raised death rates and once death rates were higher, incomes could remain at an elevated level even in a Malthusian world. High density of population, rapid urbanization and the disease-ridden European cities kept the death rates relatively high. International trade also helped spread diseases across the continent. High death rates meant labour scarcity and higher wages. This changed the nature of demand. As a result people started having more children. However, people still had above subsistence income and they spent their surplus income mainly on manufactured goods produced in cities. This reinforced urbanization and trade. Urbanization raised the risk of diseases and trade raised the risk of war which was financed by taxes from the cities. Therefore they argue that geography, diseases and war had a positive impact on living standards in pre-modern Europe. 3.7.2

Malaria in Africa

The malaria view can be categorized into two broad categories. The first deals with the economic burden of malaria in contemporary Africa, and the second deals with the historical impact of malaria and diseases on the continent’s long-run economic development. I call the former the ‘contemporary malaria view’, whereas the latter is the ‘historical malaria view’. The ‘contemporary malaria view’ is due to Sachs (2003b). According to this view, malaria dramatically lowers labour productivity and the return on foreign investment and raises transaction costs of international trade, limiting development, which is typically observed in sub-Saharan Africa. I have discussed this view in the following sections. Sachs (2003a) in a cross-country empirical study shows that institutions do matter, but not exclusively. He highlights the prevalence of malaria as another important factor which should not be underestimated. In a series of studies, Bloom and Sachs (1998), Gallup et al. (1998) and Gallup and Sachs (2001) show that infectious malaria has a debilitating effect on human productivity and directly affects economic performance. Sachs et al. (2004) argue that with malaria and subsequent low productivity of agricultural labour, sub-Saharan Africa cannot generate enough marketable

M2545 - BHATTACHARYYA PRINT.indd 37

02/03/2011 10:09

38

Growth miracles and growth debacles

surpluses which also limits the prospect of market development. Markets, even if they develop, remain concentrated at a very local level. This is indicative of a situation of ‘low level equilibrium trap’ in these economies. Acemoglu et al. (2003a) strongly disagree with Sachs’s view. They argue that the disease environment influences the balance of power between previously isolated populations when they come into contact. The local disease environment influences the colonization strategy and settlement decision, which sets up the path for future institutional development (Acemoglu et al., 2001, 2002). Therefore, the disease environment affects the level of development indirectly through institutions. I call this the ‘historical malaria view’. This view is also closely related to the historian Phillip Curtin’s work on epidemiology of the New World and Africa (Curtin, 1968). In this article, Curtin shows that epidemiological factors have influenced economic decisions and economic patterns of the New World (particularly tropical America) and Africa. Another angle within the ‘historical malaria view’ is a line of argument that is strongly pursued by a section of the historians. Miller (1982) uses Portuguese traveller’s records, missionary and church documents to show that frequent epidemics of malaria and yellow fever caused massive depopulation in the agriculturally marginal zones of West Central Africa. Dias (1981), in a similar study of nineteenth and twentieth century Angola argues the same. They observe that the effects of disease, epidemics and famines were far more powerful than the slave trade in depopulating the region. According to their view, the increase in slave trade was an outcome of local epidemiology and poor agriculture rather than strong Atlantic demand. Bhattacharyya (2009b) also shows that malaria explains longrun economic development in Africa. All other variables including institutions and the slave trade are statistically insignificant. Miller (1982) writes, ‘The slave trade appears in some ways less a cause of depopulation than a consequence of it when viewed in terms of droughts and demographic changes in West Central Africa.’ Hence, according to his view, the disease environment had direct effects on the demography and economic development of the region. Inikori (1992) and Manning (1981, 1982), however, provide evidence to show that these effects were not strong enough to have a larger impact on the African population compared to the strong Atlantic demand for slaves.

3.8 THE TRADE OPENNESS THEORY The greatest improvement in the productive power of labour, and the greater part of the skill, dexterity, and judgement with which it is any where directed,

M2545 - BHATTACHARYYA PRINT.indd 38

02/03/2011 10:09

Theories of root causes of economic progress

39

or applied, seem to have been the effects of the division of labour . . . This division of labour, from which so many advantages are derived, is not originally the effect of any human wisdom, which foresees and intends that general opulence to which it gives occasion. It is the necessary, though very slow and gradual, consequence of a certain propensity in human nature which has in view no such extensive utility; the propensity to truck, barter, and exchange one thing for another. (Adam Smith (1776 [1976]), Wealth of Nations, Chapters I & II, pp. 7 and 17)

The openness view of development goes back at least to Adam Smith. Smith (1776 [1976]) argues that openness to trade increases the size of the market, which raises the possibility of greater division of labour. Division of labour in turn improves productivity and productivity improvement induces faster economic growth.7 The neo-classical theory, however, holds a slightly different view. According to this theory, reduction in trade barriers opens up the possibility of a more efficient exploitation of comparative advantage through reallocation of factors. Labour and capital move towards their highest valued uses improving overall productivity and the welfare of the economy. Growth takes place in the economy due to transitional dynamics. In other words, growth lasts only for the duration of the transitional period and stops after the economy reaches its new steady-state levels of capital and output per worker. Another significant theory in the openness and growth literature is the technology transfer view. According to the closed economy neo-classical growth models, given the technology level at a particular point in time, a country accumulating more physical and human capital grows faster than a country accumulating less of the same, but all of them converge to the same long-run steady-state equilibrium rate of growth (Solow, 1956; Swan, 1956; Mankiw et al., 1992). In this model, income per capita can grow in the long run only when there is exogenous technological progress. In real life, we hardly observe convergence of income. In contrast, what we notice is that the rich economies are growing faster than the poor ones and the gap between the rich and the poor is widening over time.8 Coe and Helpman (1995) point out that this gap is due to the variable rate of technological progress. They find evidence that technology catch up or R&D spillover bridges this gap. Dowrick and Rogers (2002), on the other hand, report that the gap is due to variable rates of capital accumulation as well as technological progress. Howitt (2000) builds a theoretical model to show that because of technology transfer, R&D performing countries converge to parallel growth paths, whereas others stagnate. Several others, including Coe and Helpman (1995), identify trade openness as a medium of technology transfer. In other words, what these studies argue is that the follower economies adopt technology and knowledge developed in

M2545 - BHATTACHARYYA PRINT.indd 39

02/03/2011 10:09

40

Growth miracles and growth debacles

the advanced economies and the transfer takes place largely through international trade. Therefore, effective participation in international trade and opening up by reducing trade restrictions exposes economies to new knowledge and new technology that helps them to grow in the long run. Other influential studies in the openness and growth literature are Sachs and Warner (1995a) and Frankel and Romer (1999). Sachs and Warner (1995a) use the growth accounting approach to show that trade openness is the key determinant of growth. They calculate trade openness using information on average tariff levels, non-tariff barriers, black market premiums, state monopoly on major exports and so forth. Their regression results suggest, on average, ‘open’ economies grow somewhere between 2 to 3 percentage points faster than ‘closed’ economies.9 Their study, however, is not free from criticism. Rodrik and Rodriguez (2000) show that the Sachs and Warner measure of openness suffers from over-reliance on the black market premium and state monopoly on major exports criteria. Frankel and Romer (1999), on the other hand, resort to level accounting using instrumental variable regressions to conclude that trade openness is a key factor influencing economic development. They construct an instrument by separating out the influences of income, population size and land area, which they call the constructed openness. This variable they claim yields more accurate estimates of the relationship between openness and per capita income. They report a strong and positive effect of openness on income per capita. Even though there has been a fair bit of agreement about the association of trade openness with growth for at least the last couple of decades, recently these results have come under fire in a series of papers. These papers argue that trade openness loses importance once institutional quality is introduced as a control in a regression framework. Rodrik et al. (2004) reach a similar conclusion, which we discussed previously. One possible explanation is openness enhancing institutional development. Therefore, a regression model with openness and institutions as controls is not picking up any individual effect of openness on income as it is operating through the institution channel. This view is supported by Wei (2000) who suggests that more open countries face greater losses from corruption than less open ones as the loss from corruption is higher in the case of foreign transactions. Therefore, open countries have a higher incentive to develop better institutions. Another explanation comes from Dollar and Kraay (2003). They argue that cross-country regressions of log-level per capita gross domestic product (GDP) on instruments of institutions and openness do not reflect the relative importance of these

M2545 - BHATTACHARYYA PRINT.indd 40

02/03/2011 10:09

Theories of root causes of economic progress

41

variables in the long run. The high correlation between institutions and openness does not allow regression models to estimate the actual relationship. According to their analysis, regression of changes in decadal growth rates on instrumented changes in trade and changes in institutional quality show a significant effect of trade openness on growth. They report that trade openness and institutions are important for economic growth in the very long run, but trade has a relatively larger role over institutions in the short run. They do not control for geography, religion and culture. Also, their study explains changes in growth rates, rather than growth itself, and growth accelerations are hard to observe in the long run (Jones, 1995). This also brings about the question of whether the spurt in growth rate that comes out of changes in institutions and openness is a mere impulse which tapers off in the long run or something that is sustainable. In a recent study, Chang (2008) argues that overtly strong commitment to free trade is not good for growth. As an alternative strategy, he promotes ‘heterodox capitalism’, which includes a combination of protectionism, government promotion of favoured industries, strong state-owned enterprise and heavy regulation of foreign direct investments. Chang’s main piece of evidence in support of this argument is as follows. He shows that the developing countries grew at a rate two times faster during the ‘heterodox’ era of the 1960s and 1970s than during the 1980s onwards ‘free trade’ era. The average growth rate for developing countries during the 1960s and 1970s, according to Chang, was 3 per cent. In contrast, the average growth rate during the period 1980 to 2002 for the same country group was 1.7 per cent. Easterly (2009), however, finds several inconsistencies in the figures reported by Chang if the 1980 turning point is shifted to 1983 or 1982. He also shows that the average growth rate for developing countries for the period 1983 to 2008 (2.7 per cent) is in fact higher than the same for the period 1960 to 1982 (2.6 per cent). Therefore, Chang’s argument heavily relies on picking 1980 as the turning point. Easterly (2009) also shows that picking 1980 as the turning point in itself is misleading as trade liberalization was first pushed by the IMF and the World Bank in the aftermath of the Third World debt crisis in 1982. In sum, even though Chang (2008) perhaps appropriately criticizes the one size fits all free trade policy as a development strategy for all developing countries, he also appears to be guilty of moving to the other extreme of making strong claims in favour of protectionism as an alternative.

M2545 - BHATTACHARYYA PRINT.indd 41

02/03/2011 10:09

42

Growth miracles and growth debacles

3.9 THE KNOWLEDGE AND HUMAN CAPITAL THEORY Technological progress predated capitalism and credit by many centuries, and may well outlive capitalism by at least as long. Joel Mokyr (1990), The Lever of Riches, p. 6

The knowledge and human capital theory of development is due to Joseph A. Schumpeter. In his book entitled The Theory of Economic Development he describes the notion of ‘creative destruction’ and how it is central to the process of economic development. He argues that the entrepreneurs in a capitalist economy are constantly looking for new ideas which will render their rivals’ ideas obsolete. A product developed from the new idea will be able to capture the entire market and allow the firm to enjoy monopoly profit. This continuous process of ‘creative destruction’ raises a firm’s output. The possibility of enjoying monopoly profits creates an incentive for rival firms to invest in R&D and innovate, and the process continues. The increase in R&D activities in the economy creates positive externalities and reduces the cost of R&D for an individual firm. The economy grows, powered by the increase in knowledge capital and R&D. Even though Schumpeter (1934) talks about R&D, he focuses mainly on the role of innovation rather than invention. Joel Mokyr (1990), on the other hand, argues that inventions and innovations are complements and both play an important role in economic development. In his book entitled The Lever of Riches he defines invention as, ‘an increment in the set of the total technological knowledge of a given society, which is the union of all sets of individual technological knowledge’ (Mokyr, 1990, p. 10). He argues that invention in itself may not be a meaningful concept, but, without it, it is impossible to develop an application which has immediate commercial usage. At the other extreme, without innovation, an inventor will have little economic incentive to pursue new ideas. In Part III of his book he goes on to argue that the complementarity between invention and innovation explains why many societies failed to be technologically creative. Invention requires individual brilliance and an inventor expects social recognition of her work. Therefore, a society more receptive to new ideas and willing to reward human inquiry is more likely to produce more scientists and researchers. Innovation, on the other hand, requires interaction with other individuals and an in-depth knowledge of the market. Hence, it has more of a social and economic nature. Therefore, a society willing to encourage business enterprise is more likely to produce more innovations. However, neither of them can survive independently in the long run. A society has to fulfil multiple conditions to generate both

M2545 - BHATTACHARYYA PRINT.indd 42

02/03/2011 10:09

Theories of root causes of economic progress

43

invention and innovation. He supports his argument using evidence from the Chinese and the European history of technology. Even though China entered the early modern era as a technology leader of the world, soon afterwards it lost its advantage to Europe and suffered a technological slowdown. Europe, on the other hand, acquired most of the early technologies from China and the East through direct and indirect means and went on to develop them further, which led to the Industrial Revolution. Mokyr asks the question, why does this happen? His explanations are as follows. First, social groups trying to sabotage and block new technology were far less powerful in Europe than in China. Second, a market for technology developed fairly rapidly in early modern Europe, which created an economically as well as socially rewarding environment for the scientists and innovators. This encouraged private initiatives in the development of new technology. Therefore, the change in the character of the state did not affect investments in technology much. Even if it did, it led to a decline of that state in the European economic hierarchy. This made every state or monarch in early modern Europe conscious of the costs of conformism. In contrast, Chinese scientific activities and technology development were largely state sponsored. This prevented the development of private initiatives and hence a technology market. A change in the character of the state often led to hostile behaviour towards innovation and nonconformism. The abrupt ending of the famous Cheng Ho voyages in 1433 is a classic illustration. A change in the ruler and internal power structure in the Ming Dynasty in 1433 led to the closing down of shipyards and production of ocean-going junks. The continuous threat to potential innovators from rival power groups had perhaps forced them to shy away from novel ideas. Even though Mokyr identifies technology and knowledge as the central factor, he does not rule out the role of culture and institutions in shaping technology. I will discuss this further in Chapter 4. Glaeser et al. (2004) go even further to claim that human capital is more fundamental than institutions. They challenge the Acemoglu et al. (2001) institutions theory which says that in temperate conditions European settlers felt more at home and decided to settle. In these colonies, they brought Western values and institutions along with them when they migrated that were favourable to capitalism. In contrast, Glaeser et al. (2004) argue that the decision to settle in temperate conditions does not tell us anything definitive about what the Europeans brought with them when they migrated. It could well be human capital and ideas and not institutions. In other words, they argue that institutions have only a second order effect on economic performance and the first order effect comes from human and social capital, which shape both institutional and

M2545 - BHATTACHARYYA PRINT.indd 43

02/03/2011 10:09

44

Growth miracles and growth debacles

productive capacities of a society. They call this the ‘Lipset-PrzeworskiBarro’ view of the world. In related research, Allen (2009) argues that the Industrial Revolution in eighteenth century Britain propelled its economy to new heights which were never experienced before elsewhere. He argues that industrial revolution took place in Britain and not elsewhere in Europe or Asia because of two main factors. Eighteenth century Britain was a high wage economy relative to the rest of the world. Britain was also abundant in cheap coal energy and the cost of capital was relatively low. This created conditions for capital- and energy-intensive technological breakthroughs. As a result, capital- and energy-intensive technologies, such as the steam engine, the cotton mill and coal-fired metal production were all profitable. The high wage economy of pre-industrial Britain also meant that schooling and apprenticeship were affordable to the masses, which, in turn, secured the steady supply of the highly skilled labour required to operate these technologies. The Industrial Revolution would spread around the world much later and only when further advances were made by the British engineers to make these technologies affordable. Other strong supporters of the human capital view are Galor and Moav (2006). They provide a counterargument to the political power and institutions theory of Acemoglu et al. (2005a) and Acemoglu and Robinson (2000).10 Galor and Moav (2006) argue that it was not concessions by the elite that lead to the demise of the class society. It was the investments in human capital by the capitalists to sustain their profit rates that caused the gradual dismantling of the class society. Their version of the story goes as follows. After the Industrial Revolution in the late eighteenth and early nineteenth century Europe, production was heavily dependent on capital and labour. The ownership of capital was with the capitalists, whereas the ownership of labour was with the workers. This created the class society characterized by a social division in line with the ownership of the factors of production. However, in the latter half of the nineteenth century, the character of capitalist production experienced a change. Capitalists started realizing the increasing importance of technology and human capital in production. They noticed that the productivity of a skilled worker is three or four times more than an average unskilled worker. Therefore, investments in human capital and technology became crucial for the sustenance of profit rates of the capitalists. These investments in mass schooling by the capitalists paid dividends in the long run by generating positive externalities for R&D and faster economic growth. Galor and Moav (2006) also provide empirical evidence in favour of their theory using the voting patterns of the legislators on England’s

M2545 - BHATTACHARYYA PRINT.indd 44

02/03/2011 10:09

Theories of root causes of economic progress

45

education reform, the Balfour Act of 1902. This bill was meant to create a publicly supported secondary school system. They show that legislators’ support for this bill was positively correlated with the industrial skill-intensity level of the counties. Legislators representing counties with a high industrial skill-intensity level overwhelmingly supported this bill, whereas the others did not. By assuming that the legislators received financial backing from the industrialists of their respective counties, they argue that this result reflects the eagerness of the capitalists owning skillintensive industries to invest in human capital. Many others debate this view. Both Acemoglu et al. (2001) and Rodrik et al. (2004) put human capital in their list of proximate determinants rather than as a fundamental determinant of economic progress. The literature, however, is yet to arrive at a consensus.

3.10 THE STATE FORMATION AND WAR THEORY All societies must deal with the possibility of violence, and they do so in different ways . . . Most societies, which we call natural states, limit violence by political manipulation of the economy to create privileged interests. These privileges limit the use of violence by powerful individuals, but doing so hinders both economic and political development. In contrast, modern societies create open access to economic and political organisations, fostering political and economic competition. Douglass C. North, John Joseph Wallis and Barry R. Weingast (2009), Violence and Social Orders, p. 1

Max Weber argued that high density of population and scarcity of land in early modern Europe led to intense territorial competition amongst monarchs and feudal lords. Therefore, warfare for the control of scarce resources and land were extremely common and frequent. Taxation was used as the main instrument by small- and medium-sized states to finance their war expenditure. In other words, the symbiotic financial relationship between the ruler and his or her subjects became ever more powerful during this time. This relationship, in Weber’s view, is fundamental to the formation of statehood and perhaps is at the heart of the construct of a modern state. Recent research by economists and political scientists suggest that state formation and state capacity play a major role in economic development. Herbst (2000), using Weber’s thesis, also argues that state weakness explains the lack of development in Africa. As I discussed earlier, Herbst’s answer depends on the factor intensity of the continent. Historically, Africa is land abundant and labour scarce. As a result, property rights over land are weak and institutions to adequately tax land have failed

M2545 - BHATTACHARYYA PRINT.indd 45

02/03/2011 10:09

46

Growth miracles and growth debacles

to evolve. African states survived for centuries without engaging in the institution building that facilitates tax collection, investment in defence, maintaining bureaucracy and providing the rule of law. The lack of these institutions made African states extremely fragile. The trend changed little during the colonial period and the Cold War. Hence the weak institutions in Africa continued to persist. Besley and Persson (2008) and Besley and Persson (2009) also present theory and evidence on why the lack of state capacity may lead to poor development outcomes. They argue that a state’s institutional capacity to levy taxes and support markets are typically constrained by its history of investments in legal and fiscal capacity. States with weak legal systems and imperfect taxation institutions are unlikely to be capable enough to adequately levy taxes and support markets. In other words, state capacity would be fairly limited. This would lead to poor development outcomes. They also show that investments in common interest public goods, such as fighting external wars, political stability and inclusive political institutions are all conducive to building state capacity. In contrast, civil war, internal conflict and restrictive political institutions are all damaging to state capacity. In a similar vein, North et al. (2006, 2009) also argue that the nature of the social order characterized by the nature of the existing state can explain the process of modern social development. Their framework suggests that in a natural state the actors use political power within the existing political system to limit economic entry and create rents. These rents are then used to stabilize the political system and limit violence. However, restrictive economic entry and restrictive political institutions are not conducive to long-run economic growth as it suffocates enterprise, the free market and innovation. They argue that this is the natural way most human societies are organized, even today, and hence the name natural state. In contrast, a handful of developed countries have developed open access social orders where open access to economic and political organizations facilitates competition and supports the market. In other words, social order in these states is sustained by competition and not rent creation. The obvious question, however, is how the transition from ‘natural state’ to ‘open access state’ takes place. According to North et al. (2006), war and the threat of violence play a crucial role. In the event of an external war, the incumbent state requires more taxation revenue and support from its subjects in general. To increase revenue, the state resorts to increasing taxation in return for more widespread political rights. O’Brien (2001, 2005) also shows that in early modern Britain the state was borrowing heavily from its subjects to finance continental war in return for

M2545 - BHATTACHARYYA PRINT.indd 46

02/03/2011 10:09

Theories of root causes of economic progress

47

political rights. This led to the development of an efficient financial system conducive to rapid economic growth. In summary, one can argue that the development of the state has played a crucial role in prosperity. The factors that influence state development are taxation, legal institutions, external conflict and internal conflict. Evidence suggests that external conflict boosts state capacity, whereas the reverse is observed if the conflict is internal in nature.

NOTES 1. 2.

3.

4. 5. 6. 7. 8.

9. 10.

Even geography is not truly fixed as we are facing the challenge of climate change. The idea of reverse causality is from Seymor Martin Lipset’s (1959) ‘modernisation hypothesis’ which says that institutional quality, education and health improve as countries develop. However, recent research in political science and economics shows that institutions diverge not in a systematic manner as outlined by Lipset (1959), but due to exogenous shocks at critical historical junctures as outlined by the famous work of Barrington Moore (1966) on institutions (see Thelen, 2000; Acemoglu et al., 2007; and many others) . For a survey of ‘critical juncture hypothesis’, see Thelen (1999). This argument, however, is not universally applicable. In particular, this does not apply to the cross-national studies that focus on a shorter time period and also use variation in diseases that are not geography-based and can be influenced by public health intervention. Herbst’s thesis is not stand-alone. Bates (1983) and Stevenson (1968) also provide some tentative evidence within Africa which is in agreement with the population density argument of the thesis. The exceptions are the settler colonies of South Africa, Zimbabwe, Namibia and Kenya. The Fertile Crescent spans a part of modern day Israel, Palestine and the Jordan valley. For further reference, see map in Diamond (1997) p. 135. Yang and Borland (1991) use a growth model to formalize Adam Smith’s notion of division of labour. This is disputed by many if we consider the last 50 years. World income distribution has increased over the last 50 years, especially with the impressive growth performances of India and China over the last two to three decades, but the diverging trend certainly holds over a longer period. Wacziarg and Welch (2003) improve the Sachs and Warner (1995a) measure and extend it to the 1990s. Their measure however is not free from the basic criticisms of Rodrik and Rodriguez (2000). I have discussed the political power and institutions theory of Acemoglu et al. (2005a) and Acemoglu and Robinson (2000) in Section 3.5.

M2545 - BHATTACHARYYA PRINT.indd 47

02/03/2011 10:09

4.

Empirical evidence

Before introducing the unified framework in Chapter 5, let me start with a discussion of some of the empirical issues involving the root causes of economic progress literature. First, I introduce some rough correlations in the data (both diagrammatically and in tabular form). Then I report some of the issues relating to the appropriateness of the levels model in estimation. I follow it up with a discussion on identification which is taken from my earlier work on this topic (see Bhattacharyya, 2009a). In Sections 4.5 and 4.6, I provide evidence on the role of diseases (malaria in particular) in Africa. The main message is that the causal linkages between deep determinants of development and economic development are complex and multidimensional. They also vary across countries and continents. The cross-section regressions are useful in identifying correlation, but may not tell us much about the causal links. The unifying framework in Chapter 5 is an attempt to identify such causal links.

4.1 A QUICK LOOK AT THE DATA The analysis here is based on a dataset which consists of per capita GDP levels, a measure of institutions, measures of geography, a measure of openness, measures of religion and a measure of human capital in (up to) 180 countries. Since I am combining data from different sources, I have to deal with different numbers of observations for different variables. The data typically also come from different years. In case of institutional measures, I use averages in order to capture the long-run effect. The definitions and sources of all the variables are summarized in the Data Appendix. Table 4.1 presents the summary statistics of these measures. I divide the dataset into six major parts depending upon the variable that they are measuring. They are as follows: measures of economic development, measure of institutions, measures of geography, measures of religion, measure of openness and trade and measure of human capital. Following is a brief outline of each of them. We also plot some of these variables to look at the correlation. Economic development is measured by the level of per capita GDP in

48

M2545 - BHATTACHARYYA PRINT.indd 48

02/03/2011 10:09

Empirical evidence

Table 4.1

49

Summary statistics

Variables

Number of obs

Measure of institutions Rule of law index Measures of geography Distance Malaria risk Land area within tropics Soil suitability Land area within 100km of ocean or oceannavigable river Measures of religion Catholicism Islam

Mean

Standard deviation

171

0.0015

0.9441

178 160 146

19.6293 0.3678 0.4991

17.0873 0.4390 0.4779

0 0 0

63.89 1 1

154 146

13.3965 46.0343

9.8512 37.604

0 0

55.07 100

174 174

31.3805 22.0817

36.1243 34.749

0 0

97.3 99.8

4.0914

0.6384

Measure of openness and trade Log of trade share 146 Measure of human capital Enrolment ratio in 88 1900 Measures of economic development Log initial income 50 (1820) Log initial income 60 (1870) Log initial income 39 (1900) Log initial income 136 (1950) Log initial income 115 (1960) Log per capita 147 GDP in 2000

32.34

26.37

Minimum Maximum

−2.19

2.58 0.1

2.37

5.76 95

6.58

0.3893

5.98

7.52

6.88

0.5668

5.98

8.09

7.46

0.6096

6.30

8.41

7.29

0.9446

5.67

10.32

6.3153

0.8647

4.72

8.23

8.588

1.1177

6.19

10.799

Note: For a detailed discussion of the definition and source of these variables, see Data Appendix.

M2545 - BHATTACHARYYA PRINT.indd 49

02/03/2011 10:09

50

Growth miracles and growth debacles

2000. I also use the level of GDP per capita in 1960 as the initial level of development. The average rule of law index is a measure of institutional quality. In particular, it measures the overall institutional quality and the quality of governance in a particular country. The range of this index varies from −2.5 to 2.5 with higher values implying better institutional quality. I use an average value of this index from 1998 through to 2000 for each of 171 countries as the measure. The data show that Somalia has the weakest institutions in the world and Singapore has the strongest institutions. In case of geography, I use five different measures: distance, malaria risk, land area within tropics, soil suitability and land area within 100 km of the ocean or an ocean-navigable river. Distance is a measure of the distance from the equator. Following the argument of Sachs (2003a), I use this as a measure of climate. The greater the distance from the equator, the further the country is from the tropics and more temperate or cold it is. I exploit data from 178 countries. Malaria risk measures the share of population at risk from malaria in 160 countries in the year 1997. I use this as a measure of disease burden. A higher value indicates greater risk for the population. Typically, tropical and subtropical countries register higher risk of malaria and the risk of malaria declines in the temperate climates. Land area within the geographical tropics is used as a measure of the effects of geography on agricultural productivity. This is calculated as a proportion, a higher value implying that a higher proportion of the country’s land is tropical in nature and, therefore, the country’s agriculture is expected to have tropical characteristics marked by slow and low plant growth. I use soil suitability as an additional measure of agriculture. Ukraine registers as the country with highest proportion of suitable soil. Finally, I use land area within 100 km of the ocean or an ocean-navigable river as a measure of market proximity. The higher the proportion of land within 100 km of the ocean or an ocean-navigable river, the greater is the chance for an economy to participate in maritime trade and have better access to larger markets. The bulk of the sub-Saharan economies is observed to be landlocked. In measuring religion, I use the proportion of population following Catholicism and Islam in the year 1980 as a measure of the same, respectively. The data shows Spain, Ireland, Portugal and the most of Latin America to be predominantly Catholic and the Middle East, Indonesia, Pakistan and some of the ex-Soviet republics to be predominantly Islamic. I measure openness to trade by using the log value of the actual trade

M2545 - BHATTACHARYYA PRINT.indd 50

02/03/2011 10:09

Empirical evidence

51

share from Frankel and Romer (1999). They calculate the actual trade share by taking the percentage of imports plus exports to GDP in 1985 from 146 countries. The higher the trade share, the more open is the economy. Singapore is observed to be the most open economy in the world and Myanmar is the most closed. The human capital measure is from Benavot and Riddle (1988). They record data on the primary school enrolment ratio in 1900. I use this as a proxy measure of historical human capital attainment. The data shows that in the United States, 95 per cent of the relevant school going population were enrolled in primary school in 1900 whereas in Equatorial Guinea the same number was as low as 0.1 per cent. Pairwise correlations are reported in Table 4.2. It shows that better institutions, less disease, better agriculture and soil conditions, more trade and higher initial income are all positively associated with development. This is also revealed in the following scatter plots (Figure 4.1a–i). The data shows that countries that were wealthier in 1960 were also likely to be wealthier in 2000. The data also shows that institutions measured by rule of law, trade and 1900 school enrolment rate are positively correlated to the current level of development. This is in line with the prediction made by many theorists of root causes of development. In contrast, malaria risk and land area within the tropics are negatively correlated to the current levels of development. The scatter plots also show that countries further from the tropics, countries with proximity to oceans and ocean-navigable waterways, and countries with Catholicism as the dominant religion are relatively prosperous. The obvious question is, does this imply causality? I will discuss this issue in detail in what follows.

4.2 GROWTH OR LEVELS: WHICH ONE IS THE APPROPRIATE EMPIRICAL FRAMEWORK? The levels framework adopted by Rodrik et al. (2004) and other recent papers1 centres on a cross-country regression where the dependent variable is the current level of economic development, typically measured by the natural logarithm of real GDP per capita, lnyiT, where i indexes countries and T indicates the year. The explanatory variables consist of a measure of the ‘quality’ of contemporary institutions, IiT, and a vector of other explanatory variables, WiT. The other explanatory variables include log trade share in the current period and latitude: ln yiT 5 hIiT 1 FWiT 1 ei

M2545 - BHATTACHARYYA PRINT.indd 51

(4.1)

02/03/2011 10:09

52

M2545 - BHATTACHARYYA PRINT.indd 52

02/03/2011 10:09

0.6298 −0.6265 −0.4975 0.3431

0.1629 0.7783

0.2374 −0.3310

0.4758

0.2573

0.7073

0.1331 −0.3178

0.7387

0.8439

0.6931 −0.8299 −0.6210

1.0000

−0.0160 −0.2768

0.7613

0.3310

0.3896

0.6837 −0.6551 −0.5993

1.0000

−0.0857 −0.0500

0.6522

0.0907

0.2096

1.0000 −0.6647 −0.8956

Log initial Average Distance income rule of law (1960) index

0.8182

1.0000

Log per capita GDP in 2000

−0.1445 0.2190

−0.4504

−0.1831

−0.3874

1.0000 0.6493

Malaria risk

0.1473 −0.0458

−0.5681

0.0101

−0.0854

1.0000

0.1417 −0.2145

0.2025

0.3128

1.0000

Land area Land area within within tropics 100 km of ocean or oceannavigable river

Pairwise correlation of the major variables (obs 5 99)

Log per capita GDP in 2000 Log initial income (1960) Average rule of law index Distance Malaria risk Land area within tropics Land area within 100 km of ocean or oceannavigable river Log of trade share Enrolment ratio in 1900 Catholicism Islam

Table 4.2

−0.1382 −0.0062

0.1415

1.0000

Log of trade share

0.0208 −0.3452

1.0000 1.0000 −0.5104

Primary Catholicism enrolment ratio in 1900

1.0000

Islam

Empirical evidence

53

11

Fitted values/per capita GDP in 2000

LUX

10

9

8

7

USA NOR SGP CAN DNK CHE HKG IRL AUS ISL JPN NLDSWE BEL AUT FIN GER FRA GBR ITA NZL ISR ESP BRB TWN PRT GRC KOR MUS TTO ARG SYC CHL GNQ PRI URY MEX MYS ZAF BRA GAB TUR VEN IRNPAN TUN BWA THA DZA FJI CRI COL DOM ROM PRY PER NAM SLV SYR EGY JOR GTM CPVMAR PHL GUY IDN CHN LKA ECUJAM PNG GIN BOL IND ZWE COG HTI CMR PAK HND CIV NIC COM BGD SEN LSONPL MRT KEN BEN GMB GHA MOZ UGA BFA MLI RWA TGO TCD MDG NER CAF ZMB NGA MWI SLE ETH GNB BDI TZA

6 5

6

7

8

Initial income

Note: See Table 8.2 for country abbreviations.

Figure 4.1a

Log per capita GDP in 2000 and log initial income

11

Fitted values/per capita GDP in 2000

LUX USA NOR CAN CHE SG P HKG IRL AUS ISLDNK JPN BEL NLD AUT SWE FIN GBR FRA ITA NZL ISR TWN ESP PRT SVN GRC KOR CZE KNA MUS TTO SVKARG SYC HUN EST CHL PRI GNQ URY RUS MEX POL HRV MYS BLR LTU KAZ DMA ZAF VCT LVA GAB BRA TUR VEN BLZIRN BWA TUN LCA THA PAN GRDLBN DZA BGR CRI COL DOM MKD FJI GEO SWZ PRYUKRROM PER NAM SLV SYR GTM MAR CPV JOR PHL IDNGUYEGY CHN LKA ECU ALB AZE PNG JAM KGZ ARM GINBOL IND COG ZWE HTI CMR MDA HND PAK CIV COM NICBGD SEN NPL LSO MRT TJK GHA GMB STP KEN KHM YEM BEN MOZ UGA MLI TCD BFA CAF RWA TGO NER ZMB MDG NGA MWI GNB SLE ETH BDI TZA

10

9

8

7

6 –2

–1

0

1

2

Rule of law

Note: See Table 8.2 for country abbreviations.

Figure 4.1b

Log per capita GDP in 2000 and rule of law

M2545 - BHATTACHARYYA PRINT.indd 53

02/03/2011 10:09

54

Growth miracles and growth debacles 11

Fitted values/per capita GDP in 2000

LUX USA NOR CAN SGP DNK CHE HKG AUS IRL ISL JPN NLD BEL AUT SWE FIN GBR GER FRA ITA NZL ISR ESP TWN PRT SVN GRC KOR CZE MUS TTO SVK ARG HUN EST PRI CHL URY RUS POL HRV BLR LTU KAZ LVA ZAF TUR VEN BRA TUN BGR DZA LBN CRI DOM MKD GEO UKR SWZ ROM PRY PER SLV EGY GTM MAR JOR CHN ALB AZE JAM KGZ ARMBOL

10

9

8

GNQ

MYS MEX IRN PAN BWA

GAB BLZ

THA COL NAM IDN GUY LKA PNG ZWEIND

SYR ECU

PHL GIN COG HTI CMR PAK HND CIV NIC SEN GHA GMB YEMKHM BEN MOZ UGAA MLI BFA CAF RW TCD TGO NER MDG ZMB NGA MWI GNB SLE BDI TZA

MDA BGD

NPL

LSO TJK

MRT KEN

7 ETH

6 0

0.2

0.4

0.6

0.8

1

Malaria risk

Note: See Table 8.2 for country abbreviations.

Figure 4.1c

Log per capita GDP in 2000 and malaria risk

11 LUX

Fitted values/per capita GDP in 2000

USA SGP GER MAC

10

ATG SVN CZE SVK SYC EST GNQ HRV MYS BLR LTU KAZ LVA GAB

HKG TWN

BRB TTO

AUS JPN ISR

KNAMUS

NZL ESP PRT GRC KOR

ARG PRI CHL URY MEX DMA BRA ZAF VCT VEN THA IRN TUN LCA BLZ BWA PANGRD DZA LBN COL FJI CRI DOM MKD GEO UKR SWZ PER PRY NAM SLV EGY GTM SYR CPV MAR JOR PHL IDN CHN LKA ECU GUY AZE JAM PNG KGZ ARM GIN BOL IND ZWE COG CMR HND HTI MDA PAK CIV NIC COMSEN BGD NPL LSO MRT TJK GHA GMB STP KEN KHM YEM BEN MOZ UGA MLI BFA CAF RWA TGO TCD NER ZMBMWI MDG NGA SLE ETH GNB BDI TZA

9

8

7

NOR CAN CHE DNK IS L NLDIRL AUTBEL SWE FIN ITA FRAGBR

HUN POL

RUS

TUR BGR ROM

6 0

20

40

60

Distance

Note: See Table 8.2 for country abbreviations.

Figure 4.1d

Log per capita GDP in 2000 and distance

M2545 - BHATTACHARYYA PRINT.indd 54

02/03/2011 10:09

Empirical evidence

55

11 USA

Fitted values/per capita GDP in 2000

CAN

AUS

10

NOR

CHE SWE FIN

URY RUS MEX KAZ LVA ZAF BRA GAB TUR TUN VEN IRN BWA THA DZA COL MKD GEO ROM NAM PER PRY EGY SYR MARGTM JOR CHN ECU AZE KGZ ARM GIN BOL IND ZWECOG PAKCMRCIV

8

NPL LSO TJK

7

MRT KEN BEN

GHA

HUN MYSPOL BLR LTU

BGR UKR IDN

ALB

PNG

NIC SEN

UGA MLI BFA CAF RWA TCD NER ZMB MWI ETH BDI

FRA ITA PRT CZE SVK CHL EST

ARG

9

DN K HKG IRLJPN SGP NLD GBR GERBEL ISR NZLTWN SVN KORGRC MUS TTO

AUT ESP

HRV PAN LBN CRIM DO SLV PHL LKA JAM HTI A MD BGD

HND

GMB

KHM YEM MOZ

TGO NGA

MDG GNB

SLE

TZA

6 0

20

40

60

80

100

Land area within 100 km of ocean or ocean-navigable river

Note: See Table 8.2 for country abbreviations.

Figure 4.1e

Log per capita GDP in 2000 and land area within 100 km of ocean

Fitted values/per capita GDP in 2000

11 USA NOR CAN DNK CHE IRL JPN NLD BEL AUT SWE FIN GBR GER FRA ITA NZL ISR ESP PRT SVN GRC KOR CZE SVK ARG HUN EST URY RUS POL HRV BLR LTU KAZZAF LVA TUR IRN TUN BGR LBN MKD GEO UKR ROM SYR MAR JOR CHN ALB AZE KGZ ARM

10

9

8

SGP HKG

AUS TWN

MUS TTO CHL

MEX

MYS BRA GAB VEN THA PAN CRI COL DOM PER SLV GTM PHL IDN LKA ECU JAM PNG GIN BOL ZWE COG HTI CMR HND CIV NIC SEN MRT GHA GMB KEN KHM YEM BEN MOZ UGAA MLI RW BFA CAF TCD TGO NER MDG ZMB NGA MWI GNB SLE ETH BDI TZA

BWA

DZA PRY

EGY

NAM

IND

MDA PAK BGD

NPL LSO TJK

7

6 0

0.2

0.4

0.6

0.8

1

Land area within tropics

Note: See Table 8.2 for country abbreviations.

Figure 4.1f

Log per capita GDP in 2000 and land area within tropics

M2545 - BHATTACHARYYA PRINT.indd 55

02/03/2011 10:09

56

Growth miracles and growth debacles 11 LUX

Fitted values/per capita GDP in 2000

USA

NOR CAN DNK CHE HKG AUS IRL ISL JPN NLD BEL FIN SWEAUT FRA GBR ITA NZL ISR ESP PRT TWN BRB GRC KOR MUS TTO ARG HUN EST SYC PRI CHL URY RUS MEX POL MYS DMA ZAF VCT GAB TUR VEN IRN BRA BLZ TUN BWA THA PAN GRD LCA BGR DZA CRI COL DOM SWZ ROM PER PRY NAM SLV EGY GTM SYRIDN CPV JOR PHL MAR CHN GUY ECU LKA JAM PNG GIN BOL IND ZWE COG HTI CMR PAK HND CIV NIC COM BGD SEN NPL LSO MRT GHA GMB KEN YEM BEN MOZ UGA RWA MLI BFATCD CAF TGO NER MDG ZMB NGA MWI GNB SLE BDI

10

9

8

7

SGP

TZA

6 2

3

4 Natural log of trade share

5

6

Note: See Table 8.2 for country abbreviations.

Figure 4.1g

Log per capita GDP in 2000 and log trade share

11

Fitted values/per capita GDP in 2000

LUX NOR DNK HKG ISL SGP JPN SWE MACGBR FIN NZL ISR TWNBRBATG GRC KORKNA

10

USA AUS

GER

MUSTTOCZE

EST RUS MYS KAZ ZAFBLRLVA VCT TUR IRN TUN BWA THA GRD BGR DZA LBN FJI MKD GEO UKR ROM SWZ NAM EGY SYR MAR JOR IDN CHN GUY ALB LKA AZE JAM PNG KGZ ARM GIN IND ZWE CMR MDA PAK CIV BGD NPL SEN MRT TJK GHA KEN GMB KHM YEM BEN MOZ MLI BFA TCD TGOCAF NER MDG ZMB NGA MWI GNB SLE ETH

9

8

7

CAN CHE NLD

FRA

ITA

BEL AUT

IRL

ESP PRT

SVN

SVK ARG SYC HUN URY PRI CHL GNQ HRVPOL MEX LTU DMA GAB BRA VEN BLZ LCA PAN CRI DOM COL PER PRY SLV GTM CPV PHL ECU BOL HTI HND NIC

LSO

STP UGA RWA BDI

TZA

6 0

20

40

60

80

100

Catholicism

Note: See Table 8.2 for country abbreviations.

Figure 4.1h

Log per capita GDP in 2000 and Catholicism

M2545 - BHATTACHARYYA PRINT.indd 56

02/03/2011 10:09

Empirical evidence

57

11

Fitted values/per capita GDP in 2000

HKG

10

FIN

TWN

ITA

PRT MUS

GRC

ARG PRI URY RUS CHLMEX ZAF BRA VEN BLZ TUNPAN BGR DZA CRI DOM COL ROM PER PRY NAM EGY SLV GTM PHL IDNSYR LKA GUY ECU

GNQ

9

8

BOL IND ZWE CMR GHA GMB

ESP BRB

USA CAN IRL AUS FRA NZL

TTO SYC HUN VCT LCA

GRD

FJI

JAM

HND

NIC

7

NOR DNK CHE BEL NLD AUT SWEGER GBR

JPN

LSO

TGOMDG SLE

MWI

6 0

20

40

60

80

100

Enrolment ratio in 1900

Note: See Table 8.2 for country abbreviations.

Figure 4.1i

Log per capita GDP in 2000 and enrolment ratio in 1900

It is apparent that this model, which is known as the levels regression model, is nested within the growth specification reported in (2.7),2 since it can be rewritten as: ln yiT 5 (1 1 aT) ln yit 1 hIiT 1 FWiT 1 ei

(4.2)

where g 5 [ hF ] , ZiT 5 sIiT t and 1 1 aT 5 0. W iT

The key question is whether (4.1) is the appropriate empirical framework to test the role of deep structural determinants on long-term economic progress. Sachs (2003a) argues that economic growth is a dynamic process rather than a state. Therefore, it is more appropriate for empirical studies to estimate growth regressions rather than levels regression. Alternatively, Dollar and Kraay (2003) argue that during the sixteenth century, income across the world was largely similar. Therefore, the growth model with an initial income around 1500 essentially boils down to a levels model. Let me illustrate how. The existing literature implicitly starts from the following model:

M2545 - BHATTACHARYYA PRINT.indd 57

02/03/2011 10:09

58

Growth miracles and growth debacles

ln yiT 5 a 1 b ln yiT 2500 1 gIiT 1 dXiT 1 ei

(4.3)

The above is a growth regression with the current level of economic development, measured by the natural logarithm of real GDP per capita, ln yiT as the dependent variable, where i indexes countries and T indicates the year. The explanatory variables consist of the initial level of economic development (or the level of development 500 years ago), ln yiT 2500, a measure of the ‘quality’ of contemporary institutions, IiT, and a vector of other explanatory variables, XiT, which includes human capital, geography and trade. Under the assumption that the levels of development in the distant past were not too different across countries3 the long-run growth literature generally estimates the following version of the above model which is similar to (4.1): ln yiT 5 a 1 gIiT 1 dXiT 1 ei

(4.4)

To estimate the causal effect of institutions on long-run growth, the literature uses historical and geographical instruments. These instruments predict the historically and/or geographically determined component of institutions at the first stage and then they are used in the second stage to estimate the causal effect of institutions on long-run growth. Easterly (2009) also supports the levels framework over growth. He argues growth is inherently unstable and it is difficult to predict in advance what policy or set of policies would work. Therefore, with a small crossnational sample on growth, it is easy to jump into erroneous conclusions too quickly. This is what he calls the ‘law of small numbers’. He suggests that the levels framework is more appropriate as it reflects the outcome of entire previous growth and hence is free from the problems of small numbers. Nevertheless, to statistically test whether growth or levels is the appropriate framework, we use the following method. The null hypothesis, H0: (1 1 aT) 5 0, reduces (4.2) to the levels regression, whilst rejection of the null favours the growth specification. The economic intuition behind this test is that the levels regression is implicitly explaining the steady-state distribution of income levels. This assumption is explicit in the augmented Solow model derived by Mankiw et al. (1992) who use investment rates as the proximate determinants of the neo-classical steady state. Mankiw et al. (1992) go on to show that if economies are not in their steady states, the transitional dynamics of the neoclassical model are captured by the addition of the ‘initial’ income level in a growth regression.4 If economies actually are in steady state, as explained by the right-hand-side variables in Equation 4.1, then the addition of the lagged dependent variable, as in Equation 4.2, should add no explanatory power.

M2545 - BHATTACHARYYA PRINT.indd 58

02/03/2011 10:09

Empirical evidence

Table 4.3

59

Omitted variable test: OLS regressions

Dependent Variable

Log initial income (1820) Log initial income (1870) Log initial income (1900) Log initial income (1950) Rule of law (2001) Average log trade Share Latitude R2

Log per capita GDP in 2000 OLS obs 5 44

OLS obs 5 52

OLS obs 5 35

OLS obs 5 108

0.35** (0.1599) 0.36** (0.1949) 0.46** (0.1705)

0.78*** (0.0780) −0.043 (0.1150) −0.003 (0.0046) 0.8303

0.60*** (0.1345) 0.043 (0.1376) 0.001 (0.0053) 0.7522

0.399*** (0.1325) −0.133 (0.1526) 0.006 (0.0070) 0.7992

0.598*** (0.0858) 0.49*** (0.0936) 0.15* (0.1013) 0.01** (0.0040) 0.8181

Notes: ***, ** and * indicate significance levels of 1%, 5% and 10%, respectively, against a one-sided alternative. Figures in parentheses are the respective standard errors. The standard errors reported in the regressions are heteroskedasticity robust. All the regressions reported above are carried out with an intercept.

I test the null hypothesis by adding an initial income term into Rodrik et al.’s (2004) levels specification. The results are reported in Table 4.3. In all the four regressions using 1820, 1870, 1900 and 1950 levels of log per capita GDP as initial income,5 I observe that the initial income term is statistically significant. This indicates that the standard levels framework neglects the role of transitional dynamics and the coefficient estimates suffer from omitted variable bias.

4.3 IDENTIFICATION ISSUES WITH THE LEVELS FRAMEWORK The levels framework also suffers from identification issues when human capital is added into the mix. I illustrate this point in Bhattacharyya (2009a). In Table 4.4, I adopt the same strategy and add schooling into the levels model to estimate partial effects of human capital and institutions. In

M2545 - BHATTACHARYYA PRINT.indd 59

02/03/2011 10:09

60

M2545 - BHATTACHARYYA PRINT.indd 60

02/03/2011 10:09

Identification issues in long-run growth empirics: part A

F-test Sargan test (p)

Latitude

Total years of schooling (2000) Avg log trade share

Rule of law (2001)

Explanatory variables

0.10

−0.01 (0.0140)

−0.47 (0.2849)

1.55*** (0.2541)

Original model obs 5 68 (1)

0.08 (0.1219) 0.038** 0.99

1.1 (1.322) 2.47 (4.023)

3.55 (6.224)

Sargan test (p)

Latitude

Total years of schooling (1995)

Avg protection against expropriation 1985–1995

Model with Explanatory schooling obs 5 43 variables (2)

–

0.002 (0.0128)

0.96*** (0.2227)

Original model obs 5 65 (3)

0.98

0.04 (1.428)

1.35 (38.11)

−2.72 (104.2)

Model with schooling obs 5 51 (4)

Acemoglu et al. (2001) Dependent variable: log per capita GDP in 1995

Panel A: 2SLS Results

Rodrik et al. (2004) Dependent variable: log per capita GDP in 2000

Table 4.4

61

M2545 - BHATTACHARYYA PRINT.indd 61

02/03/2011 10:09

0.7311 33.57

−1.32*** (0.2243) 2.05** (0.8444) 1.44** (0.5645) 0.002 (0.0164) 0.014 (0.0199)

obs 5 56 (2)

1.29*** (0.2949) 0.8061

−0.83*** (0.1923) 1.1* (0.6263) 1.33*** (0.4949) −0.026* (0.0141) −0.01 (0.0200)

obs 5 56 (3)

Schooling (2000)

R2 F-stat

Avg. prot. against expropriation 1985–1995

Latitude

Log settler mortality L. pop. den. in 1500

Dependent variable

0.2553 10.97

0.02 (0.0155)

−0.47*** (0.1448)

Avg prot. against expropriation obs 5 67 (4)

0.7195 43.61

0.02 (0.0202)

−1.39*** (0.2179) −0.58*** (0.1333)

obs 5 55 (5)

0.7692

0.01 (0.0189) 0.49*** (0.1683)

−1.11*** (0.2153) −0.49*** (0.1283)

obs 5 51 (6)

Schooling (1995)

Notes: ***, ** and * indicate significance levels of 1%, 5% and 10%, respectively, against a two-sided alternative. Figures in parentheses are the respective standard errors. Joint F-test p-value of rule of law and schooling is reported in column 2, Panel A. The standard errors are heteroskedasticity robust. L. pop. den. in 1500 is log population density in 1500, ENGFRAC is fraction of population speaking English, EURFRAC is fraction of population speaking European languages. In Panel B the dependent variables are rule of law (2001), schooling (2000), average protection against expropriation, and schooling (1995).

Rule of law (2001) R2 F-stat

Latitude

CONST

0.5472 21.54

−0.25*** (0.1039) 0.95*** (0.3311) 0.18 (0.2197) 0.008 (0.0057) 0.02** (0.0088)

Log settler mortality ENGFRAC

EURFRAC

Rule of law (2001) obs 5 76 (1)

Dependent variable

Panel B: first stage regressions

62

Growth miracles and growth debacles

column 1 (Panel A, Table 4.4) I follow Rodrik et al. (2004)6 and estimate their preferred model using the same set of instruments as they did in their study. Rule of law is the only statistically significant variable in this model, which confirms their basic finding that institutions dominate the influence of both trade and geography as the fundamental determinant of long-run development. In column 2, I add schooling into this model and we are unable to isolate the partial effects of institutions and human capital as none of the coefficients are statistically significant. However, the F-test on the joint significance of schooling and rule of law reveals that they are jointly significant with a p-value of 0.038. To investigate the reason behind this, I look at the first stage regressions reported in columns 1 and 2 of Panel B. In column 1, I confirm that there is a strong partial correlation between the settler mortality instrument and current rule of law – a result well documented in the literature. However, I also find that the settler mortality instrument is correlated with current schooling (see column 2) and this correlation is independent of the correlation between current schooling and current rule of law (see column 3). I observe that the correlation between fitted values of rule of law and the fitted values of schooling is as high as 0.9435. This is perhaps causing a severe multicollinearity problem in the second stage regression making all coefficients statistically insignificant. In column 3 (Panel A, Table 4.4), I follow Acemoglu et al. (2001)7 and estimate their preferred model. I also confirm their finding that institutions measured by expropriation risk have a causal effect on long-run growth. In column 4, I add schooling in 1995 into this specification and I confront the same multicollinearity problem that I encountered in column 2. None of the coefficient estimates are statistically significant. I take a quick look at the first stage regressions (Panel B, columns 4–6). They reveal that the settler mortality instrument is as good a predictor of schooling as it is for expropriation risk, and this relationship is independent of the correlation between schooling and expropriation risk. Also, the correlation between fitted values of expropriation risk, and schooling is as high as 0.9036. This, again, is causing a multicollinearity problem at the second stage. In Table 4.5, I follow Acemoglu and Johnson (2005) and Easterly and Levine (2003) and estimate their preferred specifications.8 Acemoglu and Johnson (2005) evaluate the relative importance of ‘property rights institutions’ and ‘contracting institutions’ measured by ‘constraint on the executive’ and the ‘legal formalism index’, respectively, and find that the former have a first-order effect on long-run growth and the latter does not seem to matter.9 Easterly and Levine (2003) show that endowment measured by settler mortality, latitude, landlocked dummy and crops/minerals dummy affect development through institutions. I confirm their findings (see Panel A, columns 1 and 3). However, I encounter the same problem

M2545 - BHATTACHARYYA PRINT.indd 62

02/03/2011 10:09

63

M2545 - BHATTACHARYYA PRINT.indd 63

02/03/2011 10:09

Constraint on executive Total years of schooling (1995) Sargan test (p)

0.99

0.06 (0.1735) 0.77*** (0.1869)

Original model obs 5 41 (1) 0.12 (0.2755) 0.45 (1.731) 0.22 (0.6945) 0.99

Model with schooling obs 5 37 (2)

Sargan test (p)

Total years of schooling (1995) Oil

Institutions index

1.56*** (0.4275) 0.26

2.19*** (0.1735)

Original model obs 5 72 (3)

0.99*** (0.2909) 0.26*** (0.0671) 0.93** (0.3638) 0.13

Model with schooling obs 5 55 (4)

Easterly and Levine (2003) Dependent variable: log per capita GDP in 1995

Explanatory variables

Panel A: 2SLS Results

Acemoglu and Johnson (2005) Dependent variable: log per capita GDP in 1995

Identification issues in long-run growth empirics – part B

Legal formalism

Explanatory variables

Table 4.5

64

M2545 - BHATTACHARYYA PRINT.indd 64

02/03/2011 10:09

15.16

36.31

0.6565 44.79

0.7249

−1.38*** (0.2048) 0.56 (0.4413) −0.597*** (0.1323)

obs 5 55 (3) −1.32*** (0.2168) 0.38 (0.4387) −0.56*** (0.1353) 0.13 (0.1370) 0.7341

obs 5 53 (4)

Total years of schooling (1995)

Crops/minerals (10 variables) Institutions index R2 F-stat

Oil

Landlocked

Log settler mortality Latitude

Dependent variable

Panel B: First Stage Regressions

0.6379 7.17

−0.14** (0.0652) 1.21* (0.6931) −0.09 (0.1886) −0.25 (0.2028) 2.41** (0.018)

Institutions index obs 5 72 (5)

0.7053 6.84

−1.5*** (0.3061) 3.5 (2.997) −0.28 (0.8690) 0.60 (1.02) 1.01 (0.452)

obs 5 55 (6)

−1.0*** (0.2398) 1.5 (2.238) −0.26 (0.6415) 1.26* (0.7609) 1.22 (0.306) 2.4*** (0.4098) 0.8434

obs 5 55 (7)

Total years of schooling (1995)

Notes: ***, ** and * indicate significance levels at 1%, 5% and 10%, respectively, against a two-sided alternative. Figures in parentheses are the respective standard errors. For crops/minerals (10 variables) the table reports F-test of joint significance of the individual variables with p-value in parentheses. English legal origin is used as an instrument for legal formalism in Acemoglu and Johnson (2005). L. pop. den. in 1500 is log population density in 1500. In panel B the dependent variables are constraint on executive, legal formalism, institutions index, and total years of schooling (1995).

F-stat

0.3285

−0.91*** (0.2104) −0.28 (0.4617)

Log settler mortality English legal origin L. pop. den. in 1500 Constraint on executive R2

0.15 (0.1146) −2.02*** (0.2515)

Constraint Legal formalism obs 5 41 (2) on executive obs 5 41 (1)

Dependent variable

Table 4.5 (continued)

Empirical evidence

65

of multicollinearity when I add schooling into these specifications.10 The first stage regressions of panel B show that the standard instruments used in these specifications are as good a predictor of institutions as they are of schooling. This makes the fitted values of institutions and schooling for the second stage regressions correlated with each other – correlation coefficient of 0.9226 in the case of the Acemoglu and Johnson (2005) specification and correlation coefficient of 0.8562 in the case of the Easterly and Levine (2003) specification. This causes the multicollinearity problem at the second stage. Furthermore, I also notice similar problems with the model estimated in Acemoglu et al. (2003b). Acemoglu et al. (2003b) argue that the fundamental cause of post-war instability in many of the least developed countries is institutional. They show that the volatility in per capita GDP growth and other macroeconomic indicators have a strong negative relationship with institutions. They argue, using the settler mortality instrument, that this negative relationship has its roots in the colonial institutions. They also show that this finding is robust when they control for religion, latitude, initial income, log inflation and so forth. However, this finding also suffers from the identification problem that I have documented before. In Table 4.6, I outline the problem using a particular measure of volatility – the standard deviation of GDP per capita growth (1970–98). In column 1 of Panel A it can be seen that initial executive constraint, which is a measure of institutions in 1950, 1960 and 1970, negatively impacts volatility in output growth in the following two decades. This implies that institutional weaknesses in the past lead to higher volatility in the current growth rate. This result holds when I control for latitude. But if I replace the initial constraint on the executive by total years of schooling in 1960, then the same negative relationship is observed, even though this time it is between schooling and volatility. This is indicative of the identification problem similar to Acemoglu et al. (2001), Rodrik et al. (2004) and Acemoglu and Johnson (2005), which I discussed previously. The first stage regressions in Panel B show that the log settler mortality instrument is not only correlated with initial constraint on the executive, but it is also correlated with schooling. The correlation between schooling and log settler mortality is independent of the correlation between schooling and initial constraint on the executive. Therefore, it appears that the cross-sectional data is not ideal for separating out the partial effects of institutions and human capital on long-run growth. This is because the fitted values of the endogenous variables in the second stage of a 2SLS regression can be highly correlated with each other given the set of instruments used. Multicollinearity at the second stage blows up the standard errors making all coefficient estimates statistically insignificant.11 However, I may end up with an entirely different result if

M2545 - BHATTACHARYYA PRINT.indd 65

02/03/2011 10:09

66

Table 4.6

Growth miracles and growth debacles

Identification problems in Acemoglu et al. (2003)

Dependent variable

Standard deviation of GDP per capita growth (1970–98) Model (1) obs 5 71

Model (2) obs 5 52

Panel A: second stage of 2SLS Initial constraint on executive Total years of schooling (1960) Latitude

−0.46** (0.1941) −0.56*** (0.1603) 0.03 (0.0231)

−0.04 (0.0214)

Panel B: first stage for initial constraint on executive and total years of schooling (1960) Dependent variables

Initial constraint on executive obs 5 71

Log settler mortality Latitude Initial constraint on executive R2

Total years of schooling (1960) obs 5 52

obs 5 49

−0.85*** (0.1957)

−1.2*** (0.2208)

−0.94*** (0.2685)

−0.003 (0.0216)

0.05** (0.0223)

0.2590

0.5387

0.05** (0.0224) 0.23* (0.1339) 0.5718

Notes: ***, ** and * indicate significance levels of 1%, 5% and 10%, respectively, against a two-sided alternative. Figures in parentheses are the respective standard errors. All the regressions reported above are carried out with an intercept. Log settler mortality is used as an instrument for initial constraints on executive.

a different instrument set is used. Hence, the appropriate interpretation is that the regressions are simply uninformative as they provide no information that would alter our theoretical priors one way or the other. In other words, based on these results, I am unable to comment on the relative importance of institutions and human capital to long-term development. Nevertheless, we do learn from the empirical studies conducted that institutions protecting individual rights and supporting entrepreneurship matter. We learn that geography plays an important role in economic development, be it directly or through shaping incentives and economic institutions. We also learn that geography is critically important for the continent of Africa (Bhattacharyya, 2009b).

M2545 - BHATTACHARYYA PRINT.indd 66

02/03/2011 10:09

Empirical evidence

67

With this as a background, I introduce the unified framework to explain the process of development in Western Europe. I then compare and contrast that with the history of development in other parts of the world.

4.4 EXISTING RESULTS WITH THE LEVELS FRAMEWORK As I have illustrated above, the majority of the existing results in the root causes of economic progress literature indicate that institutions are a major driver of economic progress over the long run (Acemoglu et al., 2001; Rodrik et al., 2004; Acemoglu and Johnson, 2005; Easterly and Levine, 2003). Sachs (2003a) and Carstensen and Gundlach (2006) show that it is not just institutions; diseases are also important. In this section, I show that there is some heterogeneity in the institutions result. Institutions are no longer important if we divide the sample between high-income and low-income countries using the World Bank criteria. In a sample of low-income countries, malaria turns out to be the main explanator of lack of development. I use the levels framework to demonstrate this effect. Therefore, I implicitly assume that initial income during the sixteenth century was similar across countries. The results, however, are robust to the inclusion of initial income. I report the results as follows. In order to dig deeper into the malaria result discussed above, I ask a further question. Is the finding a reflection of the overall variation in the data or is it largely driven by the significant presence of the low-income economies (LIEs) in the sample? A closer look at the dataset reveals that 26 out of 39 observations used in the estimation are from LIEs.12 Estimating the same model using a sample of LIEs shows that malaria is the only variable with statistically significant coefficient estimates. This is reported in Table 4.7a. The impact of one sample standard deviation reduction in malaria (which is 0.44) on per capita GDP is 1.7 fold. This is less than the pooled sample estimate of a twofold increase. However this is the only variable that is statistically significant. Estimating the same regression on a sample of high-income economies (HIEs) fails to yield meaningful estimates because of very few degrees of freedom. The regression relies on as few as 13 observations. The ordinary least squares (OLS) estimates with the HIE sample, however, show that diseases are no longer important. It is only institutions that matter. The partial R2 of the OLS estimates reported in Table 4.7b show that in a pooled sample 28 per cent of the variation in per capita GDP is explained by institutions and about 41 per cent by malaria, whereas in a sample of LIEs, the explanatory power of institutions is statistically insignificant and 43 per cent of the variation is

M2545 - BHATTACHARYYA PRINT.indd 67

02/03/2011 10:09

68

Table 4.7a

Growth miracles and growth debacles

Determinants of level of development: the case of low-income economies (LIEs)

Dependent variable

Log per capita GDP in 2000 OLS Pooled sample obs 5 66

0.45*** (0.0947) −0.04 (0.0938) −1.03*** (0.1839) Catholicism −2e-04 (0.0012) Enrolment 3e-04 ratio in 1900 (0.0029) R2 0.8833 F-test (p-value) 0.0000 Instruments

Average rule of law index Log of trade share Malaria risk

2SLS

Restricted sample (LIEs) obs 5 32

Restricted sample (HIEs) obs 5 34

Pooled sample obs 5 39

Restricted sample (LIEs) obs 5 26

0.23 (0.2050) −0.12 (0.1585) −0.87*** (0.2011) −1.6e-04 (0.0022) 0.002 (0.0083) 0.6742 0.0000

0.41*** (0.1103) 0.02 (0.1073) −0.05 (0.6148) −0.002* (0.0013) 0.003 (0.0029) 0.7841 0.0000

0.89*** (0.3173) −0.08 (0.1685) −1.61*** (0.3238) 0.004 (0.0032) 0.004 (0.0049)

0.83 (0.7267) −0.45 (0.3345) −1.17*** (0.5040) 0.001 (0.0041) 0.02 (0.0139)

Log settler mortality, ENGFRAC, EURFRAC, Constructed openness, Malaria ecology

Notes: ***, ** and * indicate significance levels of 1%, 5% and 10%, respectively, against a one-sided alternative. Figures in parentheses are the respective standard errors. LIEs and HIEs signify low-income economies and high-income economies, respectively. For the 2SLS regression on the HIEs, malaria no longer remains significant. The 2SLS regression on the HIEs does not yield meaningful estimates due to the problem of degrees of freedom. There are only 13 observations for the HIEs and hence are not reported here. The standard errors reported in the regressions are heteroskedasticity robust. All the regressions reported above are carried out with an intercept.

explained by malaria. In an HIE sample, the situation reverses and institutions perform 34 per cent of the explaining and the contribution of malaria is statistically insignificant. Based on the evidence, therefore, it is fair to conclude that the major effect of malaria on per capita GDP comes from the LIEs. Institutions become more important in a sample of HIEs. Therefore, one can put forward the following hypothesis with some caution to explain this correlation.13 Disease is the most important factor at a low level of development. Overcoming diseases ensures a basic

M2545 - BHATTACHARYYA PRINT.indd 68

02/03/2011 10:09

Empirical evidence

Table 4.7b

69

Partial R2 of OLS estimates (reported in Table 4.7)

Dependent variable

Log per capita GDP in 2000 Pooled sample obs 5 66

Average rule of law index Natural log of trade share Malaria risk Catholicism Enrolment ratio in 1900

Restricted sample (LIEs) obs 5 32

Restricted sample (HIEs) obs 5 34

0.283***

0.048

0.344***

0.004

0.021

0.001

0.413*** 0.0004 0.0002

0.430*** 0.0002 0.002

0.0003 0.088* 0.034

Notes: ***, ** and * indicate significance levels of 1%, 5% and 10%, respectively, against a one-sided alternative. The partial R2 indicates the proportion of variation in the dependent variable that is explained by the explanatory variables when other things are held constant. Multiplying the numbers by 100 will give the percentages.

minimum level of labour productivity and from then on institutions become important. One implication could be that the data on economic development has twin peaks and there are multiple equilibria. The lower peak can be best explained by geography. Note that this is not just geography per se, but the poverty traps that emerge out of it. However, just by looking at the log settler mortality data (which is an instrument in the 2SLS regression) one may suspect that the strong negative correlation between malaria and current per capita GDP in LIEs is driven by the large proportion of high malaria and low per capita GDP African economies in the sample. To find out the true picture, I perform some quick statistical analysis, which, in fact, confirms the suspicion. There is a strong association between malaria and low per capita GDP when the sample size is reduced down to African LIEs. The next section presents this finding and discusses its possible implications.

4.5 IS THERE ANY CONNECTION BETWEEN LOWINCOME COUNTRIES AND AFRICA? A quick check of the sample that the 2SLS regression (see Table 4.7a) on LIEs uses reveals that 10 out of 26 observations are African LIEs, which is approximately 39 per cent of the sample.

M2545 - BHATTACHARYYA PRINT.indd 69

02/03/2011 10:09

70

Table 4.8

Growth miracles and growth debacles

Determinants of level of development: the case of LIEs in Africa

Dependent variable

Average rule of law index Log of trade share Malaria risk Catholicism Enrolment ratio in 1900 R2 F-test (p-value) Instruments

Log per capita GDP in 2000 OLS obs 5 15

2SLS obs 5 10

0.055 (0.2167) 0.15 (0.1981) −0.9*** (0.2653) −3e-05 (0.0080) 0.01 (0.0172) 0.9282 0.0004

−0.07 (0.1725) 0.35 (0.2452) −0.98*** (0.2902) −0.005 (0.0085) −0.04 (0.0221)

Log settler mortality, ENGFRAC, EURFRAC, Constructed openness, Malaria ecology

Notes: ***, ** and * indicate significance levels of 1%, 5% and 10%, respectively, against a one-sided alternative. Figures in parentheses are the respective standard errors. The 2SLS regression may not yield meaningful estimates due to the problem of degrees of freedom. There are only ten observations for this regression. The standard errors reported in the regressions are heteroskedasticity robust. All the regressions reported above are carried out with an intercept.

The regression estimates reported in Table 4.8 show that malaria risk is the only explanator that has a statistically significant coefficient estimate when the sample is exclusively of LIEs from Africa. Based on these results, is it fair to conclude that the root of Africa’s development problem is malaria? The answer is no. The correlation between malaria and the level of GDP can be due to some omitted factor specific to Africa which moves in the same direction as malaria. For instance, the history of the slave trade in Africa may be the true reason for the continent’s lack of development. Malaria being an outcome of the lack of development will also share a high positive correlation with the slave trade variable. Therefore, running a regression of current per capita GDP on malaria without controlling for the correlation between slave trade and current per capita GDP will lead to the error of attributing the movement in per capita GDP data that is due to slave trade to malaria. A proper

M2545 - BHATTACHARYYA PRINT.indd 70

02/03/2011 10:09

Empirical evidence

71

analysis of the true causes should involve estimation of a model which controls for all the relevant variables. In summary, this section shows that ‘institutions don’t rule’. Malaria is an important correlate of development. The explanatory power of institutions and malaria are almost equal in the pooled sample. Malaria, however, is the most importan`t factor for the LIEs. The results also show that the malaria result in an LIE sample is driven by the large proportion of African economies in the sample. This certainly makes sense given the enormity of the malaria problem of that continent. Malaria kills 1–2 million Africans every year, out of a continent-wide total of roughly 9 million deaths per year. However, this does not convincingly reject all the other competing theories, which may be correlated with both malaria and the current level of development. In order to present more convincing evidence of cause and effect, I construct a model which controls for all the variables relevant to Africa’s development or lack of development. I estimate this model to confirm that malaria does matter when a host of other relevant factors are controlled for. This is to be investigated in Section 4.6.

4.6 COLONIAL INSTITUTIONS, DISEASES AND FORCED MIGRATION: THE CASE OF AFRICA It is well known that Africa is falling behind the rest of the world in terms of economic well-being. Even though global poverty is on the decline due to rapid economic growth in India, China and other parts of the world, Africa’s contribution to this decline is disappointing. Absolute poverty in many of the African nations is, in fact, rising (Sachs, 2005). What is the fundamental cause behind this decline? This has been a topic of research for a few decades now. Even though it is extremely difficult to summarize this voluminous literature, it is perhaps fair to say that three strands of thought stand out.14 4.6.1

Towards an Empirical Question

The first is the disease view. According to this view, malaria and other infectious diseases have fatal as well as debilitating effects on the human population in Africa. It negatively influences productivity, savings and investments in physical and human capital, and directly affects the economic performance of the continent (Bloom and Sachs, 1998; Gallup and Sachs, 2001).15 According to Bloom and Sachs (1998), the high incidence of malaria in sub-Saharan Africa reduces the annual growth rate of the continent by 1.3 percentage points a year and eradication of malaria in

M2545 - BHATTACHARYYA PRINT.indd 71

02/03/2011 10:09

72

Growth miracles and growth debacles

the 1950s would have resulted into a doubling of per capita income. Sachs (2003a) and Carstensen and Gundlach (2006) using a global cross-national dataset and Lorentzen et al. (2008) using cross-national and sub-national datasets also make similar arguments about the role of diseases. Lorentzen et al. (2008), in particular, argue that higher adult mortality is associated with an increased level of risky behaviour, higher fertility and lower investment in physical and human capital. Acemoglu and Johnson (2007), however, question these results. They find that there is no statistically significant effect of improved life expectancy on GDP levels, leading them to conclude that diseases do not have a direct role in development. Despite the doubts posed by Acemoglu and Johnson (2007), a significant number of recent studies tend to support the disease view both at the macro as well as micro level. Weil (2005) and Bloom and Canning (2005) calibrating the effects of health from a range of micro estimates into a macro model show that these effects are important at the aggregate level. Kalemli-Ozcan et al. (2000) and Kalemli-Ozcan (2002) also show that lower mortality as a result of better health contributes to economic growth. In related literature, Arndt and Lewis (2000), Bell et al. (2003) and Kalemli-Ozcan (2006) find that HIV/AIDS is reversing the trends in demographic transition in Africa and is negatively affecting growth.16 At the micro level, Knaul (2000), Schultz (2002), Bleakley (2003), Behrman and Rosenzweig (2004), Miguel and Kremer (2004), and many others find that improved health leads to better individual economic outcomes.17 The second is the colonial institutions view. According to this view, the persistent effect of colonial institutions can explain the huge differences in income across all ex-colonies including Africa (Knack and Keefer, 1995; Hall and Jones, 1999; Acemoglu et al., 2001; Bhattacharyya, 2004, 2009a; Rodrik et al., 2004; Nunn, 2007). The story as outlined by Acemoglu et al. (2001) goes as follows.18 Europeans resorted to different style of colonization depending on the feasibility of settlement. In a tropical environment, the settlers had to deal with killer malaria and hence a high mortality rate. This prevented colonizers from settling in a tropical environment and they erected extractive institutions in these colonies. These colonial institutions have persisted over time and they continue to influence the economic performance of the colonies even long after independence. Hence, the Acemoglu et al. (2001) argument is that diseases affect economic performance only indirectly through institutions. Nunn (2007), using a stylized model for Africa, shows that colonial extraction, when severe enough, can cause a society to move from a high to a low production level equilibrium. Due to the stability of a low-level equilibrium, a society can remain trapped in this equilibrium even after the period of colonial extraction is over.

M2545 - BHATTACHARYYA PRINT.indd 72

02/03/2011 10:09

Empirical evidence

73

Earlier work by Easterly and Levine (1997), Sachs and Warner (1997b) and Temple (1998) also reports a strong link between quality of institutions and post-war growth (or the lack of it) in Africa.19 Easterly and Levine (1997) show that ethnic diversity in Africa has led to social polarization and the formation of several rival interest groups, which increase the likelihood of selecting socially suboptimal policies when an ethnic representative in the government fails to internalize the entire social cost of their rent-seeking policies. Sachs and Warner (1997b), on the other hand, stress Africa’s lack of openness to international markets and unfavourable geography as other contributors to poor growth, in addition to poor quality institutions. Temple (1998) emphasizes the role of social arrangements in explaining Africa’s slow growth. Finally, a third group of explanations relates to the economic impact of Africa’s engagement in the slave trade. According to this view, Africa’s engagement in the slave trade caused massive depopulation of the continent over two centuries (see Gemery and Hogendorn, 1979; Manning, 1981; Inikori, 1992). The result was a significant slowdown in division of labour, demographic transition,20 human capital accumulation and long-run economic growth (Inikori, 1992). Depopulation also resulted in an implosion of the continent’s production possibility frontier21 and an unambiguous reduction in welfare (Gemery and Hogendorn, 1979). The secular decline in welfare continued over more than two centuries, plunging the continent into economic backwardness. In a recent paper, Nunn (2008) also reports a negative causal relationship between the slave trade and current economic performance in Africa. He shows that the slave trade prevented state development, encouraged ethnic fractionalization and weakened legal institutions, and through these channels it affected economic development. These competing theories, even though plausible, do not tell us how much of the variation in income across countries in Africa they can explain. One possible way to arrive at an answer is to check the relative strengths of these theories in explaining the variation when they are pitted against each other in a regression model. In this section, I investigate their relative strength by setting up a parsimonious regression model. In the regression model I use log GDP per capita in 2000 as the dependent variable and malaria risk, institutions and log total slave exports out of Africa normalized by land area as explanatory variables. This exercise is important to demonstrate that diseases and other factors are particularly important in Africa. In other words, it is not all institutions when it comes to Africa. In fact, institutions are statistically insignificant in an African sample and malaria is the only statistically significant variable. This is contrary to the results reported by the proponents of the institutions theory using cross-national data.

M2545 - BHATTACHARYYA PRINT.indd 73

02/03/2011 10:09

74

Growth miracles and growth debacles

I deal with the complex causality issues by using appropriate exogenous instruments for malaria risk, institutions and total slave exports. Malaria ecology from Kiszewski et al. (2004) is used as an instrument for malaria risk. Given the controversy regarding exogeneity of malaria ecology, I also use rain, humidity and frost as alternative instruments. My basic result survives this test. Institutions and slave exports are instrumented by log settler mortality22 from Acemoglu et al. (2001) and distance measures from Nunn (2008), respectively. The results show that malaria matters the most and all other factors are statistically insignificant. This result survives even when I use Nunn’s econometric specification and dataset. I also show that malaria dampens savings. Increases in mortality and morbidity can be possible channels through which malaria impacts African development. Increased mortality induces households to increase current consumption and save less for the future (hence the negative relationship between savings and malaria). Increased morbidity, on the other hand, adversely affects productivity, reducing household income and savings. This slows down capital accumulation and economic development. This discussion perhaps sheds some light on why malaria is so persistent in Africa. The result is striking. Malaria is the most powerful explanator (at least statistically) of long-run economic development (or the lack of it) in Africa. None of the other factors (including institutions and the slave trade) are statistically significant. I also provide an explanation for the persistence of malaria in Africa. The benefits of looking at an Africa-only sample are twofold. First, it allows us to statistically test the strengths of competing theories of African underdevelopment. Second, it allows us to focus on a continent where the majority of the bottom billion’s countries are located (Collier, 2007). 4.6.2

Specification and Data

In order to estimate the causal effects of malaria, colonial institutions and the slave trade on Africa’s long-run economic development, I use the following model. The implicit assumption here is that initial incomes across countries in Africa going back several centuries were similar. Therefore, I no longer need to control for initial income explicitly and the model can be interpreted as a long-run growth model. logyi 5 l 1 aMALi 1 bINSi 1 gSLVXi 1 xi F 1 ei

(4.5)

where yi, MALi, INSi and SLVXi are per capita income in country i, measure of malaria, measure of institutions and measure of slave exports, respectively. xi is a row vector of additional control variables,23 F is the

M2545 - BHATTACHARYYA PRINT.indd 74

02/03/2011 10:09

Empirical evidence

Table 4.9

75

Descriptive statistics

Variable Log GDP per capita in 2000 (log yi) Malaria Risk (MALi) Expropriation Risk in 1985 to 1995 (INSi) Log total slave exports normalized by land area (SLVXi)

Number of obs

Mean

Standard deviation

Minimum Maximum

46

7.46

0.815

6.19

9.24

49 35

0.77 5.82

0.386 1.30

0 3

1 8.27

52

3.26

3.89

−2.30

8.82

vector of coefficients on other control variables denoted by the vector and ei is the random error term. We are interested in the size, sign and statistical significance of the three coefficients a, b and g. The estimation of Equation 4.5 is based on a dataset consisting of per capita GDP levels, measure of malaria risk, measure of institutions and measure of slave exports in (up to) 52 countries in Africa. Definitions and sources of all the variables used in this study are summarized in the Data Appendix. Table 4.9 presents summary statistics for the key variables of interest. GDP per capita in 2000 data is from the Penn World Tables 6.1. According to these figures, Tanzania was the poorest country in Africa in 2000. I also use per capita income data from Nunn in Table 4.15 when I check the robustness of the result using Nunn’s dataset and specification. Note that Nunn uses income data from Maddison (2004). Malaria risk is the percentage of population living in areas of high malaria risk in a country in 1994. It is calculated using GIS software from a digitized WHO map of the world distribution of malaria and a detailed database of world population distribution in 1994.24 The variable lies between 0 and 1 and a higher value indicates greater risk for the population. Most of the countries in the sample register high malaria incidence, except Algeria, Tunisia and Egypt. There are at least three measures of institutional quality that have been used in the literature. Knack and Keefer (1995), Acemoglu et al. (2001) and many others use expropriation risk averaged over 1985 to 1995 from the Political Risk Services. Rodrik et al. (2004) use the rule of law index from the World Bank. Others use the executive constraint from the Polity dataset. The expropriation risk measure is perhaps the most appropriate for my purpose as I would like to capture the variation in institutions

M2545 - BHATTACHARYYA PRINT.indd 75

02/03/2011 10:09

76

Growth miracles and growth debacles

originating from different types of colonial states and state policies (see Acemoglu et al., 2001). It is also the closest to Douglass North’s (1981) definition of good institutions25 as it captures the notion of extractive state. I also check the robustness of the results using rule of law and executive constraint measures. Slave export data is from Nunn (2008). Nunn (2008) reports the natural log of total slaves exported out of each of the African nations normalized by land area and population in 1400.26 According to Nunn, the maximum number of slaves exported was from Angola, which accounted for 23.1 per cent of the total slave exports, followed by Nigeria (12.9 per cent) and Ghana (10.2 per cent). The fewest slaves exported were from Tunisia. I follow Nunn and use log total slave exports normalized by land area as my preferred measure. Identifying good empirical proxies for each of these variables is difficult but perhaps not the most challenging part of the analysis. The major challenges are to estimate the causal effects. In order for the estimates of a, b and g to be interpreted as causal effects, they have to overcome some serious econometric challenges. I list them as follows. Endogeneity Economic development is a complex phenomenon. Given the complex nature of this process, reverse causality is a real possibility. For example, rather than malaria influencing development, the causality may run the other way round. The rich economies can afford to invest in the research and development of drugs that cure or minimize the effect of malaria. They can also invest in public health programmes to tackle malaria. A similar argument can be made about institutions. Rich nations have better institutions, not because they have grown richer due to better institutions, but because they can afford better institutions. Furthermore, there are endogeneity concerns with the slave trade. Societies that initially had poor domestic institutions may have selected into the slave trades. Therefore, the observed negative relationship between slave exports and development may not be the causal effect (Nunn, 2008). If this is the case then OLS estimates of a, b and g will be biased away from zero as we will be erroneously attributing the effects of income or other factors on endogenous variables to the direct effects of these variables on income. Measurement error The slave export data are likely to contain both classical and non-classical measurement error. One can identify the following sources. First, slave ethnicities in the dataset may have been misclassified. Slaves with similar but different ethnicities may have been classified under one ethnicity. But

M2545 - BHATTACHARYYA PRINT.indd 76

02/03/2011 10:09

Empirical evidence

77

the possibility of a bias due to errors of this nature is minimal as the data is aggregated at the country level. Second, measurement error may arise due to the under-representation of slaves from the interior or due to the assumption used in the construction of the data that slaves shipped from a port within a country are either from that country or from countries directly to the interior. In either case, OLS estimates of a, b and g will be biased towards zero – the classical measurement error (Wooldridge, 2000). Furthermore, any random measurement error present in the data will also have the same effect on OLS estimates. Moreover, it is not possible to rule out non-classical measurement error. Omitted variable bias Many of the omitted time-invariant deep factors (culture, ethnic make-up, colonial or legal origin, religion, climate) influencing long-run economic development can be correlated with malaria risk, institutions and slave exports. This has the potential to bias the OLS estimates of a, b and g away from zero. I control for regional fixed effects, colonizer fixed effects and legal origin fixed effects to tackle this problem. I also test the robustness of our estimates by controlling for additional covariates. Some of the obvious ones are trade openness, Catholicism, Islam, historical schooling, ethnic fractionalization, share of mining, foreign aid and the Gini coefficient. However, as is the case with all empirical modelling, I can never be entirely sure that I have adequately controlled for all the omitted factors. To tackle the problems of endogeneity and measurement error, I use the instrumental variable (IV) estimation. A valid instrument has to satisfy the twin conditions that it is correlated with the suspected endogenous variables (malaria, institutions and slave exports, in this case) but uncorrelated with the error term or a measurement error hidden in the error term. It is obviously a difficult task to find valid instruments. However, the literature has identified several instruments that could serve my purpose. Previous studies have used log settler mortality as an instrument for institutions (Acemoglu et al., 2001; Rodrik et al., 2004; and others). It is based on the idea that European colonizers erected good institutions only in the settlement colonies. Elsewhere they erected extractive institutions. Therefore, the settler mortality instrument is likely to be negatively correlated with the quality of institutions and also orthogonal to the random error term since it is geography-based. Recently this instrument has come under intense scrutiny. In Section 4.3, I show that it is almost impossible to separate out the effects of institutions and human capital on long-run development in a cross-section model due to multicollinearity. This is true regardless of the specifications used. The instruments also have issues with satisfying the exclusion restrictions as they are highly correlated with

M2545 - BHATTACHARYYA PRINT.indd 77

02/03/2011 10:09

78

Growth miracles and growth debacles

schooling. Albouy (2008) identifies several weaknesses with regard to the construction of the instrument. In spite of the controversy regarding this instrument, I continue to use it here to facilitate comparison of my results with previous studies in the literature. I also follow Nunn (2008) and use sailing distance from the coast to the closest market of the Atlantic slave trade, sailing distance from the coast to the closest market of the Indian Ocean slave trade, overland distance from the centroid to the closest port of export for the trans-Saharan slave trade, and overland distance from the centroid to the closest port of export for the Red Sea slave trade as instruments for slave exports. Nunn (2008) argues that the distance instruments are negatively correlated with slave exports and also exogenous. Therefore, they are valid instruments. He also uses the overland distance from the centroid to the coast and log population density in 1400 as additional instruments. However, he notes that the additional instruments may not satisfy the exclusion restrictions. Therefore, I decide not to use these additional instruments.27 Finally, I follow Sachs (2003a) and Carstensen and Gundlach (2006) and use malaria ecology as an instrument for malaria risk. Malaria ecology is an ecologically based spatial index and depends on climatic factors and biological properties of each regionally dominant malaria vector. Hence it is exogenous to public health interventions and economic conditions, and thus can serve as an instrumental variable in regressions of economic performance on malaria risk (Kiszewski et al., 2004).28 Rodrik et al. (2004) doubt the exogeneity of malaria ecology as they argue that from the little information provided by Sachs (2003a), it remains unclear whether malaria ecology can be influenced by human action. Another concern regarding malaria ecology comes from a previous version of the text describing the construction of the index as it says the calculation includes mosquito abundance. Even though both critiques are technically correct, the doubts about the exogeneity of the instrument may not be justified for the following reasons. First, the index is vector-based and not affected by human activity as public health interventions against malaria only serve to break the transmission cycle, but do not eliminate the presence of the vector itself. Even today, Anopheles mosquitoes capable of transmitting malaria can be found throughout the US and Europe, places where malaria has been largely eradicated (see Kiszewski et al., 2004). Second, observed mosquito abundance enters the index only as a screen for precipitation data, where the independently identified dominant malaria vector is assumed to be absent from the specific site under consideration if precipitation falls below a certain level per month (see Carstensen and Gundlach, 2006). Nevertheless, I also use average rainfall, average humidity and prevalence of frost as alternative instruments for malaria and the

M2545 - BHATTACHARYYA PRINT.indd 78

02/03/2011 10:09

Empirical evidence

79

results are robust to these changes. Rainfall, humidity, and lack of frost are crucial to the life cycle of the parasite and hence could serve as good instruments. They are also geography-based and hence exogenous to economic conditions. One concern is that rainfall, humidity and frost may not satisfy the exclusion restriction because they may affect development through channels other than malaria. Statistically, this will bias my estimates only if the predicted value of malaria at the second stage is correlated with the error term. The Hansen J-test for exogeneity of instruments indicates otherwise (see Table 4.11). Nevertheless, I also use them as exogenous control variables which may directly influence economic performance. My results survive this test. In IV estimation, endogenous explanatory variables are replaced by their predicted values from the first stage equations. The first stage equations are as follows: MALi 5 m 1 dMEi 1 cLSMi 1 kDCi 1 xiF 1 eMALi

(4.6)

INSi 5 ␾ 1 hLSMi 1 sMEi 1 nDCi 1 xiF 1 eINSi

(4.7)

SLVXi 5 y 1 wDCi 1 fMEi 1 pLSMi 1 xiF 1 eSLVXi

(4.8)

where MEi, LSMi, and DCi refer to malaria ecology, log settler mortality and the distance instruments from Nunn (2008). Equations 4.5–4.8 are at the core of the empirical results that I report in the next section. I also report statistical tests (Hausman test, Sargan test and Hansen test) for the validity of instruments. An additional concern with IV is the bias due to weak instruments. Staiger and Stock (1997) and others have shown that the consequence of weak instruments is a large-sample bias in IV as in effect the model becomes unidentified. Furthermore, the magnitude of the large-sample bias increases with the number of instruments. The Staiger and Stock (1997) results rely on asymptotic properties and asymptotic distribution theory may not necessarily apply for this small sample. However, the bias in 2SLS cannot be ruled out. More importantly, the limited information maximum likelihood (LIML) estimator does not have such bias. It is also more robust to the weak instruments problem than IV (Stock and Yogo, 2005). My basic results survive when I use the LIML estimator. 4.6.3

Results

Table 4.10 reports the core results. In column 1 of Panel A, I start with estimating the basic model using OLS. I find that malaria negatively

M2545 - BHATTACHARYYA PRINT.indd 79

02/03/2011 10:09

80

Growth miracles and growth debacles

Table 4.10

Malaria as a root cause of African underdevelopment: core results

Panel A: Model log yi 5 l 1 aMALi 1 bINSi 1 gSLVXi 1 ei Dependent variable

Log per capita GDP in 2000

Malaria risk (MALi) Expropriation risk in 1985 to 1995 (INSi) Log total slave exports normalized by land area (SLVXi) Log per capita income in 1960 R2 Hansen J-test (p)

Growth during 1960–2000

OLS estimate obs 5 33 (1)

2SLS estimate obs 5 27 (2)

LIML Fuller estimate obs 5 27 (3)

2SLS estimate obs 5 27 (4)

−0.86* (0.4576) 0.18* (0.0992) −0.08* (0.0451)

−4.19** (2.105) 0.29 (0.6543) 0.39 (0.3043)

−3.3** (1.758) 0.16 (0.5251) 0.25 (0.2505)

−0.04* (0.0244) 0.004 (0.0049) 0.002 (0.0033) −0.005 (0.0049)

0.59 0.92

Hausman/Sargan test (p) Cragg-Donald test (p) Additional controls Instruments

–

0.63

– 0.97

–

0.71 – – – LPDi, IDCi ME, LSM, ADC, IODC, SDC, RDC

Panel B: First stage regressions Dependent variables

MALi obs 5 27 (1)

INSi obs 5 27 (2)

SLVXi obs5 27 (3)

Malaria ecology (MEi)

0.02** (0.0080) 0.12* (0.0663) 0.09 (0.0745) −0.00004 (0.00009) −0.07 (0.0588)

−0.03 (0.0277) −0.08 (0.2906) −0.26 (0.3741) −0.002*** (0.0007) −0.12 (0.2149)

0.18** (0.0863) 0.63 (0.6501) 1.64* (0.8451) −0.001 (0.0022) −0.36 (0.6597)

Log settler mortality (LSMi) Log population density in 1500 (LPDi) Interior distance (IDCi) Atlantic distance (ADCi)

M2545 - BHATTACHARYYA PRINT.indd 80

02/03/2011 10:09

Empirical evidence

Indian distance (IODCi) Saharan distance (SDCi) Red Sea distance (RDCi) R2 F-stat

−0.02 (0.0454) 0.12 (0.0969) −0.18* (0.0912) 0.82 54.75

81

−0.13 (0.1680) −0.55* (0.3154) 0.19 (0.3920) 0.58 6.65

0.24 (0.6589) 2.2* (1.162) −1.6* (0.8680) 0.64 3.88

Panel C: Instrument redundancy tests Instruments tested

ME

LSM

IDC, ADC, IODC, SDC, RDC

LM test statistic p-value Degrees of freedom

7.7 0.05 3

6.92 0.07 3

19.57 0.08 12

Notes: ***, ** and * indicate significance levels of 1%, 5% and 10%, respectively, against a two-sided alternative. Figures in parentheses are cluster standard errors and they are robust to arbitrary heteroskedasticity and arbitrary intra-group correlation. All the regressions reported are carried out with an intercept. Fuller’s modified LIML estimator with a 5 1 (correction parameter proposed by Hausman et al., 2005) is used in Panel A, column 3. Both Hansen J-test and Hausman/Sargan test p-values are reported. In both cases, the null hypotheses are that the instruments are jointly exogenous. CraggDonald test p-values for weak instruments are also reported. The null hypothesis, in this case, is that the instruments are jointly weak. The test statistic follows F-distribution under the null with degrees of freedom 5 N-L, L1 (N, number of observations; L, total instruments; L1, excluded instruments). The LM statistic for instrument redundancy tests are distributed as chi-squared under the null hypothesis that the specified instruments are redundant with degrees of freedom equal to the number of endogenous regressors times the number of instruments being tested. The endogenous regressors are MALi, INSi and SLVXi. The abbreviations used in the table are MEi, malaria ecology; LSMi, log settler mortality; LPDi, log population density in 1500; IDCi, interior distance; ADCi, Atlantic distance; IODCi, Indian Ocean distance; SDCi, Saharan distance; and RDCi, Red Sea distance.

impacts development, institutions are good for development and slave exports are negatively correlated with development.29 I also plot the OLS partial effects (see Figure 4.2). The estimates, however, are likely to be inconsistent as OLS does not account for endogeneity or measurement error problems. In column 2, I estimate the model using IV. I notice that the negative effects of malaria survive; however, institutions and slave exports are statistically insignificant. The magnitude of the malaria effect is also large. A one standard deviation decrease in malaria risk increases the income of an average country in Africa fivefold. To put this into

M2545 - BHATTACHARYYA PRINT.indd 81

02/03/2011 10:09

82

Growth miracles and growth debacles 2

1

GIN COG AGO CMR SEN BWA DZAZAF GHA CIV TUN ZWE MOZ KEN TGO MLI BFA EGY GNB NGA MDG UGA MAR MWI SLE ZMB NER ZAR ETH TZA SDN

0

–1

–2

–1

0

1

2

GMB

1.5 1 .5 0 –.5

.5 EGY 0 ETH

–.5

AGOGINCOG SDN GHASEN TUN CMR MLI GNB BFA CIV BWA NGA MOZZWE TGO KEN MDG ZAR GMB MAR UGA SLE MWI NER

–1 TZA 3

e( expr8595 | X ) coef = .17928689, (robust) se = .09921514, t = 1.81

e( lgdppc2000 | X )

GAB DZA ZAF

1

e( lgdppc2000 | X )

e(lgdppc2000 | X )

GAB

–.6

ZMB

–.4

–.2 0 .2 .4 e( mal94p | X ) coef = –.8617306, (robust) se = .45775097, t = –1.88

GAB COG SDN GIN AGO CMR BWA ZAF SEN TUNCIV ZWE MLI UGA BFA GHA DZA MDG EGY ZAR NER KEN MOZ GNB NGA MAR GMB TGO SLE ZMB MWI

–1

TZA

ETH

–4

–2 0 2 4 e( ln_export_area | X ) coef = –.07661942, (robust) se = .0451374, t = –1.7

Note: See Table 8.2 for country abbreviations.

Figure 4.2

Partial correlation plot: root causes of African underdevelopment

perspective, the model explains approximately 92 per cent of the difference in per capita income in Namibia and Nigeria – two countries who also share approximately one standard deviation actual gap in malaria risk. The Hansen J-test30 and the first stage regressions reported in Panel B show that the instruments are valid; however, the Cragg-Donald test for weak instruments suggests that some of the instruments may be weak. Staiger and Stock (1997) and others have shown that weak instruments can cause large-sample bias in the IV estimates even when there are multiple instruments. The extent of the bias increases with the number of instruments. They suggest that an F-statistic of less than ten at the first stage is a cause of concern. They recommend that cutting down on the number of instruments may help in reducing the large-sample bias. However, this may not be a useful strategy here as the instruments pass the Hall and Peixe (2000) instrument redundancy test (see Panel C). Note that the weak instruments problem is not unique to this study and may well be a

M2545 - BHATTACHARYYA PRINT.indd 82

02/03/2011 10:09

Empirical evidence

83

general problem with the empirical comparative development literature. Stock and Yogo (2005) show that LIML estimators are more robust to weak instruments than IV. In column 3, I report Fuller’s modified LIML estimates with a 5 1 (correction parameter proposed by Hausman et al., 2005) and I get results similar to IV.31 The magnitude of the coefficient on malaria risk declines. The model now explains approximately 73 per cent of the difference in per capita income in Namibia and Nigeria. I choose the LIML as my preferred estimate since it is the lower bound. The positive correlation between malaria ecology and the slave trade at the first stage is certainly noteworthy. This is consistent with the view that the slave trade was also an outcome of local epidemiology, particularly malaria (see Dias, 1981; Miller, 1982). I also notice that the interior distance is negatively correlated with colonial institutions. This may be due to the possibility that proximity to the coast leads to more trade and more trade leads to better institutions (see Acemoglu et al. 2005b). Sachs (2003a) predicts 1.6 fold, 1.9 fold and 1.8 fold increases in per capita GDP due to one standard deviation decline in malaria risk in AJR, RST and EL samples respectively. Carstensen and Gundlach (2006) predict a 1.6 fold increase of the same. Both studies are based on a global sample and they find both institutions and malaria are statistically significant. I find that the malaria effect is even larger (my preferred LIML estimate predicts a 3.6 fold increase) in an Africa-only sample and all other factors are statistically insignificant. My results are at odds with the findings of Nunn (2008) who reports that slave exports have a causal effect on current development in Africa via state development, ethnic fractionalization and weakened legal institutions. I do not find any statistical evidence of direct and indirect effects of the slave trade on Africa’s current development. To be completely sure, I also check the robustness of my result using Nunn’s specification and dataset (see Table 4.15). The malaria result survives. I also do not find statistical support for the colonial institutions view in Africa. This is regardless of the specification and sample. In column 4, I estimate the causal effect of malaria on growth over the period 1960 to 2000. The effect is large as one standard deviation reduction in malaria yields approximately 1.5 per cent growth dividends annually to an average country in Africa. This suggests that eliminating malaria alone in 1960 would have resulted in doubling of income in Africa by now. The relationship between malaria and growth is not surprising as current income levels and growth in Africa are correlated (approximately 0.7). I fail to find evidence of causal effects of institutions and the slave trade on growth. Tables 4.11, 4.12 and 4.13 report robustness tests with alternative instruments, with fixed effects and with additional covariates. The alternative

M2545 - BHATTACHARYYA PRINT.indd 83

02/03/2011 10:09

84

Table 4.11

Dependent variable

Growth miracles and growth debacles

Malaria and African underdevelopment: robustness with alternative instruments Log per capita GDP in 2000 2SLS estimate obs 5 27 (1)

2SLS estimate obs 5 27 (2)

2SLS estimate obs 5 25 (3)

2SLS estimate obs 5 25 (4)

2SLS estimate obs 5 27 (5)

2SLS 2SLS estimate estimate obs 5 27 obs 5 25 (6) (7)

Malaria risk (MALi) Expropriation risk in 1985 to 1995 (INSi) Log total slave exports normalized by land area (SLVXi) Hansen J-test (p) Additional controls

−2.38** −3.6** −3.45** −1.95* −3.35** −6.3 (0.9327) (1.887) (1.632) (1.188) (1.425) (10.70) 0.36 0.33 0.51 0.61 0.29 1.89 (0.5166) (0.5721) (0.7613) (0.6554) (0.5764) (5.629)

Instruments

Replacing Replacing Replacing Replacing ME by ME by ME by ME by rain humidity frost rain, humidity and frost

0.11 (0.1475)

0.53

0.29 (0.2864)

0.94

0.33 (0.2358)

0.93

0.13 (0.2306)

0.64

LPDi, IDCi

0.24 (0.1989)

0.53

0.85 (2.031)

0.97

−0.26 (3.363) 0.52 (0.5102)

0.12 (0.1705)

0.28

LPDi, IDCi, rain

LPDi, LPDi, IDCi, rain, IDCi, humidity rain, humidity, frost ME, LSM, ADC, IODC, SDC, RDC

Notes: ***, ** and * indicate significance levels of 1%, 5% and 10%, respectively, against a two-sided alternative. Figures in parentheses are cluster standard errors and they are robust to arbitrary heteroskedasticity and arbitrary intra-group correlation. All the regressions reported above are carried out with an intercept. P-values of Hansen J-tests are reported. The null hypothesis is that the instruments are jointly exogenous. The endogenous regressors are MALi, INSi and SLVXi. The abbreviations used in the table are MEi, malaria ecology; LSMi, log settler mortality; LPDi, log population density in 1500; IDCi, interior distance; ADCi Atlantic distance; IODCi, Indian Ocean distance; SDCi, Saharan distance; and RDCi, Red Sea distance.

instruments strategy is to address the concern that malaria ecology is not exogenous. The fixed effects and the additional covariates strategies are to address the omitted variable problem. In Table 4.11 columns 1–4, I replace the malaria ecology instrument with geography-based instruments (rain,

M2545 - BHATTACHARYYA PRINT.indd 84

02/03/2011 10:09

Empirical evidence

Table 4.12

85

Malaria and African underdevelopment: robustness with fixed effects

Dependent variable

Malaria risk (MALi)

Log per capita GDP in 2000 2SLS estimate obs 5 27 (1)

2SLS estimate obs 5 27 (2)

2SLS estimate obs 5 27 (3)

−2.36*** (0.6808) −0.19 (0.4147) 0.09 (0.0915)

1.43 (1.654) 0.23 (0.2615) −0.03 (0.1034)

−3.99** (1.729) 0.07 (0.3751) 0.36 (0.2436)

Expropriation risk in 1985 to 1995 (INSi) Log total slave exports normalised by land area (SLVXi) Hansen J-test (p) 0.28 0.09 0.92 Additional controls LPDi, IDCi Fixed effects Colonizer Fixed Region Fixed Legal Origin Effects Effects Fixed Effects Instruments ME, LSM, ADC, IODC, SDC, RDC Notes: ***, ** and * indicate significance levels of 1%, 5% and 10%, respectively, against a two-sided alternative. Figures in parentheses are cluster standard errors and they are robust to arbitrary heteroskedasticity and arbitrary intra-group correlation. Colonizer fixed effects, region fixed effects and legal origin fixed effects are dummies representing colonial origin, region and legal origin, respectively. The endogenous regressors are MALi, INSi and SLVXi. The abbreviations used in the table are MEi, malaria ecology; LSMi, log settler mortality; LPDi, log population density in 1500; IDCi, interior distance; ADCi, Atlantic distance; IODCi, Indian Ocean distance; SDCi, Saharan distance; and RDCi, Red Sea distance.

humidity and frost) and the malaria result survives.32 One concern is that rain, humidity and frost may not satisfy the exclusion restriction as they might influence income through channels other than malaria. To address this concern, I use these variables as additional controls in columns 5–7. The malaria result survives in column 5. The large standard errors and statistical insignificance of all variables in columns 6 and 7 may be due to small sample size, degrees of freedom problems and multicollinearity. The malaria result also survives the inclusion of colonizer fixed effects and legal origin fixed effects (see columns 1 and 3, Table 4.12). However, it vanishes when regional fixed effects are added (see column 2). This is not surprising as I find that the western region indicator dummy and the eastern region indicator dummy (which are representative of tropical Africa) predict a negative impact on development. Therefore, it could very well be the case that these dummies are picking up the negative malaria effect. Multicollinearity between malaria and the regional dummies can

M2545 - BHATTACHARYYA PRINT.indd 85

02/03/2011 10:09

86

M2545 - BHATTACHARYYA PRINT.indd 86

02/03/2011 10:09

Mining

−3.29** (1.464) 0.20 (0.4987) 0.25 (0.1943) 0.67

2SLS estimate obs 5 27 (1) −1.36* (0.8251) 0.43 (0.4931) −0.01 (0.1422) 0.08

2SLS estimate obs 5 26 (3) −2.39*** (0.8453) 0.02 (0.5876) 0.09 (0.1216) 0.41 LPDi, IDCi

2SLS estimate obs 5 26 (4) −2.31*** (0.8267) 0.33 (0.4141) 0.16 (0.1341) 0.45

2SLS estimate obs 5 19 (5) −3.22 (2.199) 0.47 (0.6639) 0.32 (0.3083) 0.84

2SLS estimate obs 5 27 (6)

Ethnic Catholicism Islam Gini coefficient Foreign aid fractionalization ME, LSM, ADC, IODC, SDC, RDC

−3.2** (1.496) 0.32 (0.5814) 0.29 (0.2359) 0.61

2SLS estimate obs 5 27 (2)

Log per capita GDP in 2000

Malaria and African underdevelopment: robustness with additional covariates

Schooling in 1900

−1.51*** (0.4838) 0.06 (0.0849) −0.05 (0.0491) 0.95

2SLS estimate obs 5 11 (7)

All instruments plus CONST and IDCi

Trade share

−2.22*** (0.6315) 0.04 (0.1488) 0.09 (0.0694) 0.42 LPDi

2SLS estimate obs 5 26 (8)

Notes: ***, ** and * indicate significance levels of 1%, 5% and 10%, respectively, against a two-sided alternative. Figures in parentheses are cluster standard errors and they are robust to arbitrary heteroskedasticity and arbitrary intra-group correlation. All the regressions reported above are carried out with an intercept. The instrument CONST is constructed openness from Frankel and Romer (1999). The endogenous regressors are MALi, INSi and SLVXi. The abbreviations used in the table are MEi, malaria ecology; LSMi, log settler mortality; LPDi, log population density in 1500; IDCi, interior distance; ADCi, Atlantic distance; IODCi, Indian Ocean distance; SDCi, Saharan distance; and RDCi, Red Sea distance.

Hansen J-test Control variables Additional covariates Instruments

SLVXi

INSi

MALi

Dependent variable

Table 4.13

Empirical evidence

87

also be an issue here as I notice large standard error on the malaria estimate. Alternatively, it may be due to deep cultural or geographic factors specific to these regions influencing both malaria and income. I am unable to separate out these effects. The malaria effect also survives the additional covariates test in the majority of cases (seven out of eight) which are reported in Table 4.13. The additional covariates (mining, ethnic fractionalization, Catholicism, Islam, Gini coefficient, foreign aid, schooling, trade share)33 are chosen on the basis of previous findings in the literature. The literature identifies these variables as important correlates of growth and development. Controlling for all additional covariates together may not be an option as it weakens the power of statistical tests due to the loss of degrees of freedom. Table 4.14 tests the robustness of the malaria result with alternative measures of institutions and slave exports, and omission of influential observations. In column 1, I replace the expropriation risk measure of institutions with Rodrik et al.’s (2004) preferred measure, the rule of law index. I notice that the malaria result survives and the magnitude of the coefficient is larger than my preferred estimate. In column 2, I replace it with executive constraints – another measure of institutions used by Acemoglu et al. (2005b) and many others. The malaria result survives in this case. In column 3, I replace the log slave exports normalized by land area measure with log slave exports normalized by population. Again, I notice that the malaria result survives. In column 4, I identify influential outliers using the DFITS, Cook’s distance and Welsch’s distance formula (see Belsley et al., 1980) on the OLS regression reported in Panel A, column 1 of Table 4.10. The DFITS and Cook’s distance formula identifies Ethiopia and Gabon as influential observations whereas the Welsch’s distance formula identifies Gabon as an influential outlier. I omit these observations and estimate the model. The malaria coefficient survives the test. In column 5, I use the DFBETA formula and omit Algeria, Ethiopia, Gabon and Zambia. The malaria result survives and the coefficient becomes larger in magnitude. In Table 4.15, I test the robustness of my malaria result using Nunn’s specification and data. In column 1, I estimate Nunn’s preferred specification34 (see Table 5, column 6 of Nunn, 2008, p. 31). My estimate of −0.20 is marginally different from Nunn’s −0.188.35 In column 2, I add malaria into this specification and the statistical significance of the slave trade variable disappears. Nunn argues that the effect of the slave trade may be working through institutions. Column 3 checks this possibility by adding institutions into the mix. The malaria effect survives and neither institutions nor slave trade are statistically significant. In column 4, I replace malaria ecology with the geography-based instrument, humidity.

M2545 - BHATTACHARYYA PRINT.indd 87

02/03/2011 10:09

88

Table 4.14

Growth miracles and growth debacles

Alternative measures and influential observations tests

Dependent variable

MALi

Log per capita GDP in 2000 2SLS estimate obs 5 27 (1)

2SLS estimate obs 5 25 (2)

2SLS estimate obs 5 27 (3)

2SLS estimate obs 5 25 (4)

2SLS estimate obs 5 23 (5)

−4.08** (1.893)

−3.82*** (1.423)

−3.97* (2.262) 0.42 (0.7933)

−2.72*** (0.7063) −0.03 (0.2977)

−3.92** (1.793) 0.11 (0.3580)

0.13 (0.0969)

0.25 (0.1990)

INSi Rule of law index Executive constraint SLVXi Log total slave exports normalized by population Hansen J-test (p) Controls Omitted influential outliers Instruments

0.17 (0.7219)

0.37 (0.3051)

−0.19 (0.2040) 0.29 (0.1846) 0.48 (0.4206)

0.87

0.47

0.99 LPDi, IDCi

0.32 ETH, GAB

0.79 DZA, ETH, GAB, ZMB

ME, LSM, ADC, IODC, SDC, RDC

Notes: ***, ** and * indicate significance levels of 1%, 5% and 10%, respectively, against a two-sided alternative. Figures in parentheses are cluster standard errors and they are robust to arbitrary heteroskedasticity and arbitrary intra-group correlation. Influential observations are omitted using the following standard rules. In column 4, omit if at least 0 DFITSi 0 . 2

k 4 , 0 Cooksdi 0 . , and 0 Welschdi 0 . 3"k Ån n

holds (see Belsley et al., 1980). In column 5, an additional formula is used which is uDFBETAiu . 2/!n. Here n is the number of observations and k is the number of independent variables including the intercept. All the distance formulas are calculated from the OLS version of the model. The endogenous regressors are MALi, INSi and SLVXi. The abbreviations used in the table are MEi, malaria ecology; LSMi, log settler mortality; LPDi, log population density in 1500; IDCi, interior distance; ADCi, Atlantic distance; IODCi, Indian Ocean distance; SDCi, Saharan distance; and RDCi, Red Sea distance.

M2545 - BHATTACHARYYA PRINT.indd 88

02/03/2011 10:09

89

M2545 - BHATTACHARYYA PRINT.indd 89

02/03/2011 10:09

Controls

Ethnic fractionalization Colonizer fixed effects Sargan test YES 0.20

0.29

−0.05 (0.0624)

−1.2*** (0.4443)

obs 5 48 (2)

YES

−0.20*** (0.0429)

obs 5 52 (1)

0.13

YES

−1.67*** (0.5749) −0.04 (0.0864) 0.05 (0.0749)

obs 5 27 (3)

0.19

YES

−1.00* (0.5861) 0.12 (0.1718) −0.05 (0.0721)

0.47

YES

−0.04 (0.0567) −0.26 (0.4171)

−1.39*** (0.4587)

obs 5 46 (5)

0.57

YES

−0.77 (0.7163)

0.11 (0.1403)

−2.6* (1.611)

obs 5 27 (6)

Log income in 2000 obs 5 27 (4)

Robustness with Nunn’s specification and data

Pre-colonial state development Rule of law

SLVXi

INSi

MALi

Dependent variable

Table 4.15

0.18

0.37 (0.7170) YES

−0.04 (0.0629)

−1.42** (0.6328)

obs 5 48 (7)

Frost

0.70

YES

−0.34 (0.2126)

obs 5 43 (8)

Exact identification Frost

YES

0.58 (0.7179)

obs 5 25 (9)

90

M2545 - BHATTACHARYYA PRINT.indd 90

02/03/2011 10:09

ADC, IODC, SDC, RDC

obs 5 52 (1)

obs 5 25 (9) LSM

obs 5 43 (8) ADC, IODC, SDC, RDC

obs 5 48 (7) ME, ADC, IODC, SDC, RDC

obs 5 27 (6) ME, LSM, ADC, IODC, SDC, RDC

obs 5 46 (5) ME, ADC, IODC, SDC, RDC

obs 5 27 (4) Humidity, LSM, ADC, IODC, SDC, RDC

obs 5 27 (3) ME, LSM, ADC, IODC, SDC, RDC

obs 5 48 (2) ME, ADC, IODC, SDC, RDC

Log income in 2000

Notes: ***, ** and * indicate significance levels of 1%, 5% and 10%, respectively, against a two-sided alternative. Figures in parentheses are cluster standard errors (except for column 1) and they are robust to arbitrary heteroskedasticity and arbitrary intra-group correlation. Colonizer fixed effects are dummies representing colonial origin. The dependent variable is from Nunn who uses Maddison’s figures for per capita GDP in 2000. The endogenous regressors are MALi, INSi, SLVXi and rule of law. The abbreviations used in the table are MEi, malaria ecology; LSMi, log settler mortality; IDCi, interior distance; ADCi, Atlantic distance; IODCi, Indian Ocean distance; SDCi, Saharan distance; and RDCi, Red Sea distance.

Instruments

Dependent variable

Table 4.15 (continued)

Empirical evidence

91

The malaria result survives.36 In columns 5–7, I check whether the indirect effects of the slave trade can survive the malaria test. Nunn argues that the slave trade works through pre-colonial state development, rule of law and ethnic fractionalization (see Table 8 of Nunn, 2008, p. 37). None of these variables are statistically significant in the presence of malaria. In columns 8 and 9, I use a more direct approach to test the robustness of the slave trade result of Nunn (2008) and the institutions result of Acemoglu et al. (2001). I check what happens to these results when I use frost as an additional control variable. Note that I do not use frost as an instrument to address the concern that it may not satisfy the exclusion restriction. Also, note that I choose not to use the controversial malaria ecology variable. It appears that frost alone is enough to knock off the slave trade and institutions results. This further reinforces the point that the slave trade and institutions results are extremely weak for the continent of Africa. Malaria is the only statistically significant variable. The results are even more unfavourable for the slave trade and institutions if I control for malaria ecology, log population density in 1500, frost, rainfall and humidity. This result holds if we eliminate log population density in 1500 from the mix. The only exception is the case when log population density in 1500, frost, rainfall and humidity are used as additional controls. Slave trade is marginally significant with p-value 0.09. This is not surprising as it is very close to Nunn’s original specification. Institutions, however, are statistically insignificant. Next, I ask the question: why is malaria so persistent in Africa? The answer to this question may lie with the mechanism through which malaria impacts long-term economic performance. In Table 4.16, I report a strong negative relationship between national savings and malaria in Africa even after controlling for income. This is perhaps indicative of the fact that malaria influences long-run development in Africa through the savings channel. Malaria increases mortality and morbidity. A high mortality rate induces households to save less and consume more. Morbidity reduces productivity, shrinking the household’s income and the ability to save. The result is a low-level equilibrium trap and persistent poverty. This perhaps helps to explain the persistence of malaria in Africa and also why malaria is a root cause of African underdevelopment. Chapter 5 explains this mechanism using an OLG model.

M2545 - BHATTACHARYYA PRINT.indd 91

02/03/2011 10:09

92

Growth miracles and growth debacles

Table 4.16

Malaria and national savings S Model a b 5 ß 1 ␽MALi 1 r log yi + zi Y i

Dependent variable

S Gross savings as percentage of GDP in 2000 a b Y OLS Estimate 2SLS Estimate OLS Estimate 2SLS Estimate obs 5 42 obs 5 42 obs 5 40 obs 5 40

MALi

−15.21*** (3.674)

−12.29** (4.997)

log yi R2 F-Stat P-value Instruments

0.30 16.67 0.0002

5.76 0.0211 ME

−15.22*** (3.923) 2.58 (2.069) 0.33 7.85 0.0014

−12.56** (5.248) 2.96 (1.919) 3.74 0.033 ME

Notes: ***, ** and * indicate significance levels of 1%, 5% and 10%, respectively, against a two-sided alternative. Figures in parentheses are cluster standard errors and they are robust to arbitrary heteroskedasticity and arbitrary intra-group correlation. The endogenous regressor is MALi. The abbreviation used in the table is MEi, malaria ecology.

NOTES 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11.

Examples of papers using the levels approach are Acemoglu et al. (2001), Glaeser et al. (2004) and Sachs (2003a). The growth specification is yˆiT 5 a ln yit 1 gZiT 1 eiT where yˆiT ; 1/T [ ln yiT 2 ln yit ] . Acemoglu et al. (2002) show that around 1500 Western Europe diverged from the rest of the world in terms of standard of living. Prior to that the living standards were more or less equal across countries. A similar point has been made by Caselli et al. (1996) and by Sachs (2003a). The data on GDP per capita in 1820, 1870, 1900 and 1950 are from Maddison (2004). For more details see Data Appendix. The data used is from Rodrik et al. (2004) with data on schooling from Barro and Lee (2000). The data used is from Acemoglu et al. (2001) and schooling data is from Barro and Lee (2000). The data used are also from Acemoglu and Johnson (2005) and Easterly and Levine (2003). See Section IIA, p. 955 of Acemoglu and Johnson (2005) for an explanation on why ‘constraint on the executive’ and the ‘legal formalism index’ are good proxies of ‘property rights institutions’ and ‘contracting institutions’, respectively. The only exception is the Easterly and Levine (2003) specification where we are able to isolate the partial effects of institutions and schooling (see Panel A, column 4). In this sense, the problem is also related to a more general critique of the empirical growth literature by Levine and Renelt (1992) who show that all growth enhancing factors are correlated with each other and also with omitted factors, which makes

M2545 - BHATTACHARYYA PRINT.indd 92

02/03/2011 10:09

Empirical evidence

12.

13. 14. 15.

16.

17. 18.

19. 20.

21. 22.

23. 24. 25. 26. 27. 28. 29.

93

it difficult to estimate the separate effects of these factors on growth using a crosssectional dataset. The World Bank classifies economies into different income categories using gross national income per capita. We use this classification to divide our sample into two groups. The low-income economies (which are the low-and lower-middle-income economies according to the World Bank list) and the rest are taken to be high-income economies. The source of this classification: http://www.worldbank.org/data/about data/errata03/Class.htm. The observed correlation in the LIE sample could be due to the small sample size or the observed variation being too limited. Note that I have summarized some of this literature in Chapter 2. However, it is perhaps worth refreshing our minds to put this section’s empirical exercise into context. Earlier contributions by historians also suggest that malaria indirectly affected development of the continent by causing massive depopulation in the agriculturally marginal regions (Dias, 1981; Miller, 1982). They argue that the slave trade was also an outcome of local epidemiology (particularly malaria) and poor agriculture among other things. For an alternative view, see Young (2005), who use a calibrated simulation for South Africa to forecast that survivors of the AIDS epidemic will be economically better off than they would have been without the epidemic. The intuition in Young’s model is that women become more cautious about sex due to the fear of infection. As others die out, female labour becomes more valuable and a consequent reduction in fertility leads to higher standards of living. Acemoglu and Johnson (2007) argue that their results are not comparable with the micro studies as the micro studies do not incorporate general equilibrium effects. An alternative story of African institutions is from Herbst (2000). He argues that due to the abundance of land in Africa, there was hardly any competition among precolonial states to defend a well-defined territory. This prevented the development of state institutions (tax collection, defence, bureaucracy, rule of law and so forth). This trend of almost no external threat continued during the colonial period. Therefore the colonizers also had very little incentive to develop good institutions. After independence the situation did not change and what we observe now is the weak institutions of contemporary Africa. See Collier and Gunning (1999) for a survey of this literature. Faced with an increasing demand for slaves from the New World, African demand for slaves also increased. Africans preferred female slaves whereas young male slaves were exported across the Atlantic. The result was a huge imbalance in African sex ratio, and a slow down of demographic transition and economic progress (Manning, 1981). There wasn’t enough labour to support capital and facilitate industrialization in an already labour scarce continent. Note that the settler mortality instrument is not free from controversy either. Recently, Albouy (2008) identified several problems with the construction of the original variable in Acemoglu et al. (2001) and the revised dataset was published as an MIT mimeo by the same authors in March 2005 and September 2006. We continue to use the original variable here to facilitate comparison with all other papers that have used this variable. I use log population density in 1500 and interior distance as additional control variables as they might influence current development through other channels. For more details, see http://www.earth.columbia.edu/articles/view/1932 North (1981) defines good institutions as those that provide checks against expropriation by the government and other politically powerful groups. These numbers are the aggregate of Atlantic slave trade, Indian Ocean slave trade, Red Sea and Trans-Saharan slave trade. For more details, see Nunn (2008). My basic results remain unaffected even if I use these additional instruments. Detailed information on the construction of the instrument is available online at http:// www.earthinstitute.columbia.edu/articles/view/1932. Note that including log population density in 1500 and interior distance as additional

M2545 - BHATTACHARYYA PRINT.indd 93

02/03/2011 10:09

94

30. 31.

32.

33. 34. 35. 36.

Growth miracles and growth debacles controls does not alter the malaria result in column 1. In fact, institutions and slave exports become statistically insignificant. The Hansen J-test is preferred over the Hausman test as it is robust to random or cluster heteroskedasticity in standard errors. One could argue that the presence of both the slave trade and institutions in the model weakens the direct effect of institutions on long-run development. To allay this concern, I run a direct contest between malaria and institutions leaving out slave trade as a control. I estimate this model using LIML. Malaria is the clear winner with a coefficient estimate of −1.37 (se: 0.4510) and institutions are statistically insignificant. For curiosity’s sake I also check the robustness of our malaria result using logmort2 from Albouy (2008) as an instrument for institutions instead of Acemoglu et al.’s (2001) settler mortality. The malaria result survives and all other variables are statistically insignificant. I also use corruption and the Sachs and Warner openness index as additional covariates. The malaria result survives these tests. This specification is estimated without interior distance and log population density in 1400 instruments. Surprisingly, I get very different first stage estimates. I am somewhat puzzled with this outcome as I am using exactly the same specification, dataset and sample of countries as Nunn. Not reported here to save space, but available upon request. Note that the result is qualitatively the same if I use humidity as an additional control and not as an instrument.

M2545 - BHATTACHARYYA PRINT.indd 94

02/03/2011 10:09

5.

Root causes of economic progress: a unifying framework

The evidence that I have reviewed so far does identify the important deep factors, namely institutions and diseases that can explain the longrun difference in living standards across nations. This also gives rise to the apparent conflict between the ‘institutions view’ and the ‘disease view’. The genesis of the conflict is the statistical significance of the institutional quality variable in a cross-country regression model estimated at a particular point in time when diseases and other geographic measures are used as controls. One possible reason behind this empirical result is perhaps that institutions and diseases are important at different stages of development and the cross-section regression model is incapable of taking this into account as it solely focuses on a particular point in time. It could very well be the case that diseases are important at an early stage of development and institutions become important at an advanced stage. Bhattacharyya (2009c) presents a case for the existence of stages of development in the cross-national data. Dividing the crossnational sample into LIEs and HIEs, he finds that diseases explain the majority of the variation in per capita income in LIEs (which are at an early stage of development), whereas institutions explain the majority of the variation in the same in HIEs (which are at an advanced stage of development). In this chapter, I make an attempt to build on that finding and explain the interrelationship between institutions, diseases and economic development by using a method which may not satisfy the purists because of its somewhat speculative nature. But this method does have some advantages over the standard cross-country regression modelling approach. It can throw new light on the complex causality issues by bringing the ‘stages of development’ hypothesis, which has been somewhat ignored by the empirical studies, into the forefront. The strategy is as follows. First, I put forward a unifying framework, which describes the process of development in Western Europe. This framework lays down the different stages of development in Western Europe. Second, I compare and contrast the Western European trajectory with the trajectories in Africa, China, India, the Americas, Russia and Australia. This allows us to understand where 95

M2545 - BHATTACHARYYA PRINT.indd 95

02/03/2011 10:09

96

Growth miracles and growth debacles

it went wrong for the other continents. It also allows us to compare and contrast the stages of development across different continents.

5.1 A UNIFYING FRAMEWORK FOR WESTERN EUROPE One could divide the process of economic development in Western Europe into four different stages. The first stage is the era of the Malthusian cycle in which geography and epidemic diseases played a crucial part in determining food production. The widely known economic as well as social impact of the ‘Black Death’ and other epidemic diseases that hit Western Europe during the fourteenth century is a testimony of the power of geography and germs. The second stage is characterized by conflict, militarism and increase in food production and population density. The third stage is the period of an increased demand for wealth. This epoch is characterized by state investments into daring expeditions to acquire wealth and resources from foreign lands to finance the soaring costs of war. Finally, the fourth stage is the stage of change in the nature of the state leading to rapid technological progress, industrial revolution, mass production and the rise of the capitalist system. In order to describe the first stage, one can think of a food production function which makes use of land, labour, human capital, technology and climate. Land is fixed and is owned by the small state or group. The entire population supplies labour except the elite. Human capital is defined as the knowledge required to use the existing technology successfully. Technology signifies the development of new tools and it is spasmodic. Climate is exogenously given. Labour supply is affected by the disease environment with high incidence of diseases resulting in less labour supply. In this kind of world, an increase in food production, because of the positive effects from all of these factors, results in an increase in the population. Increased population raises the demand for food. However, there is a limit to what a fixed amount of land can yield given technology, climate and diminishing returns to labour and human capital. Therefore, a food crisis ensues and eliminates a large proportion of the population. This cycle repeats itself in the absence of technological progress. In addition to this natural cycle, food production is also constrained by epidemic diseases and natural disasters. This is what Thomas Malthus described as the principle of population in his famous essay in 1750. This process took place in Western Europe during the thirteenth and the fourteenth centuries when it was ruled by the small- and medium-sized feudal states. Robbins (1928) provides evidence of a Europe which is close to what I

M2545 - BHATTACHARYYA PRINT.indd 96

02/03/2011 10:09

Root causes of economic progress

97

have described above. In her paper on the impact of the Black Death in France and England, she records that France was hit by famine on at least fifteen occasions during the fourteenth century. In stage two, the production of food may rise due to technological progress. New technology may evolve due to indigenous effort or due to technology transfer or can be completely serendipitous. At least in the case of Western Europe, we know that most of the early technologies were acquired from the Chinese or the Arabs or from ancient Rome (Mokyr, 1990). There is support for this in Diamond (1997, pp. 409–10) as well. He writes: Until the proliferation of water mills after about AD 900, Europe west or north of the Alps contributed nothing of significance to Old World technology or civilization; it was instead a recipient of developments from the eastern Mediterranean, Fertile Crescent, and China. Even from AD 1000 to 1450 the flow of science and technology was predominantly into Europe from the Islamic societies stretching from India to North Africa, rather than vice versa.

The arrival of new technology increases agricultural production by manyfold and creates a situation of food surplus. The food surplus also increases fertility and reduces mortality, raising the total population. Rising population puts pressure on land and other resources inducing the state to get involved in territorial conflicts. This is what we observe during the age of the Crusades when Europe engaged in repeated conflicts and wars. The state also gained more in terms of tax revenue in the event of an increased agricultural yield. A significant proportion of this revenue is spent in the development of new armoury and the military. The logic is simple. More lethal weapons and a well-nourished army can win battles and winning battles is crucial to the very existence of the state. The state investment in military technology creates positive externalities for civilian R&D leading to more breakthroughs in technology for agriculture and crafts. New technology in agriculture and crafts results in steady growth in output and population, causing more territorial conflict. This pattern is observed until the fifteenth century in Western Europe when the states become stronger and stronger. However, the increased frequency of military conflict did put enormous pressure on the state’s exchequer and forced it to look for alternative and richer sources of revenue. Perhaps, this is what led Western Europe to stage three. Stage three signifies a state in which investments in maritime expeditions in order to hunt for alternative sources of wealth and resources become extremely important. It is also a state which borrows heavily from its subjects to finance expeditions and wars. So, the sometimes heavily indebted state is more eager to give away political rights to some of its subjects in

M2545 - BHATTACHARYYA PRINT.indd 97

02/03/2011 10:09

98

Growth miracles and growth debacles

return for loans. This leads to more egalitarian political institutions and also the development of financial markets. Michael Beaud (2000, p. 14) describes this process in his book. He writes, ‘Monarchs greedy for greatness and wealth, states battling for supremacy, merchants and bankers encouraged to enrich themselves: these are forces which inspired trade, conquests, and wars.’ These investments lead to the discovery and conquest of new land. The prevailing mercantilist philosophy1 induces explorers to search for bounty in these newly discovered lands and bring it back to their motherland. This is what Hernan Cortez did when his band of conquistadors came into contact with Montezuma’s Aztecs in the New World. Similar was the fate of Atahualpa’s Inca when Francisco Pizzaro’s army of 200 conquistadors defeated them in Cajamarca in 1532. The capture of Atahualpa by Pizzaro’s men yielded the largest ransom recorded in human history. The ceaseless pillage of wealth and precious stones from the new land triggers inflation in the home country as too much money chases too few goods. To counter inflation, the state imposes restrictions on imports, but encourages exports so that it does not run out of wealth. This policy leads to the expansion of maritime trade and commerce. The ‘no import’ ideology also boosts domestic manufacturing, providing it with a large domestic as well as overseas market. Outward orientation and trade in manufacturing leads to specialization, division of labour and increased gains from trade. The nature of the distribution of gains from trade changes the structure of the political economy and the distribution of political power. Two distinct patterns emerge. The first is an absolutist state which takes control of all gains from trade and concentrates political power. The second is a type of state which allows private accumulation (such as money lending, trading of precious metals, real estate and so forth) by the bourgeoisie and hence a relatively equitable distribution of wealth and political power. In an absolutist state no change occurs in the institutional structure. However, in the second type of state institutional changes take place which are favourable to capitalism. The increase in wealth of the bourgeoisie due to private accumulation provides them with de facto political power. The bourgeoisie invests in private manufacturing and trade which generates more wealth for the future. This further strengthens their de facto political power. The bourgeoisie with their new found political power start demanding institutional change by challenging the authority of the monarchs. They demand protection of private property and a more equitable distribution of political power. If the de facto political power of the bourgeoisie is greater than the de jure political power of the monarchs then the will of the bourgeoisie prevails over the will of the monarchs. This leads to the establishment of institutions which protect private property and the political rights of the bourgeoisie. Democratic

M2545 - BHATTACHARYYA PRINT.indd 98

02/03/2011 10:09

Root causes of economic progress

99

institutions are established to cement the power of the bourgeoisie and also to make sure that the monarch cannot take over power in the future. This pattern of institutional development is observed in Western Europe from the sixteenth century onwards. The Spanish and Portuguese monarchs were absolutist in nature and they centralized the process of manufacturing and trade, discouraging private enterprise (Acemoglu et al., 2005b). This prevented the development of institutions which provide incentives to private investment. However, in Britain and the Netherlands, the state allowed private enterprise, which led to the Civil War in 1642 and the Glorious Revolution in 1688 in Britain and the Dutch War of Independence which began in the 1570s. Describing the events in Britain and the Netherlands, Acemoglu et al. (2005b) writes: The victory of Parliament in the Civil War and after the Glorious Revolution introduced major checks on royal power and strengthened the rights of merchants. After the Civil War, the fraction of MPs who were merchants increased dramatically. (p. 564) . . . Dutch merchants always had considerable autonomy and access to profitable trade opportunities. Nevertheless, prior to the Dutch Revolt, the Netherlands (in fact, the entire Duchy of Burgundy) was part of the Habsburg Empire, and the political power of Dutch merchants was limited . . . The critical improvement in Dutch political institutions was therefore the establishment of the independent Dutch Republic, with political dominance and economic security for merchants, including both the established wealthy regents and the new merchants immigrating from Antwerp and Germany. (p. 566)

Stage four signifies more private as well as state investments in technology, which leads to the development of the factory system and industrial revolution. The institutional changes of stage three create the ideal incentive structure for private investments in technology development. This induces rapid technological progress. The rapid improvement in technology increases the cost of moving information relative to the cost of moving people (Mokyr, 2001). This leads to the rise of the factory system and a subsequent breakdown of cottage industry. Such a pattern is observed in Britain and other parts of Western Europe during the period of the Industrial Revolution (1760–1830). Therefore, in sum, the story that I want to get across is as follows. Western Europe managed to beat the constraints imposed by its geography, in particular diseases, on food production early on and started a journey on an independent growth trajectory. Availability of food increased population density which caused territorial conflicts and war. Ceaseless conflicts induced more investment in military technology. The

M2545 - BHATTACHARYYA PRINT.indd 99

02/03/2011 10:09

100

Growth miracles and growth debacles

conflicts also put enormous pressure on the finances of the state. The state commissioned daring naval expeditions to search for bounty so that it could finance its military expenditure and avarice. These expeditions brought wealth from overseas which also caused inflation. In order to remedy inflation and also to abide by the principles of mercantilist philosophy, the state restricted imports of foreign goods and promoted exports of domestically manufactured goods. This induced specialization and division of labour in the domestic economy. Institutional changes followed depending upon the initial distribution of the gains from trade. A non-absolutist state allowed bourgeois accumulation which increased the power of the bourgeoisie, resulting in major institutional changes favourable to capitalism. In contrast, an absolutist state allowed very little or no bourgeois accumulation which arrested the prospect of any institutional change. The states with capitalist institutions attracted private investments in production and technology building. This led to rapid technological progress, the rise of the factory system and industrial revolution. Therefore, what we learn from the unifying framework is that breaking the disease bottleneck is crucial for future institutional development, which leads to sustained technological progress and economic growth.

5.2 WHAT WAS DIFFERENT IN AFRICA, CHINA, INDIA, THE AMERICAS, RUSSIA AND AUSTRALIA? 5.2.1

Africa

Africa has a long history of diseases. Epidemic diseases such as smallpox, measles, yellow fever, cholera, tuberculosis, malaria and typhus have always been a part of African life. Many of these diseases and some new killers (HIV/AIDS is an example) play a significant role in African life to date. Africa has also been subject to huge climatic variations. Long dry seasons were followed by very humid periods with heavy rain (Miller, 1982). These factors have impacted in the past and still continue to impact Africa’s growth trajectory. If one seeks an explanation in terms of the unifying framework that I have outlined in Section 5.1, the obvious question to ask is at what stage did African economics go wrong? My answer is stage one. Why stage one? The intuitive explanation is as follows. Geography has always constrained food production in Africa. Long stretches of drought causing major reductions in cultivation have always weakened the African population by subjecting them to malnutrition. Malnutrition made them vulnerable to epidemic diseases. A return of the

M2545 - BHATTACHARYYA PRINT.indd 100

02/03/2011 10:09

Root causes of economic progress

101

rain also brought diseases along with it, further weakening the labour force – an important input in food production. Miller (1982) writes, ‘Outbreaks of diseases paralleled the chronology of drought in an epidemiological sequence familiar from many other regions. Africans weakened by malnutrition and exhausted by dispersal into the bush or by flight into lowland became particularly vulnerable to endemic pathogens’ (pp. 22–23). A Portuguese observer in eighteenth century Angola commenting on the increase in disease incidence after the rain writes, ‘Rain brings food in abundance but leaves no one alive to eat it.’2 This situation was further complicated by the African involvement in the slave trade. Africa had a long history of slavery as a social institution. However, it was never commercialized on such a large scale prior to the European engagement.3 The slave trade led to depopulation of the continent, reducing food production further (Inikori, 1992).4 However, the fact is that without depopulation, Africa struggled to produce more than a subsistence level of food grains. This restricted Africa from attaining stages two, three and four and moving towards the development of a fully home-grown capitalist system. The engagement with the Europeans during the sixteenth century and formal colonization during the nineteenth century aborted the independent trajectory of institutional development in Africa. In the colonies with a high European mortality rate, the colonizers erected extractive institutions. The slave trade encouraged the African elites to go for violent slave raids inland which institutionalized the culture of violence and lawlessness in certain parts of the continent. Many of these institutional features have persisted over time and still exist in the economic and political institutions of modern Africa. These weak institutions continue to influence the economic performance of the continent. Coupled with diseases and geographic constraints, poor institutions perhaps explain the bulk of the African growth tragedy. Let me introduce an overlapping generation model to explain the story that I have laid down above. The idea in this model is that the direct relationship between the rate of time preference and diseases at the representative household level is perhaps the best explanator of the African situation. In other words, households living in an environment where malaria incidence and death rate are typically high will choose to save less for the future and this affects economic progress adversely. I consider a closed overlapping generation economy consuming and producing a single homogeneous product. This can be thought of as a representative economy in the continent of Africa or the whole continent itself. A typical household in this economy comprises of both young and old members and each member of the household lives for only two periods.

M2545 - BHATTACHARYYA PRINT.indd 101

02/03/2011 10:09

102

Growth miracles and growth debacles

The young members of the household work in the first period and retire in the second period when they are old, and then they die. The members of this household consume in both periods and the consumption in the second period is supported by their savings in the first period. Therefore, at each point in time, members of only two generations are alive. Each individual within the household maximizes their lifetime utility, which depends on consumption in the two periods of life. In order to maintain the simplicity of the structure, I assume away the possibility of bequests or any altruistic behaviour. The lifetime utility of a representative individual of generation t can be expressed as follows: ut 5

2q c11t2q 2 1 c12t11 21 1 1 ␾a ba b, q . 0, r . 0, ␾ [ [ 0, 1 ] (5.1) 12q 11q 12q

where c1tand c2t11 are the consumption of generation t when young and old, respectively, r is the exogenously determined pure rate of time preference and q is the intertemporal elasticity of substitution. The survival probability to the old age ␾ of the representative individual depends on the unfavourable geography vector, G.5 ␾ 5 ␾ (G) [ [ 0, 1 ]

(5.2)

G is exogenous to the model and shares an inverse relationship with ␾. If G is too high then ␾ can be too low. Using Equation 5.2, one can rewrite the lifetime utility of the representative individual as follows: ut 5

2q c11t2q 2 1 c12t11 21 1 s (G) a b 12q 12q

(5.3)

where s (G) 5 ␾ (G) /1 1 r is the effective rate of time preference. If G is too high then s is too low. The representative individual supplies one unit of labour inelastically when young and receives a wage income wt. Therefore the budget constraint faced by this individual in period t is given by the following expression: c1t 1 st 5 wt

(5.4)

where st is the amount of grain saved in period t. The saved grain also grows at a rate rt11 when planted. Therefore in period t 1 1 the individual

M2545 - BHATTACHARYYA PRINT.indd 102

02/03/2011 10:09

Root causes of economic progress

103

consumes the amount saved in period t plus the growth in the grains. So the consumption in the second period is as follows: c2t11 5 (1 1 rt11) st

(5.5)

Using Equation 5.5 one can rewrite the budget constraint as c2t11 5 wt 1 1 rt11

c1t 1

(5.6)

Each individual treats wt and rt11 as given and maximizes their lifetime utility subject to the budget constraint. This yields the following Euler equation: 1 c2t11 5 [ s (1 1 rt11) ] q c1t

(5.7)

Using Equations 5.6 and 5.7 and solving for c1tand c2t11 yields 1

c1t 5 wt [ 1 2

1 2q

s c2t11 5 wt [

] and

1

(1 1 rt11) 1 2 q 1 1 1 1 rt11

1 2q

s

] 1

(5.8)

(1 1 rt11) 1 2 q 1 1

I assume that the representative household uses a Cobb-Douglas technology to produce the grains and the factors of production are paid according to their marginal product. The production function is represented as follows: yt 5 Akat

(5.9)

where yt and kt are per capita output and per capita capital stock, respectively. In this economy, net investment has to be equal to total income less consumption. Therefore, capital stock in this economy evolves as follows: Kt11 2 Kt 5 AKat L 1t 2a 2 dKt 2 c1tLt 2 c2tLt21

(5.10)

I assume that the economy starts off with an initial capital stock K1 that is owned by the L0 elderly in period 1. Also each individual wants to end

M2545 - BHATTACHARYYA PRINT.indd 103

02/03/2011 10:09

104

Growth miracles and growth debacles

up with no assets when they die. This yields the result that the savings of the young in this period are equal to the next period’s capital stock. Hence we have the following relationship: Kt11 5 stLt4t $ 2

(5.11)

In this economy, the population grows at a rate n. Then using the above relationship and logarithmic preferences6 I get the steady-state value of per capita capital stock as k* 5 c

1

(1 2 a) As 1 2a d (1 1 n) (1 1 s)

(5.12)

It is evident from the above equation that a high value of s increases the steady-state level of per capita capital stock and hence enhances growth. In Africa, due to the high incidence of killer malaria and other diseases, one would expect Gto be very high and s to be fairly low. The economy in this case is in danger of getting trapped into a low income, low capital stock equilibrium which I have described as stage one of the unified framework. This result is also in line with the empirical findings of Bloom et al. (2003). They show that economies experience lack of growth not due to geographical bottlenecks per se but due to the poverty trap situation emerging out of the bottleneck. This perhaps explains the situation in Africa which is supported by the empirics in Bhattacharyya (2009b). Another observation is that if contemporary Africa is stuck at stage one due to diseases and other geographic constraints then the data is going to show a strong correlation between the current level of development and these factors. The correlation between institutions and other factors will not be visible if it is a poverty trap situation similar to stage one. This is precisely what the data shows in Chapter 4. 5.2.2

China

The case of China is somewhat surprising. The Chinese were at the forefront of the Old World technology and knowledge till the mid-fifteenth century. Cast iron, the compass, gunpowder, paper, printing and many others were first invented in China. The Chinese also invented sophisticated irrigation canals which increased rice production by many-fold (Diamond, 1997). In light of this long list of technological breakthroughs, why have the Chinese failed to achieve the same heights as the Western Europeans? Why did they waste their early technological advantage? Why was it Britain and not China that progressed towards building an

M2545 - BHATTACHARYYA PRINT.indd 104

02/03/2011 10:09

Root causes of economic progress

105

industrial society? The answer lies with Chinese institutions. The following paragraph attempts to provide an intuitive explanation in terms of the broad structure. Food production developed in China as early as 7500 BC (Diamond, 1997, p. 100). By the start of the millennium, Chinese agriculture was able to support a large population and the hierarchical structure of Chinese society was comparable to the social institutions of stage two and three of the proposed broad structure. One can claim that by the fourteenth and the fifteenth centuries, China has taken significant steps towards reaching stage four. The treasure fleets of the early fifteenth century, the discovery of gunpowder and the compass suggest that the Chinese were incredibly close to making it to stage four. However, the question remains, what went wrong? The fate of the treasure fleet after it returned in 1433 gives us a clue to the answer. After the return of the fleet in 1433, the composition of the Chinese state changed significantly. The previously powerful eunuchs were overthrown by their opponents within the Chinese court. This was partly triggered by Li Zicheng’s rebellion and the collapse of the Ming Dynasty into the hands of the Manchu-led Qings. The eunuchs were in favour of technology, scientific discovery and daring expeditions. Ming rule, under which the eunuchs were influential, saw a rapid growth in private maritime trade, especially with Portugal and Spain,7 the size of the navy, and construction projects related to infrastructure (see Ebrey, 1999; Ebrey et al., 2006).8 The commander of the treasure fleet, Cheng Ho, was himself a eunuch. When their opponents assumed power, they aborted all the activities that the eunuchs were involved in, either directly or indirectly. Gradually, they dismantled the entire infrastructure that was put in place to encourage these activities. The absolutist nature of their regime also did not allow private initiatives in these activities. In this way the absolutist regime destroyed all the institutional incentives for technological research (Landes, 1998) and China went backwards in the next 500 years. This is a good example of the theoretical claim that bad institutions can destroy all the incentives for economic progress even when the region is endowed with the right geography. In summary, the Chinese experience shows that escaping the poverty trap is a necessary, but not sufficient, condition for development. In other words, a transition from stage two to stages three and four is not automatic. It also demonstrates that institutions are a deeper cause of development than technology as technological progress is not sustainable without strong institutions.9 The contemporary Chinese success story is perhaps another strong example of institutional reversal, where local level institutions are more favourable towards business and economic progress (Rodrik, 2000). How this change happened, however, is a different matter.

M2545 - BHATTACHARYYA PRINT.indd 105

02/03/2011 10:09

106

5.2.3

Growth miracles and growth debacles

India

Pre-modern India India escaped the strong grips of the Malthusian cycle (stage one) long before the British arrived. India was an exporter of industrial goods and an importer of primary and intermediate goods when Sir Thomas Roe visited the court of the Mughal emperor Jahangir in 1615.10 After the European conquest of the Americas, international trade increased manyfold. India also became an important part of this triangular trade network (Prakash, 2004). Europeans got access to precious metals in South America through conquest. They traded precious metals (especially silver) in return for manufactured goods from India. Initially it was the Portuguese who established a monopoly over this trade but later on they were successfully challenged by the British and the Dutch East India Companies. In spite of an increase in trade this period delivered very little in terms of overall economic growth in India. This could be explained by institutional weaknesses. Monopoly rents of the landed elite were protected by the state through violence, higher taxes and other forms of coercive instruments. The state created entry barriers for the outsiders (especially farmers, merchants and artisans) through higher taxes and violence. Revenue collected from farmers and artisans in the form of taxes was largely spent on conspicuous consumption. There was very little investment in public goods such as roads and transportation networks. This seriously constrained private investments and discouraged technological innovation. As a result, growth dividends were fairly limited in spite of an increase in trade. In summary, property rights institutions were weak because political power was concentrated in the hands of the elite and the merchants were unable to constrain the actions of the elite. Trade produced very few growth dividends in an institutional environment which was weak. Bhattacharyya (2011) presents a more detailed account of this period. In spite of sluggish growth in pre-modern India some progress was made during the Mughal period. The structure of the Mughal Empire was already very hierarchical with power concentrated in the hands of the minority elites. It also generated enormous amounts of wealth for the nobility. In support of this fact Landes (1998) writes, ‘India also had a large and skilled industrial workforce, whose products circulated throughout the region. As a result, the Indian yielded a substantial surplus that supported rulers and courts of legendary opulence’ (p. 156). Therefore, it is perhaps fair to say that the Mughal Indian society achieved living standards and institutional structure comparable to stage three of the proposed broad structure. However, this pattern reversed as

M2545 - BHATTACHARYYA PRINT.indd 106

02/03/2011 10:09

Root causes of economic progress

107

the British started gaining more political control during the late eighteenth century. The obvious question that one would like to ask is why? The answer is not as complicated as it may seem. Acemoglu et al. (2002) talk about an institutional reversal that brought about this change. Their definition of institutional reversal, however, is very broad. They argue that the British colonizers never considered India and other tropical colonies as possible settlements and therefore they erected extractive institutions in these colonies. These extractive institutions reversed the trend of economic performance. In the case of India, however, it wasn’t only the lack of settlement opportunities that persuaded the British colonizers to erect extractive institutions. It was also a direct result of the then prevailing political economy in both countries. Dutt (1992) argues that strong parliamentary lobbying by the British cotton manufacturers against the import of Indian textile forced the East India Company to resort to policies which led to a systematic destruction of the Indian textile industry. He writes, ‘Even in 1813, witness after witness in the Select Committee of the House of Lords testified that free Indian textile imports (of both finer and coarser varieties) would damage British industry’ (pp. 148–149). The British East India Company resorted to policies of imposing internal tariffs and transit duties on Indian goods, dislocation and direct exploitation of the artisans, and forceful reduction of market demand to destroy the industry.11 Indian textiles also lost their overseas market due to the imposition of high import tariffs in Britain. The company had an influence on the land tenure system and property rights during that time. In many areas the existing landlords received proprietary rights in land. The company extracted rents from them without caring much about investment. The landlords passed on this burden of rent to the farmers and the poor farmers struggled to make investments in capital and technology. This system of rent-seeking significantly reduced agricultural productivity and trapped farmers in a vicious cycle of poverty. One such institution is the Permanent Settlement concluded by the Cornwallis administration in 1793. It was a grand contract between the company, the government and the Bengal landlords. Under the contract, the landlords were admitted into the colonial state system as the absolute proprietors of landed property and the government was barred from enhancing its revenue demands from the landlords. This arrangement institutionalized the alliance between the landlords and the colonial rulers. It also legitimized rent-seeking. In a recent study, Banerjee and Iyer (2005) show that these institutional arrangements had and continue to have a significant impact on economic outcome within India. Areas where proprietary rights were given to the landlords have significantly lower agricultural investments and productivity than areas where rights were given to the cultivators.

M2545 - BHATTACHARYYA PRINT.indd 107

02/03/2011 10:09

108

Growth miracles and growth debacles

Therefore, colonization by the British led to institutional reversal which prevented India from reaching stage four and developing a home-grown efficient capitalist system. The progressive forces within Indian society which had the capacity (at least theoretically) to push the economy towards large-scale industrialization were systematically destroyed by the existing polity. The domestic extractive institutions were allowed to continue and it strengthened the feudal landlords both economically and politically. These institutional changes systematically destroyed the incentives for private investments in land, capital, and technology. As the incentives changed, so did the comparative advantages. India soon became a net exporter of raw materials and primary products and a net importer of industrial goods. What ensued is two centuries of deindustrialization and economic slowdown. Deindustrialization during the eighteenth and nineteenth centuries In recent research using the long-run relative price series, Clingingsmith and Williamson (2008) document that India suffered from deindustrialization during the eighteenth and nineteenth centuries.12 The novelty of Clingingsmith and Williamson’s approach is that they use relative price data as opposed to employment and output share numbers used in previous studies (see Clark, 1950; Thorner, 1962; Bagchi, 1976a,b; and Prakash, 2005). The key question, however, is why Indian manufacturing suffered deindustrialization. Bhattacharyya (2011) provides a detailed account of the series of events that led to deindustrialization. Without going into detail, I will summarize the economic theory here that may help explain deindustrialization in India. Indian manufacturing (especially cotton textiles) faced both supply side and demand side shocks. On the supply side, frequent climate shocks (droughts and floods) had a profound impact on productivity. On the demand side, globalization forces in the form of declining transport cost and productivity improvements in British manufacturing made it very difficult for Indian manufacturing to compete in the global market. Globalization and productivity growth in British manufacturing significantly reduced the market share of Indian manufacturing in the international market. Domestic demand for Indian manufacturing also took a hit with the disintegration of the Mughal Empire and the subsequent decline of royal courts who were major buyers of Indian manufacturing. Indian manufacturing succumbed to these shocks because of weak institutions. In the chaos that ensued after the disintegration of the Mughal Empire, local powers increasingly resorted to tax farming and expropriation of private property to finance a seemingly never ending conflict (Bhattacharyya, 2011). This led to very little investment from the

M2545 - BHATTACHARYYA PRINT.indd 108

02/03/2011 10:09

Root causes of economic progress

109

local merchants and artisans in machinery and technology to keep up with the productivity improvements in British manufacturing. As a result deindustrialization ensued in the eighteenth and nineteenth centuries. In summary, Indian manufacturing was unable to withstand shocks because of weak institutions during the late Mughal period. These rent-seeking institutions persisted throughout the period when the British East India Company took control in 1757 after the Battle of Plassey and when the British Crown took over in 1858. The Permanent Settlement of Bengal between the East India Company (headed by Lord Cornwallis) and the Bengali landlords, concluded in 1793, made tax farming legal (Banerjee and Iyer, 2005). The agreement between the two parties was that the landlords would transfer a fixed magnitude of the revenue raised from land to the company. Post-independence India After independence in 1947 India resorted to a development strategy heavily reliant on public sector investments. Private property and private investments were not illegal but they were discouraged. This model of development was inspired by the success of the Soviet Union at that time. Bhattacharyya (2011) provides a detailed account of the economic history of that time. This strategy of development was heavily reliant on command and control and it brought in layers of inefficient bureaucracy and regulatory institutions. The culture of rent-seeking became wellentrenched in the system and India experienced very slow growth during this time. Rodrik and Subramanian (2004) document that India’s growth rate in per capita income over the period 1950 to 1980 was as low as 1.7 per cent. This was mainly an outcome of poor regulatory institutions which encouraged inefficiency and corruption. 1980s policy shift and turnaround There was good news for the Indian economy after a policy shift in the 1980s. The Indian economy experienced rapid growth in per capita income in the 1980s (Rodrik and Subramanian, 2004; De Long, 2003; and Williamson and Zagha, 2002). The cause of the 1980s turnaround was mainly attitudinal (Rodrik and Subramanian, 2004). In a gesture demonstrating clear disconnect from the past, the government of the day became pro-business and encouraged private investments without any significant change in policy. This accidentally came about when Indira Gandhi was seeking support from the business community to ward off the electoral threat from the Janata Party (Kohli, 1989). This in itself was enough to kick-start growth in the 1980s. The 1990s reforms were a more serious policy shift (Bajpai and Sachs,

M2545 - BHATTACHARYYA PRINT.indd 109

02/03/2011 10:09

110

Growth miracles and growth debacles

1999; Ahluwalia, 2002; Srinivasan and Tendulkar, 2003). Trade openness, economic liberalization and privatization led to improvements in institutional quality. Better institutions and pro-business policies were able to deliver rapid growth (Bhattacharyya, 2011). 5.2.4

The Americas

When the Europeans first arrived in the Americas in the late fifteenth century, the indigenous American civilization of the Incas and the Aztecs was quite developed both economically and politically. The Incas and the Aztecs developed agriculture, which was capable of supporting large populations. Their political structures were also very advanced and somewhat similar to the Europeans. The majority of the political power was concentrated in the hands of the minority elites and the ruling nobility. If one wants to make a comparison between the then states of the Europeans and the indigenous Americans, one would be able to point out that there were certain things that the indigenous Americans were able to achieve and there were certain things that they failed to achieve. Whatever they may be, they are secondary to my focus. The important issue is that the European arrival stalled the independent process of development in the Americas. The indiscriminate massacre of the indigenous population and epidemic diseases, such as smallpox, contracted from the Europeans rapidly reduced the indigenous population to an inconsequential level. This allowed the Europeans to grab more indigenous land and erect institutions, which were along the lines of institutions in Western Europe. However, in the case of the Spanish colonies in South America, the Spanish colonial rulers continued with the Inca tribute system and other rent-seeking institutions for their own benefit. Engerman and Sokoloff (2001) argue that the institutional differences between North and South America after the European conquest stem from the factor endowment of the two continents. The following is their theory. They argue that the factor endowment in South America supported resource extraction and rent-seeking. Huge reserves of precious metals supported mining. The climate in many of the southern colonies was suited to growing sugar, which can be efficiently produced in large plantations. To enjoy economies of scale and extract maximum value, the owners of mines and plantations employed a large population of slave labour. These labourers had no rights and no assets. This contributed to the extreme differences in the distributions of landholding, wealth and political power, which shaped future institutions in the south. In contrast, the factor endowment in the northern colonies supported small family-sized farms and farming of grains and livestock. This led to the development

M2545 - BHATTACHARYYA PRINT.indd 110

02/03/2011 10:09

Root causes of economic progress

111

of a society with relatively equitable distribution of wealth and political power, and institutions that honoured private property rights. The better institutions of the North contributed to development as an advanced capitalist society, whereas for the South it was always a struggle thereafter. Robinson (2006), in an excellent survey of North American development history, also finds support for the Engerman and Sokoloff theory. He writes, British American colonies were founded by entities such as the Virginia Company and the Providence Island Company whose aim was to make profits. The model that they had in mind was not so different from that adopted by the Spanish or Portuguese (a system that other British colonizing entities, such as the East India Company, used to great effect). Yet these colonies did not make money and indeed both the Virginia Company and the Providence Island Company went bankrupt. A colonial model involving the exploitation of indigenous labor and tribute systems was simply infeasible in these places, because of lack of large indigenous population and the absence of complex societies. (p. 28)

A manorial system, as envisaged by Charles I, failed to materialize in places such as Maryland due to the acute shortage of labour. As a result, institutions in these settlements ended up giving far more economic and political rights (access to land, property rights and suffrage) to the migrants than they originally wished. Therefore, institutions encouraged private enterprise and investments, which resulted into economic growth in the long run. In contrast, the Spanish and the Portuguese colonies of the south were abundant in indigenous labour and natural resources, which the colonizers used to good effect to set up extractive colonies. These colonies ran on exploitation of indigenous labour and the native tribute system. After Pizzaro’s conquest of Peru, he set up several institutions to extract rent from the indigenous population (Robinson, 2006). Among these institutions were: (1) encomienda (forced labour), (2) mita (forced labour used in the mines), and (3) repartimiento (forced selling of goods at a higher price to the native population). Many of these institutions persisted till independence and they discouraged private enterprise and investment throughout. This is perhaps a major reason for the lack of growth and economic stagnation in Latin America. In summary, the above discussion shows that two different styles of colonization policies and, hence, institutions created two different types of capitalist societies in the Americas after the European conquest. Many of the old indigenous institutions were replaced after the conquest. The economic performance of the two continents thereafter depended on the

M2545 - BHATTACHARYYA PRINT.indd 111

02/03/2011 10:09

112

Growth miracles and growth debacles

new institutions. The Engerman and Sokoloff (2001) theory shows that the difference in living standards of the two continents can be explained by institutional differences, which have their roots in the respective factor endowments. 5.2.5

Russia

It is perhaps appropriate to say that by the start of the nineteenth century Russia had managed to successfully overcome the Malthusian disease bottleneck. Therefore, it is perhaps best described as an economy operating in stage three of our unifying framework. However, Russia experienced a decline relative to Western Europe over the next two centuries. This is somewhat puzzling given that Russia was not very far away from the technological breakthroughs and the social changes that were taking place in Western Europe during that time. Therefore, the key question is what explains this decline. To understand, we may have to go back to the nature of resource endowments that Russia had at the start of the nineteenth century. Russia, at that time, was a predominantly land-abundant and labour-scarce country with most of its population living in the European parts of the country. It was also not as mechanized as the Western European economies of that time. Therefore, the elite held most of their wealth in the form of land and serfs. The country was virtually run by an alliance between the monarch and the landed elites. In August 1812, Napoleon Bonaparte invaded Russia from the west. It is well documented, including in Tolstoy’s classic novel War and Peace, that the elite and the peasants fought side by side for the patriotic cause. This and several other factors led to the belief in some sections of the liberal nobility that the future of Russia rested on the shoulders of the serfs and the peasants. Inspired by this idea, a group of liberal noblemen (including Prince Sergey Volkonsky) led a popular, but unsuccessful, uprising on 14 December 1825. Legendary Russian poet Alexander Pushkin was also sympathetic to this idea and to part of this group, who were known as the Decemberists. The main idea behind the Decemberist uprising was to grant more democratic rights to the peasants and the serfs. The landed elite and the monarch vigorously resisted any such democratic pressure. This is not surprising in a country where the majority of wealth of the nobility came from the land. Land being an asset easy to expropriate or impose tax upon, it is not difficult to understand why the elite felt so threatened by the prospect of relinquishing political power. Therefore, the outcome was a high level of inequality and weak institutions, which persisted over a century. This led to very little investment relative to Western Europe in the creation of market-supporting public goods, such as road or rail networks.

M2545 - BHATTACHARYYA PRINT.indd 112

02/03/2011 10:09

Root causes of economic progress

113

The lack of growth and high level of inequality persisted over a century and led to the Bolshevik Revolution in 1917 and the creation of the Soviet Union. What ensued was an institution reversal of a different kind, where the state takes control of all private property and investments. These institutions were not conducive to private investments and markets. The lack of economic incentives for private individuals was crippling to growth over the long term and ultimately led to the collapse of the socialist system in 1991. In summary, it was not disease or geography that hindered economic progress in Russia. It was the crippling political institutions during the nineteenth century Czarist Russia and restrictive political and economic institutions of the Soviet Union that limited growth over two centuries. 5.2.6

Australia

The initial resource endowments in Australia during the early part of colonization in the 1820s and 1830s were similar to Argentina and other Latin American countries. There was abundant land with a very small native population and a minuscule European population consisting of convicts, ex-military servicemen and the military. Therefore, the key question is why Australia took such a different course from Argentina over the next two centuries. The answer may lie in Australian institutions. By the 1840s, the British colonizers moved towards granting self-government to Australia. This step by the colonial rulers was perhaps motivated by the earlier loss of American colonies and also the problems in Upper and Lower Canada. The decision to not grant self-rule and more representation in the North American colonies cost the British colonizers dearly. So they may have learned a lesson or two from that experience. Therefore, they did not risk it with Australia. Furthermore, the political trends in Britain may have also influenced this decision. Nevertheless, Australia established fairly representative political institutions to support its markets from a very early stage. In contrast, the landed interests in Argentina fiercely resisted democratic pressures. This may have been due to their fear of ownership change and taxation in the event of democratization (Acemoglu and Robinson, 2006). In a land-abundant economy, where the elite hold most of their wealth in the form of land, the elite may strongly resist democratic pressures. This is because of the relative ease with which land can be taxed or expropriated by the new regime in the event of a regime change. The resistance is likely to be fiercer if the agriculture sector is closely aligned with international trade as it reinforces the political power of the landed elite. This perhaps

M2545 - BHATTACHARYYA PRINT.indd 113

02/03/2011 10:09

114

Growth miracles and growth debacles

explains why Argentina did not undergo democratization at that time, which led to a relative decline in living standards, especially during the twentieth century. Therefore, to sum up, the key to Australia’s success is institutions; the geography bottleneck was not a factor. Better institutions may have also helped Australia to diversify and industrialize during the twentieth century in the event of frequent commodity price shocks, which were larger in magnitude in some cases than those that an average African commodity exporting country experienced (Bhattacharyya and Williamson, 2009).

5.3 SUMMARY In summary, the framework presented above shows that diseases are important at an early stage of development. But as technology coupled with population growth and some good luck allows a society to escape this early stage then institutions become important. The interactions of institutions, technology and trade drive the economy to a sustained growth path thereafter. This is perhaps an appropriate way of describing the process of economic development in Western Europe. In China and India, the Malthusian cycle was broken fairly early on and institutional weaknesses played a crucial role in their respective decline or stagnation. In the Americas and Australia, colonial institutions were a crucial factor. In Russia, it was the crippling political institutions of the nineteenth century and the restrictive political and economic institutions of the Soviet Union that did the damage. The African case was somewhat different from the others as the continent struggled to escape from the strong grip of the Malthusian cycle. The long history of the slave trade and colonial institutions complicated the story even more later on. Therefore, the unifying framework does show that there is a case for dealing with diseases and institutions in the same framework rather than in isolation. This is a major value added to the literature given that the existing studies tend to view these two factors as being mutually exclusive, rather than interlinked. Even though it is necessary, understanding history is not sufficient for better policy outcomes in contemporary times. In Part II of this book, I specifically look at policy outcomes over the last two to three decades. I ask the questions: whether it is possible to influence institutional quality through trade policy? Whether trade can improve growth on its own? Which institutions are important for economic growth? I provide empirical evidence and propose a roadmap for future growth.

M2545 - BHATTACHARYYA PRINT.indd 114

02/03/2011 10:09

Root causes of economic progress

115

NOTES 1. 2. 3. 4. 5.

6. 7.

8. 9. 10. 11.

12.

The major argument of the mercantilist philosophy is that a nation’s wealth depends on the amount of precious metal it has. Cited in Miller (1982), p. 23. The Islamic slave trade started in AD 700, but it never reached the epic proportion of the Atlantic slave trade. The historians are yet to reach any agreement on this. For alternative views, see Lovejoy (1982). One argument made by Chakraborty and Das (2005) in a recent paper is that the households can influence ␾ by investing in health. However, in this case we assume that increase in ␾ requires huge investment, which is often beyond the scope of a private investor or a household. This is because malaria is predominantly geographic in nature and its reduction would require considerable public health intervention. Logarithmic preference implies q 5 1. Initially, the Ming court wanted to control trade by using some formal rules. However, these rules became impossible to sustain with the advent of international trade with the Europeans. In support of this, Ebrey et al. (2006, p. 277) write, ‘In the sixteenth century, this formal system [of containing trade] proved unable to contain the emergence of an international East Asian maritime trading community composed of Japanese, Portuguese, Spanish, Dutch, and Chinese merchants and adventurers. Because the profits to be had from maritime trade were high, both open and clandestine trade took place all along the coast.’ Irrigation projects, restoration of the Grand Canal and the Great Wall are some of the construction projects that the Ming Dynasty undertook. This is somewhat similar to the Portuguese and the Spanish experience: they failed to capitalize on their initial technological advantage in maritime trade and shipbuilding largely due to absolutist institutions (Acemoglu et al., 2005b). Sir Thomas Roe was the emissary of King James I and he gained for the British the right to establish a factory at Surat, a port city where the British East India Company’s ships first arrived in India. According to Dutt (1992), many artisans were subjected to flogging, imprisonment, and worse. Cutting off the thumbs of winders of raw silk has been documented. The domestic demand for textiles also reduced significantly due to the decline of the Indian royal courts, as they were the major buyers of the quality products. Note that whether deindustrialization took place in India is a long-standing debate in economic history. We deliberately avoid getting dragged into that debate here. We start with the premise that deindustrialization took place. For a more detailed account of the deindustrialization debate, see Bhattacharyya (2011).

M2545 - BHATTACHARYYA PRINT.indd 115

02/03/2011 10:09

M2545 - BHATTACHARYYA PRINT.indd 116

02/03/2011 10:09

PART II

Promoting growth in the current environment: evidence and policies

M2545 - BHATTACHARYYA PRINT.indd 117

02/03/2011 10:09

M2545 - BHATTACHARYYA PRINT.indd 118

02/03/2011 10:09

6.

Institutions and trade: competitors or complements in economic development

6.1 THE BACKGROUND The empirical relationship between trade and development at the crossnational level has been a topic of research for several decades now. Many policymakers argue that trade openness is beneficial for growth. Outward orientation of a country promotes efficiency amongst the local firms and also facilitates technology transfer. Both of these are good for growth. The root causes literature has reinvigorated the debate over the effectiveness of policy and trade policy, in particular, in promoting growth. Until recently, it appeared that a growing academic as well as policy consensus was emerging on the positive effects of trade on development. Dollar (1992), using an ‘index of real exchange rate distortion’ and an ‘index of real exchange rate variability’, shows that outward orientation is good for economic growth. Sachs and Warner (1995a) construct an index that combines all aspects of trade policy and show that countries with an open trade regime, on average, perform better than countries with closed trade regimes. Ben-David (1993), on the other hand, shows that trade liberalization leads to less dispersion in income across countries and hence convergence. In another influential study, Frankel and Romer (1999) show that there is a positive relationship between trade volume and national income to the extent that an increase in trade volume is the result of a reduction in natural or geographical barriers to trade rather than trade policy. They use the geographical component of trade volume as an instrument to identify the effects of trade on income. In the policy arena, the World Bank (2002) emphasizes the advantages of trade openness, especially for developing economies. In their report entitled ‘Globalization, Growth and Poverty’, they write: Some 24 developing countries – with 3 billion people – have doubled their ratio of trade to income over the past two decades. The rest of the developing world trades less today than it did 20 years ago. The more globalized developing 119

M2545 - BHATTACHARYYA PRINT.indd 119

02/03/2011 10:09

120

Growth miracles and growth debacles

countries have increased their per capita growth rate from 1 per cent in the 1960s to 3 per cent in the 1970s, 4 per cent in the 1980s, and 5 per cent in the 1990s . . . much of the rest of the developing world – with about 2 billion people – is becoming marginalized. Their aggregate growth rate was actually negative in the 1990s. (pp. 4–5)

Nevertheless, this growing consensus was shattered by Rodrik and Rodriguez’s (2000) critical survey of the literature. They showed that the findings of the empirical literature are not robust due to the difficulties in measuring openness, statistically sensitive specifications, the collinearity of protectionist policies with other bad policies, and other econometric problems.1 In an empirical study using data since 1870, Vamvakidis (2002) finds no support for a positive growth-openness connection before 1970. In a recent paper, Rodrik et al. (2004) also challenge Frankel and Romer’s (1999) result. Using an instrumental variable estimation technique and a cross-country study, they show that institutions dominate the influence of both trade and geography as the fundamental determinant of long-run economic development. Rigobon and Rodrik (2005) analyse the interrelationship between rule of law, democracy, openness and income. They find that openness negatively impacts on per capita income levels. The conclusion drawn by Rodrik et al. (2004), however, is different from Dollar and Kraay (2003). Dollar and Kraay (2003) argue that crosscountry regressions of the log-level of per capita GDP on instrumented measures of trade and institutional quality are uninformative about the relative importance of trade and institutions in the long run because of the very high correlation between the latter two variables. Using an empirical growth model and panel data, they show that improvements in trade and institutions have positive effects on growth. Given the doubts that these studies have created about the empirical relationship between trade and economic development, further research on this topic is certainly called for. In this chapter we take a fresh look at the empirical relationship between trade and development in a model that also accounts for the effects of institutions on development. Our major contributions are as follows. Unlike previous studies which look at the partial effects of trade and institutions in a linear regression model and hence treat them as competitors in economic development, I join forces with Steve Dowrick and Jane Golley and consider here the potential for complementarities between these two variables (see Bhattacharyya et al., 2009). We do this by introducing an interactive variable in the model which is the product of our measures of institutional quality and trade openness. We observe that the coefficient on the interactive variable is positive and statistically significant, which is indicative of the complementary effects of trade and

M2545 - BHATTACHARYYA PRINT.indd 120

02/03/2011 10:09

Institutions and trade

121

institutions on development. We also observe that in order for a country to benefit from trade, its institutional quality needs to be above a certain threshold level. This is indicative of the fact that to reap the full rewards of trade liberalization, countries should also focus on improving their institutional quality. Using a similar framework, we also test the interrelationship between trade policy openness, institutions and economic development.

6.2 EMPIRICAL STRATEGY To examine the potential for complementarities between institutions and trade share in economic development, we estimate an equation of the form: log y srt 5 ar 1 bt 1 g1TRsrt 1 g2INSsrt 1 g3TRsrt*INSsrt 1 XrsrtL 1 esrt logy (6.1) where log ysrt is the natural logarithm of real GDP per capita purchasing power parity (PPP) in country s in region r averaged over years t24 to t, ar represents regional dummy variables controlling for region-specific timeinvariant unobserved heterogeneity,2 bt represents year dummy variables controlling for time-varying global shocks, TRsrt is the trade share of GDP in country s in region r averaged over years t24 to t, INSsrt is the quality of institutions in country s in region r averaged over years t24 to t, Xsrt is a vector of other control variables, L is a vector of coefficients on additional control variable Xrsrt and e is a random error term. One of the major challenges in obtaining unbiased estimates of the above model is the potential for endogeneity or two-way causality. That is, while it is a possibility that better institutions and an open trade regime cause economic development, it is also possible that economic development triggers improvements in institutional quality and trade openness (see Rodrik et al., 2004). If the latter is true, then a regression based on Equation 6.1 would spuriously attribute a direct effect of trade openness and institutions on income that is really due to income affecting them instead. To tackle this problem, we use the instrumental variable method of estimation. An instrumental variable has to satisfy the twin conditions that it is correlated with the suspected endogenous variables (trade share and institutions) but uncorrelated with the error term in the levels regression.We use log settler mortality, log population density in 1500, fraction of population speaking English (ENGFRAC), fraction of population speaking other European languages (EURFRAC), Frankel and Romer

M2545 - BHATTACHARYYA PRINT.indd 121

02/03/2011 10:09

122

Growth miracles and growth debacles

(1999) constructed openness (CONST), landlocked dummy and land area as instruments. We confirm below through Hausman and overidentification tests that these are appropriate instruments for the task at hand. The point estimate of the direct impact of trade openness on development is (g1 1 g3INSsrt) . We expect the coefficient g3 to be positive if there are complementarities between institutions and trade openness. If g3 is positive then there are two possible scenarios. First, for a positive estimate of g1 the net impact of trade openness on development is always positive given positive values of INSsrt. Second, for a negative estimate of g1 the net impact of trade openness on development will only be positive when the value of INSsrt is above a certain threshold level. In other words, only countries with institutional quality above the threshold level will benefit from trade. We estimate another model to uncover potential complementarities between trade policy and institutions in economic development. The equation that we estimate is as follows: log ysrt 5 ar 1 bt 1 y1posrt 1 y2POsrt 1 y3INSsrt 1 y4POsrt*INSsrt logy 1 XrsrtL 1 xsrt

(6.2)

where posrt is the trade policy openness in country s in region r averaged over years t24 to t, POsrt is the longer term trade policy measured by the fraction of open trade policy years in country s in region r since 1950 until year t and xsrt is a random error term. We also estimate this equation using the instrumental variable method. The interpretation of the partial effects in this case is similar to Equation 6.1 discussed above.

6.3 DATA We use panel data. The master dataset consists of over 209 countries and covers the period 1950 to 2004. However, our preferred models (columns 4 and 12, Table 6.2) cover 58 countries and the time period 1980 to 1995. A list of 58 countries is provided in Table 6.5 and there is a maximum of four data points (1980, 1985, 1990 and 1995) for each country. The major variables that we use in this study are: log GDP per capita, trade share, short-run trade policy openness, long-run trade policy openness and expropriation risk. Table 6.1 presents summary statistics of the major variables and the Data Appendix reports definitions of all variables used here. The natural logarithm of GDP per capita PPP (in current international

M2545 - BHATTACHARYYA PRINT.indd 122

02/03/2011 10:09

Institutions and trade

Table 6.1

123

Summary statistics

Variable

Log GDP per capita (log ysrt) Trade share (TRsrt) Trade policy openness since t24 (posrt) Trade policy openness since 1950 (POsrt) Expropriation risk (INSsrt)

Number Mean Standard of obs deviation (overall)

Standard Standard Minimum Maximum deviation deviation (between) (within)

1684

7.67

1.36

1.06

0.87

4.08

10.87

1425

74.20

43.73

39.58

19.25

1.53

362.53

1406

0.46

0.48

0.33

0.36

0

1

1384

0.31

0.38

0.34

0.18

0

1

445

6.81

2.30

1.84

1.45

1

10

Notes: For a detailed discussion of the definition and source of these variables, see Data Appendix.

dollars) is used as a proxy measure of economic development. According to our sample, the country with the highest per capita income is Luxembourg in 2004 and the country with the lowest per capita income is China in 1950. Measuring trade openness is always difficult. We use two measures of openness in our study. First, the trade share of GDP is used to reflect the degree of a country’s engagement in trade. This has the obvious advantage of being clearly defined and well measured. However, it does not tell us anything about why some countries trade more. For example, large countries have larger internal markets and, therefore, they are more likely to trade less externally. Therefore, the trade share itself does not tell us whether this is the case. In our sample, Hong Kong has the highest trade share and Myanmar has the lowest trade share. Second, a trade policy measure is used to try and capture a range of policies that explain why some countries trade more than others. Sachs and Warner’s (1995a) trade policy index, which runs from 1950 to 1990, is the most well-known attempt to quantify trade openness along these lines. We use this index through to 1990. Wacziarg and Welch (2003) update the Sachs and Warner (1995a) index and extend it to 2000, which we use for the 1990s. The Sachs and Warner index is a dummy variable which classifies a country closed (and hence takes the value 0) if any of the following conditions apply: (1) its average tariff rate on imports of capital or intermediate goods is above 40 per cent; (2) its non tariff barriers cover 40 per cent or more of its imports of capital and intermediate goods; (3) its black

M2545 - BHATTACHARYYA PRINT.indd 123

02/03/2011 10:09

124

Growth miracles and growth debacles

market premium is 20 per cent or more; (4) it has a socialist economic system; or (5) it has a state monopoly on major exports. Rodrik and Rodriguez (2000) provide the most well-known critique of Sachs and Warner’s openness index. They show that the index suffers from measurement problems and is also correlated with other non-trade-related bad policies, which make any econometric estimation of its effect on economic development unreliable. However, several recent studies including one by Rodrik use this index (see Hausmann et al., 2005; Giavazzi and Tabellini, 2005 and many others).3 We use the Sachs and Warner index because of its wide coverage (both cross-section and time series) and easy availability. We also use average tariff, black market premium and real exchange rate distortions as proxy measures of trade policy. Our results are not robust to these measures. This is hardly surprising as these measures are only imperfect proxies of trade policy openness (see Wacziarg, 2001; Dollar and Kraay, 2003). They also have a very limited country and time coverage. In comparison, the Sachs and Warner index appears to be superior in capturing trade policy openness (Wacziarg, 2001). Also note that Wacziarg and Welch (2003) accounted for some of the old weaknesses in the index and significantly improved it in the updated version – the version that we use. Using the Sachs and Warner index we calculate two indicators of trade policy openness indicator. First, is an indicator of short-run policy openness which is constructed by dividing the number of years of trade policy openness between t24 and t by 5. Second, is an indicator of long-run policy openness which is constructed by dividing the number of years of trade policy openness between 1950 and t by t21950. The first measure is expected to be endogenous so we use instrumental variables in our estimation. In contrast, we treat the second measure as exogenous, as one would expect long-term trade policy going back to the 1950s not to be influenced by the current level of GDP per capita. Finally, we use the Political Risk Services index of expropriation risk as our measure of institutional quality. The measure ranges from zero to ten where higher values indicate a lower probability of expropriation of private property by the state. There are other measures of institutions (such as rule of law, repudiation of contracts, executive constraints, corruption and democracy) used in the literature. However, none of these measures is close to Douglass North’s notion of good institutions. North (1981) defines good institutions as those that provide checks against expropriation by the government and other politically powerful groups (see North, 1981, pp. 20–27). Expropriation risk is perhaps the closest to North’s definition as it captures the notion of an extractive state. Hence we use expropriation risk as our measure of institutions. Nevertheless, I

M2545 - BHATTACHARYYA PRINT.indd 124

02/03/2011 10:09

Institutions and trade

125

will also present the empirical estimates of the role of different types of institutions (rule of law, repudiation of contracts, executive constraints, corruption and democracy) on development.

6.4 EVIDENCE This section systematically tests whether institutions and trade are complements in economic development. We first provide the basic results and then conduct some robustness tests. 6.4.1

Basic Results

Table 6.2 presents the basic results. In column 1 we look solely at the partial relationship between contemporaneous trade shares and log GDP per capita. We observe that a one sample standard deviation (43.7 per cent) increase in the trade share of an average country results in a 1.2 fold increase in per capita GDP.4 This is undoubtedly a large effect and is statistically significant. However, the coefficient estimate of this model is a suspect of omitted variable bias. Trade may be correlated with other factors and with institutions, in particular, which also influence income, in which case our estimate will show an inflated effect of trade. In order to tackle this issue, we add institutions in column 2. We observe that both trade and institutions have positive effects on income and are statistically significant. The effect of trade on income diminishes from the unconditional estimate of column 1. However, there are still issues of reverse causality that this model does not address. In column 3 we estimate the model in column 2 using the IV method to tackle this problem. Our use of the IV method in this case is statistically valid on the grounds that institutions and trade share variables fail the Hausman test of exogeneity and the instruments pass the overidentification (OID) test of exogeneity. Note that institutional quality is the only variable that is statistically significant, while the coefficient of trade share is insignificantly different from zero. This confirms previous findings of Rodrik et al. (2004) that institutional quality is the dominant explanator of long-run variations in economic development. The problem with this specification and the specifications adopted by previous studies is that it implicitly assumes institutions and trade are competitors in economic development. In column 4 we attempt to address this issue by introducing an interactive term which is the scalar product of each country’s trade share and institutional quality to capture any possible complementarities between the two. This is our preferred model with trade

M2545 - BHATTACHARYYA PRINT.indd 125

02/03/2011 10:09

126

M2545 - BHATTACHARYYA PRINT.indd 126

02/03/2011 10:09

YES YES

181

Controls Region dummies Year dummies

Countries

POsrt*INSsrt

posrt*INSsrt

TRsrt*INSsrt

(2)

(3)

(4)

117

YES YES

0.22*** (0.0222)

58

YES YES

0.78*** (0.1198)

58

YES YES

0.43** (0.2171) 0.008** (0.0037)

0.005*** 0.003*** 0.002 −0.057** (0.0005) (0.0008) (0.0026) (0.0285)

(1)

138

YES YES

1.27*** (0.1062)

(5)

(7)

106

YES YES

0.21*** (0.0214)

106

YES YES

0.13*** (0.0374)

0.15*** (0.0266)

−0.02 −0.88*** (0.0908) (0.2602)

(6)

(8)

138

YES YES

1.79*** (0.1299)

Log GDP per capita (log ysrt) (9)

(10)

(12)

0.04 0.68 (0.1039) (0.9398)

(11)

106

YES YES

0.21*** (0.0202)

106

YES YES

0.17*** (0.0501)

0.17*** (0.0228)

106

YES YES

58

YES YES

0.18*** 1.11** (0.0507) (0.5064)

0.17*** 0.35* (0.0245) (0.2054)

−0.03 −1.22*** −1.27*** −8.2** (0.1385) (0.3740) (0.3965) (3.608)

Institutions and trade: competitors or complements in economic development

Trade share (TRsrt) Trade policy openness since t24 ( posrt) Trade policy openness since 1950 (POsrt) Expropriation risk (INSsrt)

Dependent variable

Table 6.2

127

M2545 - BHATTACHARYYA PRINT.indd 127

02/03/2011 10:09

1334 0.9928

393 0.9950

216 0.000 0.92

216 0.000 0.21

1258 0.9652

374 0.9955

374 0.9956

1246 0.9666

374 0.9955

374 0.9956

374 0.9956

0.000 0.31

219

Notes: ***, **, and * indicate significance levels at 1%, 5%, and 10%, respectively, against a two-sided alternative. Figures in parentheses are the respective standard errors. Columns 3, 4, and 12 report instrumental variable (IV) estimates. The instruments used are log settler mortality, log population density in 1500, fraction of population speaking English (ENGFRAC), fraction of population speaking other European languages (EURFRAC), Frankel and Romer (1999) constructed openness (CONST), landlocked dummy and land area. The sample years are every fifth year from 1950 to 2004. All variables are five-year averages except POsrt, which shows policy openness since 1950. Regressions involving INSsrt cover the period 1980 to 1995 because of data limitations.

Observations Adjusted R2 Hausman test OID test

128

Growth miracles and growth debacles

share as the measure of trade.5 We notice that the coefficient estimate of the trade share is negative and the coefficient estimate of the interactive term is positive. Both coefficients are statistically significant. The positive coefficient on the interactive term indicates that institutions and trade are complements in economic development. The negative and positive signs on the coefficient estimates of the trade share and the interactive term, respectively, are indicative of the fact that the relationship between trade and development is non-linear. There is a threshold level of institutional quality for an average country beyond which the partial effect of trade on development is positive. For countries with institutions below this threshold, the partial effect of trade on development is in fact negative. Our model predicts that the threshold level of institutional quality is 7.1 which sits well within the sample range of 1 and 10.6 The point estimate suggests that a one sample standard deviation increase in trade share in a country with an average institutional quality of 7.2 over the period 1980 to 1995 will lead to a 1.03 fold increase in per capita GDP. To put this into perspective, the model explains 2.2 fold of the tenfold per capita GDP difference between India (with a trade share of 21.78 and an institutional quality of 9.9) and the UK (with a trade share of 55.9 and an institutional quality of 10) in 1995. The model predicts a positive impact of trade on development in the United States, Canada, New Zealand, Australia, Hong Kong, Gambia, Malaysia, India, Brazil, Papua New Guinea, Chile, South Africa, Ivory Coast and Mexico in a sample of 58 former colonies. Gabon and Indonesia are predicted to have small negative effects and Haiti, Mali, Sudan and Zaire register large negative effects. Table 6.5 and Figure 6.1 show predicted values for all 58 countries (see column TRhat). In columns 5–12 we explore the relationship between trade policy openness and economic development. Point estimates reported in columns 5–7 are unreliable due to endogeneity. To tackle the endogeneity problems associated with the trade policy measure, we adopt two strategies. First, we use a long-run trade policy measure (fraction of openness years since 1950) instead of a contemporaneous trade policy measure (fraction of openness years since t 2 4). We assume the long-run trade policy measure to be exogenous. Second, we estimate our model using the instrumental variable method. In column 8 we look at the unconditional correlation between long-run trade policy openness and economic development. The point estimate is positive and statistically significant. In column 9 we add institutions into the specification and we observe that the long-run trade policy openness variable is no longer statistically significant. In column 10 we note that the coefficient estimate of the interactive term is positive and statistically significant. In column 11 we add the contemporaneous

M2545 - BHATTACHARYYA PRINT.indd 128

02/03/2011 10:09

129

M2545 - BHATTACHARYYA PRINT.indd 129

02/03/2011 10:09

0

.02

ZAR

4

HTI MLI SDN

NGA BOL HND BGD COG NIC SLV GTM BFA MDG UGA

INS

IND BRA PNG CHL

7.7

ZAF CIV MEX GAB IDN TTO COL VEN ECU PRY TZA CRI JAM GIN URY DZA CMR EGY LKA KEN DOM TUN PER TGO MAR SLE GHA PAN SEN ARG AGO PAK NER

Partial effect of trade on development

See Table 8.2 for country abbreviations.

–.04

Figure 6.1

Note:

Partial effect of TR on log y

GMB MYS

HKG

AUS

10

CAN NZL

US A

130

Growth miracles and growth debacles

trade policy openness variable. The model suggests that the long-run trade policy openness matters and the contemporaneous trade policy openness variable is no longer significant. This model may suffer from endogeneity. This is also revealed by the Hausman test of exogeneity of explanatory variables reported in column 12. The test rejects the null of exogeneity of the explanatory variables. Therefore, we estimate the model using the instrumental variable method in column 12. The estimates predict that the threshold level of institutional quality required for a positive partial effect of long-run trade policy openness on per capita GDP is 7.4.7 A one standard deviation increase in long-run trade policy openness in a country with institutional quality 7.5 will result in a 1.1 fold increase in its per capita income. In this case the model predicts a 2.8 fold difference in per capita income between India (open trade regime for 2 out of 45 years) and the UK (always open) in 1995, which is also less than the actual difference of tenfold. South Africa, Ivory Coast and Mexico are omissions from the list of countries with a positive partial effect when we use the long-run trade policy measure of openness. Table 6.5 and Figure 6.2 report predicted values for all 58 countries (see column POhat). In Table 6.3 we report the first stage regressions of the IV estimates reported in columns 4 and 12 of Table 6.2. The instruments that we use for the IV estimation are valid as they are correlated with the suspected endogenous variables and also exogenous to the model. We perform an overidentification test to check the exogeneity of the instruments. The test p-values are reported in columns 4 and 12 of Table 6.2. The high p-values indicate that we fail to reject the null of exogeneity of the instruments. Countries with higher constructed openness (CONST) are likely to trade more (see column 1, Table 6.3 and also Frankel and Romer, 1999, for detailed explanation). Furthermore, countries with a higher density of population generally trade less externally as there is a larger market for internal trade (note negative coefficient on LPOPDEN). A similar explanation may apply for area. The negative partial correlation between EURFRAC and trade share may be driven by low trade share in Argentina, the United States and Uruguay. In column 2 we report the first stage of the institutional quality variable. Countries with high European settler mortality in the past inherited extractive institutions from their colonizers and hence the negative relationship between log settler mortality and institutional quality (see Acemoglu et al., 2001). Furthermore, countries with a high proportion of population speaking a Western European language (high EURFRAC) are observed to have strong institutions (see Hall and Jones, 1999). The major partial correlations reported in columns 3, 4 and 5 can also be explained using the abovementioned arguments.

M2545 - BHATTACHARYYA PRINT.indd 130

02/03/2011 10:09

131

M2545 - BHATTACHARYYA PRINT.indd 131

02/03/2011 10:09

0

4.6

ZAR

4

HTI MLI SDN

NGA BOL HND BGD COG NIC SLV GTM BFA MDG UGA

INS

7.3

ZAF CIV MEX GAB IDN TTO COL VEN ECU PRY TZA CRI JAM GIN URY DZA CMR EGY LKA KEN DOM TUN PER TGO MAR SLE GHA PAN SEN ARG AGO PAK NER

Partial effect of long-run trade policy on development

See Table 8.2 for country abbreviations.

–6.2

Figure 6.2

Note:

Partial effect of PO on log y

IND BRA PNG CHL

GMB MYS

HKG

AUS

10

CAN NZL

USA

132

Table 6.3

Growth miracles and growth debacles

Determinants of institutions and trade: first stage for the core specifications

Dependent variables:

Constructed openness (CONST) Landlocked dummy Log settler mortality (LSM) Log population density in 1500 (LPOPDEN) ENGFRAC EURFRAC AREA

Trade Expropriation TRsrt*INSsrt share risk (INSsrt) (TRsrt)

Trade POsrt*INSsrt policy openness since t24 (posrt)

(1)

(2)

(3)

(4)

(5)

2.05*** (0.2382)

0.015 (0.0115)

16.1*** (1.973)

0.0002 (0.0012)

0.01** (0.0044)

−7.67 (6.989) −1.04 (2.555)

−0.53 (0.3379) −0.21* (0.1235)

−74.6 (57.88) −29.9 (21.16)

0.05 (0.0326) 0.008 (0.0155)

−0.25* (0.1393) −0.03 (0.0539)

−6.66*** (2.052)

−0.13 (0.0992)

−66.4*** (16.99)

−0.02* (0.0106)

−0.15*** (0.0414)

4.48 (9.948) −23.2** (11.07) −0.004*** (0.0016)

0.37 (0.4809) 0.98* (0.5354) 0.0001 (0.0001)

11.5 (82.38) −163.7* (91.70) −0.03** (0.0131)

−0.01 (0.0611) 0.10 (0.0632) 0.000005 (0.000009) 0.88*** (0.0389)

0.12 (0.1966) 0.02 (0.2204) 0.00002 (0.00003) 6.92*** (0.1710)

YES

YES

YES

YES

YES

YES

YES

YES

YES

YES

59 219 0.5365

59 219 0.6100

59 219 0.5820

66 705 0.6742

60 226 0.9535

Trade policy openness since 1950 (POsrt) Controls Region dummies Year dummies Countries Observations Adjusted R2

Notes: ***, **, and * indicate significance levels at 1%, 5%, and 10%, respectively, against a two-sided alternative. Figures in parentheses are the respective standard errors. The variables used as instruments are correlated with the suspected endogenous variables which make them valid instruments.

M2545 - BHATTACHARYYA PRINT.indd 132

02/03/2011 10:09

Institutions and trade

Table 6.4

133

How much does within-country variation matter?

Dependent variable: Trade share (TRsrt) Trade policy openness since t24 (posrt) Trade policy openness since 1950 (POsrt) Expropriation risk (INSsrt) TRsrt*INSsrt

Log GDP per capita (log ysrt) (1)

(2)

0.002 (0.0013)

0.0004 (0.0129)

(3)

−0.42*** (0.1116)

−0.009 (0.0122) −0.00003 (0.0001)

0.42*** (0.0971) 0.0003 (0.0018)

0.01 (0.0136)

−0.16*** −0.41 (0.0481) (0.6185)

0.016 (0.0129)

0.34*** (0.1187)

0.05** (0.0230)

1.13** (0.4760) 2.12*** (0.6060)

0.05*** (0.0159)

POsrt*INSsrt Latitude

Countries Observations Adjusted R2

(5)

0.83*** −7.6** (0.2843) (3.423)

posrt*INSsrt

Controls Country dummies Region dummies Year dummies

(4)

2.31*** (0.4664) YES NO YES

NO YES YES

YES NO YES

YES NO YES

NO YES YES

117 393 0.9997

28 108

106 374 0.9995

106 374 0.9995

29 112

Notes: ***, **, and * indicate significance levels at 1%, 5%, and 10%, respectively, against a two-sided alternative. Figures in parentheses are the respective standard errors. Columns 2 and 5 report instrumental variable (IV) estimates. The instruments used are log settler mortality, log population density in 1500, fraction of population speaking English (ENGFRAC), fraction of population speaking other European languages (EURFRAC), Frankel and Romer (1999) constructed openness (CONST), landlocked dummy and land area.

In Table 6.4 we explore in greater detail where the complementarities between trade and institutions are coming from by looking at the source of identification of the complementarities. If most of the identification is due to cross-sectional differences between countries that are permanent in nature then we will not find anything in fixed effect regressions. However,

M2545 - BHATTACHARYYA PRINT.indd 133

02/03/2011 10:09

134

Table 6.5

Growth miracles and growth debacles

Partial effects of trade on development given institutional quality

ISO code

Country

INS

TRhat

POhat

USA CAN NZL AUS HKG GMB MYS IND BRA PNG CHL ZAF CIV MEX GAB IDN COL TTO VEN ECU PRY TZA CRI JAM GIN URY DZA CMR EGY LKA KEN DOM TUN PER TGO MAR SLE GHA PAN SEN ARG

United States Canada New Zealand Australia Hong Kong Gambia, The Malaysia India Brazil Papua New Guinea Chile South Africa Côte d’Ivoire Mexico Gabon Indonesia Colombia Trinidad and Tobago Venezuela Ecuador Paraguay Tanzania Costa Rica Jamaica Guinea Uruguay Algeria Cameroon Egypt Sri Lanka Kenya Dominican Republic Tunisia Peru Togo Morocco Sierra Leone Ghana Panama Senegal Argentina

10 9.7 9.7 9.4 8.6 8.3 8.3 7.9 7.8 7.7 7.6 7.3 7.2 7.2 7.1 7.1 7 7 6.9 6.8 6.8 6.8 6.7 6.7 6.6 6.6 6.5 6.3 6.3 6.3 6.2 6.1 6.1 6 6 5.9 5.9 5.8 5.8 5.8 5.7

0.023 0.0206 0.0206 0.0182 0.0118 0.0094 0.0094 0.0062 0.0054 0.0046 0.0038 0.0014 0.0006 0.0006 −0.0002 −0.0002 −0.001 −0.001 −0.0018 −0.0026 −0.0026 −0.0026 −0.0034 −0.0034 −0.0042 −0.0042 −0.005 −0.0066 −0.0066 −0.0066 −0.0074 −0.0082 −0.0082 −0.009 −0.009 −0.0098 −0.0098 −0.0106 −0.0106 −0.0106 −0.0114

2.9 2.567 2.567 2.234 1.346 1.013 1.013 0.569 0.458 0.347 0.236 −0.097 −0.208 −0.208 −0.319 −0.319 −0.43 −0.43 −0.541 −0.652 −0.652 −0.652 −0.763 −0.763 −0.874 −0.874 −0.985 −1.207 −1.207 −1.207 −1.318 −1.429 −1.429 −1.54 −1.54 −1.651 −1.651 −1.762 −1.762 −1.762 −1.873

M2545 - BHATTACHARYYA PRINT.indd 134

02/03/2011 10:09

Institutions and trade

ISO code

Country

PAK NER NGA BOL HND BGD COG NIC SLV GTM BFA MDG UGA HTI MLI SDN ZAR

Pakistan Niger Nigeria Bolivia Honduras Bangladesh Congo, Republic of Nicaragua El Salvador Guatemala Burkina Faso Madagascar Uganda Haiti Mali Sudan Congo, Democratic Republic

135

INS

TRhat

POhat

5.6 5.5 5.2 5.1 5.1 5 5 5 4.9 4.8 4.7 4.6 4.6 4.1 4 3.9 3.6

−0.0122 −0.013 −0.0154 −0.0162 −0.0162 −0.017 −0.017 −0.017 −0.0178 −0.0186 −0.0194 −0.0202 −0.0202 −0.0242 −0.025 −0.0258 −0.0282

−1.984 −2.095 −2.428 −2.539 −2.539 −2.65 −2.65 −2.65 −2.761 −2.872 −2.983 −3.094 −3.094 −3.649 −3.76 −3.871 −4.204

Notes: TRhats is the partial effect (20.057 1 0.008 3 INSs) of a percentage point increase in trade share in country s where INSs is the average expropriation risk over the period 1980 to 1995. Similarly, POhati is the partial effect (28.2 1 1.11 3 INSs) of a percentage point increase in PO in country s.

it could also be because of some common omitted factors such as culture or geography driving both the complementarities effect and per capita income. We try to explore this by introducing country fixed effects into our preferred models. Column 1 of Table 6.4 reports a regression involving the trade share when country fixed effects are introduced. We observe that the interactive term TRsrt*INSsrt is no longer statistically significant which may indicate that permanent cross-country differences are driving the complementarities effect and within-country differences over time do not seem to matter that much. Geography can be a possible source of identification. In column 2 we test this by replacing country fixed effects with latitude. We notice that the coefficient on latitude is positive and statistically significant and the interactive term is no longer significant. This is perhaps indicative of the fact that geography is driving both the differences in living standards and complementarities between ‘trade and institutions’ over the very long run. This is consistent with previous findings in the literature that geography and disease environment shape the long-run evolution of institutions and trade (see Gallup et al., 1998; Acemoglu et al., 2001;

M2545 - BHATTACHARYYA PRINT.indd 135

02/03/2011 10:09

136

Growth miracles and growth debacles

Rodrik et al., 2004). Column 3 checks the impact of country fixed effects on the complementarities between contemporaneous trade policy openness and institutions. We find that this effect survives even in the presence of country fixed effects since the coefficient on posrt*INSsrt is positive and statistically significant. This holds even when we add long-run trade policy openness. In column 4 we see that the coefficient on POsrt*INSsrt is positive and statistically significant. In column 5 we check how much of this effect is due to permanent cross-country differences such as geography or culture and how much is due to within-country differences. We do this by replacing country fixed effects with latitude. We observe that the complementarities effect survives, which indicates that some part of this effect is coming from the within-country variation. The fundamental difference in the results between the trade share openness model and the trade policy openness models with the inclusion of country fixed effects and latitude could be because of the difference in the construction of these two measures. Trade shares perhaps reflect a part of openness which is deeper and largely time-invariant8 whereas trade policy is more, influenced by short-term changes in the policy environment. However, we concede that in the absence of truly exogenous variation, our analysis does not fully resolve all the identification issues. 6.4.2

Robustness

We check the robustness of our basic results by using additional control variables (schooling, investment, foreign aid, ethnic fractionalization, black market premium and mining share of GDP) that are reported by previous studies to be correlated with development. We also check the robustness of our basic results across different continental sub-samples. Our results are reasonably robust with both the trade share and the longrun trade policy as measures of openness. Results are reported in Tables 6.6 and 6.7.

6.5 SUMMARY The results suggest that there is a reasonably robust correlation between trade openness and economic development when the complementarities between institutions and trade are taken into account. This is done by introducing an interactive variable which is the scalar product of institutional quality and trade into the model. The coefficient on the interactive variable is positive and statistically significant which is indicative of the complementary effects of trade and institutions on development.

M2545 - BHATTACHARYYA PRINT.indd 136

02/03/2011 10:09

Institutions and trade

Table 6.6

Robustness check: trade share

Dependent variable Trade share (TRsrt) Expropriation risk (INSsrt)

TRsrt*INSsrt

Log GDP per capita (log ysrt) (1)

(2)

(3)

(4)

−0.04 (0.0301) 0.48* (0.2852) 0.01** (0.0039)

−0.08* (0.0499) 0.90** (0.4149) 0.01* (0.0066)

−0.071** (0.0309) 0.25 (0.2994) 0.01** (0.0040)

−0.074* (0.0452) 0.78** (0.3314) 0.01* (0.0059)

YES YES

YES YES

YES YES

YES YES

Controls Region dummies Year dummies Additional controls

137

(5)

(6)

−0.09* −0.08* (0.0583) (0.0478) 0.82** 0.69** (0.4262) (0.3612) 0.01* 0.01* (0.0074) (0.0063) YES YES

YES YES

Schooling Investment Foreign Ethnic Black Mining aid fractional- market ization premium

Countries Observations

50 189

58 216

55 204

59 219

55 184

59 219

Notes: ***, **, and * indicate significance levels at 1%, 5%, and 10%, respectively, against a two-sided alternative. Figures in parentheses are the respective standard errors. All regressions are estimated using the instrumental variable estimation method. The instruments used are log settler mortality, log population density in 1500, fraction of population speaking English (ENGFRAC), fraction of population speaking other European languages (EURFRAC), Frankel and Romer (1999) constructed openness (CONST), landlocked dummy and land area.

However, questions regarding identification remain due to the absence of truly exogenous variation in macro trade data. The co-movement in trade institutions and development may be due to other factors (such as culture, geography or both) which are driving all three of them. Still, the fact that the results are robust to the inclusion of schooling, investment, foreign aid, ethnic fractionalization, black market premium and share of mining to GDP is encouraging enough for an interpretation of this effect as the combined impact of trade and institutions on economic development. It is also observed that in order for the partial effect of trade on development to be positive, a country’s institutional quality has to be above a threshold level. This is indicative of the fact that trade alone may not promote development. Institutions play a crucial role.

M2545 - BHATTACHARYYA PRINT.indd 137

02/03/2011 10:09

138

Table 6.7

Growth miracles and growth debacles

Robustness check: trade policy

Dependent variable

Log GDP per capita (log ysrt)

Trade policy openness since

(1)

(2)

(3)

(4)

(5)

(6)

−0.37 (0.8357)

1.73 (2.402)

0.21 (1.326)

0.72 (1.434)

1.04 (1.964)

1.07 (1.521)

−9.1*** (3.458)

−23.3* (12.72)

−16.4*** (4.974)

−13.1* (7.788)

−17.8** −13.2** (6.951) (5.802)

0.34 (0.2676) 1.36*** (0.4814)

0.51 (0.4754) 3.3* (1.773)

0.10 (0.3078) 2.43*** (0.7132)

0.57 (0.3812) 1.8* (1.076)

0.29 0.53 (0.4641) (0.3240) 2.51** 1.8** (1.004) (0.8010)

YES YES

YES YES

YES YES

YES YES

t24 ( posrt) Trade policy openness since 1950 (POsrt) Expropriation risk (INSsrt)

POsrt*INSsrt Controls Region dummies Year dummies Additional controls

Schooling Investment

Countries Observations

50 190

58 217

Foreign aid 55 205

YES YES

YES YES

Ethnic Black Mining fractional- market ization premium 59 220

55 182

59 220

Notes: ***, **, and * indicate significance levels at 1%, 5%, and 10%, respectively, against a two-sided alternative. Figures in parentheses are the respective standard errors. All regressions are estimated using the instrumental variable estimation method. The instruments used are log settler mortality, log population density in 1500, fraction of population speaking English (ENGFRAC), fraction of population speaking other European languages (EURFRAC), Frankel and Romer (1999) constructed openness (CONST), landlocked dummy and land area.

NOTES 1. Lee et al. (2004) answer Rodrik and Rodriguez (2000). They show that trade has a small but positive effect on growth when the reverse causality issues are handled appropriately. 2. The region dummies cover Europe and Central Asia, East Asia and the Pacific, Latin America, the Middle East and North Africa, South Asia and sub-Saharan Africa. Timeinvariant factors such as geography and culture are often region specific. 3. Also see Warner (2003) for a reply to Rodrik and Rodriguez (2000). 4. The partial effect of a one sample standard deviation increase in trade share is Dlog y 55log y11 2 y00 5 Dlogy logy 2log logy 5 log loga

M2545 - BHATTACHARYYA PRINT.indd 138

y1 y0

b 5 0.005 3 43.7 < 0.22,

02/03/2011 10:09

Institutions and trade

5. 6.

7.

8.

139

which can also be expressed as y1 5 e0.22y0 < 1.2y0 where y1 and y0 are final and initial incomes, respectively. We also estimate our preferred model using log trade share and the results are qualitatively similar. The partial effect of trade on development is given by 20.057 1 0.008INST, where INST is the threshold level of institutional quality. The threshold is calculated by equating the partial effect to zero and solving for INST. We perform an F-test to check the statistical significance of the threshold estimate and we fail to reject the null (p-value 5 0.55) of H0: 2g1 /g3 5 7.1 against a two-sided alternative. We also reject the H0: g1 1 g3 5 0 (pvalue 5 0.048). The coefficient estimate is 20.05 (se: 0.0248) suggesting that an average country in the sample is not benefiting from trade. We perform an F-test to check the statistical significance of the threshold estimate and we fail to reject the null (p-value 5 0.97) of H0: 2y2 /y4 5 7.4 against a two-sided alternative. We also reject the H0: y2 1 y4 5 0 (p-value 5 0.023). The coefficient estimate is 27.1 (se: 3.114) suggesting that an average country in the sample is not benefiting from trade policy openness. See Frankel and Romer (1999) for a discussion on this; they argue that an increase in trade volume reflects a reduction in natural or geographical barriers to trade.

M2545 - BHATTACHARYYA PRINT.indd 139

02/03/2011 10:09

7.

Improving institutions with trade policy: myth or a possibility

7.1 INSTITUTIONS AND TRADE POLICY As I described in Part I of this book, a growing number of economists argue that institutional divergence across countries can be best explained by colonial history. In brief, the story is as follows. The Europeans resorted to different styles of colonization in different parts of the world depending on the feasibility of settlement. In tropical climates, the mortality rates among European colonizers were extremely high which prevented them from settling there and they erected extractive institutions characterized by weakly defined property rights and weak enforcement of contracts. However, in temperate climates, the mortality rates among colonizers were low, which made them ideal for settlement and they erected strong institutions characterized by well-defined property rights and enforcement of private contracts. These institutions persisted over time and they continue to influence the current institutional and economic performance of these countries. Therefore, colonial history shaped institutions in most countries around the globe and perhaps policy has very little or no role to play. But the fact that we occasionally do notice improvements in institutional quality due to good policy suggests that good institutions are not entirely determined by history. This is an important policy question for growth given that institutions are a crucial determinant. A good illustration of this is perhaps post-independence India.1 India inherited relatively good institutions from the British in 1947. A democratic polity, an independent judiciary and secure property rights were among many other positives that were enshrined into the constitution of independent India. High scores in executive constraint (consistently around 7) and democracy (consistently around 9) from Polity data during the 1950s are indicative of the fact that the institutions were strong, at least on paper. India also embarked on an import substituting industrialization policy during this time, relying on high tariffs and quantitative restrictions to prevent imports. This led to the well-known problem of a ‘rent-seeking society’ (see Bhagwati and Desai, 1970; Krueger, 1974). As a consequence of the widespread culture 140

M2545 - BHATTACHARYYA PRINT.indd 140

02/03/2011 10:09

Improving institutions with trade policy

141

of rent-seeking and bad policy, good institutions, on paper, often yielded poor institutional outcomes over the next three decades. A quick comparison of the Polity score (which measures institutions on paper) and the ICRG score (which measures institutional outcome) during 1980 to 2000 supports this view.2 The Polity executive constraint index remained consistently at 7 throughout the period. In contrast the ICRG expropriation risk index was as low as 6 in 1982 when the import licensing system was fully operational and it became as high as 10 in 1993 and thereafter when India liberalized its economy.3 In this chapter, I explore this topic in more detail. I start with a discussion of some of the existing theories linking trade and institutional development. This is followed by evidence from both cross-section and panel data. The chapter concludes with a discussion of what this may mean for the policymakers.

7.2 THEORIES OF TRADE AND INSTITUTIONAL DEVELOPMENT The theoretical literature on institutional change over time perhaps originates from North (1981). North (1981) emphasizes that a change in per capita capital stock due to population growth and technological progress brings about institutional change over time. What North does not mention is the impact of international trade on population and technological progress. International trade increases the size of the market which is equivalent to an increase in the size of domestic population (see Smith, 1776 [1976]). It is also a widely accepted view that trade induces technological progress via technology transfer (see Romer, 1990; Coe and Helpman, 1995). Therefore, potentially, engagement in international trade can bring about institutional change in a country. In related research, Rogowski (1989) also shows that trade affects domestic political alignments through changes in factor prices. He, however, does not focus on the impact of trade on institutions. Acemoglu et al. (2005b) document historical evidence in favour of the trade induced institutional change view. Their hypothesis, however, is different from North’s (1981) capital stock theory as they focus on trade’s impact on the distribution of political power and subsequent institutional change. They show that Western Europe’s engagement in Atlantic trade induced institutional change by strengthening commercial interests which resulted in rapid economic growth in countries where the initial political institutions were non-absolutist. In a related study, Acemoglu and Robinson (2006) show that trade

M2545 - BHATTACHARYYA PRINT.indd 141

02/03/2011 10:09

142

Growth miracles and growth debacles

induced transfer of skill-biased technology increases the income share of the middle class. This increases their political power relative to the rest of society and they impose checks and balances on the existing institutions to protect their property rights and contracts. With a larger share of income, the powerful middle class also favours taxation, which is less redistributive. This makes the elite more willing to accept the checks and balances on institutions imposed by the middle class. Without doubt, there is enough theory floating around to believe that trade liberalization may have an impact on institutions. The key issue, however, is how much of it is supported by the data. I report some empirical results supporting the link in Section 7.3.

7.3 EVIDENCE A simple plot of the data in Figure 7.1 shows that indeed there is a positive relationship between long-run trade policy openness and institutional quality. Long-run trade policy openness is measured in the same manner as in Chapter 6. It is the fraction of years a country has been open since 1950. 2

CHE FIN DNK NOR SWE AUT AUS NLD GBR CAN DEU IRL SGP USA

NZL

BEL JPN FRA HKG PRT ESP TWN MUS ITA GRC

Fitted values/Rule of law

CHL SVN ISR HUN EST CZE CRI POL URY LTU LVA NAM SVK MNG TTO TUN LKA ZAF HRV MAR EGY IND BGR TUR LSO ROM GHA MDG SEN CHN MEX GAB BRA MRT MWI VNM MKD SYR BEN PER DOM ARM ETH SLV MDA TZA NPL PHL GMB ERI ZMB DZA MLI BFA IRN NIC MOZ TGO PAK ARG COL GIN BGD NER RUS AZE HND UKR PNG KGZ GTM UGA CAF KAZ ALB TCD CUB YUG GNB RWA LAO VENKEN BLR PRY TKM UZB GEO CIV COG SLE TJK CMR ZWE NGA SDN LBR BDI AGO BUR HTI ZAR

1

0

–1

–2

KOR BWA

MYS JOR THA

JAM ECU

BOL IDN

SOM

0

0.2

0.4 0.6 Trade openness

0.8

1

Note: See Table 8.2 for country abbreviations.

Figure 7.1

Long-run trade policy openness and institutions

M2545 - BHATTACHARYYA PRINT.indd 142

02/03/2011 10:09

Improving institutions with trade policy

Table 7.1

143

Summary statistics

Variable Rule of law Trade policy openness Total years of schooling Gini decadal average Log GDP per capita in 1960

Number of obs

Mean

Standard deviation

175 137

−0.002 0.255

1.002 0.399

90

3.489

2.531

111 115

43.24 6.32

10.65 0.865

Minimum Maximum −2.05 0 0.074 21.46 4.72

2.03 1 9.56 64.29 8.23

Hence, by construction, the value of this variable varies between 0 and 1. Institutional quality, on the other hand, is measured by the rule of law index. Table 7.1 reports summary statistics of all the important variables. To explore further, I also estimate a model regressing long-run trade policy openness on institutional quality measured by rule of law. But there are, of course, technical challenges to appropriately estimate the effect of trade policy on institutional quality. It is possible that institutional quality influences trade policy rather than the other way round. This is commonly known as the problem of reverse causation. This would lead to inaccurate estimates from the model. In order to tackle reverse causation, I also estimate the model using the instrumental variable method. Instrumental variables are exogenous to the model but correlated to trade policy. Therefore, this allows us estimate the effect reasonably accurately. Frankel and Romer constructed openness is one such instrumental variable. A detailed discussion on this variable is given in Chapter 4. The results are reported in Tables 7.2 and 7.3. We notice that one standard deviation increase in trade policy openness would lead to one standard deviation increase in institutional quality. This implies that if Tanzania were to become open then her institutional quality would improve by one-quarter of the institutional quality of Botswana. This result also holds when I consider a sample of former colonies. In a related study using panel data covering the periods 1865–1940 and 1980–2000, and 31 and 103 countries, respectively, I also find that the variation in economic institutions (namely property rights and contracts) can be explained by trade liberalization (Bhattacharyya, 2010). To address the reverse causality concern I use export partner growth and rainfall as instruments for trade liberalization and log GDP per capita, respectively, and estimate our model using the LIML Fuller estimation method. I find

M2545 - BHATTACHARYYA PRINT.indd 143

02/03/2011 10:09

144

Table 7.2

Growth miracles and growth debacles

Trade policy and institutions Dependent variable: rule of law Full sample (obs 5 72) OLS estimate

Trade policy openness Other controls Total years of schooling Gini decadal average Log GDP per capita in 1960 R2 F-test (p-value) Hausman test for endogeneity (p-value)

1.05*** (0.1784) 0.18*** (0.0449) −0.003 (0.0072) 0.12 (0.1214) 0.7974 0.0000***

Full sample (obs 5 72) 2SLS estimate

AJR sample AJR sample without Neo(obs 5 42) Europe (obs 5 38) 2SLS 2SLS estimate estimate

2.05*** (0.5753)

2.93** (1.302)

2.87** (1.219)

0.091 (0.0716) 0.013 (0.0122) 0.11 (0.1474)

−0.039 (0.1798) 0.012 (0.0179) 0.44 (0.3857)

−0.076 (0.2049) 0.001 (0.0170) 0.52 (0.3901)

0.0000*** 0.022**

0.0001***

0.013***

Notes: ***, ** and * indicate significance levels of 1%, 5% and 10%, respectively, against a two-sided alternative. Standard errors are reported in parentheses. The p value of the Hausman test for endogeneity is also reported. It is a t-test in this case. Rejection of the null in this case indicates that trade policy openness is endogenous. Trade policy openness is instrumented by Frankel and Romer (1999) constructed openness (CONST) and landlocked dummy. The AJR sample without neo-Europe excludes Australia, Canada, New Zealand and United States from the AJR sample.

that one sample standard deviation increase in trade liberalization would lead to a 2.1 point increase in the property rights institutions index. To put this into perspective, let us focus on Myanmar. Myanmar is a virtually closed economy with very poor institutions during our sample period. The average trade liberalization (poit) score and the average property rights (PRINSit) score over the sample period in Myanmar are 0 and 5.7, respectively. If the trade liberalization index in Myanmar is to increase by 0.5, then the corresponding increase in the institutional score would be 37 percentage points. I also find similar results for contracting and regulatory institutions. The basic result holds after controlling for country fixed effects, time-varying common shocks and various additional covariates. It is also robust to various alternative measures of liberalization and institutions, as well as across different samples.

M2545 - BHATTACHARYYA PRINT.indd 144

02/03/2011 10:09

Improving institutions with trade policy

Table 7.3

145

Trade policy and institutions: first stage regressions Dependent variable: trade policy openness Full sample (obs 5 72) OLS estimate

AJR sample (obs 5 42) 2SLS estimate

AJR sample without Neo-Europe (obs 5 38) 2SLS estimate

0.01** (0.0031) −0.07 (0.1043) 0.098*** (0.0273) −0.02*** (0.0042) −0.04 (0.0811) 0.5650 0.0000

0.011** (0.0052) 0.035 (0.1775) 0.13** (0.0389) −0.01* (0.0057) −0.16 (0.1249) 0.3904 0.0024

0.01** (0.0052) 0.05 (0.1753) 0.14*** (0.0490) −0.01 (0.0061) −0.19 (0.1244) 0.2763 0.0552

Constructed openness (CONST) Landlocked dummy Total years of schooling Gini decadal average Log GDP per capita in 1960 R2 F-test (p-value)

Notes: ***, ** and * indicate significance levels of 1%, 5% and 10%, respectively, against a two-sided alternative. Standard errors are reported in parentheses. The p-value of the Hausman test for endogeneity is also reported. It is a t-test in this case. Rejection of the null in this case indicates that trade policy openness is endogenous. Trade policy openness is instrumented by Frankel and Romer (1999) constructed openness (CONST). The AJR sample without Neo-Europe excludes Australia, Canada, New Zealand and United States from the AJR sample.

7.4 SUMMARY Many studies argue that institutional variation across countries originates from colonial history. Without disagreeing with that view, I show here that trade policy may also influence institutional quality within a country. One could, of course, point towards anecdotal evidence on how trade policy reforms led to institutional development. This may work through realignment of political power within the country. This may also work through transfer of institutions. In this chapter, I present evidence using cross-section as well as panel data showing that trade liberalization indeed impacts positively on institutions. The challenge, however, is to take this beyond the broad framework and work out a detailed understanding of the channels through which trade liberalization impacts institutions. The next step would be to identify several channels through which trade liberalization leads to importation of institutions from abroad. Nevertheless,

M2545 - BHATTACHARYYA PRINT.indd 145

02/03/2011 10:09

146

Growth miracles and growth debacles

we do learn that institutional quality could be improved through good policy. Better institutional quality, of course, could significantly improve the growth prospects of a country. Institutions, however, are by definition a broad construct and it is often not obvious in policy terms what is meant by improving institutional quality. In the next chapter, I try to deconstruct institutions into four different types – market creating, market regulating, market stabilizing and market legitimizing institutions. I also identify proxy measures of each of these institutions and estimate their impact on growth since 1980.

NOTES 1. Other good examples are the Philippines and Mexico. In post-independence Philippines, restrictive trade policy led to institutional decline till the late 1980s when it was gradually reversed. See Shepherd and Alburo (1991) for a summary of trade policy in postindependence Philippines. Haber (2006) documents that import substitution policies were adopted in Mexico in the early twentieth century to protect the businesses of the rich elite aligned with the government of the dictator Porfirio Diaz. This damaged subsequent institutions. Haber (2006) also reports similar consequences of import substitution in Brazil. 2. In both the Polity and the ICRG indices, higher scores signify better institutional quality. 3. India liberalized its economy in 1991 and the positive effect on expropriation risk started to show in 1992 when the index score jumped from 6.2 to 8.2.

M2545 - BHATTACHARYYA PRINT.indd 146

02/03/2011 10:09

8.

Which institutions matter most for economic growth?

8.1 UNBUNDLING INSTITUTIONS Institutions are often defined as a collection of many attributes. The most commonly used measure of institutions in the institutions and development literature is the rule of law. The implicit assumption made in these studies is that better implementation of the rule of law is an indication of the overall quality of institutions. However, from the point of view of policy, it remains crucial to find out which type of institutions are important for growth and development. In this section, I make an attempt to unbundle institutions and estimate their impact on growth. I adopt Rodrik’s (2005) four-way classification (market creating, market regulating, market stabilizing and market legitimizing) of institutions and include them as explanatory variables of growth in my model. This is in contrast to other studies in the contemporary literature that focus on the importance of property rights and contracts in creating the right incentive structure for investment and entrepreneurship. But long-run economic growth requires more than just a boost in investment or entrepreneurship. To sustain growth momentum requires institutions that can facilitate exchange in a world of imperfect information, handle random shocks to the economy and facilitate socially acceptable burden sharing in the event of a shock. The four-way classification attempts to cover all aspects of institutions and also distinguish them on the basis of their functions. Market-creating institutions are those that protect property rights and ensure contract enforcement. They are called market creating since, in their absence, either the markets do not exist or they perform very poorly. Ideally one would expect expropriation risk or executive constraint to be good measures of market-creating institutions, but they are not entirely suitable for our purpose, as we will see below. As an alternative, I choose the ICRG law and order index. The underlying assumption is that a country with strong law and order is expected to have better property rights and contract enforcement. The ‘law’ subcomponent of the measure assesses the strength and impartiality of the legal system and the ‘order’ subcomponent is an assessment of popular observance of the law. The 147

M2545 - BHATTACHARYYA PRINT.indd 147

02/03/2011 10:09

148

Growth miracles and growth debacles

assessment is made on a six point scale with a high score implying better law and order. I observe that 79 per cent of the variation in law and order is between countries and the remaining 21 per cent is within countries. Sri Lanka ranks lowest on law and order with a score of 0.3 in 1990 and Luxembourg ranks highest with a score of 6 throughout the period. One would expect better law and order would lead to less risk of expropriation and better contract enforcement. There are at least four other measures that have been used as a proxy for market-creating institutions in the literature – the rule of law index of Rodrik et al. (2004), the expropriation risk of Acemoglu et al. (2001), the executive constraint and the legal formalism index of Acemoglu and Johnson (2005).1 Acemoglu and Johnson (2005) use executive constraint as a proxy for property rights institutions and legal formalism as a proxy for contracting institutions to separately estimate their effects on long-run growth. Therefore, the obvious question is why I chose the ICRG law and order index as the preferred measure. Even though far from perfect, there are at least two clear-cut advantages of using the law and order index. First, it has a long time dimension running from 1984 to 2004, which is useful for the dynamic panel data estimation. In contrast, the rule of law index of Rodrik et al. (2004) and the legal formalism index of Acemoglu and Johnson (2005) are only available in cross-section and the expropriation risk measure of Acemoglu et al. (2001) only covers the period 1982 to 1997 which is too short a time dimension for estimating our dynamic panel model. Second, it can be counted upon as a measure that represents both property rights and contracting institutions. I do, however, check the robustness of the result using alternative measures of institutions. The other question is why I chose not to separate out the effects of property rights and contracts. This is largely due to the unavailability of a reasonable proxy for contracting institutions. Acemoglu and Johnson (2005) argue that an ideal measure of contracting institutions is the cost of enforcing private contracts. They use the legal formalism index as a proxy. This measure, however, is not available in a panel. As an alternative, I use repudiation of government contracts, which measures the cost of enforcing government contracts and is far from ideal.2 It primarily focuses on institutions that define the relationship between the state and its subjects and not on institutions that provide the legal framework that enables private contracts to facilitate economic transactions. This measure also drastically reduces the sample size. Nevertheless, I report these findings in Section 8.3. Market regulating institutions are those that prevent market failure and help to sustain growth momentum over the long run. Rodrik and Subramanian (2003) mention regulatory agencies in telecommunication, transport and financial services as examples of market-regulating

M2545 - BHATTACHARYYA PRINT.indd 148

02/03/2011 10:09

Which institutions matter most for economic growth?

149

institutions. I proxy this by using Gwartney and Lawson’s (2005) composite index of regulation (MR) in the credit market, labour market and business in general since this is the closest to Rodrik and Subramanian’s examples. The index operates on an 11 point scale ranging from 0 to 10 with a high score implying fewer regulations. I notice that 67 per cent of the variation in the measure is between countries. Romania records the lowest level of regulation and New Zealand records the highest level of regulation in the dataset. This is the only proxy of regulatory institutions that I could locate which has the desirable time dimension. Market-stabilizing institutions are those that build resilience towards shocks, reduce inflationary pressure, minimize macroeconomic volatility and avert financial crises. Examples of market stabilizing institutions include central banks, exchange rate regimes and budgetary and fiscal rules. Finding a suitable proxy for this is a challenge. I think that it is not unreasonable to assume that good market-stabilizing institutions (independent central banks) do look to minimize inflationary pressure and volatility in the long run.3 Hence, I use Gwartney and Lawson’s (2005) sound money index (SM) as a proxy which takes into account: (1) average annual growth of the money supply in the last five years minus average annual growth of real GDP in the last ten years, (2) standard inflation variability in the last five years, and (3) recent inflation rate. This index also ranges from 0 to 10 with a high score implying better market-stabilizing institutions. I observe that 56 per cent of the variation in market-stabilizing institutions is between countries. Canada, France, Denmark and Switzerland are among countries with very strong ‘market-stabilizing institutions’, whereas some of the Latin American countries, including Argentina and Brazil, fare among the worst especially during the 1980s. Market-legitimizing institutions are those that handle redistribution, manage social conflict and provide social protection and insurance in the event of a shock. In other words, they help to minimize the idiosyncratic risk to economic growth and employment in a modern market economy. Rodrik (2005) suggests democracy as a proxy for market-legitimizing institutions as there is a positive relationship between the effectiveness of democratic institutions and the quality of social insurance. Therefore, the Polity IV democracy index suits such a purpose as it measures the effectiveness of democratic institutions by capturing different shades of democracy and it ranges from 0 to 10 with a high score implying a more democratic system. Among the most democratic nations in the dataset are the United States, Japan and the United Kingdom, and among the least democratic are Morocco, Nigeria and Equatorial Guinea. A total of 78 per cent of the variation in this series is between countries and 22 per cent of the variation is within countries. Several existing studies (see

M2545 - BHATTACHARYYA PRINT.indd 149

02/03/2011 10:09

150

Growth miracles and growth debacles

Barro, 1996; Rodrik, 1999; Tavares and Wacziarg, 2001; Acemoglu et al., 2005c; Acemoglu et al., 2008; Bhattacharyya and Hodler, 2010) use this as a measure of democracy. I also use the freedom house political rights index as an alternative measure of market-legitimizing institutions.4 There is also a strong view in this literature that a simple dichotomy between democracy and non-democracy is the most appropriate empirical definition (see Przeworski et al., 2000). However, this definition is not suitable for our purpose as it does not provide any information on the effectiveness of democratic institutions. Recently, Persson and Tabellini (2006) used a similar empirical definition and looked at the impact of regime change and different styles of democracy on within-country growth. They found that countries liberalizing their economy before extending political rights do better in terms of growth than countries following the opposite sequence. Their measure also focuses on democratic transitions and hence is not suitable for our purpose. Finally, there is a view in the literature that corruption should be treated as an institution. Hall and Jones (1999) and Knack and Keefer (1995) include corruption in their overall measure of institutions. I am, however, slightly sceptical about classifying corruption as an institution. In my view, corruption is the outcome of poor market-creating institutions rather than an institution itself. This is perhaps reflected by the measure of corruption that Mauro (1995) uses. Mauro (1995) uses an indicator that reflects experts’ assessments of the degree to which business transactions involve corruption or questionable payments averaged over the period 1980–83. This is an index of perception and is reflective of institutional outcome rather than institutions themselves. Nevertheless, I do include ICRG index of corruption as an additional control variable (see Table 8.6, column 3) and my basic results are robust to this inclusion.

8.2 DATA The dataset includes measures of schooling, institutions, per capita GDP growth, per capita GDP levels and trade share. There are six time points in the data, spanning the period 1980–2004, with each time point being approximately five years on average. There are missing observations in the data and hence I have an unbalanced panel. I follow the literature and proxy human capital using schooling data from Barro and Lee (2000). It measures the average schooling years in the total population. A total of 96 per cent of the variation in schooling is between countries. As indicated in the previous section, I use four different measures of

M2545 - BHATTACHARYYA PRINT.indd 150

02/03/2011 10:09

Which institutions matter most for economic growth?

Table 8.1

151

Summary statistics

Variable

Number of obs

Economic growth and development Growth (yˆiT) 629 Initial income (yiT25) 629 Market creating institutions Law and order (LO) 507 Market regulating institutions Regulation of credit, 616 labour, business (MR) Market stabilizing institutions Sound money Index (SM) 634 Market legitimizing institutions Democracy index 659 (DEMOC) Schooling Total years of schooling 482 (TYS) Trade openness Log trade share of GDP 754 (LTRS)

Mean

Standard Minimum Maximum deviation

0.014 8.4

0.031 1.13

−0.10 6.01

0.35 10.88

3.7

1.5

0.3

6

5.6

1.1

2.5

8.8

7.5

2.5

0

9.8

5.0

4.1

0

10

5.4

2.9

0.4

12.3

4.19

0.59

2.43

6.01

institutions. There are a number of conceptual and empirical challenges that I have to face in the analysis. It is hard to rule out potential overlaps between these measures of institutions. The pairwise correlations reported in Table 8.3 show that the measures of institutions are correlated. The correlation between law and order and democracy is 0.5321 and the correlation between democracy and the regulation of credit, labour and business is 0.5066. However, none of the correlations is large enough to cause any serious problems of multicollinearity. The GDP per capita PPP measured at a constant 2000 international dollars is obtained from the World Development Indicator (WDI). The WDI data provides larger time coverage and allows me to extend the study to 2004. Annualized growth rates of GDP per capita are calculated using the formula yˆiT ; 1/5 ( ln yiT 2 ln yiT 25) . The trade share measure is also from WDI and is widely used (see Dollar and Kraay, 2003; Rodrik et al., 2004). It is perhaps the simplest measures of trade openness where trade is expressed as a share of GDP. Table 8.1 presents the descriptive statistics for the important variables used in the study and Table 8.2 presents a list of countries.

M2545 - BHATTACHARYYA PRINT.indd 151

02/03/2011 10:09

152

Table 8.2

Growth miracles and growth debacles

List of countries

Code

Country

Code

Country

Code

Country

AGO ARG ATG AUS AUT BDI BEL BEN BFA

Angola Argentina Antigua Australia Austria Burundi Belgium Benin Burkina Faso Bangladesh Belize Bolivia Brazil Barbados Botswana Central African Republic Canada Switzerland Chile China Côte d’Ivoire Cameroon Congo, Republic of Colombia Comoros Cape Verde Costa Rica Cyprus Dominica Denmark Dominican Republic Algeria Ecuador Egypt Spain Ethiopia Finland Fiji France Gabon United Kingdom Germany Ghana

GIN GMB

Guinea Gambia, The GuineaBissau Equatorial Guinea Greece Grenada Guatemala Guyana Hong Kong Honduras Haiti Hungary Indonesia India Ireland Iran Iceland Israel Italy Jamaica Jordan Japan Kenya St. Kitts & Nevis Korea, Republic of St. Lucia Sri Lanka Lesotho Luxembourg Morocco Madagascar Mexico Mali Mozambique Mauritania Mauritius Malawi Malaysia Namibia Niger Nigeria Nicaragua Netherlands

NOR NPL NZL

Norway Nepal New Zealand Pakistan Panama Peru Philippines Papua New Guinea Poland Puerto Rico Portugal Paraguay Romania Rwanda Senegal Singapore Sierra Leone El Salvador Sao Tome and Principe Slovak Republic Sweden Swaziland Seychelles Syria Chad Togo Thailand Trinidad &Tobago Tunisia Turkey Taiwan Tanzania Uganda Uruguay USA St.Vincent & Grenadines Venezuela South Africa Congo, Dem. Rep. Zambia Zimbabwe

BGD BLZ BOL BRA BRB BWA CAF CAN CHE CHL CHN CIV CMR COG COL COM CPV CRI CYP DMA DNK DOM DZA ECU EGY ESP ETH FIN FJI FRA GAB GBR GER GHA

M2545 - BHATTACHARYYA PRINT.indd 152

GNB GNQ GRC GRD GTM GUY HKG HND HTI HUN IDN IND IRL IRN ISL ISR ITA JAM JOR JPN KEN KNA KOR LCA LKA LSO LUX MAR MDG MEX MLI MOZ MRT MUS MWI MYS NAM NER NGA NIC NLD

PAK PAN PER PHL PNG POL PRI PRT PRY ROM RWA SEN SGP SLE SLV STP SVK SWE SWZ SYC SYR TCD TGO THA TTO TUN TUR TWN TZA UGA URY USA VCT VEN ZAF ZAR ZMB ZWE

02/03/2011 10:09

Which institutions matter most for economic growth?

153

8.3 EVIDENCE Table 8.4 reports the dynamic regressions. They are estimated using the Generalized Method of Moments (GMM) Blundell and Bond estimator. In column 1, I start with a simple specification regressing law and order and schooling on growth. The other control variables are initial income and trade. In this specification, law and order is statistically significant, but schooling is not. This may be due to endogeneity of the instruments as they fail the Hansen J-test (p-value 0.006). In column 2, I replace law and order by MR and I observe that schooling is statistically significant, but MR is statistically insignificant. In column 3, I replace MR with SM. In this specification, both schooling and SM are statistically significant. In column 4, I replace SM with democracy. Schooling is still statistically significant, but democracy is statistically insignificant. In column 5, I estimate my preferred model with a full set of control variables. I add schooling, law and order, MR, SM, democracy, trade and initial income into the specification. I find that schooling, law and order and SM are statistically significant and all institution measures are jointly significant. One sample standard deviation increase in schooling increases the annual growth rate in an average country by 1.7 percentage points. MR and democracy do not seem to matter. Comparable increases in law and order and SM both have growth effects of 0.75 percentage points. In column 6, I make an attempt to unbundle market creating institutions and separate out the effects of property rights and contracting institutions. I use executive constraint from the Polity IV dataset as a measure of property rights institutions. In the absence of time series data on the legal formalism index, I use repudiation of government contracts as an alternative measure of contracting institutions, which, of course, is far from ideal. I notice that both property rights and contracts have positive and statistically significant effects on growth. This is in contrast to Acemoglu and Johnson (2005), who report that property rights matter more than contracts for growth in the long run. The magnitude of the coefficient on schooling remains unchanged and is statistically significant. However, a major drawback of this specification is that it drastically reduces the sample size. In this case, T 5 2 and the estimates are equivalent to 2SLS and they can be poorly identified or suffer from weak instrument problems (Wooldridge, 2002). In order to explore the possibility of non-linearity in MR, I introduce MR2 in column 7. The public interest theory of Pigou (1938) suggests that unregulated markets are comparatively more prone to failure and should be associated with socially inferior outcomes. In contrast, the public choice theory of Shleifer and Vishny (1998) suggests that regulation leads to corruption and hence is harmful for development.5 Therefore, it is possible that there

M2545 - BHATTACHARYYA PRINT.indd 153

02/03/2011 10:09

154

M2545 - BHATTACHARYYA PRINT.indd 154

02/03/2011 10:09

1.0000 0.7243 0.6824 0.2753 0.4437

0.8511 0.1250

0.3075

0.1812

0.3071

0.1729

Initial income (yiT25)

1.0000 0.1887 0.3631 0.1639

Growth (yˆiT)

0.1969

0.7217

0.4287

0.3770

1.0000 0.5321

0.0626

0.6237

0.5066

0.1796

1.0000

Law and Democracy order (LO) (DEMOC)

0.2707

0.2746

0.4089

1.0000

Sound money index (SM)

Unbundled institutions, human capital and growth: pairwise correlation

Growth (yˆiT) Initial income (yiT25) Law and order (LO) Democracy (DEMOC) Sound money index (SM) Regulation of credit, labour, business (MR) Total years of schooling (TYS) Log trade share of GDP (LTRS)

Table 8.3

0.2740

0.4899

1.0000

0.1497

1.0000

Regulation Total years of credit, of schooling (TYS) labour, business (MR)

1.0000

Log trade share of GDP (LTRS)

155

M2545 - BHATTACHARYYA PRINT.indd 155

02/03/2011 10:09

Other controls Initial income (yiT25)

DEMOC2

Executive constraint (EXCONST) Repudiation of govt contracts (REPU) Regulation of credit, labour, business (MR) Sound money index (SM) Democracy index (DEMOC) Total years of schooling (TYS) MR2

YES

0.004 (0.0029)

0.004** (0.0020)

(1)

YES

0.007** (0.0034)

−0.001 (0.0029)

(2)

YES

0.005* (0.0028)

0.003*** (0.0009)

(3)

YES

0.0004 (0.0011) 0.01** (0.0038)

(4)

YES

0.003*** (0.0009) −0.0003 (0.0009) 0.006** (0.0026)

−0.003 (0.0024)

0.005** (0.0020)

(5) 0.005** (0.0020)

(7)

YES

0.002 (0.0013) −0.007*** (0.0026) 0.006* (0.0034)

−0.003 (0.0025)

0.005*** (0.0018)

(8)

0.023 (0.0148)

0.005*** (0.0018)

(9)

YES

YES

YES

0.002** 0.003*** 0.002** (0.0009) (0.0009) (0.0009) −0.001 0.001 0.0003 (0.0010) (0.0021) (0.0021) 0.007*** 0.005** 0.006*** (0.0024) (0.0025) (0.0024) −0.002* −0.002* (0.0013) (0.0013) −0.0001 −0.0001 (0.0002) (0.0002)

0.012*** (0.0042) 0.007*** (0.0017) −0.006* 0.023 (0.0036) (0.0153)

(6)

Dependent variable: annualized growth (yˆiT)

Unbundled institutions, human capital and growth: dynamic regressions

Law and order (LO)

Table 8.4

156

M2545 - BHATTACHARYYA PRINT.indd 156

02/03/2011 10:09

(continued)

363/93

0.325

0.291

348/87

0.109 0.001

YES

YES

0.060 0.001

(2)

(1)

369/93

0.032

0.009 0.000

YES

(3)

370/95

0.251

0.157 0.001

YES

(4)

323/83

0.174

0.351 0.001

0.0027

YES

(5)

242/83

–

0.480 0.034

0.0000

YES

(6)

(7)

323/83

0.166

0.670 0.002

0.0079

YES

Dependent variable: annualized growth (yˆiT)

323/83

0.175

0.776 0.001

0.0030

YES

(8)

323/83

0.174

0.866 0.002

0.0044

YES

(9)

Notes: ***, ** and * indicate significance levels of 1%, 5% and 10%, respectively, against a two-sided alternative. Standard errors are reported in parentheses. Hansen test is the test of the H0: the instruments as a group are exogenous. Hansen test statistic from two-step Arellano and Bond estimations is reported, which is robust to heteroskedasticity or autocorrelation. Arellano and Bond AR(1) and AR(2) tests in residuals are also reported. Note that to pass these tests, one has to reject the null of no AR(1) and fail to reject the null of no AR(2).

Observations/countries

Specification tests (P-values) Joint F-test (Institutions) Hansen test Test for AR(1) in residuals Test for AR(2) in residuals

Log trade share (LTRS)

Table 8.4

Which institutions matter most for economic growth?

157

is an optimum level of regulation. In other words, the relationship between regulation and growth is non-linear and hence it justifies our inclusion of MR2 in column 7. I observe that the coefficient on MR2 is negative and statistically significant, but the coefficient on MR is positive and statistically insignificant. This is perhaps because of the non-linear effect of regulation. In other words, perhaps there is a particular level of regulation that maximizes growth. This level turns out to be 4.9 when I equate the partial derivative of growth with respect to MR to zero.6 One possible interpretation of this result is that too much or too little regulation is not good for growth. Too little regulation encourages anti-competitive behaviour among firms and can lead to market failure. Too much regulation, on the other hand, can lead to red tape which has tangible costs to the economy. However, I treat this result with caution as the coefficient on MR is statistically insignificant. In column 8, I make an attempt to check for non-linearity in democracy by adding democracy2 into the preferred model. I find that the statistical significance of schooling, law and order and SM survives, but both democracy and democracy2 are statistically insignificant. In column 9, I add both MR2 and democracy2 into the preferred model. The effects of schooling, law and order and SM survive. The non-linear effect of MR also survives. To summarize, I find that both human capital proxied by schooling and institutions have positive and statistically significant effects on growth. The effect of human capital is relatively larger in size compared to the effect of institutions. Among institutions, market-creating institutions proxied by law and order and market-stabilizing institutions proxied by SM have positive and equal effects on growth. Regulatory institutions proxied by MR seem to matter only to a certain extent and marketlegitimizing institutions proxied by democracy do not seem to matter. This holds even when I use the Freedom House Political Rights index as a measure of democracy. The democracy result is in conformity with the existing evidence in the literature. Previous studies have documented that the evidence in favour of democracy yielding subsequent growth is at best weak (see Barro, 1996; Przeworski et al., 2000; Acemoglu et al., 2008). In Table 8.5, I deal with an important technical issue. Is the use of Blundell and Bond estimation technique appropriate? Nickell (1981) shows that when fixed effects are correlated with explanatory variables then OLS overestimates the effect of the lagged dependent variable, fixed effect underestimates it and system GMM should be in between. In a recent Monte Carlo study, Hauk and Wacziarg (2004) showed that fixed effect and Arellano and Bond GMM can in fact overestimate the effect of the lagged dependent variable and bias towards zero the effect of other variables in the presence of measurement error. They showed that OLS

M2545 - BHATTACHARYYA PRINT.indd 157

02/03/2011 10:09

158

Table 8.5

Growth miracles and growth debacles

Is the Blundell and Bond method appropriate? Dependent variable: annualized growth (yˆiT)

Initial income (yiT25) Law and order (LO) Regulation of credit, labour, business (MR) Sound money index (SM) Democracy index (DEMOC) Total years of schooling (TYS) Log trade share (LTRS) Specification tests (P-values) Hansen test Test for AR(1) in residuals Test for AR(2) in residuals F-test that all μi50 F-test for overall significance R2 Observations/ countries

(2)

GMM Arellano and Bond (3)

GMM Blundell and Bond (4)

−0.01*** (0.0025) 0.005*** (0.0012) −0.001 (0.0015)

−0.05*** (0.0059) 0.004*** (0.0013) 0.001 (0.0018)

−0.08*** (0.0117) 0.004** (0.0015) 0.004** (0.0021)

−0.01* (0.0063) 0.005** (0.0020) −0.003 (0.0024)

0.002*** (0.0005) 3e-04 (0.0005) 0.003*** (0.0009) 0.003 (0.0024)

8.5e-04 (0.0006) 4e-04 (0.0005) 0.003* (0.0017) 0.024*** (0.0057)

Pooled OLS

Fixed effect

(1)

1e-05 (0.0007) 8e-05 (0.0007) 0.002 (0.0029) 0.022*** (0.0070)

0.003*** (0.0009) −3e-04 (0.0009) 0.006** (0.0026) 0.008 (0.0058)

0.0004 0.0191

0.351 0.001

0.344

0.174

240/83

323/83

0.0000 0.0000 0.2211 323/83

323/83

Notes: ***, ** and * indicate significance levels of 1%, 5% and 10%, respectively, against a two-sided alternative. Standard errors are reported in parentheses. Hansen test is the test of the H0: the instruments as a group are exogenous. Hansen test statistic from twostep Arellano and Bond estimations is reported which is robust to heteroskedasticity or autocorrelation. Arellano and Bond AR(1) and AR(2) tests in residuals are also reported. Note that to pass these tests, one has to reject the null of no AR(1) and fail to reject the null of no AR(2).

M2545 - BHATTACHARYYA PRINT.indd 158

02/03/2011 10:09

Which institutions matter most for economic growth?

159

and Blundell and Bond perform the best in this situation. The advantage of Blundell and Bond over OLS is that it tackles the endogeneity problem better. My results match with Hauk and Wacziarg (2004), as I notice in columns 2 and 3 of Table 8.5 that fixed effect and Arellano and Bond overestimate the effect of the lagged dependent variable (20.05 and 20.08, respectively) and underestimate the contribution of schooling (0.003 and 0.002, respectively). In the case of Arellano and Bond, the effect of schooling is statistically insignificant. Furthermore, the Hansen test statistic in this case also documents the weak instrument problem with Arellano and Bond when there is a short panel with persistence in the data. Therefore, the evidence suggests that GMM Blundell and Bond is the appropriate way to go. Table 8.6 reports robustness results of the preferred model with additional covariates. I add additional covariates (investment and population growth, corruption, foreign direct investments (FDI), foreign aid, real exchange rate distortions and credit to the private sector)7 and I find that my schooling, law and order and SM result survives. Column 1 reports the preferred specification. In column 2, I add Solow style variables (investment share and population growth) as additional controls. The law and order and schooling result survives as these variables register positive and statistically significant effects on growth. The magnitude of the effects on growth, however, reduces by 0.1 percentage points in the case of law and order and by 0.58 percentage points in the case of schooling. This implies that there is an upward bias in my estimates in the absence of these variables. However, the extent of the bias is not severe enough to alter the direction and statistical significance of the effects. In column 3, I control for corruption. Coefficients on law and order, schooling and SM survive this test. This indicates that including corruption in our set of institutions will not alter the major result. In columns 4 to 7, I examine the effect of adding FDI, foreign aid, real exchange rate distortion and credit to the private sector, respectively. Credit to the private sector is used as a proxy for financial liberalization in the literature (Levine et al. 2000). In all the cases, my main result survives. In column 8, I control for all covariates taken together. Only the effect of schooling survives. However, controlling for all factors drastically reduces the sample size. Also, it suffers from the weak instrument problem and perhaps multicollinearity, and hence is unreliable. In column 9, I replace law and order by expropriation risk – an alternative measure of market-creating institutions. This measure has been used by Acemoglu et al. (2001). I find that expropriation risk, SM and schooling are statistically significant with positive effects on growth. The impact of schooling on growth increases by 0.6 of a percentage point as a result of this change. The impact of market-creating institutions proxied

M2545 - BHATTACHARYYA PRINT.indd 159

02/03/2011 10:09

160

M2545 - BHATTACHARYYA PRINT.indd 160

02/03/2011 10:09

0.0004 (0.0008) 0.001 (0.0007) 0.004** (0.0021)

0.003*** (0.0009) −0.0003 (0.0009) 0.006** (0.0026)

0.351

Specification tests (P-values) Hansen test 0.891

0.002 (0.0025)

0.004** (0.0016)

−0.003 (0.0024)

0.005** (0.0020)

0.715

0.002*** (0.0009) −0.0004 (0.0009) 0.005** (0.0025)

−0.003 (0.0022)

0.006*** (0.0020)

Preferred Investment Corruption specification share and population growth (2) (3) (1)

0.868

0.002*** (0.0009) 0.0002 (0.0008) 0.006** (0.0024)

−0.003 (0.0024)

0.998

0.002** (0.0007) 0.0005 (0.0007) 0.006** (0.0026)

−0.002 (0.0027)

0.007*** (0.0019)

(5)

(4) 0.004** (0.0017)

Foreign aid

FDI

0.851

0.003*** (0.0008) 0.0004 (0.0008) 0.004* (0.0023)

−0.0002 (0.0021)

0.003* (0.0017)

Real exchange rate distortions (6)

0.877

0.002*** (0.0009) 0.0001 (0.0009) 0.005** (0.0024)

−0.004* (0.0023)

0.004** (0.0020)

Credit to the private sector (7)

Dependent variable: annualized growth (yˆiT)

0.468

0.001 (0.0011) 0.001 (0.0009) 0.005* (0.0030)

−0.001 (0.0032)

0.002 (0.0022)

(8)

−0.002 (0.0022)

0.007** (0.0019)

(10)

Alcala Ciccone openness

0.230

0.328

0.003** 0.002*** (0.0013) (0.0008) −0.001 −0.0002 (0.0012) (0.0008) 0.008** 0.007*** (0.0041) (0.0025)

0.003** (0.0014) −0.005 (0.0032)

(9)

All With covariates EXPR

Unbundled institutions, human capital and growth: robustness with additional covariates

Law and Order (LO) Expropriation Risk (EXPR) Regulation of Credit, Labour, Business (MR) Sound Money Index (SM) Democracy Index (DEMOC) Total Years of Schooling (TYS)

Additional Covariates

Table 8.6

161

M2545 - BHATTACHARYYA PRINT.indd 161

02/03/2011 10:09

LTRS,

yiT25

LTRS,

yiT25

319/82

0.462

0.174

323/83

0.008

0.001

yiT25

LTRS,

322/83

0.179

0.002

yiT25

LTRS,

319/82

0.116

0.001

yiT25

LTRS,

231/61

0.036

0.008

yiT25

LTRS,

300/77

0.098

0.002

yiT25

LTRS,

319/82

0.146

0.001

yiT25

LTRS,

209/55

0.044

0.058

323/83

0.156

0.002

yiT25

yiT25

LTRS, LROPEN,

242/83

–

0.013

Notes: ***, ** and * indicate significance levels of 1%, 5% and 10%, respectively, against a two-sided alternative. Standard errors are reported in parentheses. Hansen test is the test of the H0: the instruments as a group are exogenous. Hansen test statistic from two-step Arellano and Bond estimations are reported, which are robust to heteroskedasticity or autocorrelation. Arellano and Bond AR(1) and AR(2) tests in residuals are also reported. Note that to pass these tests, one has to reject the null of no AR(1) and fail to reject the null of no AR(2). In each regression, the standard controls are log trade share (LTRS) and Initial income (yiT25) except column 10 where Alcalá Ciccone real openness (LROPEN) is used.

Other controls

Observations/ Countries

Test for AR(1) in residuals Test for AR(2) in residuals

162

Growth miracles and growth debacles

by expropriation risk is 0.11 percentage points less than the impact of market-creating institutions proxied by law and order. The impact of SM, however, remains unchanged. A major drawback of using expropriation risk is that it drastically reduces the sample size. Finally, in column 10, I use Alcalá and Ciccone’s (2004) ‘log real openness’ index instead of the log trade share measure. This index is a ratio of trade to PPP GDP.8 My major findings survive even when I use this measure of trade. In Table 8.7, I check the robustness of my basic finding in alternative samples. I find that the schooling, law and order and SM result survives in the base sample without British legal origin countries, the base sample without French legal origin countries, the base sample without Africa, the base sample without Neo-Europe9 and the base sample without oil exporters. The magnitude of the coefficient on law and order varies from 0.004 to 0.007, which is 0.2 per cent of its sample standard deviation. The coefficient on schooling varies from 0.005 to 0.008, which is 0.1 per cent of its sample standard deviation. All other variables except law and order become statistically insignificant in the Africa sample. This is perhaps emphasizing the importance of market-creating institutions in Africa.

8.4 SUMMARY It is well documented that institutions are important for growth. In this chapter, I make an attempt to go beyond the broad institution continuum. I use a four-way classification of institutions (market-creating, marketregulating, market-stabilizing and market-legitimizing institutions) and identify a proxy for each of them. This allows me to unbundle institutions and take the analysis beyond property rights and contracts. I estimate the contributions of market-creating, market-regulating, market-stabilizing and market-legitimizing institutions to growth. I find that strong marketcreating institutions and market-stabilizing institutions are good for growth. There exists a growth-maximizing level of market regulation beyond which red tape creates disincentives for investment. Marketlegitimizing institutions do not seem to matter for growth. Our basic result survives across different samples and the additional covariate test. This exercise allows us to learn about the statistical importance of these institutions on growth. However, these models, in spite of their sophistication, are not free from limitations. One needs to bear in mind while interpreting these coefficients that they are average relationships across countries. In order to draw policy lessons, one needs to look into more detailed analysis of the role of these factors. I embark on such an exercise in Chapter 9.

M2545 - BHATTACHARYYA PRINT.indd 162

02/03/2011 10:09

163

M2545 - BHATTACHARYYA PRINT.indd 163

02/03/2011 10:09

Specification tests (P-values) Hansen test Test for AR(1) in residuals Test for AR(2) in residuals 0.987 0.002 0.610

0.351 0.001

0.174

0.006** (0.0024) −0.002 (0.0022) 0.002* (0.0009) −0.001 (0.0008) 0.008*** (0.0022)

(2)

(1)

0.005** (0.0020) Regulation of credit, −0.003 labour, business (MR) (0.0024) Sound money index 0.003*** (SM) (0.0009) Democracy index −0.0003 (DEMOC) (0.0009) Total years of schooling 0.006** (TYS) (0.0026)

Base sample without British legal origin countries

Base sample

0.152

1.000 0.087

0.002 (0.0022) −0.01*** (0.0032) 0.006*** (0.0014) −0.0004 (0.0008) 0.005* (0.0027)

Base sample without French legal origin countries (3)

0.929

1.000 0.021

0.327

0.935 0.002

0.007*** (0.0022) −0.002 (0.0023) 0.002** (0.0010) −0.0004 (0.0009) 0.007*** (0.0025)

(5)

(4) 0.007*** (0.0025) 0.005 (0.0036) 0.0009 (0.0007) 0.0005 (0.0008) −0.0007 (0.0032)

Base sample without Africa

African countries

Dependent variable: annualized growth (yˆiT)

0.165

0.500 0.002

0.005** (0.0020) −0.003 (0.0026) 0.003** (0.0010) −0.0003 (0.0009) 0.007** (0.0027)

(6)

Base sample without Neo-Europe

Unbundled institutions, human capital and growth: robustness with alternative samples

Law and order (LO)

Table 8.7

0.102

0.554 0.002

0.004** (0.0021) −0.003 (0.0024) 0.002** (0.0010) 0.0001 (0.0010) 0.006** (0.0027)

Base sample without oil exporters (7)

164

M2545 - BHATTACHARYYA PRINT.indd 164

02/03/2011 10:09

LTRS, yiT25

LTRS, yiT25

157/40

Base sample without French legal origin countries (3)

LTRS, yiT25

LTRS, yiT25

247/63

(5)

(4) 76/20

Base sample without Africa

African countries

LTRS, yiT25

311/80

(6)

Base sample without Neo-Europe

LTRS, yiT25

304/78

Base sample without oil exporters (7)

Notes: ***, ** and * indicate significance levels of 1%, 5% and 10%, respectively, against a two-sided alternative. Standard errors are reported in parentheses. Hansen test is the test of the H0: the instruments as a group are exogenous. Hansen test statistic from two-step Arellano and Bond estimations is reported which is robust to heteroskedasticity or autocorrelation. Arellano and Bond AR(1) and AR(2) tests in residuals are also reported. Note that to pass these tests, one has to reject the null of no AR(1) and fail to reject the null of no AR(2). In each regression, the standard controls are log trade share (LTRS) and initial income (yiT25) . Neo-Europe includes Australia, Canada, New Zealand and United States.

LTRS, yiT25

Other controls

(2)

(1) 210/54

Base sample without British legal origin countries

Dependent variable: annualized growth (yˆiT)

Base sample

323/83

(continued)

Observations/countries

Table 8.7

Which institutions matter most for economic growth?

165

NOTES 1. Note that Hall and Jones (1999) use the social infrastructure index as an overall measure of institutions. They combine law and order, bureaucratic quality, corruption, risk of expropriation and government repudiation of contracts from ICRG and the Sachs and Warner (1995a) index of openness to construct the index. This, however, is not suitable for the purpose of unbundling institutions. Also, the Sachs and Warner openness index used by Hall and Jones is not a measure of institutions. 2. Knack and Keefer (1995) use this as a measure of contracting institutions. 3. Barro (1995) reports a negative relationship between inflation and central bank independence in a cross-section of countries. 4. This measure is suitable since it ranges from 1 to 7 and distinguishes between different shades of democracy. 5. See Djankov et al. (2002) for a survey of this literature. 6. The value 4.9 lies well within the sample range of MR (which is 0 to 10). 7. Previous studies have reported strong correlation between these variables and growth. 8. Alcalá and Ciccone (2004) argue that this measure performs better than the standard measure in the presence of trade-driven productivity change. However, this measure is not free from controversy. Rodrik et al. (2004) show that the ‘real openness’ index (Ropen) and trade volume (Open) are linked by the identity logRopen 5 logOpen 1 logP, where P is a country’s price level. Also, from the Balassa-Samuelson argument P is well known to be closely associated with a country’s income/productivity level. Rodrik et al. (2004) plot logRopen 2 logOpen on logGDP and find a positive relationship. They also find very little correlation between logP and logOpen. Based on this, they argue that augmenting P (which has very little or no correlation to Open) into the standard trade volume measure is likely to spuriously attribute the correlation between logP and logGDP on the correlation between openness and logGDP. 9. Neo-Europe includes Australia, Canada, New Zealand and United States.

M2545 - BHATTACHARYYA PRINT.indd 165

02/03/2011 10:09

9.

Making policy work: a road map for future growth

9.1 WHAT HAVE WE LEARNED SO FAR? Part I of this book focuses on the historical determinants of long-run economic development. In Chapter 2, I document wide variations in growth rates and living standards across countries over the post-war period. The key questions, however, are what explains this variation and what explains its persistence. Looking further back, I document that this variation is not just attributable to how countries performed over the post-war period but also from the start of the sixteenth century. Some countries diverged from the rest in terms of living standards starting from that period. Chapter 2 presents a detailed account of this process of divergence and the data involved. In Chapter 3, I present some of the existing theories explaining this divergence. I present a case for exploring the root causes. I also review theories of institutions, geography, human capital, trade, religion and culture, and state formation and war. Chapter 4 presents evidence and illustrates the difficulties in quantifying the relationships between root causes and economic development. It also shows that the empirical results are strikingly different in Africa compared to the rest of the world. Diseases (malaria, in particular) are the only significant explanator of economic development in Africa. All other variables including institutions are statistically insignificant. This is, however, not the case in a more general sample where both malaria and institutions are important for development. Chapter 5 makes an attempt to relate these empirical results to economic history. Empirical results are in reduced form and, therefore, are not able to explain the dynamics of economic development across continents over centuries. In order to make sense out of them, one needs to study them in conjunction with historical narratives. Keeping this in mind, Chapter 5 presents a unifying framework to explain the process of development in Western Europe. This framework is then compared with the growth narrative in other countries and continents namely Africa, China, India, the Americas, Russia and Australia. A narrative style is adopted throughout this section to create a bridge between the empirical literature and history. 166

M2545 - BHATTACHARYYA PRINT.indd 166

02/03/2011 10:09

Making policy work

167

The framework is also an attempt to merge all seemingly disparate theories of economic progress over the long term. This, without doubt, is the main value added of this book. The main lesson of Chapter 5 is that diseases and geography matter at an early stage of development. Institutions, however, become much more important as the economy develops. Geography, and in particular disease epidemics, are a crucial explanator of the lack of development in Africa. Diseases impact African development through discouraging economic agents to save for the future. A high discount rate on future consumption due to diseases and the resulting lack of saving translates into a poverty trap. In contrast, in China and India, the Malthusian cycle was broken fairly early and institutional weaknesses played a crucial role in their respective declines. In the Americas and Australia, colonial institutions were a crucial factor. In Russia, it was the crippling political institutions of the nineteenth century and restrictive political and economic institutions of the Soviet Union that did the damage. Part II of the book focuses on empirical evidence using contemporary data. The aim is to identify policies that have been successful in delivering growth in the recent past. Chapter 6 focuses on the effectiveness of trade policy and trade openness in general. The results suggest that trade on its own is not capable of delivering growth dividends. The effect of trade is conditional on the quality of institutions. In other words, I identify that there is a threshold level of institutional quality below which trade is not beneficial for growth. I also present a list of countries that are most likely to benefit from trade at the margin, given their institutional quality. In Chapter 7, I deal with a related question. What is the likelihood of improving institutional quality through trade? I empirically show that indeed institutional quality improves with trade. The estimates suggest that if the trade liberalization index in Myanmar (almost a closed economy) was to increase by 0.5, then the corresponding increase in the property rights institutions score would be 37 percentage points. I also find similar results for contracting and regulatory institutions. The effect could work through realignment of local political forces. Alternatively, it could also work through institutional transfer. Chapter 8 explores the empirical question, which institutions matter for growth. I use a four-way classification of institutions (market-creating, market-regulating, market-stabilizing, and market-legitimizing institutions) and identifying a proxy for each of them. Market-creating institutions are characterized by the adequate protection of private property and contract enforcement. The main role of market-stabilizing institutions is to ensure macroeconomic stability and avoid undertaking policies that lead to market distortions. Market-regulating institutions tackle market failure and restrict anti-competitive behaviour by firms. Market-stabilizing

M2545 - BHATTACHARYYA PRINT.indd 167

02/03/2011 10:09

168

Growth miracles and growth debacles

institutions, on the other hand, take care of redistributive policies to make sure that the benefits of growth are distributed fairly across society. The four-way classification allows me to unbundle institutions and take the analysis beyond property rights and contracts. I estimate the contributions of market-creating, market-regulating, market-stabilizing and marketlegitimizing institutions to growth. I find that strong market-creating institutions and market-stabilizing institutions are good for growth. There exists a growth-maximizing level of market regulation beyond which red tape creates disincentives for investment. Market-legitimizing institutions do not seem to matter for growth. In summary, we have learned that institutions and geography are important factors influencing economic development over the very long run. Evidence suggests that geography is particularly important in an early stage of development. In fact, in Africa, malaria is the only significant explanator of lack of development. We also learn that policies such as trade openness could work only in an environment where institutions are strong. However, trade has the potential to influence institutional quality. Even though informative, the empirical results reported and discussed, so far, are average trends. Therefore, they may not be accurate in individual country situations. In order to inform policymakers, one needs to go beyond the average results of cross-national regressions and focus on details. This is what I aim to do in Section 9.2. I start with a discussion of the first principles of institutions and policies for development. Then I discuss aberrations and what could be done in such situations.

9.2 POLICIES FOR THE FUTURE 9.2.1

First Principles: Setting up Conditions for Growth

Property rights Protection of private property is essential for the development of any entrepreneurial activity. In situations where private property rights are weak or non-existent, it is almost impossible to attract investments from private investors. Predatory behaviour either by the state or other non-state actors on private investment creates uncertainty for the investors. The risk of predation makes the cost of investment prohibitively high. This leads to very low levels of investment and poor growth outcomes. Therefore, it is absolutely essential that private property rights are guaranteed by law and also properly enforced by the government in case of disputes. In many instances, the law guaranteeing property rights may exist, but predation by private individuals or non-state actors, or expropriation by the state

M2545 - BHATTACHARYYA PRINT.indd 168

02/03/2011 10:09

Making policy work

169

or bureaucrats may be rife. In such instances, long-term growth outcomes may be poor because of investment risks. As a result, the true economic potential of that country may not be realized. Rule of law Like property rights, the rule of law is also a key ingredient for growth. The rule of law ensures that private individuals are protected by law and makes it difficult for a state to actively engage in expropriation or predatory behaviour. In other words, it makes it difficult for powerful individuals within the state to unlawfully use the state apparatus against private individuals. If such a situation arises, then there are provisions to hold that individual to account in a court of law. Therefore, it provides an important layer of protection for investors. In the absence of rule of law, investment risks multiply and investors only invest in low-risk low-return projects (for example, in the subsistence sector). Furthermore, violent conflict may arise making investments unworthy. Therefore, the upshot is a poor growth outcome in the absence of the rule of law. Contract enforcement A strong contracting environment is essential not only for investments but also for the development of financial markets. Business and enterprise flourish in an environment where it is easier for two completely unrelated private individuals to draft contracts and take action in situations where these contracts are reneged. The law of the land should facilitate this process. The easier and faster the process, the better it is for new investments and innovations. In situations where this is not the case, growth dividends are relatively smaller. However, one could argue that formal contracting institutions are redundant in an environment of high social capital and trust. In reality, trust and social capital could operate within relatively small groups. Therefore, the cost of seeking a contract or getting access to credit becomes difficult for individuals who are not part of this network. In such an environment, potentially high pay-off projects may fail to get funded if they come from an individual who is not part of the network. Hence, by construction such institutions limit growth as they fail to ensure free entry. Regulatory institutions Strong regulatory institutions are important for growth. Private economic agents in the market are driven by greed and animal spirit. Therefore, if left unregulated, economic agents may engage in predatory and anticompetitive behaviour, which may lead to market failure and net welfare loss to society. Recent events in financial markets across the world, which

M2545 - BHATTACHARYYA PRINT.indd 169

02/03/2011 10:09

170

Growth miracles and growth debacles

led to the global financial crisis, underscore the importance of regulation. Regulators in the previous decade were increasingly of the view that selfregulation is Pareto superior to regulation by the state or independent bodies. In retrospect, one could now clearly identify that error. Therefore, it is important to put in place independent regulatory bodies overseeing the behaviour of market actors to limit anti-competitive behaviour, monopoly practices and collusion. It is important for the regulators to be independent of the government so that they are relatively free from political pressures and also independent of market actors so that they are relatively free from industry pressures. A strong regulatory environment may foster investments and economic growth. However, there is a risk of over-regulation. Excessive regulation, on the other hand, may lead to red tape, increased cost of investment and low growth. Therefore, in an ideal world, regulatory institutions should optimize the level of regulation. Macroeconomic stabilization Macroeconomic stabilization is crucial for growth over the long term. Investment, by definition, is a risky activity. In most cases, there is a gap between investment in a project and the point in time when it starts delivering returns. This is commonly known as the gestation gap. In an environment of macroeconomic volatility, it becomes difficult for investors to foresee costs and future returns. Price volatility could impact on both input costs as well as future returns. Interest rate volatility could also impact on costs through the credit channel. Therefore, it is crucial for macroeconomic institutions such as the central bank and the treasury to deliver a stable macroeconomic environment for growth. Central banks are best able to do that when they are independent. In most successful economies, central banks are mandated to tackle inflation by using interest rates as an instrument. However, government borrowing also has an impact on interest rate. If a government is borrowing from the market for government spending, it is likely to make credit dearer and increase the interest rate. This would have an impact on the balance sheet of private businesses seeking credit from the market. The cost of credit for private businesses would go up, crowding out investments. Therefore, the treasury’s actions would also have an impact on investments and the economy. To sum it up, the use of a central bank’s monetary policy to keep inflation low and the treasury’s efforts to keep debt low are important for a stable macroeconomic environment and growth. Representative politics The process of economic growth can be harsh at times. Growth in a capitalist economy, by its nature, picks winners and losers. Rapid growth

M2545 - BHATTACHARYYA PRINT.indd 170

02/03/2011 10:09

Making policy work

171

driven by technological progress or expansion of a particular sector in an economy might make other sectors unprofitable. This would lead to redundancies, unemployment and factory closures. Therefore, redistribution would be an important policy tool to help the unemployed receive training and re-engage with the workforce. Examples of such support mechanisms are social safety nets, unemployment benefits and social insurance. These policies are capable of minimizing social unrest and help to maintain a climate conducive to further investments. Representative political institutions are best to deliver on redistribution. Some have argued that democratic institutions are the most efficient form of representative political institution. However, there are examples of other types of institutions that are also capable of delivering redistributive outcomes as well as a democracy. Of course, China comes to mind. Even though inequality in China has risen over the last three decades, significant steps have been taken by the Chinese authorities to reduce the gap through retraining and re-employing the unemployed workforce, allowing migration and providing unemployment benefit. Nevertheless, it is perhaps fair to conclude that the more representative the political institutions are, the better it is for redistribution. Human capital investments Having an educated population is also crucial for growth. Without an educated population it would be difficult to properly run institutions. Running institutions would require properly trained bureaucrats. With a low level of schooling in a country, there would be a limited supply of properly trained bureaucrats to run the institutions important for growth. It would also be difficult for businesses to recruit an adequately skilled workforce. This may lead to low levels of investment and also investments in low-return projects. On the other hand, in a country with high levels of schooling, the workforce would be skilled enough to handle multiple tasks and hence would be much more employable. Therefore, they would be extremely attractive to investors in more than one sector. Hence, it is crucial for countries to improve the level of schooling in order to run national institutions, attract investments and deliver economic growth. Access to markets Infrastructure bottlenecks in developing countries often scuttle economic growth. Lack of roads, railway links, seaports and ocean-navigable waterways limit access to markets and thereby restrict growth. Landlocked economies in Africa and Central Asia suffer from the lack of access to international trade routes. Lack of telecommunication networks also limit business growth and reduce productivity as sellers struggle to contact

M2545 - BHATTACHARYYA PRINT.indd 171

02/03/2011 10:09

172

Growth miracles and growth debacles

potential buyers of their products. This could cost economies dearly in terms of future economic growth. In contrast, an economy well connected to the global markets both physically as well as through telecommunication networks can engage in internal trade and international trade much more easily. The positive externalities from a functioning road and railway network could be large. They facilitate institutional delivery at the local level, in addition to helping business and trade. It makes it much easier for the central government to keep in contact with local level institutions and deliver public goods such as health services and materials for schools effectively. Therefore, infrastructure investments to overcome geographic bottlenecks are essential for growth. International trade International trade increases the size of the market. It can also induce technology transfer and transfer of institutions. All of these factors are potentially growth enhancing. Therefore, trade in general is believed to be good for growth even though hard evidence from macro data is difficult to find. However, there is a caveat for developing economies. Trade and thereby specialization in one or two products may make a developing economy vulnerable to volatility in international prices. If the specialization is in primary products and natural resources, the risk of Dutch Disease also multiplies. In such a situation, a developing economy may not benefit from trade over the long term. It may also have an adverse impact on its institutions as natural resource exports are likely to increase inequality and increase social friction. Diversification, expanding export product mix and curbing over-reliance on natural resource exports are perhaps the best way forward in such situations. Therefore, trade openness is perhaps best for developing countries when the aim is to diversify their exports. 9.2.2

Recent Success Stories and First Principles

Recent growth successes in China and India have allowed these countries to significantly reduce absolute poverty and improve the quality of life for millions. How much of their success relies upon the first principles described above? This could be useful for countries that have not been able to enjoy such success so far. In this section, I make an attempt to find out. First I discuss the case of China. This is followed by the case of India. China The Chinese economy has grown rapidly over the last three decades. In particular, China has been the fastest growing economy since the 1990s.

M2545 - BHATTACHARYYA PRINT.indd 172

02/03/2011 10:09

Making policy work

173

What is the reason behind China’s phenomenal growth? Many economists argue that trade orientation and FDI are central to China’s success. However China’s economic reforms cover something more than that. China embarked on economic reforms in 1978, led by Deng Xiaoping, the then leader of the Communist Party. The main idea was to generate enough economic surplus to eradicate imbalances across regions in the mainland Chinese economy. To this effect, the Chinese government embarked on a large-scale modernization project for the economy. The first step was to institute household responsibility systems in the farms. The idea was to transfer responsibility for cultivation and, in effect, property rights from the Soviet-style collective farms to the households. Under this system, the farmers were also allowed to sell their surplus crop on the open market. This immediately delivered significant productivity dividends to Chinese agriculture. This reform was followed by the establishment of Township Village Enterprises (TVEs). TVEs were mainly small- to medium-sized industries owned by townships and villages. TVEs were also a success story. Throughout this period, the Chinese government invested in modernizing the infrastructure. New roads, highways, airports, sea ports and telecommunication systems were being built. During the late 1980s, the Chinese government introduced an open door policy towards international trade and FDIs. Export processing zones were set up along the coast and foreign manufacturing firms were allowed to set up factories using inexpensive Chinese labour. This turned out to be huge success and helped lift millions out of poverty. As a result, the Chinese economy, at present, is dominated by the private sector. However, state controlled enterprises dominate utilities and the resources sector. Without doubt, China did not follow the first principles of growth to the letter. There were no private property rights in China. Property rights in TVEs and household run farms were shared. Nevertheless, it managed to create enough incentives for long-term investments. The emphasis on FDI in manufacturing to exploit low labour costs was also successful and generated employment opportunities for millions. This was made possible because of an unwritten guarantee of non-predatory behaviour by the state. Even though there is little rule of law in the Western sense of the word, the Chinese state has developed a symbiotic relationship with the multinationals investing in China through mutual trust. Economic progress is crucial for the legitimacy of the one-party Chinese state and the state is well aware that the multinational firms are important contributors to that process. Furthermore, Chinese political institutions remained far from democratic throughout this process. However, they are likely to remain reasonably representative as long as the Communist Party is in touch with what is happening on the ground. In spite of rising inequality,

M2545 - BHATTACHARYYA PRINT.indd 173

02/03/2011 10:09

174

Growth miracles and growth debacles

it is perhaps fair to say that the Chinese state has managed it effectively through redistribution programmes and minimized social unrest. Therefore, what we notice in retrospect is that the Chinese Communist Party perhaps has used the first principles as a guideline and not as gospel. In the process, they have developed institutions which are rooted in local customs and culture and best represent local conditions to deliver economic growth. India India’s economic reforms started in the 1980s with privatization of some state-owned enterprises and unilateral reductions in tariffs and quantitative restrictions. However, the reform process gained momentum in 1991 following a debt crisis. Growth in per capita income experienced a significant boost following the 1991 reforms. However, the institutional change, on paper, was not significant. Post-independence India always had property rights and laws to enforce private contracts. She also had the rule of law and constitutional rights for individuals to buy and sell goods and services on the open market. However the sociopolitical environment was not pro business till the late 1980s. A commonly held view amongst the public and politicians pre-reform was that a socialist system is best suited for India to deliver shared prosperity. As a result, the economy was held hostage to numerous regulations and licensing procedures commonly known as the ‘licence raj’. With reforms, the entire discourse shifted and the government and its institutions became more pro business. This delivered improvements in institutional quality and rapid growth. The economy experienced expansion in both manufacturing and services at the expense of agriculture. However, problems remained at the front of public goods delivery and infrastructure. With a rapidly expanding economy, the demand for infrastructure and other public goods quadrupled. Laying out infrastructure efficiently within the current institutional framework is always a challenge in India. The rapidly growing demand for land to lay out infrastructure and other projects in a growing economy, such as India, is often faced with challenges of land disputes. The sheer number of these disputes poses challenges for the court system and other institutions. Nevertheless, in spite of all these hurdles, India’s progress over the last two decades has been impressive. In summary, the paths of development chosen by India and China are distinctly different. Both countries have successfully utilized local conditions and developed institutions that are well aligned with their initial conditions. However, there is little debate on the fact that they are built on the appropriate foundation of incentives. In other words, even though the first principles are not applied to the letter in both of these countries, they are applied to create the right incentives for growth.

M2545 - BHATTACHARYYA PRINT.indd 174

02/03/2011 10:09

Making policy work

9.2.3

175

Growth Amidst Chaos: The Case for Multilateral Interventions

A prerequisite for economic growth is a functioning state which is able to tax its citizens and provide public goods in return. Therefore, it is extremely difficult to start growth in an environment of state failure and chaos. In recent times, conflict-ridden parts of Africa, Central Asia and the Middle East have experienced such situations. The principal challenge in such situations is to form a functioning state. In this section, I analyse whether multilateral interventions can help in a situation of state failure. Internal conflict and lack of effective government are a serious developmental challenge in many parts of Africa, Central Asia and the Middle East. What is the best possible way to react in such situations so that an effective government can be formed? History tells us that conflict and often violent conflict is integral to institution formation. Conflict shapes the distribution of economic and political power, which determines institutional outcomes over the long term. One could argue that the contemporary economic and political institutions in Europe are a product of two centuries of internal as well as external conflict. The North American economic as well as political institutions are an outcome of the American War of Independence and the American Civil War. Many historians argue that contemporary Australian political institutions also have their roots in the American War of Independence (see Section 5.2). British colonizers gave in to democratic pressures in Australia because of their fear of a repeat of the North American war and losing the Australian colonies. However, violent conflict is not an option in the contemporary world as it leads to loss of life and property damage. In most cases, what we observe are interventions by the international community to contain violence and build peace. Forming a functioning government and economic reconstruction are integral parts of that process. The key question, therefore, is how to make interventions work. Intervention, by definition, is risky. It distorts the distribution of political and economic power in a country across different power groups by bringing in an external player. Often the external players are not impartial and are driven by their own security and economic interests. This has an influence on local distribution of power and makes institution building even more challenging. In order to form an effective government and other ancillary institutions, all these interests need to converge. This is, of course, easier said than done. In situations when they do not converge, interventions may end up making the situation even more unstable and prolonging the conflict. Without going into detail, in the following paragraphs I construct a list of strategies that may be useful in such situations. However, note that these strategies are far from comprehensive and are perhaps a topic of research for another book.

M2545 - BHATTACHARYYA PRINT.indd 175

02/03/2011 10:09

176

Growth miracles and growth debacles

Ideally, institution building in a conflict-ridden country should be based on representative politics and also raising the cost of conflict for the parties involved. This could be done by making the state as representative as possible and also providing military support to the state. Multilateral military presence may have an advantage over unilateral military presence. Both the locals and the international community may perceive multilateral military presence as more legitimate than unilateral military presence. While constructing a state, legitimacy is a key issue. At the local level, efforts should be made so that all sections of society are well represented in any form of government. This, on one hand, may involve creating incentives for warring groups to work together and, on the other hand, may seek representation of all ethnic groups and tribes in the government. The more broad-based the government is, perhaps the better it is for legitimacy. Efforts should be made to create a fair tax system and provide services such as policing. This may also help the quest for legitimacy of any young government. Multilateral interventions may have advantages over unilateral interventions for the following three reasons. First, the perceived legitimacy of multilateral interventions is, in general, higher than country A- or country B-led interventions. Second, multilateral interventions are more likely to be relatively impartial and, as a result, any distortions in the local distribution of power would be less damaging. At least, distortions would not be driven by security or economic interests of any individual intervening country. This would also help achieve legitimacy. Third, multilateral interventions guided by institutions such as the United Nations are more likely to be willing to make long-term commitments, if required. This is unlikely to be the case with unilateral interventions as the government involved is more likely to be driven by their own country’s short-term political cycle. One could, of course, point towards British success in Sierra Leone as an example of successful unilateral intervention. But these examples are extremely rare and multilateral interventions are perhaps the best way forward. Nevertheless, there is no denying the fact that these situations are extremely challenging and very difficult to handle. It is also important to note that these efforts are extremely costly in terms of human lives and require long-term commitment. 9.2.4

Diseases and Foreign Aid

Debilitating and killer diseases often limit an individual’s ability to work. Therefore, they have a direct effect on productivity and economic growth. In a poor country, often the economic agent may not have the resources to fight such conditions on their own. Fiscally strapped poor country

M2545 - BHATTACHARYYA PRINT.indd 176

02/03/2011 10:09

Making policy work

177

governments also may not have the resources to engage in large-scale investment in public goods to tackle such conditions. Therefore, what ensues is a poverty trap. Foreign aid is vital in such situations. The dominance of disease, especially malaria, in Africa described in Chapters 4 and 5 does make a strong case for aid. However, aid itself in these circumstances is not an engine of growth. Aid should be used in a targeted fashion to tackle diseases and create a healthy workforce capable of undertaking work in the agriculture or manufacturing sectors. Injection of foreign aid to improve health should be combined with creating the right incentives for investments in sectors that have potential for growth. Some have argued in favour of a big-push-style policy of aid to pull poor countries out of the poverty trap. Any big-push-style policy relies on planning and any planning exercise disregards the often unsolvable information and incentive problems attached to it (Easterly, 2006). In particular, the incentive of the aid worker and the information problems associated with aid infrastructure are often overlooked by macro-level planning. Therefore, a step by step approach to create the right incentives for investments seems to be the appropriate way forward, rather than a big-push-style approach funded by foreign aid. 9.2.5

The Role of Technical Assistance

Technical assistance is crucial for the development of a well-functioning bureaucracy in a developing country. Countries in their early stage of development may not have the talent pool to choose from for recruitment of government officials, university teachers, school teachers and police officers. These services are essential for economic growth and also to supply the next generation of smart individuals in the country capable of tackling governance and creating a favourable environment for business. Technical assistance from developed countries may go a long way in helping young countries and governments to address these issues. Creating university places for developing countries’ citizens and also creating training opportunities for their government officials are desirable. This would boost the countries’ ability to provide public goods in an efficient manner, which is vital for economic growth.

9.3 EPILOGUE This book is an attempt to understand the process of economic progress and why it varied across countries and continents over the last two centuries. The principal message is that history and geography are important

M2545 - BHATTACHARYYA PRINT.indd 177

02/03/2011 10:09

178

Growth miracles and growth debacles

in explaining this difference. The limited data that we have indicate that economic progress is an outcome of a combination of multiple factors over a very long period of time. Geography shapes history and also shapes incentives that affect economic development. Geography also affects development directly. It determines factor endowments and natural resource availability. Incentives remain crucial for economic development, and geography and resource endowment often set the tone of what type of incentives are going to evolve over the long term. The historical narratives chapters (Part I) show that these incentives evolved in a varied way across continents and countries, which gave rise to different types of institutions. Hence, the development outcome varies across continents. The second part of the book focuses on methods that could be used to influence incentives and institutions. First, it documents some macro evidence on the role of policy in shaping these incentives. It shows that trade can benefit nations in situations where institutions are adequate. Property rights and contracting institutions are good for growth. Market-stabilizing institutions are good for growth and regulation is important only to a certain extent. Second, it also outlines policies that may be beneficial in promoting growth. It outlines the ‘first principles of growth’ namely property rights, contracts, regulatory institutions, the rule of law, macroeconomic stabilization, representative politics, human capital investments, market access and international trade. Third, it describes the cases of India and China, two recent success stories. It shows how these countries have preserved incentives for private investments even without following the first principles. But more importantly they have been able to create institutions which are well grounded in local traditions and culture and also able to create appropriate incentives for investments. Fourth, the book also outlines steps that could be taken to facilitate growth in situations of state failure, disease trap, poverty trap and scarcity of skilled workers. The main contribution of the book is to show that economic growth in the past has been a product of multiple factors. Institutions and geography are perhaps the most important ones that create incentives for investment and growth. Theories are presented and empirical results are discussed to argue in favour of such a viewpoint. It also provides a list of policies that countries could follow to bring about growth and reduce poverty. Of course, the book does not provide us with a magic formula to bridge the gap between the rich and the poor. However, it does flag the importance of history, geography, politics and culture in economic development. It also presents a strong case for studying these issues in combination with each other and not as disparate theories. A better understanding of the interconnection between these factors may increase our knowledge of how countries grow in the future.

M2545 - BHATTACHARYYA PRINT.indd 178

02/03/2011 10:09

Data appendix CHAPTER 4 Measures of Economic Development Log per capita GDP in 2000: Natural log of real GDP per capita in 2000. Real GDP figures are measured in US$ in current prices and the figures are PPP converted. For Botswana, Cambodia, Fiji, Guyana, Mauritania, Namibia and Papua New Guinea, I use 1999 values as an approximation. For Central African Republic, Haiti, Puerto Rico and Taiwan, I use 1998 values. Source: Penn World Table (PWT) 6.1, Heston et al. (2002). Log initial income (1820): Natural logs of per capita GDP (1820) in 1990 international Geary-Khamis dollars. Source: Maddison (2004). Log initial income (1870): Natural logs of per capita GDP (1870) in 1990 international Geary-Khamis dollars. Source: Maddison (2004). Log initial income (1900): Natural logs of per capita GDP (1900) in 1990 international Geary-Khamis dollars. Source: Maddison (2004). Log initial income (1950): Natural logs of per capita GDP (1950) in 1990 international Geary-Khamis dollars. Source: Maddison (2004). Log initial income (1960): Initial level of per capita GDP (1960) in natural logs and PPP figures. Source: Penn World Table (PWT) 6.1, Heston et al. (2002). Measure of Institution Average rule of law index: This variable measures the quality of public service provision, the quality of the bureaucracy, the competence of civil servants, the independence of the civil service from political pressures, and the credibility of the government’s commitment to policies. The main focus of this index is on ‘inputs’ required for the government to be able to produce and implement good policies and deliver public goods. This variable ranges from 22.5 to 2.5, where higher values equal higher government effectiveness. This variable is measured as the average from 1998 through 2000. Source: Kaufman et al. (2002).

179

M2545 - BHATTACHARYYA PRINT.indd 179

02/03/2011 10:09

180

Growth miracles and growth debacles

Measures of Religion Catholicism: Identifies the percentage of population of each country that is Catholic in 1980. Source: LaPorta et al. (1999). Islam: Identifies the percentage of population of each country that was Muslim in 1980. Source: LaPorta et al. (1999). Measure of Openness and Trade Log of trade share: Natural log of trade share calculated by taking log values of figure obtained by dividing volume of trade with GDP. Source: Frankel and Romer (1999). Measure of Human Capital Enrolment ratio in 1900: It is the ratio of the number of students enrolled at the primary level and the relevant school age population. Source: Benavot and Riddle (1988). Measures of Geography Distance: Absolute distance from the equator measured in latitude. Source: Hall and Jones (1999). Malaria risk: Percentage of the population at risk of malaria transmission in 1994. Source: Glaeser et al. (2004). Land area within tropics: The proportion of a country’s land area within the geographical tropics. Source: Gallup et al. (1998). Land area within 100 km of ocean or ocean-navigable river: Proportion of country’s total land area within 100 km of the ocean or ocean-navigable river, excluding coastline above the winter extent of sea ice and the rivers that flow to this coastline. Source: Gallup et al. (1998).

SECTION 4.6 Log per capita GDP in 2000 (log yi): Penn World Table (PWT) 6.1. Log income in 2000: Nunn (2008), originally from Maddison (2004). Expropriation risk (INSi): Risk of ‘outright confiscation and forced nationalization’ of property. Source: ICRG. Executive constraint in 2000: A seven category scale, 1 to 7, with a higher score indicating more constraint. Source: Polity IV. Rule of law index: See Rodrik et al. (2004) for details.

M2545 - BHATTACHARYYA PRINT.indd 180

02/03/2011 10:09

Data appendix

181

Pre-colonial state development: Nunn (2008). Malaria risk: Percentage of the population at risk of malaria transmission in 1994. Source: CID datasets, Harvard University. Malaria ecology (ME): Kiszewski et al. (2004). Log total slave exports normalized by land area (SLVXi): See Nunn (2008). Log total slave exports normalized by population: See Nunn (2008). Log settler mortality (LSM): Acemoglu et al. (2001). Log population density in 1500 (LPD): Acemoglu et al. (2001). Interior distance (IDCi), Atlantic distance (ADCi), Indian distance (IODCi), Saharan distance (SDCi) and Red Sea distance (RDCi): Nunn (2008). Frost: Masters and McMillan (2001), see Carstensen and Gundlach (2006) for details. Rain: Minimum of monthly average rainfall, Nunn (2008). Humidity: Maximum of monthly afternoon average humidity (%), Nunn (2008). Legal origin: LaPorta et al. (1999). Schooling in 1900: Benavot and Riddle (1988). Log trade share in 2000: WDI online. CONST: Constructed openness, Frankel and Romer (1999). Ethnic fractionalization: Alesina et al. (2003). Mining: Share of mining in GDP, Hall and Jones (1999). Catholicism and Islam: LaPorta et al. (1999). Gini coefficient: World Bank. Foreign aid and national savings: WDI online. Corruption: ICRG, PRS dataset. Sachs and Warner openness: Sachs and Warner (1997).

CHAPTER 6 Dependent and Explanatory Variables Log GDP per capita (log ysrt): Natural log of GDP per capita PPP (current international dollars), CGDP. Source: Penn World Table, PWT 6.1. Trade share (TRsrt): (exports 1 imports)/GDP. Source: WDI Online, The World Bank Group. Trade policy openness since t24 (posrt): Fraction of years open between t and t24. Source: Sachs and Warner (1997). Trade policy openness since 1950 (POsrt): Fraction of years open between t and t21950. Source: Sachs and Warner (1997). Expropriation risk (INSsrt): Risk of ‘outright confiscation and forced

M2545 - BHATTACHARYYA PRINT.indd 181

02/03/2011 10:09

182

Growth miracles and growth debacles

nationalization’ of property. It covers the period 1982 to 1997. Average of 1982 to 1983 is used as a proxy for 1980, average of 1984 to 1987 as a proxy for 1985, average of 1988 to 1992 as a proxy for 1990, and average of 1993 to 1997 as a proxy for 1995. Source: ICRG. Instruments Constructed openness (CONST): Frankel and Romer (1999). Landlocked dummy: Sachs and Warner (1995). Log settler mortality (LSM): Acemoglu et al. (2001). Log population density in 1500 (LPOPDEN): Acemoglu et al. (2002). ENGFRAC: Fraction of the population speaking English. Source: Hall and Jones (1999). EURFRAC: Fraction of the population speaking Western European languages (English, French, German, Portuguese, Spanish). Source: Hall and Jones (1999). AREA: CID Harvard Geography datasets.

CHAPTER 8 Growth (yˆiT) : Calculated using the formula yˆiT ; 1/5 (yiT 2 yiT 25) for 127 countries. Source: WDI Online, The World Bank Group. Initial income (yiT 25) : Log GDP per capita PPP (constant 2000 international $). Source: WDI Online, The World Bank Group. Law and Order (LO): Source: ICRG, The PRS Group. Regulation of credit, labour and business (MR): Source: Gwartney and Lawson (2005). Sound money index (SM): Source: Gwartney and Lawson (2005). Democracy (DEMOC): Source: Polity IV. Rule of law (RULE): Source: Rodrik et al. (2004). Expropriation risk (EXPR): Source: ICRG, The PRS Group. Executive constraint (EXCONST): Source: Polity IV. Repudiation of government contracts: Source: ICRG, The PRS Group. Total years of schooling (TYS): Source: Barro and Lee (2000). Log trade share (LTRS): Source: WDI Online, The World Bank Group. Log real openness (LROPEN): Source: Alcalá and Ciccone (2004). Investment share: Source: PWT 6.2. Population growth: Source: PWT 6.2. Corruption: Source: ICRG, The PRS Group. FDI: Foreign direct investment as a share of GDP. Source: WDI Online, The World Bank Group.

M2545 - BHATTACHARYYA PRINT.indd 182

02/03/2011 10:09

Data appendix

183

Foreign aid: Foreign aid as a share of GDP. Source: WDI Online, The World Bank Group. Real exchange rate distortions: Real overvaluation. Source: WDI Online, The World Bank Group. Credit to private sector: Domestic credit to private sector as a share of GDP. Source: WDI Online, The World Bank Group.

M2545 - BHATTACHARYYA PRINT.indd 183

02/03/2011 10:09

References Acemoglu, D. and S. Johnson (2005), ‘Unbundling Institutions’, Journal of Political Economy, 113(5), 949–995. Acemoglu, D. and S. Johnson (2007), ‘Diseases and Development: The Effect of Life Expectancy on Economic Growth’, Journal of Political Economy, 115(6), 925–985. Acemoglu, D. and J. Robinson (2000), ‘Why Did the West Extend the Franchise? Democracy, Inequality and Growth in Historical Perspective’, Quarterly Journal of Economics, 115, 1167–1199. Acemoglu, D. and J. Robinson (2006), ‘Economic Origins of Dictatorship and Democracy’, New York: Cambridge University Press. Acemoglu, D., S. Johnson and J. Robinson (2001), ‘The colonial origins of comparative development: an empirical investigation’, American Economic Review, 91(5), 1369–1401. Acemoglu, D., S. Johnson and J. Robinson (2002), ‘Reversal of Fortune: Geography and Institutions in the Making of the Modern World Income Distribution’, Quarterly Journal of Economics, 117, 1231–1294. Acemoglu, D., S. Johnson and J. Robinson (2003a), ‘Disease and Development in Historical Perspective’, Journal of the European Economic Association, 1, 397–405. Acemoglu, D., S. Johnson, J. Robinson and Y. Thaicharoen (2003b), ‘Institutional Causes, Macroeconomic Symptoms: Volatility, Crises and Growth’, Journal of Monetary Economics, 50, 49–123. Acemoglu, D., S. Johnson and J. Robinson (2005a), ‘Institutions as the Fundamental Cause of Long-Run Growth’, Handbook of Economic Growth, 1, 385–472. Acemoglu, D., S. Johnson and J. Robinson (2005b), ‘The Rise of Europe: Atlantic Trade, Institutional Change and Economic Growth’, American Economic Review, 95, 546–579. Acemoglu, D., S. Johnson, J. Robinson and P. Yared (2005c), ‘From Education to Democracy?’, AEA Papers and Proceedings, 95(2), 44–49. Acemoglu, D., S. Johnson, J. Robinson and P. Yared (2006), ‘Income and Democracy’, American Economic Review, forthcoming. Acemoglu, D., S. Johnson, J. Robinson and P. Yared (2007), ‘Reevaluating the Modernization Hypothesis’, NBER Working Paper No. 13334, August. 184

M2545 - BHATTACHARYYA PRINT.indd 184

02/03/2011 10:09

References

185

Ahluwalia, M. (2002), ‘Economic Reforms in India Since 1991: Has Gradualism Worked’, Journal of Economic Perspectives, 16(7), 67–88. Albouy, D. (2008), ‘The Colonial Origins of Comparative Development: An Investigation of the Settler Mortality Data’, NBER Working Paper No. 14130, June. Alcalá, F. and A. Ciccone (2004), ‘Trade and Productivity’, Quarterly Journal of Economics, 119(2), 612–645. Alesina, A. and A. Drazen (1991), ‘Why Are Stabilizations Delayed?’, American Economic Review, LXXXI, 1170–1188. Alesina, A. and G. Tabellini (1989), ‘External Debt, Capital Flight and Political Risk’, Journal of International Economics, 27, 199–220. Alesina, A., A. Devleeschauwer, W. Easterly, S. Kurlat, and R. Wacziarg. (2003), ‘Fractionalization’, Journal of Economic Growth, 8, 155–194. Allen, R. (2008), ‘A Review of Gregory Clark’s A Farewell to Alms: A Brief Economic History of the World’, Journal of Economic Literature, 46(4), 946–973. Allen, R. (2009), ‘The British Industrial Revolution in Global Perspective’, Cambridge: Cambridge University Press. Arndt, C. and J. Lewis (2000), ‘The Macro Implications of HIV/AIDS in South Africa: A Preliminary Assessment’, South African Journal of Economics, LXVIII, 856–887. Bagchi, A. (1976a), ‘Deindustrialisation in India in the Nineteenth Century: Some Theoretical Implications’, Journal of Development Studies, 12, 135–164. Bagchi, A. (1976b), ‘Deindustrialisation in Gangetic Bihar 1809–1901’, in B. De et al. (eds), Essays in Honour of Prof. S.C. Sarkar, New Delhi: People’s Publishing House. Bairoch, P. (1982), ‘International Industrialization Levels from 1750 to 1980’, Journal of European Economic History, 11, 269–333. Bajpai, N. and J. Sachs (1999), ‘Fiscal Policy in India’s Economic Reforms’, in J. Sachs, A. Varshney, and N. Bajpai (eds), India in the Era of Economic Reforms, New Delhi: Oxford University Press. Banerjee, A. (2007), ‘Making Aid Work’, Cambridge, MA: Boston Review, MIT Press. Banerjee, A. and L. Iyer (2005), ‘History, Institutions, and Economic Performance: The Legacy of Colonial Land Tenure Systems in India’, American Economic Review, 95, 1190–1213. Barro, R. (1995), ‘Inflation and Economic Growth’, NBER Working Paper No. 5326, October. Barro, R. (1996), ‘Democracy and Growth’, Journal of Economic Growth, 1, 1–27. Barro, R. and J.W. Lee (2000), ‘International Data on Educational

M2545 - BHATTACHARYYA PRINT.indd 185

02/03/2011 10:09

186

Growth miracles and growth debacles

Attainment: Updates and Implications’, Manuscript, Harvard University. Bates, R. (1983), ‘The Centralisation of African Societies’, Essays on the Political Economy of Rural Africa, Berkeley: University of California Press, Chapter 2. Beaud, M. (2000), A History of Capitalism: 1500–2000, New York: Monthly Review Press. Behrman, J. and M. Rosenzweig (2004), ‘The Returns to Birthweight’, Review of Economics and Statistics, 86(2), 586–601. Bell, C., S. Devarajan and H. Gersbach (2003), ‘The Long-Run Economic Costs of AIDS: Theory and an Application to South Africa’, Working Paper, June. Belsley, D., E. Kuh and R. Welsch (1980), ‘Regressions Diagnostics: Identifying Influential Data and Sources of Collinearity’, New York: John Wiley & Sons. Benavot, A. and P. Riddle (1988), ‘The Expansion of Primary Education, 1870–1940: Trends and Issues’, Sociology of Education, 61(3), 191– 210. Ben-David, D. (1993), ‘Equalizing Exchange: Trade Liberalization and Income Convergence’, Quarterly Journal of Economics, 108(3), 653–679. Besley, T. and T. Persson (2008), ‘War and State Capacity’, Journal of the European Economic Association, 6(2–3), 522–530. Besley, T. and T. Persson (2009), ‘The Origins of State Capacity: Property Rights, Taxation, and Politics’, American Economic Review, 99(4), 1218–1244. Bhagwati, J. and P. Desai (1970), ‘Planning for Industrialization: A Study of India’s Trade and Industrial Policies Since 1950’, Cambridge: Cambridge University Press. Bhattacharyya, S. (2004), ‘Deep Determinants of Economic Growth’, Applied Economics Letters, 11(9), 587–590. Bhattacharyya, S. (2009a), ‘Unbundled Institutions, Human Capital, and Growth’, Journal of Comparative Economics, 37, 106–120. Bhattacharyya, S. (2009b), ‘Root Causes of African Underdevelopment’, Journal of African Economies, 18(5), 745–780. Bhattacharyya, S. (2009c), ‘Institutions, Diseases and Economic Progress: A Unified Framework’, Journal of Institutional Economics, 5(1), 65–87. Bhattacharyya, S. (2010), ‘Trade Liberalization and Institutional Development’, Unpublished Manuscript. Bhattacharyya, S. (2011), ‘Five Centuries of Economic Growth in India: The Institutions Perspective’, in R. Jha (ed.), Routledge Handbook of South Asian Economics, New York: Routledge, forthcoming.  Bhattacharyya, S., S. Dowrick and J. Golley (2009), ‘Institutions and

M2545 - BHATTACHARYYA PRINT.indd 186

02/03/2011 10:09

References

187

Trade: Competitors or Complements in Economic Development?’, Economic Record, 85(270), 318–330. Bhattacharyya, S. and R. Hodler (2010), ‘Natural Resources, Democracy and Corruption’, European Economic Review, 54, 608–621. Bhattacharyya, S. and J. Williamson (2009), ‘Commodity Price Shocks and the Australian Economy since Federation’, Australian Economic History Review, forthcoming. Bleakley, H. (2003) ‘Disease and Development: Evidence from the American South,’ Journal of the European Economic Association, 1(2–3), 376–386. Bloom, D. and D. Canning (2005), ‘Health and Economic Growth: Reconciling the Micro and Macro Evidence’, Working Paper, February. Bloom, D. and J. Sachs (1998), ‘Geography, Demography, and Economic Growth in Africa’, Brookings Papers on Economic Activity, 2, 207–295. Bloom, D., D. Canning, and J. Sevilla (2003), ‘Geography and Poverty Traps’, Journal of Economic Growth, 8, 355–378. Carstensen, K. and E. Gundlach (2006), ‘The Primacy of Institutions Reconsidered: Direct Income Effects of Malaria Prevalence’, World Bank Economic Review, 20, 309–339. Caselli, F., G. Esquivel, and F. Lefort (1996), ‘Reopening the Convergence Debate: A New Look at Cross-Country Growth Empirics’, Journal of Economic Growth, 1(3), 363–389. Chakraborty, S. and M. Das (2005), ‘Mortality, Human Capital and Persistent Inequality’, Journal of Economic Growth, 10, 159–192. Chang, H.-J. (2008), Bad Samaritans: The Myth of Free Trade and the Secret History of Capitalism, New York: Bloomsbury Press. Clark, C. (1950), Conditions of Economic Progress, London: Macmillan. Clark, G. (2007), A Farewell to Alms: A Brief Economic History of the World, Princeton and Oxford: Princeton University Press. Clingingsmith, D., and J. Williamson (2008), ‘Deindustrialization in 18th and 19th century India: Mughal decline, climate shocks and British industrial ascent’, Explorations in Economic History, 45, 209–234. Coe, D., and E. Helpman (1995), ‘International R&D Spillovers’, European Economic Review, 39(5), 859–887. Collier, P. (2007), The Bottom Billion: Why the Poorest Countries Are Failing and What Can Be Done About It, New York: Oxford University Press. Collier, P. and J.W. Gunning (1999), ‘Explaining African Economic Performance’, Journal of Economic Literature, 37(1), 64–111. Curtin, P. (1968), ‘Epidemiology and the Slave Trade’, Political Science Quarterly, 83(2), 181–216. Curtin, P. (1975), ‘Economic Change in Precolonial Africa: Senegambia in the Era of the Slave Trade’, Madison: University of Wisconsin Press.

M2545 - BHATTACHARYYA PRINT.indd 187

02/03/2011 10:09

188

Growth miracles and growth debacles

De Long, B. (2003), ‘India Since Independence: An Analytical Growth Narrative’, in D. Rodrik (ed.), In Search of Prosperity: Analytical Narratives on Economic Growth, Princeton, NJ: Princeton University Press, Princeton. Diamond, J. (1997), Guns, Germs and Steel: The Fate of Human Societies, New York: W.W. Norton & Co. Dias, J. (1981), ‘Famine and Disease in the History of Angola, 1830–1930’, Journal of African History, 21, 349–378. Djankov, S., R. LaPorta, F. Lopez-de-Silanes, and A. Shleifer (2002), ‘The Regulation of Entry’, Quarterly Journal of Economics, 117(1), 1–37. Dollar, D. (1992), ‘Outward Oriented Developing Economies Really Do Grow More Rapidly: Evidence from 95 LDCs, 1976–1985’, Economic Development and Cultural Change, 40, 523–544. Dollar, D., and A. Kraay (2003), ‘Institutions, Trade, and Growth’, Journal of Monetary Economics, 50, 133–162. Dowrick, S. and M. Rogers (2002), ‘Classical and Technological Convergence: Beyond the Solow-Swan Growth Model’, Oxford Economic Papers, 54, 369–385. Dutt, A. (1992), The Origins of Uneven Development: The Indian Subcontinent, American Economic Review, 82, 146–150. Easterly, W. (2001), The Elusive Quest for Growth: Economists’ Adventures and Misadventures in the Tropics, Cambridge, MA: The MIT Press. Easterly, W. (2006), ‘The Big Push Déjà Vu: A Review of Jeffrey Sachs’s The End of Poverty: Economic Possibilities for Our Time’, Journal of Economic Literature, XLIV, 96–105. Easterly, W. (2009), ‘The Anarchy of Success’, New York Review of Books, LVI(15), 28–30. Easterly, W. and R. Levine (1997), ‘Africa’s Growth Tragedy: Policies and Ethnic Divisions’, Quarterly Journal of Economics, 112(4), 1203–1250. Easterly, W. and R. Levine (2001), ‘It’s Not Factor Accumulation: Stylized Facts and Growth Models’, World Bank Economic Review, 15(2), 177–219. Easterly, W. and R. Levine (2003), ‘Tropics, Germs, and Crops: How Endowments Influence Economic Development’, Journal of Monetary Economics, 50(1), 3–39. Ebrey, P. (1999), The Cambridge Illustrated History of China, Cambridge: Cambridge University Press. Ebrey, P., A. Walthall and J. Palais (2006), East Asia: A Cultural, Social, and Political History, Boston: Houghton Mifflin. Engerman, S. and K. Sokoloff (1994), ‘Factor Endowments: Institutions, and Differential Paths of Growth Among New World Economies: A

M2545 - BHATTACHARYYA PRINT.indd 188

02/03/2011 10:09

References

189

View from Economic Historians of the United States’, NBER Working Paper No.w066. Engerman, S. and K. Sokoloff (2001), ‘Inequality, Institutions, and Differential Paths of Growth Among New World Economies’, Working Paper, University of California, Los Angeles. Fage, J. (1978), A History of Africa, Hutchinson & Co: London. Frankel, J. and D. Romer (1999), ‘Does trade cause growth?’, American Economic Review, 89(3), 379–399. Gallup, J. and J. Sachs (2000), ‘Agriculture, Climate, and Technology: Why are the Tropics Falling Behind?’, American Journal of Agricultural Economics, 82, 731–737. Gallup, J. and J. Sachs (2001), ‘The Economic Burden of Malaria’, American Journal of Tropical Medicine and Hygiene, 64(1–2), 85–96. Gallup, J., J. Sachs and A. Mellinger (1998), ‘Geography and Economic Development’, NBER Working Paper No.w6849, December. Galor, O. and O. Moav (2006), ‘Das Human-Kapital: A Theory of the Demise of the Class Structure,’ Review of Economic Studies, 73, 85– 117. Gemery, H. and J. Hogendorn (1979), ‘The Economic Costs of West African Participation in the Atlantic Slave Trade: A Preliminary Sampling for the Eighteenth Century’, in H. Gemery, and J. Hogendorn (eds), The Uncommon Market: Essays in the Economic History of the Atlantic Slave Trade, New York: Academic Press. Giavazzi, F. and G. Tabellini (2005), ‘Economic and political liberalizations’, Journal of Monetary Economics, 52, 1297–1330. Glaeser, E., R. LaPorta, F. Lopez-de-Silanes and A. Shleifer (2004), ‘Do Institutions Cause Growth?’, Journal of Economic Growth, 9, 271–303. Grief, A. (1994), ‘Cultural Beliefs and the Organization of Society: A Historical and Theoretical Reflection on Collectivist and Individualist Societies’, Journal of Political Economy, 102(5), 912–950. Gwartney, J., and R. Lawson, (2005), Economic Freedom of the World: 2005 Annual Report, Vancouver: The Fraser Institute. Data retrieved from www.freetheworld.com. Haber, S. (2006), ‘It Wasn’t All Prebisch’s Fault: The Political Economy of Latin American Industrialization’, in Victor Bulmer-Thomas, John Coatsworth, and Roberto Cortes Conde (eds), The Cambridge Economic History of Latin America, Cambridge: Cambridge University Press. Hall, A., and F. Peixe (2000), ‘A consistent method for the selection of relevant instruments’, Econometric Society World Congress. Hall, R., and C. Jones (1999), ‘Why do some countries produce so much more output per worker than others?’, Quarterly Journal of Economics, 114(1), 83–116.

M2545 - BHATTACHARYYA PRINT.indd 189

02/03/2011 10:09

190

Growth miracles and growth debacles

Hauk, W., Jr., and R. Wacziarg (2004), ‘A Monte Carlo Study of Growth Regressions’, NBER Working Paper No. T0296, January. Hausman, J., J. Stock, and M. Yogo (2005), ‘Asymptotic Properties of the Hahn-Hausman Test for Weak Instruments’, Economics Letters, 89(3), 333–342. Herbst, J. (2000), ‘States and Power in Africa: Comparative Lessons in Authority and Control’, Princeton: Princeton University Press. Heston, A., R. Summers and B. Aten (2002), ‘Penn World Table Version 6.1’, Center for International Comparisons (CICUP), University of Pennsylvania, October. Howitt, P. (2000), ‘Endogenous Growth and Cross-Country Income Differences’, American Economic Review, 90(4), 829–846. Inikori, J. (1977), ‘The Import of Firearms into West Africa, 1750–1807: A Quantitative Analysis’, Journal of African History, 18, 339–368. Inikori, J. (1992), The Chaining of a Continent: Export Demand for Captives and the History of Africa South of the Sahara, 1450–1870, Mona, Jamaica: Institute of Social and Economic Research, University of the West Indies. Inikori, J. (2000), ‘Africa and the Trans-Atlantic Slave Trade’, In T. Falola (eds), Africa Volume 1: African History before 1885, Durham, North Carolina: Carolina Academic Press. Jones, C. (1995), ‘Time Series Tests of Endogenous Growth Models’, Quarterly Journal of Economics, 110(2), 495–525. Jones, E. (1981), The European Miracle: Environments, Economies and Geopolitics in the History of Europe and Asia, Cambridge: Cambridge University Press. Kalemli-Ozcan, S. (2002), ‘Does the Mortality Decline Promote Economic Growth?’, Journal of Economic Growth, 7(4), 311–349. Kalemli-Ozcan, S. (2006), ‘AIDS, Reversal of the Demographic Transition and Economic Development: Evidence from Africa’, University of Houston, mimeo. Kalemli-Ozcan, S., H. Ryder, and D. Weil (2000), ‘Mortality Decline, Human Capital Investment, and Economic Growth’, Journal of Development Economics, 62, 1–23. Kaufmann, D., A. Kraay and P. Zoido-Lobaton (2002), ‘Governance matters II – updated indicators for 2000/01’, World Bank Policy Research Department Working Paper No. 2772, Washington D.C. King, R., and R. Levine (1994), ‘Capital Fundamentalism, Economic Development, and Economic Growth’, Carnegie-Rochester Conference Series on Public Policy, 40, 259–292. Kiszewski, A., A. Mellinger, A. Spielman, P. Malaney, S. Sachs and J. Sachs (2004), ‘A Global Index Representing the Stability of Malaria

M2545 - BHATTACHARYYA PRINT.indd 190

02/03/2011 10:09

References

191

Transmission’, American Journal of Tropical Medicine and Hygiene, 70(5), 486–498. Klein, M. (2001), ‘The Slave Trade and Decentralized Societies’, Journal of African History, 42, 49–65. Knack, S., and P. Keefer (1995), ‘Institutions and Economic Performance: Cross-country Tests using Alternative Institutional Measures’, Economics and Politics, 7(3), 207–227. Knaul, F. (2000), ‘Health, Nutrition, and Wages: Age at Menarche and Earnings in Mexico’, in W. Savedoff and T. Schultz (eds), Wealth from Health: Linking Social Investments to Earnings in Latin America, Washington, DC: Inter-American Development Bank. Kohli, A. (1989), ‘Politics of Economic Liberalization in India’, World Development, 17(3), 305–328. Krueger, A. (1974), ‘The Political Economy of the Rent-Seeking Society’, American Economic Review, 64(3), 291–303. Kuran, T. (1997), Private Truths, Public Lies: The Social Consequences of Preference Falsification, Cambridge: MA: Harvard University Press. Landes, D. (1998), The Wealth and Poverty of Nations: Why Some Are So Rich and Some So Poor, London: Abacus. Landes, D. (2000), ‘Culture Makes Almost all the Difference’, in L. Harrison and S. Huntington (eds), Culture Matters: How Values Shape Human Progress, New York: Basic Books. Landes, D. (2003), The Unbound Prometheus: Technological Change and Industrial Development in Western Europe from 1750 to the Present, Cambridge: Cambridge University Press. LaPorta, R., F. Lopez-de-Silanes, A. Shleifer and R. Vishny (1999), ‘The Quality of Government’, Journal of Law, Economics and Organization, 15(1), N1, 222–279. Lee, H. Y., L. Ricci and R. Rigobon. (2004), ‘Once again, is openness good for growth?’ Journal of Development Economics, 75, 451–472. Levine, R., and D. Renelt (1992), ‘A Sensitivity Analysis of Cross-Country Growth Regressions’, American Economic Review, 82(4), 942–963. Levine, R., N. Loayza and T. Beck. (2000), ‘Financial intermediation and growth: Causality and causes’, Journal of Monetary Economics, 46, 31–77. Lipset, S. (1959), ‘Some Social Requisites of Democracy: Economic Development and Political Legitimacy’, American Political Science Review, 53, 69–105. Lorentzen, P., J. McMillan and R. Wacziarg (2008), ‘Death and Development’, Journal of Economic Growth, 13, 81–124. Lovejoy, P. (1982), ‘The Volume of the Atlantic Slave Trade: A Synthesis’, Journal of African History, 23, 473–501.

M2545 - BHATTACHARYYA PRINT.indd 191

02/03/2011 10:09

192

Growth miracles and growth debacles

Lovejoy, P. (1994), ‘Background to Rebellion: The Origins of Muslim Slaves in Bahia’, Slavery and Abolition, 15(2), 151–180. Maddison, A. (2004), The World Economy: Historical Statistics, Paris: OECD. Mankiw, G., D. Romer, D. Weil (1992), ‘A Contribution to the Empirics of Economic Growth’, Quarterly Journal of Economics, 107(2), 407– 437. Manning, P. (1981), ‘The Enslavement of Africans: A Demographic Model’, Canadian Journal of African Studies, 15(3), 499–526. Manning, P. (1982), ‘Slavery, Colonialism and economic Growth in Dahomey, 1640–1960’, Cambridge: Cambridge University Press. Masters, W., and M. McMillan (2001), ‘Climate and Scale in Economic Growth’, Journal of Economic Growth, 6(3), 167–186. Mauro, P. (1995), ‘Corruption and Growth’, Quarterly Journal of Economics, 110, 681–712. Meillassoux, C. (1976), ‘The Role of Slavery in the Economic and Social History of Sahelo-Sudanic Africa’, in J. Inikori (ed.), Forced Migration: The Impact of the Export Slave Trade on African Societies, London: Hutchinson & Co., pp. 74–99. Miguel, E., and M. Kremer (2004), ‘Worms: Identifying Impacts on Education and Health in the Presence of Treatment Externalities’, Econometrica, 72(1), 159–217. Miller, J. (1982), ‘The Significance of Drought, Disease and Famine in the Agriculturally Marginal Zones of West Central Africa’, Journal of African History, 23, 17–61. Miller, J. (1988), Way of Death: Merchant Capitalism and the Angolan Slave Trade, 1730–1830, Madison: University of Wisconsin Press. Mokyr, J. (1990), The Lever of Riches: Technological Creativity and Economic Change, New York: Oxford University Press. Mokyr, J. (2001), ‘The Rise and Fall of the Factory System: Technology, Firms, and Households since the Industrial Revolution’, CarnegieRochester Conference Series on Public Policy, 55, 1–45. Montesquieu, C. (1748 [1989]), The Spirit of Laws, New York: Cambridge University Press. Moore, B. (1966), Injustice: The Social Bases of Obedience and Revolt, New York: White Plains. Nickell, S. (1981), ‘Biases in Dynamic Models with Fixed Effects’, Econometrica, 49(6), 1417–1426. North, D. (1981), Structure and Change in Economic History, New York: Norton. North, D. (1989), ‘Institutions and Economic Growth: An Historical Introduction’, World Development, 17(9), 1319–1332.

M2545 - BHATTACHARYYA PRINT.indd 192

02/03/2011 10:09

References

193

North, D. (1994), ‘Economic Performance Through Time’, American Economic Review, 84(3), 359–368. North, D., J. Wallis and B. Weingast (2006), ‘A Conceptual Framework for Interpreting Recorded Human History’, NBER Working Paper No. 12795, December. North, D., J. Wallis and B. Weingast (2009), ‘Violence and Social Order’, New York: Cambridge University Press. Nunn, N. (2007), ‘Historical Legacies: A Model Linking Africa’s Past to its Current Underdevelopment’, Journal of Development Economics, 83(1), 157–175. Nunn, N. (2008), ‘The Long-Term Effects of Africa’s Slave Trades’, Quarterly Journal of Economics, 123(1), 139–176. O’Brien, P. (2001), ‘Fiscal Exceptionalism: Great Britain and its European Rivals’, Department of Economic History Working Paper No. 65/01, London School of Economics, October. O’Brien, P. (2005), ‘Fiscal and Financial Preconditions for the Rise of British Naval Hegemony 1485–1815’, Department of Economic History Working Paper No. 91/05, London School of Economics, November. Parker, P. (2000), Physioeconomics: The Basis for Long-Run Economic Growth, Cambridge, MA: The MIT Press. Persson, T. and G. Tabellini (2006), ‘Democracy and Development: The Devil in the Details’, NBER Working Paper No. 11993. Pigou, A. (1938), The Economics of Welfare, London: Macmillan and Co. Prakash, O. (2004), Bullion for Goods: European and Indian Merchants in the Indian Ocean Trade, 1500–1800, New Delhi: Manohar. Prakash, O. (2005), The Great Divergence: Evidence from Eighteenth Century India, Paper presented at the GEHN Conference on Imperialism, Istanbul, September 11–12. Przeworski, A., M. Alvarez, J. Cheibub, and F. Limongi (2000), Democracy and development: Political institutions and material well-being in the world, 1950–1990, New York: Cambridge University Press. Putnam, R. (1993), Making Democracy Work: civic traditions in modern Italy, Princeton: Princeton University Press. Rigobon, R. and D. Rodrik (2005), ‘Rule of Law, Democracy, Openness, and Income: Estimating the Interrelationships’, Economics of Transition, 13(3), 533–564. Robbins, H. (1928), ‘A Comparison of the Effects of the Black Death on the Economic Organization of France and England’, Journal of Political Economy, 36(4), 447–479. Robinson, J. (2002), ‘States and Power in Africa by Jeffrey I. Herbst: A Review Essay’, Journal of Economic Literature, 60, 510–519.

M2545 - BHATTACHARYYA PRINT.indd 193

02/03/2011 10:09

194

Growth miracles and growth debacles

Robinson, J. (2006), ‘Equity, Institutions and the Development Process’, Nordic Journal of Political Economy, 32, 17–50. Rodney, W. (1966), ‘African Slavery and Other Forms of Social Oppression on the Upper Guinea Coast in the Context of the Atlantic Slave Trade’, Journal of African History, 7, 431–443. Rodrik, D. (1999), ‘Democracies Pay Higher Wages’, Quarterly Journal of Economics, 114, 707–738. Rodrik, D. (2000), ‘Institutions for High-Quality Growth: What They Are and How to Acquire Them’, Studies in Comparative International Development, 35(3), 3–31. Rodrik, D. (2005), ‘Growth Strategies’, Handbook of Economic Growth, 1(1), 967–1014. Rodrik, D. and F. Rodriguez (2000), ‘Trade Policy and Economic Growth: A Skeptics Guide to Cross-National Evidence’, in B. Bernanke and K. Rogoff (eds), NBER Macroeconomics Annual 2000, Cambridge, MA: MIT Press. Rodrik, D. and A. Subramanian (2003), ‘The Primacy of Institutions and what this does and does not mean’, Finance & Development, 40(2), 31–34. Rodrik, D. and A. Subramanian (2004), ‘From “Hindu Growth” to Productivity Surge: The Mystery of the Indian Growth Transition’, NBER Working Paper No. 10376, March. Rodrik, D., A. Subramanian and F. Trebbi (2004), ‘Institutions Rule: the Primacy of Institutions over Geography and Integration in Economic Development’, Journal of Economic Growth, 9, 131–165. Rogowski, R. (1989), Commerce and coalitions: How trade affects Domestic Political Alignments, Princeton: Princeton University Press. Romer, P. (1990), ‘Endogenous Technological Change’, Journal of Political Economy, 98(5), S71–S102. Sachs, J. (2003a), ‘Institutions Don’t Rule: Direct Effects of Geography on Per Capita Income’, NBER Working Paper No.w9490, February. Sachs, J. (2003b), ‘Institutions Matter, but Not for Everything’, Finance and Development, 40(2), 38–41. Sachs, J. (2005), The End of Poverty: How We Can Make it Happen in our Lifetime, London: Penguin Books. Sachs, J. and A. Warner (1995a), ‘Economic reform and the process of global integration’, Brookings Papers on Economic Activity, 1, 1–118. Sachs, J. and A. Warner (1995b), ‘Natural Resource Abundance and Economic Growth’, NBER Working Paper No.w5398, December. Sachs, J. and A. Warner (1997a), ‘Fundamental sources of long-run growth’, American Economic Review, 87(2), 184–188.

M2545 - BHATTACHARYYA PRINT.indd 194

02/03/2011 10:09

References

195

Sachs, J. and A. Warner (1997b), ‘Sources of Slow Growth in African Economies’, Journal of African Economies, 6(3), 335–376. Sachs, J., J. McArthur, G. Schmidt-Traub, M. Kruk, C. Bahadur, M. Faye, and G. McCord (2004), ‘Ending Africa’s Poverty Trap’, Brookings Papers on Economic Activity, 1, 117–240. Samuelson, P. (1976), ‘Illogic of Neo-Marxian Doctrine of Unequal Exchange’, in David A. Belsley et al. (eds), Inflation, Trade and Taxes: Essays in Honor of Alice Bourneuf, Columbus: Ohio State University Press, pp. 96–107. Samuelsson, K. (1993), Religion and Economic Action: the Protestant Ethic, the Rise of Capitalism, and the Abuses of Scholarship, Toronto: University of Toronto Press. Schultz, P. (2002), ‘Wage Gains Associated with height as a Form of Health Human Capital’, American Economic Review, 92(2), 349–353. Schumpeter, J. (1934), The Theory of Economic Development, Cambridge, MA: Harvard University Press. Shepherd, G. and F. Alburo (1991), ‘The Philippines’, in D. Papageorgiou, M. Michaely and A. Choksi (eds), Liberalizing Foreign Trade, Cambridge, MA: Basil Blackwell. Shleifer, A. and R. Vishny (1998), The Grabbing Hand: Government Pathologies and their Cures, Cambridge, MA: Harvard University Press. Smith, A. (1776), An Inquiry into the Nature and Causes of the Wealth of Nations, reprinted in E. Cannon (ed.) (1976), The Wealth of Nations, Chicago: University of Chicago Press. Solow, R. (1956), ‘A Contribution to the Theory of Economic Growth’, Quarterly Journal of Economics, 70(1), 65–94. Srinivasan, T. and S. Tendulkar (2003), Reintegrating India with the World Economy, Washington D.C.: Institute for International Economics. Staiger, D. and J. Stock (1997), ‘Instrumental Variables Regression with Weak Instruments’, Econometrica, 65, 557–586. Stevenson, R. (1968), Population and Political Systems in Tropical Africa, New York: Columbia University Press. Stock, J. and M. Yogo (2005), ‘Testing for Weak Instruments in Linear IV Regression’, in D. Andrews and J. Stock (eds), Identification and Inference for Econometric Models: Essays in Honor of Thomas Rothenberg, Cambridge: Cambridge University Press, pp. 80–108. Stulz, R. and R. Williamson (2001), ‘Culture, openness and finance’, NBER Working Paper No.w8222, April. Swan, T. (1956), ‘Economic Growth and Capital Accumulation’, Economic Record, 32, 334–361. Tavares, J. and R. Wacziarg (2001), ‘How Democracy Affects Growth’, European Economic Review, 45(8), 1341–1379.

M2545 - BHATTACHARYYA PRINT.indd 195

02/03/2011 10:09

196

Growth miracles and growth debacles

Tawney, R. (1926), Religion and the Rise of Capitalism, London: John Murray. Temple, J. (1998), ‘Initial Conditions, Social Capital and Growth in Africa’, Journal of African Economies, 7(3), 309–347. Thelen, K. (1999), ‘Historical Institutionalism in Comparative Politics’, American Review of Political Science, 2, 369–404. Thelen, K. (2000), ‘Timing and Temporality in the Analysis of Institutional Evolution and Change’, Studies in American Political Development, 14, 101–108. Thorner, D. (1962), ‘Deindustrialisation in India 1881–1931’, in D. Thorner and A. Thorner (eds), Land and Labour in India, Bombay: Asia Publishing House. Thornton, J. (1980), ‘The Slave Trade in the Eighteenth-Century Angola: Effects on Demographic Structures’, Canadian Journal of African Studies, 14(3), 417–427. Vamvakidis, A. (2002), ‘How Robust is the Growth-Openness Connection? Historical Evidence’, Journal of Economic Growth, 7, 57–80. Voigtländer, N. and H.-J. Voth (2008), ‘The Three Horsemen of Growth: Plague, War and Urbanization in Early Modern Europe’, Working Paper, Department of Economics, Universitat Pompeu Fabra. Wacziarg, R. (2001), ‘Measuring the Dynamic Gains from Trade’, World Bank Economic Review, 15(3), 393–429. Wacziarg, R. and K. Welch (2003), ‘Trade Liberalization and Growth: New Evidence’, NBER Working Paper No. 10152, December. Warner, A. (2003), ‘Once more into the breach: economic growth and integration’, Center for Global Development Working Paper No. 34, December. Weber, M. (1930), The Protestant Ethic and the Spirit of Capitalism, London: Allen and Unwin. Wei, S.-J. (2000), ‘Natural Openness and Good Government’, NBER Working Paper No.w7765, June. Weil, D. (2005), ‘Accounting for the Effects of Health on Growth’, NBER Working Paper No. 11455, July. Williamson, J., and R. Zagha (2002), ‘From Slow Growth to Slow Reform’, Washington, DC: Institute for International Economics. Wooldridge, J. (2000), Introductory Econometrics: A Modern Approach, Cincinnati: South-Western Publishing. Wooldridge, J. (2002), ‘Econometric Analysis of Cross Section and Panel Data’, Cambridge, MA: MIT Press. World Bank (2002), Globalization, Growth and Poverty, New York: Oxford University Press. Yang, X. and J. Borland. (1991), ‘A Microeconomic Mechanism for Economic Growth’, Journal of Political Economy, 99, 460–482.

M2545 - BHATTACHARYYA PRINT.indd 196

02/03/2011 10:09

References

197

Young, A. (2005), ‘The Gift of the Dying: The Tragedy of AIDS and the Welfare of Future African Generations’, Quarterly Journal of Economics, CXX(2), 423–466.

M2545 - BHATTACHARYYA PRINT.indd 197

02/03/2011 10:09

M2545 - BHATTACHARYYA PRINT.indd 198

02/03/2011 10:09

Index access to markets 171–2 Acemoglu, D. 12, 14, 18, 22, 24, 25, 27, 28, 38, 44, 45, 62, 65, 72, 74, 75, 76, 87, 99, 107, 141, 148, 159 agriculture 50, 97 China 105 economic growth and development and 34–5 aid programmes 176–7 Alberdi, Juan Bautista 31 Albouy, D. 78 Alcalá, F. 162 Alesina, A. 26 Allen, R. 33, 44 Angola 101 impact of slavery on 30, 76 Argentina economic growth and development 20 institutions 113–14, 149 international trade 130 Arndt, C. 72 Australia 9 industrial production 13 institutions 113–14, 175 international trade 128 Banerjee, A. 107 Barro, R. 150 Beaud, Michael 98 Behrman, J. 72 Bell, C. 72 Benavot, A. 51 Ben-David, D. 119 Besley, T. 46 Bhattacharyya, S. 18, 25, 30, 59, 95, 106, 108, 109 Bleakley, H. 72 Bloom, D. 35, 37, 71, 72 Botswana 143

Brazil 9 industrial production 13 institutions 149 international trade 128 Calvinism 31–2 Canada industrial production 13 institutions 149 international trade 128 Canning, D. 72 Carstensen, K. 35, 72, 78, 83 Caselli, F. 17 causes of economic growth see root causes of economic growth Chang, H.-J. 41 Chile 128 China agriculture 105 economic growth and development 9, 10, 11, 71, 172–4 institutions 104–5, 114 technology in 43, 104 urbanization 12 Ciccone, A. 162 Clark, Gregory 33 climate 34, 50 Clingingsmith, D. 108 Coe, D. 39 Colombia 20, 24 colonialism 14, 22, 25, 27, 71 colonial institutions view 72–3, 77–8, 107–9, 110–12 decolonization 27 conditional convergence 17 contract enforcement 169 Cortez, Hernan 98 Côte d’Ivoire (Ivory Coast) 20, 128 culture and economic growth 31–3 Curtin, Philip 28, 29

199

M2545 - BHATTACHARYYA PRINT.indd 199

02/03/2011 10:09

200

Growth miracles and growth debacles

Dahomey 30 decolonization 27 Deng Xiaoping 173 Denmark 149 depopulation, slavery and 28–9 development aid 176–7 Diamond, Jared 34, 35, 97 Dias, J. 38 diseases 96 economic growth and development and 35, 37–8, 114 Africa 100–104 empirical evidence 70, 71–2, 73, 74, 78–9, 81, 83, 87, 91 foreign aid and 176–7 division of labour 39 Dollar, D. 18, 40, 57, 119 Dowrick, Steve 39, 120 Drazen, A. 26 Dutt, A. 107 Easterly, W. 17, 23, 26, 41, 58, 62, 65, 73 economic growth see individual topics education 62 empirical evidence on economic growth 48 appropriate empirical framework 51–9 colonial institutions, diseases and forced migrations 71–91 connection between low-income countries and Africa 69–71 data 48–51 existing results with levels framework 67–9 identification issues with levels framework 59–67 Engerman, S. 23, 24, 30, 110 English language 22 Equatorial Guinea 51, 149 exchange rates 119 factor accumulation 17–18 Fage, J. 28 foreign aid 176–7 France institutions 149 urbanization 12 Frankel, J. 40, 50, 119, 120, 143

M2545 - BHATTACHARYYA PRINT.indd 200

future of economic growth 166–8 case for multilateral intervention 175–6 diseases and foreign aid 176–7 recent success stories and first principles 172–4 role of technical assistance 177 setting up conditions for growth 168–72 Gabon 128 Gallup, J. 34, 35, 37 Galor, O. 25–6, 44 Gambia 128 Gemery, H. 28, 29, 30 geography theory of economic growth 34–8 geography theory of economic growth and development 34–8 Ghana 76 Glaeser, E. 18, 43 Golley, Jane 120 Grief, Avner 32 Gundlach, E. 35, 72, 78, 83 Gwartney, J. 149 Haiti 128 Hall, A. 82 Hall, R. 17, 21, 22, 150 Hauk, W. 157, 159 Heavily Indebted Poor Country (HIPC) Program 1 Helpman, E. 39 Herbst, J. 26, 27, 45 heterodox capitalism 41 history of economic growth since AD 1000 9–14 Hogendorn, J. 28, 29, 30 Hong Kong 128 Howitt, P. 39 human capital 51, 62, 96, 171 economic growth and development and 42–5 India colonialism in 107–9 economic growth and development 9, 10, 11, 71, 174 industrial production 13 institutions 106–10, 114, 140

02/03/2011 10:09

Index international trade 128 urbanization 12 Indonesia 50, 128 industrial production 12–13 infrastructure bottlenecks 171–2 Inikori, Joseph 28, 29 institutions 14, 20 building 176 improving institutions with trade policy 140–41, 145–6 evidence 142–5 theories of trade and institutional development 141–2 institutions and trade as competitors or complements in economic development 119–21, 136–8 data 122–5 empirical strategy 121–2 evidence 125–36 institutions theory of economic growth and development 21–6, 114 Africa 28–30, 73, 75, 100–104 Americas 110–12 Australia 113–14 China 104–5 colonial institutions view 72–3, 77–8, 107–9, 110–12 empirical evidence 58, 62, 65–6, 67 India 106–10 institutional weakness in Africa 26–8 Russia 112–13 Western Europe 98–100 measure of institutional quality 50, 75–6 unbundling 147–50 which institutions matter most for economic growth 147–50, 162 data 150–52 evidence 153–62 international trade 172 improving institutions with trade policy 140–41, 145–6 evidence 142–5 theories of trade and institutional development 141–2 institutions and trade as competitors or complements in economic development 119–21, 136–8

M2545 - BHATTACHARYYA PRINT.indd 201

201

data 122–5 empirical strategy 121–2 evidence 125–36 measurement of openness to 50–51 slave trade 28–30, 73, 75 trade openness theory of economic growth and development 38–41 interventions 175–6 Ireland 50 Italy 11 Ivory Coast 20, 128 Iyer, L. 107 Japan 10 institutions 149 urbanization 12 Johnson, S. 62, 65, 72, 148 Jones, C. 17, 21, 22, 150 Kalemli-Ozcan, S. 72 Keefer, P. 21, 75, 150 King, R. 17 Klein, M. 29 Knack, S. 21, 75, 150 Knaul, F. 72 knowledge, economic growth and 42–5 Kraay, A. 18, 40, 57, 120 Kremer, M. 72 Kuran, T. 33 labour supply 96 Landes, D. 12, 32, 106 languages 22 Lawson, R. 149 Lee, J. W. 150 Levine, R. 17, 23, 26, 62, 65, 73 Lewis, J. 72 living standards 13 Lorentzen, P. 72 Lovejoy, P. 28, 29 Luxembourg 148 economic growth and development 1, 2, 9 macroeconomic stabilization 170 Maddison, Angus 10, 75 malaria 35, 37–8, 50, 177 economic growth and development and 35, 37–8

02/03/2011 10:09

202

Growth miracles and growth debacles

empirical evidence 70, 71–2, 73, 74, 78–9, 81, 83, 87, 91 Malaysia 128 Mali 128 Malthus, Thomas 96 Mankiw, G. 15, 17, 58 Manning, Patrick 28, 29, 30 markets access to 171–2 economic growth and development and market proximity 35 market-creating institutions 147–8 market-legitimizing institutions 149–50 market-regulating institutions 148–9 market-stabilizing institutions 149 Mauro, P. 150 Meillassoux, C. 29 Mexico 24, 98, 110, 128 Miguel, E. 72 Miller, J. 29, 30, 38, 101 Moav, O. 25–6, 44 Mokyr, Joel 18, 42, 43 Montesquieu, Baron de 34 Morocco 149 Morogoro Shoe Factory 1 multilateral interventions 175–6 Myanmar 51, 144 Namibia 82, 83 neoclassical growth model 15–17, 39 factor accumulation and measure of our ignorance 17–18 ‘root’ and ‘proximate’ causes of development 19–20 Netherlands 10, 99 colonialism and 25 urbanization 11 New Zealand industrial production 13 international trade 128 Nickell, S. 157 Nigeria 82, 83 impact of slavery on 76 institutions 149 North, Douglass 21, 45, 46, 76, 124, 141 Nunn, N. 30, 72, 74, 75, 76, 78, 83, 87, 91 O’Brien, P. 46

M2545 - BHATTACHARYYA PRINT.indd 202

Pakistan 50 Papua New Guinea 128 Parker, P. 34 Peixe, F. 82 Persson, T. 46, 150 Peru 24, 98, 110 Pigou, A. 153 Pizarro, Francisco 98 political power theory 25 Polity IV Project 19–20 population 96 slavery and depopulation 28–9 Portugal 50, 99 colonialism and 25, 111 predestination 31–2 property rights 168–9, 173 public choice theory 153 public interest 153 Pushkin, Alexander 112 Putnam, R. 33 Quesnay, François 2 regulatory institutions 148–9, 169–70 religion 50 economic growth and development and 31–3 representative politics 170–71 research and development (R&D) 42 Riddle, P. 51 Rigobon, R. 120 Robbins, H. 96 Robinson, J. 25, 26, 27, 28, 111, 141 Rodney, W. 29 Rodriguez, F. 40, 120, 124 Rodrik, D. 18, 22, 40, 45, 51, 58, 59, 75, 78, 87, 109, 120, 124, 125, 147, 148, 149 Roe, Thomas 106 Rogers, M. 39 Romer, D. 40, 50, 119, 120, 143 Romer, Paul 15 root causes of economic growth 95–6 Africa 100–104 Americas 110–12 Australia 113–14 China 104–5 India 106–10 Russia 112–13 Western Europe 96–100

02/03/2011 10:09

Index Rosenzweig, M. 72 rule of law 62, 169 Russia economic growth and development 20 institutions 112–13, 114 urbanization 12 Sachs, J. 18, 34, 35, 37, 40, 50, 57, 71, 72, 73, 78, 119, 123, 124 Samuelsson, K. 31 Schultz, P. 72 Schumpeter, Joseph A. 42 Shleifer, A. 153 Sierra Leone 176 Singapore 50, 51 slavery 27, 38, 70, 78, 101 effects of institutions of slavery and slave trade in Africa 28–30 empirical evidence 73, 75 measurement problems 76–7 Smith, Adam 21, 39 Sokoloff, K. 23, 24, 30, 110 Solow, Robert 15 Somalia 50 South Africa 128 Spain 50, 99 colonialism and 25, 111 Sri Lanka 148 Staiger, D. 79, 82 state failure 175 state formation theory of economic growth and development 45–7 Stock, J. 79, 82, 83 structured growth 18 theories of deep structural determinants 20–21 Subramanian, A. 109, 148 Sudan 128 Swan, Trevor 15 Switzerland 149 Tabellini, G. 26, 150 Tanzania 1–2, 9, 75, 143 Tawney, R. 32 taxation 45–6, 97 technology 43, 96, 97, 104 technical assistance 177 transfer 39–40

M2545 - BHATTACHARYYA PRINT.indd 203

203

Temple, J. 73 Thornton, J. 28 trade see international trade Tunisia 76 Ukraine 50 unbundling institutions 147–50 United Kingdom 99 colonialism and 25, 107–9, 113 economic growth and development 9, 10, 11, 44–5 industrial production 13 institutions 149 international trade 128 urbanization 11–12 United Nations 27 United States of America 51 economic growth and development 9, 10, 11 industrial production 13 institutions 149 international trade 128, 130 urbanization 11 colonialism and 22 Uruguay 130 Vamvakidis, A. 120 Vishny, R. 153 Voigtländer, N. 37 Voth, H.-J. 37 Wacziarg, R. 123, 124, 157, 159 Wallis, John Joseph 45 war 175 economic growth and development and 45–7, 99–100 Warner, A. 18, 35, 40, 73, 119, 123, 124 Weber, Max 27, 31–2, 45 Wei, S.-J. 40 Weil, D. 72 Weingast, Barry R. 45 Welch, K. 123, 124 Williamson, J. 108 World Bank 119–20 Yogo, M. 83 Zaire 128

02/03/2011 10:09