Globalization and Economic Integration: Winners and Losers in the Asia-Pacific

Globalization and Economic Integration

Globalization and Economic Integration Winners and Losers in the Asia-Pacific

Edited by

Noel Gaston Bond University, Australia and

Ahmed M. Khalid Bond University, Australia

Edward Elgar Cheltenham, UK • Northampton, MA, USA

© Noel Gaston and Ahmed M. Khalid 2010 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: 2009943890

ISBN 978 1 84844 861 2

02

Printed and bound by MPG Books Group, UK

Contents List of contributors PART I 1

3

4

6

7 8

3

GLOBALIZATION AND THE PROSPECTS FOR INTEGRATION

The twilight of globalization? Marcus Noland Economic integration in extended East Asia: toward a new trade regime Fukunari Kimura The politics of (anti-)globalization: what do we learn from simple models? David Greenaway and Douglas Nelson

PART III

5

INTRODUCTION

Globalization and integration in the Asia-Pacific: prospects and risks Noel Gaston and Ahmed M. Khalid

PART II

2

vii

25

42

69

MACROECONOMIC COORDINATION AND FINANCIAL MARKET INTEGRATION

Asian financial integration Jennifer Corbett Forecasting international financial prices with fundamentals: how do stocks and exchange rates compare? Robert P. Flood and Andrew K. Rose Trade policy, exchange rate adjustment and unemployment Yoshiyasu Ono Trade and wage inequality in a world of incomplete diversification Christis G. Tombazos

v

95

119 153

173

vi

Globalization and economic integration

PART IV

9

10 11

12

GLOBALIZATION, IMMIGRATION AND LABOR MARKET ISSUES

Globalization and labor markets: developments in the Asia-Pacific Ian Harper Aging and migration in Japan Junichi Goto Labor market transitions for female workers in Japan: the role of global competition Tomoko Kishi and Noel Gaston The effects of overseas operations on home employment of Japanese multinational enterprises Nobuaki Yamashita and Kyoji Fukao

Index

185 205

223

254

283

Contributors Jennifer Corbett, Professor of Economics and Executive Director, Australia–Japan Research Centre, Crawford School of Economics and Government, Australian National University, Canberra, Australia, and Reader in the Economy of Japan at the University of Oxford, UK. Robert P. Flood, Research Department, International Monetary Fund, Washington, DC, USA. Kyoji Fukao, Professor of Economics, Institute of Economic Research, Hitotsubashi University, Tokyo, Japan. Noel Gaston, Professor of Economics and Director, Globalisation and Development Centre, Bond University, Gold Coast, Queensland, Australia. Junichi Goto, Professor of Economics, Faculty of Policy Management, Keio University, Tokyo, Japan. David Greenaway, Professor in the School of Economics and Director, Leverhulme Centre for Research on Globalisation and Economic Policy, University of Nottingham, UK. Ian Harper, Professor of Economics and Chairman, Australian Fair Pay Commission until 2009. Since then he has been at Access Economics, Melbourne, Australia. Ahmed M. Khalid, Professor of Economics and Finance and Co-Director, Globalisation and Development Centre, Bond University, Gold Coast, Queensland, Australia. Fukunari Kimura, Professor of Economics, Faculty of Economics, Keio University and Chief Economist, Economic Research Institute for ASEAN and East Asia, Tokyo, Japan. Tomoko Kishi, Professor of Economics, Nanzan University, Nagoya, Japan. Douglas Nelson, Professor of Economics, Murphy Institute of Political Economy, Tulane University, New Orleans, LA, USA and Professorial

vii

viii

Globalization and economic integration

Research Fellow, Leverhulme Centre for Research on Globalisation and Economic Policy, University of Nottingham, UK. Marcus Noland, Senior Research Fellow, Peter G. Peterson Institute for International Economics (IIE) and East-West Center, Washington, DC, USA. Yoshiyasu Ono, Professor of Economics and Director, Institute of Social and Economic Research, Osaka University, Osaka, Japan. Andrew K. Rose, B.T. Rocca Jr. Professor of International Business in the Economic Analysis and Policy Group, Haas School of Business, University of California, Berkeley, CA, USA. Christis G. Tombazos, Associate Professor of Economics, Monash University, Clayton, Vic, Australia. Nobuaki Yamashita, Research Fellow, School of Business, La Trobe University, Melbourne, Vic, Australia.

PART I

Introduction

1.

Globalization and integration in the Asia-Pacific: prospects and risks Noel Gaston and Ahmed M. Khalid

INTRODUCTION Globalization is very much part of the modern vernacular. Arguably, it was first used by McLuhan and Fiore (1968) when they introduced the concept of the “global village”. Globalization is characterized by the growth of the international trade of goods and services, the growth in foreign direct investment (FDI) as well as the political and social linkages that accompany growing economic integration. Outwardly, the driving forces seem to be the decline in administrative barriers to trade, sharp falls in the costs of transportation and communication, fragmentation of production processes and the development in information and communication technology (ICT). Arguably, the perceived neoliberal shift to greater market orientation in domestic economies has been accelerated by financial sector liberalization. Since the collapse of Lehman Brothers in September 2008, the backlash to globalization now seems to be in full swing. In more normal times, the issue of macroeconomic stability is about riding the wave of business cycles with short-run or more immediate policy considerations at play. However, the current crisis will undoubtedly have effects that are both powerful and longlasting. The hopes of countries in the Asia-Pacific region are transparently pinned on a successful return to business-as-usual for China. China’s re-emergence as a major player in world merchandise trade and the impending emergence of India is bringing hundreds of millions of people into the modern world. It is also fostering substantial uncertainty in some quarters. More generally, globalization is unsettling; the uneven global pattern of development has implications for capital markets and, quite possibly also, for migration flows, forcing national policy makers to confront issues as fundamental as national identity. The “problem”, of course, is that globalization is neither universally good nor universally bad. For example, as we discuss below, it increases economic growth. However, it also seems positively associated with 3

4

Globalization and economic integration

income inequality in some countries. In all likelihood, globalization does yield positive net returns for the globalizers; however, it also produces losers. By the early 1980s, the more mature developed economies in the Organization for Economic Cooperation and Development (OECD) had continued the long-run process of deindustrialization with real income growth well below 4 percent per annum. In turn, capital arbitrage stimulated FDI by these countries. There was also a substantial increase in portfolio investment.1 The economic success of the wealthiest developed countries provided lessons on how developing countries could use more market-oriented economic strategies for their own development. This approach was reinforced and sanctified by the so-called “Washington Consensus” (Williamson, 1989). The economic reforms needed to be accompanied by establishing the necessary institutional structures. While many East Asian economies introduced sweeping economic reforms and liberalization policies from the mid-1980s to the early 1990s, investment in the institutional infrastructure was sometimes lacking. It has been an arduous period of trial and error in the economic and development policies in the emerging economies. For example, the East Asian experience suggests that the more significant mistakes may have been: (i) the benign neglect of banking reforms and (ii) an over-exposure to short-term foreign capital flows when declining exports were dramatically weakening the currencies of South Korea, Thailand, the Philippines and Indonesia starting from October 1994. McKinnon observed: For many years, diverse financial institutions in each of the five (crisis) countries had struggled with festering bad-loan problems from over-investment in real estate: lending to profitless real industry; government-sponsored mega projects; subsidized rural lending; and so on . . . undermined the capital position of banks. (McKinnon 1998, p. 96)

In addition, many emerging economies, impressed by the exceptional performance of the East Asian economies, adopted aggressive exportoriented policies. In some cases, these policies were poorly implemented and there was a general lack of investment in infrastructure and commitment to a legal framework which would foster greater investment certainty. Corruption has been endemic and a major problem (Ariff and Khalid, 2005). The variation in growth patterns across countries has been significant and large. A potential downside of greater financial integration is the more rapid transmission of the spillover effects of crisis in any one economy in the now more integrated bloc of countries. Many studies find evidence of the “contagion” effects of crisis, both within and across regions. For instance,

Globalization and integration in the Asia-Pacific

5

a comprehensive study by the International Monetary Fund (IMF) suggests that 10 countries experienced substantial currency pressures during the Asian financial crisis.2 It has been déjà vu with the current global financial crisis. This financial tsunami originated in the United States, the largest and most economically developed economy. As the saying goes, “when the US sneezes, everyone else catches a cold”. An unfortunate feature of most of these contemporary financial crises has been the inability of policy makers and economists to predict not only the occurrence of crises, but also to foresee their strength and depth. The observations made by Rodrigo de Rato, Managing Director of the IMF in August 2007 (in the early phase of the sub-prime loan crisis) are a good example of how little we know about these crises: Actually, what we are seeing is a test of new markets and instruments under less flexible conditions than those that prevailed over the last few years. However, the spillover to noncredit markets – including currencies and equities – has so far been manageable, helped by solid fundamentals not only in more developed markets but also in emerging economies. And, if history is any guide, even when shocks in one region spill over into other markets, these effects are typically short-lived and sometimes even provide a necessary wake-up call to investors who may be underestimating the risks.

Of course, reality bites; the current crisis is fully global and unlikely to be short-lived. Mishkin (2007) identified four factors that could impede financial development due to asymmetric information in credit markets: the lack of sufficient collateral, inefficient enforcement of restrictive covenants, government-directed credit, and lack of transparency and an underdeveloped legal system. It is widely accepted that one or more of these factors were responsible for the major financial crises in the 1990s (Latin American crisis, Asian financial crisis, and Russian crisis). Major factors that academics identify as contributing causes of the current crisis also include a poor regulatory system (possibly induced by deregulations in the United States since 1994) and a lack of enforcement of prudential regulations. The deregulation which allowed banks to make greater use of off-balance sheet vehicles and the pooling of credit assets into complex structured financial products seem to have been the major factors that contributed to the US crisis. Europe has been a major casualty of the crisis due to the similarities in financial market development and the use of innovative financial products. Most of the innovative products issued by US financial institutions, such as securitized contracts, were sold in European markets. Somewhat fortuitously, the majority of emerging economies, including

6

Globalization and economic integration

newly developed East Asian economies, still lack a market for innovative products. In some cases, these latter economies may have actually had better regulatory structures that discouraged market participants, especially banks, from taking excessive risks. Relatively speaking, the less-developed economies may have been spared the worst of the global financial crisis (with the obvious caveat that the crisis is still playing itself out). One lesson learnt from the multiple episodes of financial crises is that financial globalization, while being a strong force for economic development, can be disastrous if not managed properly. The turmoil experienced by the Asian economies in the aftermath of the Asian financial crisis created the impetus for the formation of a new financial architecture which defined financial rules and prudential regulations. Basel II and its underlying principles can be considered building blocks of this new architecture. Eatwell and Taylor (2000) argue that the conventional practice of devising microeconomic regulations is inefficient in the current environment. In particular, the role of macroeconomic linkages in the formation of regulatory policy is far more important. The IMF and the World Bank also initiated the Financial Sector Assessment Program (FSAP) involving the microeconomic appraisal of financial markets and the regulatory institutions. The FSAP requires countries to adhere to regulatory requirements under the Basel Accord. In the wake of the global financial crisis, further questions have been raised about the effectiveness of supervisory controls. Ito (2002) observes that a new international financial architecture should focus on an appropriate exchange rate regime and some measure of capital controls as part of crisis prevention and management. Studies suggest that many of the currency crises during the 1990s were associated with either pegged or soft-pegged exchange rate regimes (for example, Bubula and Otker-Robe, 2004). Moreover, Fisher (2001) argues that pegs are unlikely to be sustainable in the current environment of financial market integration. Another interesting development in the globalizing process which emerged in the aftermath of a series of crises in the 1990s was the active consideration of currency blocs and greater regional monetary integration (for example, Mussa et al., 2000). These developments were encouraged by the success of monetary integration and the adoption of a single currency in Europe. An important argument in support of greater regional monetary and exchange rate coordination in Asia is the protection of emerging economies from speculative attacks which were largely believed to have exacerbated the Asian financial crisis. Some international finance economists suggested that world currencies would converge to a tripolar regime where the US dollar, the euro and the

Globalization and integration in the Asia-Pacific

7

20% 18% 16% 14% 12% 10% 8% 6%

United States Source:

Japan

Europe

M1 2008

M7 2007

M1 2007

M7 2006

M1 2006

M7 2005

M1 2005

M7 2004

M1 2004

M7 2003

M1 2003

M7 2002

M1 2002

M7 2001

M1 2001

M7 2000

M1 2000

M7 1999

M1 1999

4%

China

Barbera et al. (2009).

Figure 1.1

East Asia weighted average trade, by major trading partner, 1999 to 2008 (%)

Japanese yen would dominate the global currency market (for example, Mundell, 2003). It was argued that most of the countries in the American region would link their currency to the US dollar while the euro would dominate Europe. Given Japan’s position as a major trading partner in the East Asian region and the yen as a stable currency, the yen was considered to be the obvious choice as the anchor currency for the region. The latter view has looked more suspect with the emergence of China and India as major players in the global economy. In particular, China has become increasingly dominant in the entire Asian region. Meissner and Oomes (2008) argue that the choice of a particular anchor currency depends on the amount of trade with countries that use that anchor. This argument is far more interesting with the China factor at play. The changing trade patterns can be seen in Figure 1.1. China’s share of trade with six major East Asian countries surpassed the trade shares of the United States and Japan by 2006. It is also important to note that China had a heavily managed exchange rate regime until quite recently; it moved to a managed float in January 2006. China’s heavy dependence on FDI and trade with the United States has seen it effectively link its currency to the US dollar

8

Globalization and economic integration

and not to the yen. Consequently, for East Asia, the US dollar still reigns supreme.

A SNAPSHOT OF GLOBALIZATION The “new” debate in the corridors of the IMF and the World Bank is the role of development as a “pro-poor growth” strategy. This leads to an important question: does globalization contribute to economic growth? The answer to the question posed in this simple form is a qualified “yes”. However, there are reservations about whether globalization leads to more economic volatility and vulnerability to external shocks. Another related issue, and perhaps politically more fundamental, is whether globalization can help the “working poor” or reduce “inequality”. Here the answer is less encouraging (see, for example, Dreher and Gaston, 2008). To provide a snapshot of globalization, we use the most widely used index of globalization. To be in a position to evaluate the consequences of globalization in a rational and scientific manner, objective indicators are needed. To assess the extent to which any country is more (or less) globalized at any particular point requires much more than employing data on flows of trade or FDI. We use a measure of globalization that is defined in a very broad manner; specifically, the KOF index.3 It is based on 25 variables that relate to different dimensions of globalization – economic, political and social. The sub-index on actual economic flows includes data on trade, FDI and portfolio investment. Trade is the sum of a country’s exports and imports and portfolio investment is the sum of a country’s assets and liabilities (all standardized by GDP). The KOF index includes the sum of gross inflows and outflows of FDI and the stocks of FDI (again, both standardized by GDP). While these variables are standard measures of globalization, income payments to foreign nationals and capital are included to proxy the extent to which a country employs foreign labor and capital in its production processes. The second sub-index refers to restrictions on trade and capital using hidden import barriers, mean tariff rates, taxes on international trade (as a share of current revenue) and an index of capital controls. Given a certain level of trade, a country with higher revenues from tariffs is less globalized. To proxy restrictions on the capital account, the KOF index includes data on 13 different types of capital controls. The KOF index classifies social globalization in three categories. The first covers personal contacts, the second includes data on information flows and the third measures cultural proximity. The index on personal contacts measures the direct interaction of people living in different

Globalization and integration in the Asia-Pacific

9

countries. It includes international telecom traffic (outgoing traffic in minutes per subscriber) and the extent of tourism (incoming and outgoing). Government and workers’ transfers received and paid (as a percentage of GDP) measure the extent to which countries interact, while the stock of foreign population is included to capture existing interactions with people from other countries. Finally, the average cost of a phone call to the United States measures the cost of international interaction. While personal contact data are meant to capture interactions among people from different countries, the sub-index on information flows measures the potential flow of ideas and images. It includes the number of internet hosts and users, telephone mainlines, cable television subscribers, number of radios (all per 1,000 people) and sales of daily newspapers. Cultural proximity is arguably the dimension of globalization most difficult to grasp. One indicator is the number of McDonald’s restaurants located in a country. For many people, the global reach of McDonald’s is symbolic of globalization itself. To proxy the degree of political globalization in each country, the KOF index includes the number of embassies and high commissions, the number of international organizations in which a country has membership and the number of United Nations peace missions in which a country has participated. The variables are combined into six groups: actual flow of trade; actual flow of investment; restrictions; variables measuring the degree of political integration; data quantifying the extent of personal contact with people living in foreign countries; data measuring transborder flows of information; and a proxy for cultural integration. These dimensions are then combined into an overall index of globalization with an objective statistical method.4 The KOF index is shown in Figure 1.2. As can be seen from the top panel, high-income OECD countries are, on average, the most globalized. Low-income countries are the least globalized. Richer countries seem to be, on average, more globalized than the poorer ones. On the face of it, globalization seems beneficial. The second panel shows that in the last 30 years, globalization has been pronounced in all regions. Western European and other industrialized countries are the most integrated; South Asia and Sub-Saharan Africa are the regions least globalized. We can further refine the focus by examining which regions have benefited most from globalization (see Figure 1.3). Using the same data we split the world into three groups – the OECD, the Asian newly industrializing economies (NIEs) and all other countries.5 It is obvious that the Asian NIEs have grown the most rapidly – from 1970 to 2003 their income per head quintupled. This is, at least partly,

10

Globalization and economic integration

2007 KOF Index of Globalization

High income: OECD

High income: non-OECD 80

80

50

50

50

20

20

20 1970

2004

1970

Lower middle income

2004

1970

2004

Upper middle income

80

80

50

50

20

20 1970

2004

1970

East Asia & Pacific

2007 KOF Index of Globalization

Low income

80

Year

2004

East Europe & Central Asia

Latin America & Caribbean

80

80

80

50

50

50

20

20

20

1970

2004

1970

Middle East & North Africa

2004

1970

South Asia

80

80

80

50

50

50

20

20 1970

2004

2004

Sub-Saharan Africa

20 1970

2004

1970

2004

Western Europe & Industrialized 80 50 20 1970

Source:

2004

Year

Dreher et al. (2008, p. 67).

Figure 1.2

KOF index of globalization, by income and by region

attributable to their export orientation and globalization. The clear losers are members of the group that are the least globalized; they have made only very modest economic progress over the past four decades. Given the importance of globalization today, the Globalisation and

Globalization and integration in the Asia-Pacific

11

(a) Increase in globalization 220

All countries

OECD

Asian NIEs

Other countries

200 180 160 140 120

2002

2000

1998

1996

1994

1992

1990

1988

1986

1984

1982

1980

1978

1976

1974

1972

1970

100

(b) Increase in real GDP 550 All countries

OECD

Asian NIEs

Other countries

500 450 400 350 300 250 200 150

Note:

2002

2000

1998

1996

1994

1992

1990

1988

1986

1984

1982

1980

1978

1976

1974

1972

1970

100

1970 = 100, 104 countries.

Source: Author calculations based on data from http://globalization.kof.ethz.ch/static/ rawdata/globalization_2007_short.xls.

Figure 1.3

Effects of globalization

12

Globalization and economic integration

Development Centre (GDC) at Bond University organized an international conference in September 2008 with the theme “How Globalization is Shaping the Asia-Pacific”. This book includes some of the papers from this conference. Those included all have relevant themes to the above discussion on globalization namely, regional trade issues, globalization and FDI, financial integration, the labor market consequences of globalization and the economic and political economic consequences of globalization.

AN OVERVIEW OF THE FOLLOWING CHAPTERS The authors of the chapters in this book take some of these facts as the starting point for their analyses. In many cases, we shall see that the portents for globalization are not particularly rosy. In Chapter 2, Marcus Noland starkly describes the vulnerability of globalization. Ambivalence to globalization has been fueled by skepticism about major international institutions, such as the World Trade Organization (WTO) and the IMF. In addition, in some countries the political consensus supportive of a more open system has been disintegrating. Of course, these challenges are made more acute by the global recession. Aggravating the situation is that international trade negotiations have been made more difficult as the substantive agenda has grown more complex and the politics have become pluralistic and increasingly partisan. Depending on how the global downturn plays out, Noland warns that this reaction against globalization could broaden into a backlash against capitalism itself. Noland highlights the increasing frailty of the WTO system and the likelihood that it will fail to deliver much in terms of further liberalization. The weaknesses of both the multilateral trade and financial institutions have provided an impetus for regional initiatives in both trade and finance. Accelerated fragmentation is a distinctly possible outcome. If the political will is there, Asia has the financial wherewithal to go its own way. In fact, Noland sees Asia as being the key actor in undermining, or strengthening, global institutions. In Chapter 3, Fukunari Kimura provides an in-depth look at the growing proliferation of free trade agreements (FTAs) in Asia. The backdrop is the bleak prospect of any further progress being made on the WTO’s Doha Development Agenda. However, Kimura strikes an optimistic tone. He views favorably what Baldwin (2006) refers to as “multilateralizing regionalism”. In particular, it is argued that East Asian FTA “networking” has been an effective driving force in promoting freer trade and investment, particularly through the facilitation of international production networks.

Globalization and integration in the Asia-Pacific

13

This process has been driven by a number of factors. In the initial instance, it was the regional response to the Asian financial crisis; more recently, it is the specter of competing with China, the “elephant in the room”, and the growing importance of FDI, international production networks and the pressure for reform applied by multinational corporations. Like Noland in the previous chapter, Kimura argues that East Asia and the Asia-Pacific have become not only the focal point of multilateral regionalism, but also the primary vehicle for the advancement of trade liberalization. In Chapter 4, David Greenaway and Douglas Nelson also warn of the reversibility of globalization. Their focus is on the pivotal role of politics in supporting, or possibly even undermining, globalization. In particular, they contrast the politics of (anti-)immigration and the politics of free trade in the United States. With regard to the latter, they argue that while reduced average levels of protection may well reflect the public interest, the cross-sectional variation or structure of protection is best understood from a group politics perspective. In contrast to trade politics, immigration has been a more ephemeral issue. While the politics over trade is played by insiders with fairly predictable public policy preferences, in the case of immigration politics, the preferences and location of individuals within the public discourse are not immutable. In fact, rather than any tangible adverse economic impact attributable to greater immigration, the documented resistance to immigration by native-born residents appears to be the main driver of more restrictive immigration regimes in contemporary times. More worrisome is the fact that in the case of the public politics of immigration, and anti-globalization politics in general, there appears to have been a strong link between poor macroeconomic conditions and the attractiveness of restrictive policies. Unfortunately, it may be that globalization politics is more like the politics of immigration than the politics of trade (with its implied negotiability). Hence, rather than the pure economic self-interest of identifiable groups, public attitudes will play an increasingly important role in setting the terms of the politics of globalization. In Chapter 5, Jennifer Corbett questions the meaning of Asian financial integration and its presumed desirability. She first reminds us that financial market openness, financial engagement and financial integration are distinct, albeit related, concepts. Financial policies run the risk of being misconstrued, if we are not clear on what financial integration really means. Economic theory does not provide any clear-cut indication of how greater integration would be beneficial. Until 2008, however, policy makers in many regions, including Asia, saw benefits from greater integration. One suspects that this view may have moderated in the midst of the global

14

Globalization and economic integration

financial crisis, with its non-Asian origin. While Asia might well benefit from having financial development, Corbett’s assessment of the recent estimates is that the gains from greater regional integration seem modest. Furthermore, it would appear that more financially integrated economies have more (and not less) synchronized GDP fluctuations. Given that one objective of integration may be to smooth out consumption fluctuations over time for each nation, this dependent and correlated nature of output shocks would seem to be problematic. In their classic paper, Meese and Rogoff (1983) found that an atheoretical random walk “model” of the exchange rate consistently outperformed structural models, despite the latter’s being given the advantage of using actual future values of market fundamentals. This finding revealed how very little is known about exchange rate determination. In Chapter 6, Robert Flood and Andrew Rose use the Meese–Rogoff methodology to study international equity markets. In a similar fashion, the authors find that none of the structural models based on market fundamentals, such as earnings and dividends, performs better than a simple random walk model of the aggregate stock market. Since the integration of international financial markets is a key part of globalization, understanding how these markets operate in practice is crucial. What the results of this chapter highlight is how little we know about the key drivers of international foreign exchange and equity markets. The key first step in the further integration of financial markets, and the implied setting of basic norms or rules to govern international financial markets, is presumably a deep understanding of how such markets operate. Flood and Rose’s results suggest that our understanding of these markets is superficial at best. The last decade of the last century and the early years of the new millennium were trying times for Japan. Within Japan, the 1990s are widely referred to as the “lost decade”. The developments in Japan have been attracting renewed academic attention; with the collapse of the “bubble” economy, the share market crash and the banking and financial crisis, the parallels to the current global financial malaise seem altogether obvious. Along with Paul Krugman (1999), Yoshiyasu Ono has helped to popularize the idea that the presence of a liquidity trap is the primary reason why Japan got “stuck” in a stagnation equilibrium. An implication, of course, is that with near zero short-term nominal interest rates in Japan, conventional monetary policy is impotent and fiscal stimulus is required. In Ono (2006), it was also observed that exchange rate movements during the 1990s saw the Japanese yen depreciate (appreciate) against the US dollar when the Japanese economy improved (deteriorated) compared to the US economy. This asymmetry in international business activity and the appreciation of the stagnant country’s currency is the background

Globalization and integration in the Asia-Pacific

15

for Ono’s contribution in Chapter 7. He shows that if a country’s liquidity preference strengthens, then domestic consumption and employment decline and its currency appreciates. The message of this chapter is that, when faced with rising domestic unemployment, governments are often tempted to raise trade barriers to protect their industries. Ono shows that this is likely to make the situation even worse. In the presence of persistent unemployment and a shortfall of aggregate demand, while higher tariffs do lead to improved terms of trade, they also lower international competitiveness, thereby reducing employment. The risk, of course, is that this could lead to yet further rounds of increases in trade barriers. The potential for an escalation of protection constitutes yet another serious threat to greater global integration. The relationship between growing trade with developing countries (less-developed countries: LDCs) and adverse labor market outcomes for unskilled workers in developed countries has without doubt been the topic attracting the greatest attention of labor economists since the early 1980s. Labor earnings and income inequality grew in most OECD countries during the 1980s and 1990s. Until recently, the academic consensus seemed to be that skill-biased technological change, rather than trade, was the most likely culprit responsible for the adverse affects on the leastskilled and most-vulnerable workers. However, more recent opinion is more divided. This appears to be the case for a couple of reasons. First, technical progress may have been accelerated by import competition. Hence, technological progress may have been overly ascribed as the major cause of adverse (unskilled) labor market effects. Second, there have been a growing number of studies using data on trade in intermediates which indicate that trade may be having significant effects on the labor market. International outsourcing as well as the imports of intermediate goods may have extensive effects at a more disaggregated level. In other words, there may be considerable distributional implications with, for example, trade in services or the increased sourcing of inputs from LDCs. Identifying the downstream (and upstream) effects of such trade is important in knowing how different types of workers may be affected. Over the last three decades the United States has experienced a dramatic increase in its imports, followed by an equally dramatic increase in the wage gap between skilled and unskilled labor. A lot has been said, but little is known, about the relationship between these developments. In Chapter 8, Christis Tombazos argues that imports may not only displace domestic output (and, hence, domestic labor), but may also be used as inputs in domestic production and subject to extensive downstream handling by unskilled workers. The latter effect stimulates domestic labor demand and may be particularly important when there is specialization in the

16

Globalization and economic integration

production of some of these intermediate goods. While the displacement effect has been thoroughly explored, the stimulation effect has been largely neglected in the literature. Using a sophisticated empirical framework and disaggregated data on imports of both goods and services by kind, the author finds that the labor demand stimulation effect is significant, that is, imports actually stimulate unskilled labor demand and reduce the wage gap between skilled and unskilled workers. In Chapter 9, Ian Harper surveys labor market developments in the Asia-Pacific region. He notes the impressive economic growth rates, which have led to rapidly growing employment in many countries in the region. These findings seem to reaffirm both the benefits of greater globalization and the shift in the world’s economic center of gravity towards Asia. However, Harper notes that while some countries in the region have witnessed a reduction in unemployment over the past two decades, others have seen unemployment increase dramatically. Moreover, while globalization does bring substantial economic benefits, these gains have not been shared by all labor market participants. The costs of adjustment disproportionately fall on the shoulders of low-skilled workers. Globalization also seems to be associated with increased uncertainty about job security for the least-skilled and most-vulnerable workers. Against this backdrop, Harper then looks at labor market regulation, particularly the successful implementation of minimum wage policies. The setting of an appropriate level of the regulated wage is described as a delicate balancing act. On the one hand, minimum wages can be an effective tool for poverty reduction in countries disrupted by rapid globalization. On the other, wages must not be set “too high”, that is, so as to minimize its disemployment effects. In the current environment, Harper advocates erring in favor of the former objective, that is, on sharing the benefits, and reducing the risks, of globalization as much as possible. While Japan has long been the dominant player in Asia, the lost decade and the growing economic importance of China have placed a spotlight on its domestic and international policy options. The long slump and the rapid aging of its population have left Japan with little choice but to attempt to restructure basic features of its economy. The demographic problem has seen pension liabilities increase; and the historically high post-war unemployment and flat economy have reduced the tax base. Unfortunately, reform is difficult when there is little or no economic growth. Aggravating this situation is that the desirability of internationalization is controversial in Japan. In Chapter 10, Junichi Goto examines the merits of increased immigration as the best means to address a declining fertility rate, rapid aging, and a future of chronic labor shortages in Japan. He classifies the effects

Globalization and integration in the Asia-Pacific

17

of immigration into three parts. There is a beneficial effect of cheaper foreign labor and workers willing to do the jobs that Japanese workers are unwilling to do. This is offset by a trade barrier effect. Since prices for a country’s “scarce factor” are higher with high tariffs, workers – native and immigrant – are essentially “overpaid”. Complicating the mix is a nontradable good effect which could operate in either direction due to a positive consumption effect and a negative income effect for native workers. Given that what the author argues is the highly uncertain dividend from increased immigration, he examines the alternatives. The obvious internationalist alternative is further trade liberalization and structural reform. The alternatives in the domestic setting are measures to raise labor productivity, encouraging greater female labor force participation and using policy to address the low fertility rate. A phenomenon being experienced by many economies during the last decade has been the strong growth in part-time employment. In fact, in Japan the new jobs added during the recovery from the lost decade were mainly part-time jobs. In Chapter 11, Tomoko Kishi and Noel Gaston study the labor market transitions of female workers during the period of economic stagnation for Japan. They examine the commonly made claim that globalization has been responsible for the changing nature of the employment relationship. Controlling for a number of the usual demographic and sociological determinants, they examine whether workers in trade-exposed industries are more likely to move from full- to part-time jobs. In fact, they find that the firms in the most internationally exposed and competitive sectors of the economy were the ones most under pressure to forgo traditional lifetime employment practices. In contrast, while all firms were under financial pressure during the economic slowdown, firms in non-tradable goods sectors seemed better able to resist the need for more flexible workplace arrangements. However, using the KOF index the authors find that globalization may be a savior, rather than a villain, for workers during economic hard times. The growing integration of Japan into the world economy appears to have softened the blow of a severe recession. The rapid growth of ICT and high-tech service sector industries, inwards FDI, financial sector liberalization and political internationalization offer workers full-time employment opportunities; they do not destroy them. Jagdish Bhagwati (1999) notes the ironic “about face” in policy-making circles concerning the impact of globalization in the last 20 or so years of the 20th century. Post-Second World War concerns about neo-colonialism and the dependency of developing countries on developed countries, raised questions for the poorer countries about the desirability of increased integration and trade. This view has been supplanted, almost completely, by

18

Globalization and economic integration

developing-country enthusiasm for trade and inward foreign investment. The reservations are now expressed by many developed countries, which worry about the consequences and perils for their domestic workers if integration via trade, migration or investment in LDCs continues unabated. Fears in many developed countries about domestic firms “sending jobs overseas”, whether to other developed countries or to LDCs, are an extension of the more traditional concerns about the “hollowing out” of manufacturing industry and the loss of “key” manufacturing activities and “good jobs” overseas. Finally, in Chapter 12, Nobuaki Yamashita and Kyoji Fukao address whether the employment of domestic workers falls when the overseas affiliates of Japanese multinationals expand their operations. They focus on disentangling the employment creation or scale effect of expanded operations and the substitution of cheaper foreign labor for domestic workers. This is a valuable study, because most recent studies have focused on either US or Swedish multinationals.6 The authors find evidence against the usual concerns expressed about the adverse effects of FDI investment; the evidence does not support the view that overseas operations expand at the cost of home employment in Japan. Jobs are not “exported”, in fact, consistent with the findings in the previous chapter, Yamashita and Fukao’s findings suggest that overseas operations may have helped to maintain the level of home employment in Japanese manufacturing.

CONCLUSION Globalization is indeed a broad issue that cannot be fully addressed in one volume. However, during the dark days of the global financial crisis and the threat posed by protectionist rhetoric and posturing and the possible retreat from further globalization and internationalization, the issues discussed here are salutary and timely. It is important that market participants and regulators clearly understand the risks associated with the present stage of globalization. Policy makers need to reconfigure the regulatory framework having learned the necessary lessons from the episodes of financial crisis experienced in various parts of the world in the last two decades. Financial literacy is essential for market participants, especially in emerging economies. Some of the issues discussed in this book are useful for the development of a new international financial architecture. These measures will help to reap the full benefits of globalization. As stated earlier, the chapters included in this book were selected from the contributions at the GDC conference in September 2008. We would like to thank our colleagues at the GDC at Bond University and

Globalization and integration in the Asia-Pacific

19

particularly Kyona Box for her support with the GDC Conference. Generous financial support was provided by the Japan Foundation, the Queensland state government and the Faculty of Business, Technology and Sustainable Development, Bond University. Our special thanks to all contributors without which this book would not have been completed. We thank Arsalan Khalid for editorial and research assistance. Finally, we are grateful and indebted to Alex Pettifer of Edward Elgar for his support and advice in helping us prepare the manuscript.

NOTES 1. “The volume of world financial transactions is usually measured in US dollars. A million dollars is a lot of money for most people. Measured as a stack of thousand dollar notes, it would be eight inches high. A billion dollars – in other words, a million million – would be over 120 miles high, 20 times higher than Mount Everest. Yet far more than a trillion dollars is now turned over each day on global currency markets, a massive increase from only 10 years ago, let alone the more distant past” (Giddens, 1999). 2. Results are based on a study involving 60 industrialized and emerging economies, see IMF (1999). The study also details how the stock markets in Brazil and Hong Kong fell by 30 percent, India by 17 percent, while losses in the stock markets in Indonesia, Malaysia, South Korea and Thailand were approximately 40 percent. 3. The KOF index (from the German Konjuncturforschungsstelle) has been increasingly used in the economics literature. It is probably the best-known and mostly widely used index of globalization. Dreher et al. (2008, pp. 75–8) list 36 journal articles published between 2003 and 2008 that employ the KOF index in statistical analyses. 4. Dreher et al. (2008) describe the method in more detail. The annual data are publicly available at: http://globalization.kof.ethz.ch/static/rawdata/globalization_2007_short. xls. 5. (Old) OECD countries: Australia, Austria, Belgium, Canada, Denmark, Finland, France, Germany, Greece, Iceland, Ireland, Italy, Japan, Luxembourg, the Netherlands, New Zealand, Norway, Portugal, Spain, Sweden, Switzerland, the UK, and the USA. Asian NIEs: China, India, Indonesia, South Korea, Malaysia, Singapore, Thailand and Papua New Guinea. Other countries: comprise a very mixed group of 73 countries. 6. Brainard and Riker (2001) is a well-known study. They find that US multinationals do substitute labor at home with labor abroad, although the substitution is greater between affiliates in countries at similar levels of development. Blomström and Kokko (2000) identify large changes in Swedish employment, with an astonishing 80 percent of jobs disappearing each year from Swedish multinationals, but an almost equivalent number being created via acquisitions of new plants.

REFERENCES Ariff, M. and A. Khalid (2005), Liberalization and Growth in Asia: 21st Century Challenges, Cheltenham, UK and Northampton, MA, USA: Edward Elgar. Baldwin, R.E. (2006), “Multilateralizing regionalism: spaghetti bowls as building blocks on the path to global free trade”, The World Economy, 29 (11), 1451–518.

20

Globalization and economic integration

Barbera, F., A.M. Khalid and R. Gulasekaran (2009), “It’s not yen bloc or koala bloc. Greenback is still dominant in East Asia”, paper presented at the 2009 Singapore Economic Review Conference, Singapore, August 6–8. Bhagwati, J. (1999), “Globalization: who gains, who loses?”, in Globalization and Labor, edited by H. Siebert, Tübingen: Mohr Siebeck, pp. 225–36. Blomström, M. and A. Kokko (2000), “Outward investment, employment, and wages in Swedish multinationals”, Oxford Review of Economic Policy, 16 (3), 76–89. Brainard, S. and D. Riker (2001), “Are US multinationals exporting US Jobs?”, in Globalization and Labour Markets (Vol. II), edited by D. Greenaway and D. Nelson, Cheltenham, UK and Northampton, MA, USA: Edward Elgar, pp. 410–26. Bubula, A. and I. Otker-Robe (2004), “The continuing bipolar conundrum”, Finance and Development, IMF, Washington, DC, March, pp. 32–5. de Rato, R. (2007), “Economic growth and financial market development: a strengthening integration”, speech by the Managing Director of the IMF at the 3rd International Derivatives and Financial Market Conference, Campos do Jordão, Brazil, August 22. Dreher, A. and N. Gaston (2008), “Does globalization increase inequality?”, Review of International Economics, 16 (3), 516–36. Dreher, A., N. Gaston and P. Martens (2008), Measuring Globalization: Gauging Its Consequences, New York: Springer. Eatwell, J. and L. Taylor (2000), Global Finance at Risk: The Case for International Regulations, New York: New Press. Fisher, S. (2001), “Exchange rate regimes: is the bipolar view correct?”, Journal of Economic Perspectives, 15 (1), 3–14. Giddens, A. (1999), “Runaway world”, 1999 Reith Lectures, available at: http:// news.bbc.co.uk/hi/english/static/events/reith_99/week1/week1.htm (accessed September 1, 2009). International Monetary Fund (IMF) (1999), World Economic Outlook 1999, Washington, DC: IMF, pp. 66–87. Ito, T. (2002), “Toward new international financial architecture: an Asian perspective”, working paper, Victoria University of Wellington, February. Krugman, P. (1999), “It’s Baaack: Japan’s slump and the return of the liquidity trap”, Brookings Papers on Economic Activity, Issue 2, 188–206. McKinnon, R.I. (1998), “Exchange rate co-ordination for surmounting the East Asian currency crisis”, Asia Pacific Journal of Economics and Business, 2 (1), 95–103. McLuhan, M. and Q. Fiore (1968), War and Peace in the Global Village, New York: Bantam Books/Random House. Meese, R. and K. Rogoff (1983), “Empirical exchange rate models of the seventies: do they fit out of sample?”, Journal of International Economics, 14 (1), 3–24. Meissner, C. and N. Oomes (2008), “Why do countries peg the way they peg? The determinants of anchor currency choice”, IMF Working Paper WP/08/132, International Monetary Fund, Washington, DC. Mishkin, F. (2007), “Is financial globalisation beneficial?”, Journal of Money, Credit and Banking, 39 (2–3), 259–94. Mundell, R. (2003), “Prospects for an Asian currency area”, Journal of Asian Economics, 14 (1), 1–10. Mussa, M., P. Masson, A. Swoboda, E. Jadresic, P. Mauro and A. Berg (2000),

Globalization and integration in the Asia-Pacific

21

“Exchange rate regimes in an increasingly integrated world economy”, IMF Occasional Paper 193, International Monetary Fund, Washington, DC. Ono, Y. (2006), “International asymmetry in business activity and appreciation of a stagnant country’s currency”, Japanese Economic Review, 57 (1), 101–20. Shaw, E.S. (1973), Financial Deepening in Economic Development, New York: Oxford University Press. Williamson, J. (1989), “What Washington means by policy reform”, in Latin American Readjustment: How Much has Happened, edited by J. Williamson, Washington, DC: Institute for International Economics, pp. 5–20.

PART II

Globalization and the prospects for integration

2.

The twilight of globalization? Marcus Noland*

INTRODUCTION Although frequently invoked, “globalization” is an imprecisely defined term subject to multiple interpretations. Typically it signifies an expansion of cross-border economic interaction, though the word also embodies less precise though important, non-economic connotations relating to the erosion of local control, autonomy, and identity in the organization of political, social, and cultural life. At the popular level, these meanings are sometimes conflated with anti-capitalist, anti-Western, or anti-American impulses. Globalization has ebbed and flowed, and today’s experience, driven by technological change, supportive institutions, and the relative absence of international conflict, is not historically unprecedented. Relative to the size of the global economy, only recently have the levels of international trade and investment surpassed those reached at the end of the 19th century, and this still may not be true for labor markets: globally, the extent of crossborder labor migration has yet to re-attain the levels observed at the end of the 19th century. The US today, where both legal and illegal immigration are relatively high compared to past levels, is something of an exception in this regard. That earlier 19th-century “Golden Age” of globalization was built by a combination of technological innovations, such as the telegraph and the steamship, which drove down transaction costs, together with supportive institutions, including the gold standard, Pax Britannica, and European colonialism, which facilitated the global diffusion of those technological innovations. Of course, this Golden Age was more golden for some than for others, at the levels of both the individual and the nation-state. This system began to fray at the end of the 19th century, and effectively came to an end with the First World War. Post-war attempts to reestablish the institutional underpinning of this regime were unsuccessful. As the world entered the Great Depression, countries engaged in beggarthy-neighbor protectionism in a desperate bid to maintain aggregate

25

26

Globalization and economic integration

April

rch

Ma

M ay

1929

Jun

bru ary

1930

Fe

e

1931 1932

2,739 2,998

1,206 1,839

July

January

1933 992

Au

gu

st

ber cem De

No

vem

October Source:

r

be

ber

m pte

Se

Kindleberger (1986).

Figure 2.1

The Kindleberger spiral

demand and employment at home. Cross-border commerce imploded, as illustrated by the economic historian Charles Kindleberger’s memorable “Spiral” (Figure 2.1). The rise of illiberal trading blocs and the onset of the Second World War further reduced cross-border economic integration. Yet the war also spurred technological innovations, particularly in information technology and air transportation, which once again abruptly reduced transaction costs and facilitated international commerce once hostilities ended. In the context of the post-war East–West divide, the noncommunist world, now under the leadership of the United States, began to reassemble the pieces of a liberal trading order: a key currency system

The twilight of globalization?

27

Trillions, USD 40 Nominal Real 35 30 25 20 15 10 5 0 1980 Figure 2.2

1985

1990

1995

2000

2005

Global trade value

based on the US dollar and supportive multilateral institutions, including the General Agreement on Tariffs and Trade (GATT) and its eventual successor, the World Trade Organization (WTO), as well as other multilateral institutions, such as the World Bank and the International Monetary Fund (IMF) in the financial sphere. The result was a rising volume of international trade and investment made possible by falling transaction costs and relative international stability (admittedly in the context of the Cold War) (Figure 2.2). The subsequent transformation of the Chinese economy from a closed, centrally planned system to a more open, marketoriented one, and the eventual end of the Cold War and the collapse of the Eastern Bloc and the concomitant spread of capitalism, gave further impetus to this second Golden Age of globalization. Globalization’s sustainability is in doubt, however. Presumably technological advances will continue to drive down cross-border transaction costs. Indeed, with greater and greater numbers of human beings achieving advanced education, if anything one would expect the rate of technological advance to accelerate. The weakness, rather, is in the supportive institutions. Globalization creates winners and losers, begetting social tensions that must be managed at both the national and international levels. At the

28

Globalization and economic integration

national level, for example, in the US, the public is split largely along educational lines – those with only a high school education, or less, fear globalization, while those with some post-high school education tend to view it benignly (Blonigen, 2008). While it is possible that this difference in outlook is due to the ignorance of poorly educated people, their assessment may also reflect a rational appraisal of the impact of globalization on their earnings and relative status. At the nation-state level, there is a concern that the major international institutions, such as the IMF and the WTO, do not adequately reflect the interests of poor countries. The situation is further complicated by the fact that globalization may be conflated with American hegemony. While the US is truly becoming globalized – no single country or region is predominant in the rapid increase in the international integration of American goods, capital, and labor markets – this is often not the case for other countries. The centripetal pull of the US economy and US political and cultural influence is such that in many parts of the world, globalization occurs with an American accent. The US itself becomes quite naturally the focal point of apprehensions regarding globalization. Key elements of US foreign policy, including military activities in both Afghanistan and Iraq, have exacerbated such tendencies in many corners of the world (Pew Global Attitudes Project, 2008). Hence globalization faces at least two sources of potential vulnerability. The first challenge is at the diplomatic level. Globalization thrives on order. The issue today is that the objections to globalization on current terms by a growing number of systemically important participants may degrade the existing commercial regime significantly without constituting a consensus sufficient either to reform the existing system or to construct a coherent alternative. The second vulnerability is that systemically important governments will be unable to sustain an internal political consensus supportive of an open system. Of course these two challenges are interrelated – changes to rules may be politically controversial and acceptable in one polity but not another, so the sustainability of globalization at its current level will reflect both the interplay of inter-state diplomacy and domestic politics. Management of these challenges is likely to be made more acute by the global recession and the political stresses entailed, at least in the short run. Trade volumes declined in the second half of 2008, and the World Bank is predicting that trade will shrink in 2009, falling for the first time since 1982, and exhibiting its weakest performance since the Bank began keeping records in 1970 (World Bank, 2009). The remainder of this chapter will review the diplomatic and domestic political challenges to globalization over the medium term.

The twilight of globalization?

29

COMPLEXITY, PLURALISM, AND LEGITIMACY When the liberal international trade system was being reconstructed following the Second World War, participants concentrated on dismantling wartime restrictions on cross-border trade and reducing border barriers that had grown since the 1920s. Multiple “rounds” of multilateral trade negotiations under the GATT focused on the elimination of quantitative restrictions and the reduction in tariffs on trade in industrial goods, with the universally politically sensitive agricultural sector largely excluded from this liberalization process. Two developments over the last quartercentury greatly complicated this process. The first is that the successful conclusion of trade negotiations has been made more difficult as the substantive agenda has broadened and grown more complex. Changes in the underlying economy, including the rise of multinational firms with global production and sales operations, the shift in the developed countries toward a “post-industrial economy”, and the growth of tradable services, all contributed to a shift away from the old tariff-cutting paradigm of trade negotiations. As difficult as agreeing on tariff-cutting formulas may have seemed to the negotiators of the Kennedy or Tokyo Rounds, the conceptual complexity and political sensitivity of these efforts paled beside the emerging “behind the border” issues such as competition policy and intellectual property rights (IPR) protection, to say nothing of the hot-button “social agenda” of labor and environmental standards. The climate change–trade linkage is likely to rise in importance in coming years, further increasing complexity. The length of time needed to complete a round has steadily lengthened, with the Uruguay Round, which greatly broadened disciplines on services protection, introduced IPR obligations into the system, and established the WTO and its strengthened dispute settlement system, taking from 1986 to 1994 to conclude. Its successor, the Doha Development Round, launched in 2001, remains in suspended animation. There are already calls to shift the negotiating agenda toward more uncharted territory or to de-emphasize multilateral negotiations through the WTO (Crook, 2008; Mattoo and Subramanian, 2009). At the same time that the trade agenda was expanding and impinging more on what had been purely domestic policy concerns, the number of negotiating parties was growing in both nominal and real terms. The GATT had effectively been a US-European Union (EU) condominium. Ironically the very successes of the Uruguay Round – the extended obligations, single undertaking, and strengthened dispute settlement – contained the seeds of the old paradigm’s destruction. In the past, practices such as “special and differential” treatment, which effectively absolved developing

30

Globalization and economic integration

countries from most obligations, and the use of opt-in “codes” on issues such as government procurement, which allowed countries to take on differing levels of obligations, created automatic political safety valves for developing-country participants. The Uruguay Round’s single undertaking at once enhanced the obligations embodied in the system and their enforceability. At the same time, the technical and financial aid promised to developing countries to support taking on these heightened obligations was largely not forthcoming. This experience left many developing-country governments embittered and skeptical of entering into new obligations through the WTO system. By the time of the Doha Development Round, launched in 2001, India, China, and other “new” players were demanding far greater influence on the negotiations than they had been accorded in the past. And while this did not represent a recrudescence of the 1970s North–South split, some of the growing chorus brought with them distinct viewpoints on the role of international trade and the appropriate rules and procedures for its governance. The naming of the round itself was a mistake both conceptually and politically. Trade negotiations can contribute to development, but they are not fundamentally about development – they are about reciprocal trade liberalization. Politically the elevation of “development” raised expectations among developing countries without addressing the reality that the US and the EU would have to be onboard for the round to reach a successful conclusion – which required that their core trade objectives be met.1 Thus an increasing complexity and political sensitivity coincided with growing pluralism, presenting an obvious, and possibly insurmountable, challenge for policy management.2 In the WTO, in principle each member state has an equal voice. But in reality, many smaller and poorer states have inadequate resources to support permanent teams in Geneva. Moreover, with more than 150 members, a negotiation in which representatives of all the members literally sat around a table and hashed out text, is effectively impossible. What has developed instead is a process in which a subset of informally selected major countries engages in negotiations and then in effect presents the rest of the membership with a fait accompli. The current practice of the agreement consisting of a “single undertaking”, in which reservations or opt-outs are not permitted, means that some democratically elected governments are being asked to sign on to agreements that they had no say in formulating – and then to deal with the domestic consequences – underscoring not only the democratic deficit embodied in the decision making of the WTO, but also the problematic legitimacy of agreements reached in this fashion.

The twilight of globalization?

31

Such governance problems are not limited to the WTO. The governance of the global international financial institutions (IFIs) more closely reflects the political realities of the mid-20th century than of the 21st. Take the IMF. Board representation is a function of quota allocations, in principle reflecting the importance of individual countries in the world economy, which set the amount of foreign exchange countries make available for Fund use and notionally determine the level of borrowing for which a country is eligible. The United States, with a quota share of 17 percent, is the only member that can exercise sole veto power over Executive Board decisions under the 85 percent qualified-majority voting system. Over the past several years, the IMF Board has authorized two quota reallocations, which have increased the influence of countries such as China, Mexico, Turkey, and South Korea. The trouble is that there is a zero-sum element to Board representation, and the two overweighted regions, Western Europe and Africa, are reluctant to see a diminution of their representation. The case of Africa is particularly problematic – given the ubiquity of Bank and Fund programs in Africa, the region’s overburdened executive directors already have difficulty adequately representing the interests of their client governments. In light of the influence that the Bretton Woods institutions wield in Africa, many would regard any reduction in Africa’s modest influence within these institutions as fundamentally unfair and illegitimate, and in any event African quotas are really too small to matter in terms of rebalancing. Western Europe, which is overweighted, has blocked more radical recalibration of quota allocations. The United States, which is slightly underweighted, has allowed its own quota to decline (though not enough to imperil its effective veto over Executive Board decisions). It might be possible to combine Western Europe into a single European Union or Eurozone quota, above the 15 percent needed to block Executive Board decisions, and reallocate to underrepresented areas the remaining quota freed by the aggregation of Europe into a single voting entity. Yet to accomplish this, Western European governments would have to be willing to sacrifice national prerogatives, and the United States would have to surrender its veto monopoly. Thus far European intransigence has freed the US from the necessity of confronting this challenge. In Asia, there is neither a single dominant economy equivalent to the US nor a degree of formal regional integration like the EU or the Eurozone that could make cumulation into a single regional voice possible. Governance problems are evident with respect to the selection of senior leadership as well. By tradition the president of the World Bank is an American, the managing director of the IMF a European, and the head of the Asian Development Bank (ADB) Japanese, reflecting the political

32

Globalization and economic integration

legacies of 1944 and, in the case of the ADB, the power of the purse. And despite some problematic choices, neither the US nor the EU appears willing to acquiesce to a reform of this anachronistic system. The growing importance of countries beyond the small group of traditional powers was evident in the November 2008 meeting of the G-20 hosted by the US government in response to the global financial crisis. While the meeting accomplished little, it at least signaled the recognition that the “steering committee” for the world economy would have to be broadened beyond the G-7.

WANING MASS AND ELITE SUPPORT The difficulties of managing globalization at the international level are mirrored at the level of national politics, where there is evidence of waning support for globalization at both the mass and elite levels. In the late 1990s, high-level multilateral economic conclaves such as the annual G-7 heads of government meetings, the joint World Bank–IMF annual meetings, and WTO ministerials were increasingly accompanied by mass, and sometimes even violent, protests culminating in the infamous 1999 “Battle of Seattle” WTO ministerial. The 9/11 terrorist attack and the subsequent public response at least temporarily cowed the anti-globalists, but as the world converges back to some kind of normalcy, it would not be surprising if the globalization issue heated up again at the mass level, particularly in the context of a global recession. Public opinion survey data suggest that support for free markets is relatively ubiquitous, though there is some variation by region in the degree of enthusiasm for markets (Table 2.1). Support declines, however, when respondents are asked about globalization or specific manifestations of globalization such as the multilateral economic organizations discussed in the preceding section. Support for globalization declines even further when one moves from economics to cultural issues. In short, the survey shows general comfort with markets, less comfort with international exchange, and discomfort with the social and cultural impact of globalization. The United States is a case in point. A 2008 poll by the Pew Center for the People and the Press concluded: Americans express increasingly negative opinions toward the World Trade Organization (WTO) and free trade agreements such as the North American Free Trade Agreement (NAFTA). In the current survey, a 48% plurality says that free trade agreements are a bad thing for the country, compared with 35% of the public who call them a good thing. Last November [that is, 2007], opinion about free trade’s impact on the country was evenly split; for the previous

33

Source:

Note:

37.5

23.7

42.3

66.8 61.0 64.3

Western Europe

28.5

49.5 46.0 57.7

Eastern Europe

25.4

49.0 56.1 54.3

Latin America

17.0

67.3 67.7 71.7

Africa

22.2

69.2 71.8 66.5

East Asia

“Views of a Changing World, June 2003”, http://people-press.org/report/185/views-of-a-changing-world-2003.

*MENA = Middle East and North Africa.

66.5 65.0 62.0

North America

59.1 57.3 59.9

Global average

Regional Pew survey responses (percentage of people agreeing)

a. Support free markets b. Globalization, good c. International organizations, good d. Do not protect against foreign influence

Table 2.1

9.0

49.6 41.7 44.8

24.7

58.3 35.5 38.0

South MENA* and West Asia

34

Globalization and economic integration

(%) 60 50 40 30 20 10 Good thing

Bad thing

Don’t know

0 Sep-97 Sep-01 Dec-03 Jul-04 Oct-05 Dec-06 Nov-07 Apr-08 Source:

Pew Research Center (2008).

Figure 2.3

Opinions on impact of free trade agreements on the country

decade, modest pluralities said that free trade agreements were good for the country . . . There is now a broad agreement that free trade negatively affects wages, jobs, and economic growth in America. (Pew Center 2008, section 4).

The Pew Center results of deteriorating popular support for trade are generally reinforced by the limited amount of longitudinal polling on the topic (Figure 2.3).3 It is possible that this decline in support for trade represents a temporary dip rather than a permanent shift in attitudes; there is some evidence to suggest similar trade skepticism during the 1992 recession, which was followed by the congressional passage of NAFTA and the Uruguay Round agreement. Given this popular skepticism, it is not surprising that political decision makers have evinced trepidation. Trade policy has become increasingly partisan in the United States,4 and while trade concerns were probably a minor contributor, the Democratic Party, the more trade skeptical of the two major US political parties, won the White House in the 2008 elections with the election of Senator Barack Obama as president, and increased its majority in both houses of the US Congress as well.5 The 2008 Democratic Party statement on trade reaffirms the party’s commitment to achieving a successful completion of the Doha Round and strengthening the

The twilight of globalization?

35

rules-based multilateral system. But the document says very little about liberalization per se; instead, its focus is on emphasizing enforcement of foreign government obligations of existing agreements, including provisions on labor, environment, and safety standards in trade agreements, and combating currency manipulation. It says nothing about either free trade agreements (FTAs) that have been signed, but await congressional ratification (Colombia, Korea, and Panama), or other potential regional initiatives with Asia and Latin America.6 Such evident apprehensions inevitably will be intensified by financial market turmoil and deteriorating macroeconomic performance in the United States.7 Operationally, the near-term prospects for US trade liberalization at either the global or regional level have been dimmed by the expiration of “fast-track” negotiating authority. The fast-track procedure pre-commits the Congress to a simple up/down vote on implementing legislation – without amendment and within a specified timeframe. Given increasingly fractious US trade politics, it is highly unlikely that – in the absence of such expedited procedures – trade accords with major partners could be successfully concluded and enacted. Nonetheless, if the Obama administration wants to have a significant proactive trade policy, it will need to turn to the Congress to renew fast-track authority, at least for the WTO negotiations, so that it can attempt to salvage the Doha Round. It will not be easy. US credibility was dealt a potentially fatal blow in April 2008 by the congressional decision to alter the fast-track rules ex post in the case of the US-Colombia FTA (Bergsten, 2008b). Some in Congress argue that the Colombian case is unique and should not set a precedent for other fast-track cases, including pending consideration of the Korea–US FTA. However, what matters in this context is not attitudes on Capitol Hill, but rather whether foreign governments will hold back in trade negotiations with the United States for fear that the negotiated deal will be reopened before a congressional ratification vote. In other words, US trading partners will be the ultimate arbiters of how badly the Congress has damaged US negotiating credibility. It is by no means assured that the US political system embodies enough flexibility after the November 2008 elections to actually deliver trade liberalization. In light of the deterioration in the trade policy environment in the United States, it is unclear when the pending FTAs will be ratified, or even if they will come to a vote. Though in fairness, it should be added that the US Congress has never failed to ratify a bilateral trade pact. There is evidence of some modest improvement in world public opinion toward the US, and Barack Obama, who seems to have captivated nonAmericans even more than his fellow countrymen, may enjoy a “honeymoon” giving the US additional room to maneuver (Kull, 2007; BBC,

36

Globalization and economic integration

2008; Harris Interactive, 2008). But confronting this situation may not be at the top of the agenda of the incoming administration. Following the election, Representative Xavier Becerra (D-CA), a protégé of Speaker of the House Nancy Pelosi, who had opposed the Central America FTA, was rumored to be Obama’s choice for US Trade Representative. Becerra declined consideration for the position, however, telling a Spanishlanguage newspaper that “My concern was [over] how much weight the position would have. I reached the conclusion that it would not be priority number one and perhaps not even priority number two or three”.8 Representative Becerra may have assessed the situation correctly: the economics plan of the incoming Obama administration mirroring the Democratic Party platform discusses renegotiating NAFTA but makes no reference to any ongoing trade negotiation or even pending agreements awaiting ratification (Office of the President Elect, n.d.). Similar tales of policy drift could be told with respect to other major players such as Japan and the EU. Japan appears to be stuck in a kind of long-term domestic political stalemate, though the current economic downturn could provoke a tectonic shift in Japanese politics that could break the logjam with unpredictable results. At present, however, due to the influence of the agricultural lobby, Japan appears to be more or less incapable of entering into any trade agreement that involves significant agricultural liberalization. Europe’s situation is complicated by the ongoing process developing modalities of governance at the supra-national level and its own electoral calendar. As in the case of Japan, at present the EU has difficulty making commitments that mobilize significant opposition from national governments, agriculture being a case in point. Europe’s trade commissioner, Peter Mandelson, resigned in October 2008 to join the cabinet in the United Kingdom. (In February 2010, the European parliament elected to give a second term to President José Manuel Barroso and also elected a new Commission.)9

POST-PAX AMERICAN GLOBALIZATION? The question remains whether the “bicycle theory” holds; that is, whether in the absence of forward movement, the trade system inevitably falls backwards. The ongoing global financial crisis will certainly test the resolve of trade policy makers. At the G-20 meeting on 15 November 2008 (as well as at the Asia-Pacific Economic Cooperation (APEC) forum summit the following week) participants committed to a loose standstill, abjuring the introduction of new protections or derogations from existing

The twilight of globalization?

37

liberalization commitments. A number of countries appear to be honoring the commitment in the breach, however: Russia has imposed new tariffs on cars, India levied a new duty on soybean oils, Indonesia has slapped restrictions on a range of products including steel, Brazil and Argentina are considering raising some barriers jointly, and some observers have termed automobile bail-out measures in the US “protectionist”. Apart from the resort to protection in the face of shrinking aggregate demand, the longer-run issue is whether the WTO system has simply become too unwieldy to deliver much in terms of liberalization. The weaknesses of both the multilateral trade and financial institutions have provided impetus for regional and plurilateral initiatives in both the trade and financial fields. Accelerated fragmentation is a distinctly possible outcome. The story with respect to trade is well known: recent years have seen a proliferation of possibly welfare-reducing preferential agreements and a shift in policy attention away from the multilateral system and toward bilateral and regional alternatives. What is less clear is the actual impact of these agreements: researchers are just beginning to get a handle on how comprehensive are their provisions and how widely these preferences actually are used by firms engaged in cross-border trade and investment. In certain respects the situation in finance mirrors that in trade. The global institution with primary responsibility for addressing macroeconomic, balance of payments, and financial crises, the IMF, is widely reviled by developing-country governments and almost wholly ignored by developed-country governments. Given its reputation, governments only go to the Fund under significant duress. A possible turning point came in the Asian financial crisis of 1997–98. Rightly or wrongly many Asians regard the Fund’s performance during this episode as incompetent if not malicious. (In Korea, the crisis is referred to as “the IMF crisis”.) In the aftermath of the crisis, the Asians in effect said “never again”, building up colossal official reserves as a form of self-insurance while exploring reinvigorated Asian regional financial initiatives. Privately, Asian policy makers indicate that any government today that went back to the Fund would be stigmatized if not turned out of office. What sets the Asians apart from other regions of the world where the IMF is regarded with similar antipathy is that the Asians have the financial wherewithal to construct a regional alternative. The 2008 financial crisis emanating from the US has further stimulated interest in regional cooperation. The Chiang Mai Initiative (CMI) is a three-part cooperation framework instituting a network of bilateral, medium-term, foreign-exchange credit arrangements among Asian central banks, undertaking regional

38

Globalization and economic integration

macroeconomic surveillance, and committing to technical assistance. Sixteen bilateral swap agreements, amounting to as much as US$83 billion, have been concluded. In December 2008 the Japanese, Chinese, and Korean governments expanded their existing swaps to $30 billion, in effect matching the agreement that Korea had earlier entered into with the US and delinking their use to any arrangement with the IMF. Under ADB stewardship, CMI participants are in the process of “multilateralizing” their existing bilateral swap networks, and further growth in swap commitments is expected. The Asians are also pursuing other regional initiatives, such as the promotion of a regional bond market and the adoption of a common basket currency peg, with the Japanese-led ADB effectively serving as the secretariat. To some observers, the CMI appears to be an embryonic Asian Monetary Fund (AMF). A key issue is the degree of coordination of lending conditionality between the IMF and a potential AMF. If the AMF were to lend under loose or absent conditionality, the large pool of public money could fuel moral hazard and eventually contribute to the collapse of the globally oriented IMF. Currently, only 20 percent of CMI funds can be drawn before triggering IMF linkage. Nevertheless, this means theoretically that the finance some Asian countries could access through the CMI mechanism now exceeds their IMF quotas. With the Fund itself effectively abandoning conditionality with its new Short-Term Liquidity Facility, it is likely that this linkage to the IMF in the use of CMI funds eventually will be abandoned altogether. Asian countries possess nearly US$4 trillion in official reserves, more than half the world total. If the political will is there, Asia has the financial wherewithal to go its own way. Whether it will depends significantly on the capacity of Japan and China to act cooperatively.

CONCLUSIONS The sustainability of the current second Golden Age of globalization is by no means assured. While technological advances will presumably continue to reduce the costs of moving labor, capital, and goods across borders, the political prerequisites may not survive, victim of both interstate and domestic politics. A global economic downturn is fueling a backlash against the Washington Consensus, the Anglo-American model of capitalism, and the multilateral institutions of economic integration. In this context American leadership is problematic, discredited by its role as the epicenter of the global financial crisis. An absence of constructive leadership may contribute to an erosion of adherence to systemic norms

The twilight of globalization?

39

reminiscent of the system-fraying observed in the first decade of the 20th century as the first Golden Age came to a close. Depending on how the global downturn plays out, this reaction against globalization could broaden into a backlash against capitalism itself. Even short of that apocalyptic development, reform of the existing global regime will prove politically challenging. One clear alternative would be further fragmentation into competing regional blocs. Asia holds the key, combining both dissatisfaction with existing global arrangements with the resources to reconstitute, at least at the regional level, an alternative set of institutions and practices. How Asia handles this situation, acting to strengthen reformed global institutions or undermine them in favor of regional alternatives, will in large part depend on the stances of its two most systemically important countries, China and Japan – as well as the policies of the dominant global power, the United States.

NOTES * I would like to thank Jennifer Lee for research assistance. 1. As an illustration, after a particularly obstreperous performance that contributed to the collapse of the Doha Round negotiations in July 2007, Indian Commerce and Industries Minister Kamal Nath implied that poverty alleviation, not trade liberalization, was the purpose of trade negotiations (Rama Lakshmi, “Hard Line at WTO Earns Indian Praise”, Washington Post, August 1, 2008). 2. One observer singled out China (and India) as behaving particularly irresponsibly in this regard: “China, far from supporting liberalization, used its newfound clout to join India in seeking new protection beyond the red lines of most of the other participants, including many developing countries. Doha may thus become the first global trade negotiation to fail since the 1930s, when protectionism erupted everywhere and brought on a worldwide depression. . . . Ironically, it was China and India – the largest and most successful developing countries – that triggered Doha’s demise because they were unwilling to open their own markets sufficiently to permit an agreement. . . . China’s (and India’s) refusal to accept a significant share of systemic responsibility for the global economy also bodes ill for other upcoming international negotiations – particularly on climate change. China recently replaced the United States as the world’s leading polluter, and Beijing’s refusal to play an active and positive role in climate negotiations would be devastating for the entire planet . . . the China–India alliance that emerged in Geneva could fundamentally alter the politics of global economic negotiations” (Bergsten, 2008a). 3. See also: PollingReport.com http://www.pollingreport.com/trade.htm. 4. See Destler (2005); Layman et al. (2006); Kupchan and Trubowitz (2007). 5. Democrats are considerably more trade-skeptical than Republicans, with solid majorities holding negative views. Republicans are essentially evenly split, 43 versus 42 percent in support of free trade agreements (Pew Research Center, 2008). 6. By contrast, the Republican Party’s platform essentially endorses current US trade policy and calls for prompt congressional action on the pending FTAs. 7. The nonpartisan Trade Policy Study Group (2008) provides additional implicit evidence of these apprehensions: in a sophisticated document presenting a cogent case for trade liberalization, trade policy itself is not addressed until page 10 of a 13–page document, being preceded by discussions of “a new narrative” regarding globalization,

40

Globalization and economic integration

enhancing US competitiveness, and designing a national strategy for promoting economic adjustment. 8. Daniel Dombey, “Obama’s commitment to trade reform thrown into question”, Financial Times, December 18, 2008. 9. The same thing happened at the WTO: Director General Pascal Lamy’s four-year term ended in 2009, but despite grumbling about his leadership and the failure to bring the Doha Round to a successful conclusion, he was reappointed for another term starting September 2009.

REFERENCES BBC Poll (2008), “Global views of US improve”, available at: http://www. worldpublicopinion.org/pipa/articles/views_on_countriesregions_bt/463. php?lb=btvoc&pnt=463&nid=&id, (accessed December 31, 2008). Bergsten, C.F. (2008a), “China and the collapse of Doha”, foreignaffairs.org, updated August 27, http://www.foreignaffairs.org/20080827faupdate87576/cfred-bergsten/china-and-the-collapse-of-doha.html. Bergsten, C.F. (2008b), “The Democrats’ dangerous trade games”, The Wall Street Journal, May 20, p. A23. Blonigen, B. (2008), “New evidence on the formation of trade policy preferences”, NBER Working Paper 14627, Cambridge, MA: National Bureau of Economic Research. Crook, C. (2008), “Obama has to lead the way on trade”, Financial Times, December 22. Destler, I.M. (2005), American Trade Politics, 4th edn, Washington, DC: Institute for International Economics. Harris Interactive (2008), “Financial Times/Harris Poll Monthly Opinions of Adults from Five European Countries and the United States”, December 21, available at: http://www.harrisinteractive.com/news/FTHarrisPoll/HI_ FinancialTimes_HarrisPoll_December2008.pdf (accessed January 7, 2009). Kindleberger, C. (1986), The World in Depression, rev. and enlarged edn, Berkeley, CA: University of California Press. Kull, S. (2007), “America’s image in the world”, Testimony before the House Committee on Foreign Affairs, Subcommittee on International Organizations, Human Rights and Oversight, US House of Representatives, Washington, DC, March 6. Kupchan, C.A. and P.L. Trubowitz (2007), “Dead center: the demise of liberal internationalism in the United States”, International Security, 32 (2), 7–44. Layman, G.C., T.M. Carsey and J.M. Horowitz (2006), “Party polarization in American politics: characteristics, causes, and consequences”, Annual Review of Political Science, 9, 83–110. Mattoo, A. and A. Subramanian (2009), “From Doha to the next Bretton Woods”, Foreign Affairs, 88 (1), 15–26. Office of the President-Elect (n.d.), “The Obama–Biden Plan”, available at: http:// change.gov/agenda/economy_agenda/ (accessed January 7, 2009). Pew Global Attitudes Project (2008), “Global Public Opinion in the Bush Years (2001–2008)”, Washington, DC, December 18. Pew Research Center for the People and the Press (2008), “Trade and Economy:

The twilight of globalization?

41

Obama’s Image Slips, His Lead over Clinton Disappears”, Washington, DC: Pew Research Center for the People and the Press, May 1. Trade Policy Study Group (2008), “A new trade policy for the United States”, Washington, DC, November 25, available at: http://www.petersoninstitute.org/ publications/papers/20081217presidentmemo.pdf (accessed January 5, 2009). World Bank (2009), Global Economic Prospects: Commodities at the Crossroads, Washington, DC: World Bank.

3. Economic integration in extended East Asia: toward a new trade regime Fukunari Kimura* INTRODUCTION East Asia, or extended East Asia, used to be relatively slow in the worldwide boom of regionalism.1 By the year 2000, only two free trade agreements (FTAs) had been successfully concluded in extended East Asia: Australia–New Zealand Closer Economic Relations (CER) and the ASEAN Free Trade Area (AFTA). However, since then, the area has become one of the most active regions in FTA networking. Compared with other regions such as Europe, North America, and Latin America, FTA networking in extended East Asia has three distinctive characteristics. First, with respect to the structure, bilateral and plurilateral FTA networking have developed in an open setting, rather than limiting membership and deepening integration. FTAs are highly flexible policy tools in terms of (i) the speed in concluding negotiations, (ii) the scope of policy modes included in agreements, and (iii) the sequencing in concluding multiple FTAs. Extended East Asia has taken full advantage of such properties of FTAs and has developed FTA networking from ASEAN to multiple ASEAN-plus-one FTAs, and FTAs with countries beyond the region. Second, regarding the nature of its participants, East Asia is a mixture of countries differing widely in income levels and stages of development. The recent participation by Australia, New Zealand, and India has further enhanced diversity. Consequently, FTAs in this region are not simple legal concessions for the symmetric liberalization of the trade regime but are accompanied by asymmetric elements regarding development issues. Deepening economic integration and narrowing development gaps are essential and are at the core of FTAs. Third, with regard to the content, East Asia has developed unprecedented international production networks, particularly in machinery industries. FTAs have been utilized for further activating such networks. FTAs in East Asia thus often include practical policy measures, inside or outside of the formal agreements, to improve the business environment for 42

Economic integration in extended East Asia

43

international production networks, rather than pursuing a rule-oriented comprehensive coverage of various policy modes. These characteristics of FTAs in extended East Asia may become a trigger for pushing the frontier of novel trade regimes in the world. Negotiations on the Doha Development Agenda (DDA) under the World Trade Organization (WTO) are now proceeding very slowly, with a shrinking agenda. While countries demand a political agenda for freer trade, practical regionalism in an open setting with a mixture of developed and developing countries, rather than involving the WTO, is likely to be at the center stage of international trade regime. Academic support for this argument has recently appeared. Mainstream trade economists used to be skeptical about regionalism (since Viner, 1950), preferring unilateral or multilateral trade liberalization to preferential/discriminatory arrangements. However, admitting the reality of regionalism, scholars, including Richard Baldwin, have started examining the political economy of regionalism carefully. They find strong forces heading for freer trade in regionalism, if it is accompanied by proper arrangements (see Baldwin, 2006). It is further asserted that the immediate task is to seek ways to realize “multilateralizing regionalism”, and extended East Asia may be a focal point in this new wave. The next section of this chapter provides an overview on the recent development of FTA networking in extended East Asia and the surrounding Asia-Pacific region. The third section assesses the quality of FTA networking in extended East Asia with novel information on the utilization of FTAs, rules of origin, WTO plus elements, and so on. The fourth section provides some thoughts for interpreting FTA networking in East Asia in the context of new political economy literature on regionalism as a trade liberalizing force. The final section concludes.

FTA NETWORKING IN EXTENDED EAST ASIA FTA networking in extended East Asia is a very recent phenomenon. Table 3.1 presents the development of hub-and-spoke FTA systems centered by ASEAN. Four ASEAN+1 FTAs, namely ASEAN-China, ASEANJapan, ASEAN-Korea, and ASEAN-CER (Australia and New Zealand), have been concluded. ASEAN-India has recently completed negotiations. Although the contents of these FTAs widely vary, a hub-and-spoke FTA system has almost been completed in extended East Asia. In terms of the relative economic size as well as investing capability, not every member of ASEAN is a big player. Much larger economies are at the other ends of the spokes. One obvious condition for ASEAN to be a hub

44

Table 3.1

Globalization and economic integration

FTA networking in extended East Asia (as of March 2009) Japan

Japan Korea

Korea

China

(suspended)

: 2008–

(suspended)

Brunei : 2008

Indonesia Malaysia Philippines : 2008

: 2006

: 2008

: 2007– : 2005–

China ASEAN

ASEAN

: 2008 –

: 2007 –

: 2005 –

: 1993–

(1992)

(1992)

(1992)

(1992)

(1992)

(1992)

(1992)

(1992)

(1992)

Brunei

: 2008

(1992)

Indonesia

: 2008

(1992)

(1992)

Malaysia

: 2006

(1992)

(1992)

(1992)

Philippines

: 2008

(1992)

(1992)

(1992)

(1992)

Singapore

: 2002

(1992)

(1992)

(1992)

(1992)

(1992)

Thailand

: 2007

(1992)

(1992)

(1992)

(1992)

(1992)

(1995)

(1995)

(1995)

(1995)

(1995)

: 2006

: 2009

Vietnam CLM

(1992)

(LM:1997/ (LM:1997/ (LM:1997/ (LM:1997/ (LM:1997/ C:1999) C:1999) C:1999) C:1999) C:1999) *

India Australia New Zealand

: 2008

: 2006

Notes: *CLM = Cambodia, Laos and Myanmar. : signed or being effective, : under negotiation or agreed to negotiate, *: negotiation completed, : feasibility study or preparatory talks. The year indicates when the concerned FTA was in force. “–” after the year means that some ASEAN countries are under the corresponding FTAs in force and other countries follow later. Dark grey indicates FTAs signed before or in the 1990s, grey indicates FTAs signed in the first half of the 2000s, and light grey indicates FTAs signed in the second half of the 2000s. For some FTAs, their status in this table is based on the agreement of trade in goods; negotiations may be still ongoing over other areas such as investment and services even if the agreements are identified as those signed or being effective here. The year in parenthesis shows the year for the corresponding ASEAN country to be a member of ASEAN/AFTA. Brunei, Indonesia, Malaysia, Philippines, Singapore, Thailand, Vietnam and CLM are members of ASEAN. Sources: Websites of trade ministries in each country and others including JETRO website (http://www.jetro.go.jp/world/).

Economic integration in extended East Asia

45

Table 3.1 Singapore : 2002

Thailand

Vietnam

CLM*

India

Australia

New Zealand

: 2007

: 2006 : 2009

: 2008

(1992)

(1992)

(1995)

(1992)

(1992)

(1995)

(1992)

(1992)

(1995)

(1992)

(1992)

(1995)

(1992)

(1992)

(1995)

(1992)

(1995)

(1992)

(1995)

(1995)

(1995)

(LM:1997/ C:1999)

(LM:1997/ C:1999)

(LM:1997/ C:1999) (LM:1997/ C:1999) (LM:1997/ C:1999) (LM:1997/ C:1999) (LM:1997/ C:1999) (LM:1997/ C:1999) (LM:1997/ C:1999) (LM:1997/ C:1999)

* : 2006

: 2005

: 2003

: 2001

: 2005

: 2005

(LM:1997/ C:1999)

: 2005 : 2003

: 2005

: 2001

: 2005

: 1983 : 1983

of FTA networking is due to a delay in FTA connections among China, Japan, and Korea, for longlasting political and historical reasons. ASEAN has taken advantage of this and has made a tremendous effort to stay in the driver’s seat of East Asian economic integration. ASEAN concluded the AFTA in the early 1990s and accelerated trade liberalization after the Asian currency crisis in the latter half of the 1990s. ASEAN now seeks deeper economic integration under the initiative of the ASEAN Economic Community (AEC). At the same time, ASEAN has provided efficient venues for ministerial and summit meetings for neighboring extended East Asian countries.2 There are a number of initiatives that pursue region-wide economic integration. The East Asian Free Trade Area (EAFTA) and the Comprehensive Economic Partnership in East Asia (CEPEA) are examples which are

46

Globalization and economic integration

currently under expert group studies and heading for ASEAN+3 and ASEAN+6 economic integration, respectively. However, due to difficulties for China, Japan, and Korea to come together in the short run, such initiatives for region-wide FTA formation seem to have lost a bit of steam recently. ASEAN may have no great enthusiasm for forming a regionwide FTA either, because the hub-and-spoke system is even more comfortable for them in playing as a hub despite its smaller size. The concept of “East Asia” as a unit of formal economic integration seems to be gradually blurring. Countries in extended East Asia are also active in negotiating and concluding FTAs with countries outside the region. The overlapping region in active FTA networking is the Asia-Pacific. Figure 3.1 shows the current status of FTAs among nine advanced APEC (Asia-Pacific Economic Cooperation) countries: seven OECD (Organization for Economic Cooperation and Development) countries, Singapore, and Chile. As of December 2008, out of 36 bilateral combinations, 20 pairs have already concluded FTAs, and six pairs are under negotiation. At the recent APEC meetings, the US proposal on FTAAP (APEC-wide FTA) has been on the agenda. Although an FTA that covers all 21 APEC member economies does not seem to be feasible in the short run, it may be possible for likeminded countries to get together and start forming a plurilateral FTA under a “dock-and-merge” strategy, gradually increasing the number of participants. The Trans-Pacific Partnership (TPP) including the United States and the P4 (the “Pacific Four” of Brunei, Chile, New Zealand, and Singapore) initiated a full-scale negotiation in September 2008. Australia and Peru also announced their participation in November 2009. These moves may develop the core of Asia-Pacific economic integration. Unlike economic integration in Europe, FTA networking in extended East Asia and the Asia-Pacific has developed as an open-ended system and has gradually weakened a solid regional concept with exclusive membership. As the WTO loosens its grip in the international trade regime, new types of regionalism in extended East Asia and the Asia-Pacific may become a prototype for “multilateralizing regionalism” to promote freer trade in the world.

THE EVALUATION OF FTA NETWORKING IN EXTENDED EAST ASIA Let us now turn to an evaluation of the FTAs in extended East Asia. First, we review the background of FTA networking in the region, particularly from the historical viewpoint of the interaction between de facto and de

Economic integration in extended East Asia

47

Japan

Canada

Korea

Singapore

US

Australia

Mexico

NZ

Chile

Signed and/or in force Under negotiation or being agreed to be negotiated

Source:

Web pages of trade ministries in each country and others.

Figure 3.1

FTA networking among Asia-Pacific developed countries (as of December 2008)

jure economic integration. We then assess various aspects of the liberalization of trade in goods and other policy modes. Finally, the interpretation in the context of political economy will be discussed. De Facto and De Jure Economic Integration in East Asia In East Asia, de facto economic integration started before de jure economic integration. The most significant event on the side of de facto economic integration was the formation of international production networks from the beginning of the 1990s. Although cross-border production sharing and off-shoring/outsourcing to less-developed countries (LDCs) are observed in the US–Mexico relationship, Western and Eastern Europe, and other regions, international production networks in East Asia are distinctive in

48

Globalization and economic integration

(i) their significance for each economy in the region, (ii) their extensive coverage of many countries and regions at the same time, and (iii) their sophistication in combining various types of intra-firm and arm’s-length (that is, inter-firm) transactions.3 The formation of international production networks was backed up by a rich series of piecemeal policy reform. In the mid-1980s, Thailand and Malaysia made a significant step of policy changes for inward foreign direct investment (FDI) during a recession. Other ASEAN countries followed them after several years. In order to attract FDI, these countries needed various requests raised by multinationals for trouble-shooting and accumulated piecemeal investment liberalization and facilitation. In addition, the initiative of information technology agreements (ITAs) supported by APEC and the WTO realized free trade in semiconductor-related electronic parts and components in the latter half of the 1990s. Note that these policy reforms were based not on regionalism but primarily on unilateral liberalization. AFTA was concluded in 1992, but provided a mere advertisement aimed at attracting FDI in order to compete with China emerging as a strong FDI attractor. The actual trade liberalization based on AFTA was minimal until the end of the 1990s. Regionalism in East Asia was at center stage after the Asian currency crisis. ASEAN started to make a collective effort to retain inward FDI by accelerating its integration process and incorporating latecomers in ASEAN. East Asia as a whole acted to establish an anti-crisis vehicle in international financial cooperation and together established the Chiang Mai Initiative. The effort of forming FTAs was launched by Japan–Korea talks in 1998, followed by the formation of Northeast Asia and ASEAN FTAs. The last three countries in extended East Asia, namely Australia, New Zealand, and India, recently deepened their relationships with ASEAN. FTA negotiations were largely motivated by the existing de facto economic integration in the region. In the negotiation process of these FTAs, the major agenda became (i) the restructuring of import-substituting industries such as automobiles, domestic electric appliances, iron and steel, and petrochemicals by removing remaining trade barriers and (ii) the further activation of intra-regional production networks by conducting trade/FDI liberalization and facilitation. These will reflect the content as well as the usage of FTAs in the region.

Economic integration in extended East Asia

49

Liberalization of Trade in Goods Liberalization coverage One of the obvious criteria for evaluating the quality of FTAs is the degree of the “cleanness” of liberalization for trade in goods. Reflecting the huband-spoke system of FTAs with ASEAN at the center, the liberalization coverage of FTAs varies with AFTA the highest. AFTA was concluded as an FTA under the enabling clause of the WTO and did not follow all the disciplines that the WTO imposed. A major deficiency is the length of the interim agreement. The WTO asks countries to substantially complete all trade liberalization within 10 years.4 In the case of AFTA, since the initiation of tariff reductions in the early 1990s, more than 15 years have passed. However, the liberalization coverage of the Common Effective Preferential Tariff (CEPT) scheme of AFTA that specifies gradual tariff reduction schedule is quite high. Under the CEPT, each member country classified traded commodities into the inclusion list (IL), the temporary exclusion list (TEL), the general exception list (GEL), and the sensitive/highly sensitive list (SL/HSL) and gradually moved items from TEL, GEL, or SL/HSL to IL. To date, the original member countries, that is, Brunei, Indonesia, Malaysia, the Philippines, Singapore, and Thailand, have eliminated TEL and have retained GEL and SL/HSL only for very limited commodities (less than 1 percent). Commodities in IL now have 0–5 percent tariffs, which are supposed to be zero by 2010.5 Although AFTA has been long criticized as a lax FTA, it turns out to be a clean FTA in terms of the liberalization coverage.6 In addition, ASEAN recently harmonized traded commodity classifications at the most detailed level. ASEAN-China FTA (ACFTA) and ASEAN-Korea FTA (AKFTA) apply similar tariff reduction schemes to CEPT though they are less clean than AFTA in their liberalization coverage. ACFTA started lowering tariffs under the interim agreement in July 2005 while the so-called Early Harvest Program for agricultural and fishery products (HS01–08) was implemented from January 2004. The interim agreement classified commodities other than those under the Early Harvest Program into (i) normal track 1, (ii) normal track 2 (less than 150 items), (iii) sensitive track (less than 400 items and less than 10 percent of trade values), and (iv) highly sensitive track (less than 100 items and less than 40 percent of items in the sensitive track). The due dates for tariff elimination are 2010 and 2012 for (i) and (ii), respectively. For (iii), the existing tariffs can be retained until the end of 2011, will be reduced to less than 20 percent by 2012 and 0–5 percent by 2018. As for (iv), tariffs should be reduced to less than 50 percent by the beginning of 2015. Items classified in sensitive

50

Globalization and economic integration

and highly sensitive lists differ across countries, though some important electric machinery and transport equipment are included. AKFTA has a similar level of liberalization coverage. Japanese bilateral FTAs with ASEAN countries set up a higher standard for ASEAN countries than ACFTA or AKFTA. For Japanese bilateral FTAs with Brunei, Indonesia, Malaysia, the Philippines, Singapore, Thailand, and Vietnam, the zero-tariff coverage after 10 years in terms of trade values with respect to ASEAN is 99.94, 90 (96 percent including iron and steel for specific use), 99, 97, 100, 97, and 88 percent, respectively. The zero-tariff coverage after 10 years with respect to Japan is often lower though: 99.99, 93, 94, 92, 95 (97 percent after the five-year review), 92, and 95 percent, respectively.7 The lower liberalization coverage for Japan is due to heavy protection of agricultural commodities.8 The asymmetric liberalization commitments are a reflection of Japan’s negotiating power in Southeast Asia as well as the existence of side-payments in the form of investment promotion and economic/technical cooperation from the Japanese side. The recently concluded ASEAN-Japan FTA (AJCEP) applies the CEPT-style tariff reduction scheme. With respect to Japan, 90 percent of commodities (in terms of trade values) will have immediate tariff removal, an additional 3 percent will have within-10–year gradual tariff removal, and the rest will be excluded from liberalization or have certain reduction of tariffs. As for ASEAN-6, 90 percent (in terms of both trade values and the number of tariff lines) will have immediate tariff removals or within10–year gradual tariff removals, and the rest will be excluded from liberalization or have certain tariff reductions. ASEAN latecomers will have a looser schedule of tariff removals or reductions. In summary, AFTA is completing a clean FTA in terms of the liberalization coverage for trade in goods, but other FTAs in East Asia still have dirty aspects. Although manufactured goods are widely covered in liberalization schemes, some specific items, particularly agriculture-related commodities in Japan, retain substantial protection. The recent entry of Australia and New Zealand in FTA networking in the region has provided a pressure on protectionism, although completely clean trade liberalization in East Asia as a whole is yet to come. FTA utilization Tariff reduction or removal does not automatically mean freer trade. Only after utilizing preferential tariffs, are trade liberalization effects realized. FTAs in East Asia, particularly AFTA, have long been criticized for their low levels of utilization. The situation, however, has recently changed drastically.

Economic integration in extended East Asia

51

Thailand and Malaysia disclose data of FTA utilization on an official customs data basis. Table 3.2 presents two countries’ exports, utilizing the AFTA’s CEPT scheme. As of 1998, CEPT was barely utilized, which confirms the old criticism. However, the utilization ratios have increased substantially since then. In 2007, 31 percent of Thailand’s intraASEAN exports and 19 percent of Malaysia’s utilize CEPT where exports to Singapore are excluded because most-favored-nation (MFN)-applied import tariffs in Singapore are zero for almost all products. These ratios are not small because the denominator, total intra-ASEAN exports, includes exports of commodities for which MFN import tariffs are already zero or very low, particularly under ITA, and for which a duty-drawback system is applied as an investment incentive. Table 3.3 shows exports utilizing various FTAs by Thailand and Malaysia. ACFTA and AKFTA do not seem to be well utilized to date, perhaps due to the slow liberalization process or the low public awareness. On the other hand, the Thailand–Australia FTA and the Early Harvest Scheme items in the Thailand–India FTA present very high utilization ratios, 66 and 98 percent, respectively, in 2007. Table 3.4 presents FTA utilization in Thai imports. While FTAs with China and India are barely used, AFTA, Thailand–Australia FTA, and Thailand–New Zealand FTA are relatively well utilized. The Japan External Trade Organization (JETRO) annually conducts an extensive questionnaire survey on foreign affiliates of Japanese firms, which now includes questions related to FTA utilization. The new results (JETRO, 2009, pp. 22–30) show that among manufacturing affiliates of Japanese firms in ASEAN conducting exporting activities, 23.0 percent use FTAs, and 23.3 percent consider using FTAs. Among those with importing activities, 19.7 percent use FTAs, and 24.4 percent consider using FTAs. The questionnaire further asks affiliates not even considering using FTAs for reasons why they do not. Among exporting affiliates without any intention of utilizing FTAs, 37.6 percent of them say that there is a duty-drawback system with respect to imports; 22.9 percent claim that there is no FTA with trading partners; and 19.9 percent state that MFN tariffs at the destination are low, so FTAs are not advantageous. A very small proportion of exporting affiliates raise troublesome administrative procedures or their ignorance of FTAs as reasons for not utilizing FTAs. Similarly, among importing affiliates with no intention of utilizing FTAs, 48.9 percent say that the duty-drawback system for imports is applied; 13.4 percent claim that domestic sales on which tariffs are imposed is small; 13.1 percent state that there is no FTA with trading partners; and 12.8 percent advocate that MFN tariffs are already low.

52

Total Total (excl. Singapore)

Total Total (excl. Singapore)

Thailand

Malaysia

214 206

391 383

7 179 99 212 91 0 17 0 0 0 606 589

1998

1,382 1,242

2,561 2,454

632 748 913 801 594 2 247 4 2 0 3,942 3,696

2003

2,921 2,731

5,146 4,942

1,343 1,333 2,468 1,270 1,227 5 393 22 6 1 8,066 7,673

2005

3,071 2,898

5,509 5,299

1,763 1,529 2,231 1,363 1,270 14 382 23 4 1 8,580 8,198

2006

Exports utilizing CEPT

3,924 3,736

7,865 7,609

2,772 1,928 3,530 1,850 1,206 15 445 30 13 1 11,789 11,345

2007

1.2 3.8

4.0 7.4

0.8 9.3 5.0 11.9 3.9 0.1 0.1 0.0 0.0 0.0 2.2 5.6

1998

5.3 13.2

15.5 23.0

30.3 24.9 20.6 20.7 13.0 0.7 1.1 0.9 0.4 0.0 9.3 18.4

2003

7.9 18.5

21.5 30.0

38.3 33.2 33.9 22.4 16.2 1.3 1.3 2.8 0.6 0.1 13.3 24.6

2005

7.3 16.9

20.2 28.2

36.3 32.0 30.1 20.5 14.9 3.3 1.2 2.3 0.4 0.1 12.4 22.8

2006

Share in total exports

8.7 19.1

22.5 30.9

43.2 34.1 34.3 22.1 13.8 3.0 1.2 2.1 1.0 0.1 14.7 25.7

2007

Sources: Malaysia Ministry of International Trade and Industry, Thailand Ministry of Commerce, trade statistics of Thailand and Malaysia, all cited in JETRO (2008, Table II-12).

Vietnam Philippines Indonesia Malaysia Thailand Brunei Singapore Laos Myanmar Cambodia Total Total (excl. Singapore)

Export destination country/region

Exports utilizing AFTA (CEPT) and their shares in total exports in Thailand and Malaysia (millions of dollars, %)

Total for Thailand and Malaysia

Table 3.2

53

ASEAN (excl. Singapore) China China-ASEAN (excluding Singapore)

Total

7,673 888 8,561

11,345 3,398 14,743

3,736 1,629 403

4,067

7,609 1,769 399 399

2007

24.6 4.8 17.2

18.5 2.9

67.3

30.0 6.7 17.6 79.0

2005

22.8 10.6 18.0

18.4 8.9

62.6

28.2 12.3 18.1 89.1

2006

Share in total exports

25.7 10.8 19.5

19.1 10.0 11.1

66.2

30.9 11.1 14.0 98.1

2007

Sources: Malaysia Ministry of International Trade and Industry, Thailand Ministry of Commerce, trade statistics of Thailand and Malaysia, all cited in JETRO (2008, Table II-13).

8,197 2,493 10,690

2,898 1,043

2,746

2,122 2,731 274

5,299 1,450 328 328

2006

4,942 614 267 267

2005

Exports utilizing FTAs

Malaysia’s trade with South Korea is for June–December 2007.

ASEAN (excl. Singapore) China South Korea

Malaysia

Note:

ASEAN (excl. Singapore) China India (82 items in the Early Harvest Scheme) Australia

Export destination country/region

Exports utilizing FTAs and their shares in total exports in Thailand and Malaysia (millions of dollars, %)

Thailand

Table 3.3

54

Globalization and economic integration

Table 3.4

FTA utilization in imports by Thailand, 2007 (millions of dollars, %)

Exporting country/region

ASEAN China India Australia New Zealand Total Note:

Imports utilizing FTA

Ratio of imports utilizing FTA to the total dutiable imports

3,053 379 35 437 156 4,060

20.5 3.7 3.3 31.4 44.4 14.5

“The total dutiable imports” include imported goods with positive MFN tariffs.

Sources: Thailand Ministry of Commerce and trade statistics of Thailand, cited in JETRO (2008, Table II-15).

The questionnaire asks some additional questions. One is the minimal preferential margin with which exporting affiliates stop using MFN tariffs and start utilizing FTAs. The average margin across exporting affiliates located in ASEAN is 5.2 percent. Another is the preferential tariff rate equivalent to the administrative cost of obtaining the duty-drawback system. The average across importing affiliates located in ASEAN is 1.9 percent. Hayakawa et al. (2008) employ micro data from the JETRO survey and regress the utilization of FTAs on individual affiliates’ characteristics. They find that the utilization of FTAs or the intention to utilize FTAs is positively associated with the size of affiliates and negatively associated with the number of commodity items with zero tariffs. The relationship with the proportion of local procurement presents an inverted-U pattern. Overall, considering other policy arrangements to avoid being taxed such as zero MFN tariffs, the duty-drawback system, and others, the utilization of FTAs seems to be fairly high in ASEAN. However, further facilitation in utilizing FTAs may be required, particularly for small and medium-sized enterprises. Rules of origin (RoO) Possible negative consequences of RoO are one of the major concerns in regionalism. The so-called “spaghetti bowl” or “noodle bowl” phenomenon refers to the trade-deterrent effects that are generated by the complication of trade regimes, particularly regarding RoO, due to the

Economic integration in extended East Asia

55

unorganized proliferation of bilateral/plurilateral FTAs. However, the logic of trade deterrence due to additional FTAs is not very clear. Adding another FTA on top of existing FTAs would certainly complicate a trade regime. However, if private agents think that a new preferential tariff system is too complicated, they will simply continue to use the MFN tariff system or other FTAs. It is very unlikely that additional FTAs reduce trade; instead, the issue we should be concerned with is whether an additional FTA promotes trade or not. In that sense, RoO may indeed work as a counteracting force against trade liberalization by FTAs. Strict and unfriendly RoO may act as protectionism by nullifying the usage of preferential arrangements. Estevadeordal et al. (2007) provide an extensive survey on RoO in FTAs worldwide. They conclude that RoO in intra-Asian FTAs tend to be less restrictive and complex than their counterparts in Europe and the Americas. Sample firm surveys in East Asian countries conducted by Kawai and Wignaraja (2009) suggest unexpectedly few spaghetti bowl phenomena, though further facilitation seems to be needed. Evidence is accumulating that RoO in FTAs in East Asia do not work as a major obstacle to promoting freer trade. Medalla and Balboa (2009) carefully examine RoO in FTAs in East Asia, review best practices in applying RoO, and propose a direction for improvement. First, they claim that an alternative or a co-equal system of RoO is less restrictive than other arrangements and is thus to be promoted. RoO are classified by the testing methodology in identifying the origin of goods. Frequently used tests are the value-added measure test, the tariff heading criterion test, the specified processes test, and a combination of these. The value-added measure test looks simple, but it is not user friendly for some products such as machinery consisting of numerous parts and components. A practical way of avoiding unnecessary user cost as well as saving the cost of negotiation is an alternative or co-equal system in which meeting one of the designated tests, for example, either the value-added measure test or the tariff heading criterion test, may suffice. Table 3.5 tabulates the number of tariff lines applying various types of RoO in AFTA, ACFTA, AKFTA, and AJCEP. ACFTA reflects an old style of RoO that applies the value-added measure test or the regional value content (RVC) test for a large number of tariff lines. AFTA used to have a similar pattern but recently switched to a co-equal system applying either the RVC test or the tariff heading criterion test (CC, CTH, or CTSH in the table) for a large number of tariff lines. AKFTA and AJCEP also apply a co-equal system extensively. Second, Medalla and Balboa recommend wider application of the

56

Globalization and economic integration

Table 3.5

RoO in AFTA, ACFTA, AKFTA, and AJCEP

RoO type WO CC CTH CTSH RVC(>40) RVC(40) RVC(<40)

AFTA

ACFTA

169

8 1

146

4,659

CC + RVC(40) CTH + RVC 564

7

2,583 689

122

AKFTA 465 61 2 36 22 2

AJCEP 3 1,344 434 8 219

2 4

1

487 4 4,078 61

126 3,056 33

CC or RVC(40) CTH or RVC(>40) CTH or RVC(40) CTSH or RVC(40) RVC(40) or Textile Rule CC or RVC(40) or Textile Rule CTH or RVC(40) or Textile Rule Total with alternative rules

300 327 4,463

556

4,630

3,215

NA Total

446 5,224

5,224

5,224

5,224

427

Notes: WO: wholly obtained. CC: change in commodity classification. CTH: change in tariff heading. CTSH: change in tariff subheading. RVC: regional value content. Source:

Medalla and Balboa (2009).

de minimis principle. This principle specifies a maximum percentage of non-originating material to be used without affecting the origin, which can substantially reduce the cost of proving the origin of products in the value-added measure test. Third, although RoO in East Asia seem to be relatively simple and liberal, they recommend further facilitation in the procedure to obtain the certificate of origin. In summary, RoO are certainly important in order to capture the benefit of liberalization effort in FTAs, and there remains room for further facilitation. However, negative consequences of the complication of RoO seem to be limited in East Asia.

Economic integration in extended East Asia

57

Regionalism promoting multilateral liberalization There has been a long debate on whether trade liberalization in regionalism is a building block or a stumbling block to worldwide trade liberalization. Various political economy models can justify both points of view; the issue is thus empirical. In this context, the papers by Estevadeordal et al. (2008a, 2008b) are pathbreaking. They employ extensive time-series data on tariff levels in selected Latin American countries, on both an FTA and an MFN basis. This proves that tariff reduction in FTAs tends to be followed by tariff reductions at the MFN level. Calvo-Pardo et al. (2009) replicate the exercise for ASEAN and find the same pattern. Trade liberalization in FTAs seems to promote multilateral trade liberalization. As pointed out by Ando (2007), in East Asia and other parts of the world we observe that MFN-based liberalization often surpasses gradual liberalization in FTAs. Trade liberalization on an FTA basis seems to be an effective trigger for trade liberalization at the MFN level, particularly in East Asia. Liberalization in Other Policy Modes Taking advantage of their flexibility, FTAs worldwide have increasingly included various policy modes other than policies on trade in goods. Trade in services is a natural extension on which GATS (General Agreement on Trade in Services) Article V imposes a certain discipline. The actual liberalization of trade in services in intra-East-Asian FTAs, however, is relatively modest because countries in the region do not have strong international competitiveness in most of the services sectors. ASEAN has ambitiously set the target of an ASEAN Economic Community (AEC) in 2015. The liberalization of trade in services is one of the major efforts. The ASEAN Framework Agreement on Services (AFAS) was signed in December 1995, and with seven sequential rounds of negotiations between 1996 and 2009 under the purview of ASEAN Economic Ministers (AEMs), the path of liberalization toward “substantially eliminating restrictions to trade in services among ASEAN countries” has gradually been specified. Air travel, healthcare, e-ASEAN (telecommunications and IT services), and tourism as well as logistics are set as priority sectors to realize liberalization earlier, and all the other sectors will follow by 2015 with services negotiations every two years. As a result, ASEAN is supposed to achieve a free flow of services by 2015 with flexibility. In addition, seven mutual recognition agreements (MRAs) have been concluded for professional services. ASEAN’s effort is certainly ambitious, though how far the actual liberalization is realized is still to be tested.

58

Globalization and economic integration

ACFTA and AKFTA include agreements on trade in services, both of which were signed in 2007. However, the structure of the articles closely resembles GATS, and the contents do not extensively explore GATS plus. Bilateral FTAs between Japan and ASEAN member countries include a number of GATS plus, due to sector-by-sector negotiations. However, agreements are not entirely comprehensive, which reflects the relatively weak services sectors in Japan. As for investment, ASEAN concluded the ASEAN Comprehensive Investment Agreement (ACIA) in February 2009, which is an upgraded version of the ASEAN Investment Area (AIA) in 1998, as a part of the comprehensive efforts toward AEC. ACIA includes liberalization, promotion, facilitation, and protection with application of a negative list approach for reservations. How far the reservations will be eliminated is as yet uncertain. ACFTA and AKFTA are supposed to include investment after additional negotiations, though their contents have not yet been disclosed. As for Japan, although AJCEP does not include a meaningful article on investment, bilateral FTAs between Japan and ASEAN countries as well as bilateral investment treaties with Cambodia (signed in June 2007) and Laos (signed in January 2008) deal with investment. They will explore investment liberalization including pre- and post-entry national treatment, bans on some performance requirements, and investment facilitation and protection. These obviously reflect the interests of Japanese firms extending business in East Asia. Other elements in intra-East Asian FTAs reflect development stages and the private sector’s interests of each country in the region. ASEAN has pursued AEC under the scheme of the AEC Blueprint (ASEAN, 2008) in which various policy areas and topics other than policies on goods, services, and investment are listed (Table 3.6). We observe that the content that seems to be workable is highly practical and relevant to the political and economic conditions of ASEAN. FTAs concluded by Japan in the region are also highly pragmatic. For example, the Japan-Indonesia EPA concluded in August 2007 as well as related documents include practical items, in addition to trade in goods, services, and investment, such as energy and mining resources, movement of natural persons and related cooperation, customs procedures, government procurement, competition, intellectual property rights, and cooperation. ACFTA and AKFTA also reflect the status of international relations as well as industrial connections; economic cooperation is always an important sub-element of FTAs. In East Asia, WTO+ works strongly. However, there is no pursuit of the legal comprehensiveness of economic integration. Rather, the motivation

Economic integration in extended East Asia

Table 3.6

59

Characteristics and elements of AEC Highlighted topics

A. Single market and production base A1. Free flow of goods

A2. Free flow of services

A3. Free flow of investment

A4. Freer flow of capital

A5. Free flow of skilled labour A6. Priority integration sectors A7. Food, agriculture and forestry B. Competitive economic region B1. Competition policy B2. Consumer protection B3. Intellectual property rights (IPR) B4. Infrastructure development

B5. Taxation B6. E-commerce C. Equitable economic development C1. SME development C2. Initiative for ASEAN Integration (IAI)

Elimination of tariffs, elimination of nontariff barriers, RoO, trade facilitation, customs integration, ASEAN Single Window, standards and technical barriers to trade Services liberalization under AFAS, mutual Recognition arrangements (MRAs), financial services sector Investment protection, facilitation and cooperation, promotion and awareness, liberalization Strengthening ASEAN capital market development and integration, allowing greater capital mobility, FDI, portfolio investment, other types of flows, capital account transactions, facilitation

Twelve sectors Enhancing competitiveness, cooperation, agricultural cooperatives

Transport cooperation, land transport, maritime and air transport, information infrastructure, energy cooperation, mining cooperation, financing of infrastructure projects

60

Globalization and economic integration

Table 3.6

(continued) Highlighted topics

D. Integration into the global economy D1. Coherent approach toward external economic relation D2. Enhanced participation in global supply networks Source:

ASEAN (2008).

for introducing WTO+ is pragmatic, that is, for serving diplomatic purposes or responding to requests of the private sector extending international production networks. Ultimately, facilitation and cooperation are often emphasized more than liberalization.

POLITICO-ECONOMIC INTERPRETATION The recent political economy literature on trade policies challenges the issue of why countries promote freer trade despite the existence of strong resistance from import-competing sectors. East Asia has obviously made big steps toward freer trade. The East Asian economic integration is certainly a relevant topic in the political economy literature. Following Baldwin (2006), we can present a list of episodes where a politico-economic interpretation seems to work well. Race to the Bottom The politico-economic logic of a “race to the bottom” claims that developing countries may rush into trade liberalization while competing with each other, for example, in attracting FDI by multinationals. As pointed out by Baldwin (2006), trade liberalization in ASEAN before the Asian currency crisis seems to fall into this category. From the latter half of the 1980s and the early 1990s, ASEAN forerunners drastically changed their FDI-hosting policies from “selective” to “open in principle”, recognizing multinationals as essential elements in their industrialization. The emergence of China as a massive FDI attractor in 1992, intensified ASEAN’s sense of alarm about the risk of losing inward

Economic integration in extended East Asia

61

FDI. At that time, import-substituting industries run by multinationals or local firms were still protected by trade barriers, but aggressive unilateral tariff reduction or removal started from parts and components in order to attract network-forming FDI. Shifts toward freer trade were initiated in the first half of the 1990s and were followed by more comprehensive trade liberalization under the initiative of ITA in the latter half of the 1990s. Note that, in this process, neither the GATT nor the WTO took many initiatives. Tariffs were largely reduced and removed unilaterally in terms of applied MFN tariff rates rather than GATT or WTO committed concessional tariff rates, which generated a huge “tariff bindings overhang”. The grip of the WTO on ASEAN is still weak; in a number of countries in the region, bound ratios (the ratio of bound tariff lines) are still substantially lower than 100 percent, and the tariff bindings overhang (gaps between bound MFN tariff rates and applied MFN tariff rates) is large (Table 3.7). Policy changes toward freer trade driven by regionalism started after the Asian currency crisis. Before the crisis, AFTA was used as a type of collective advertisement and did not work as an effective collective forum for trade liberalization. Unilateral trade liberalization based on the “race to the bottom” is the most convincing interpretation for trade liberalization by ASEAN before the Asian crisis. Domino Effect There are at least two episodes in which the “domino effect” of participating in regionalism by multiple countries seems to work well. The first is the expansion of ASEAN. ASEAN started with Indonesia, Malaysia, the Philippines, Singapore, and Thailand in 1967; Brunei joined in 1984. Although the original motivation for ASEAN was rather political, because of the Vietnam War, AFTA was launched in 1992 as a core agreement aimed at regional integration. The following adoption by Vietnam (1995), Laos (1997), Myanmar (1997), and Cambodia (1999) was obviously motivated by the fear of being left out when other neighboring countries were taking steps towards regional integration. The original member countries were also generous in accepting these countries with special treatment. The second is the formation of ASEAN+1 times X hub-and-spoke system of FTAs. Japan and China competed with each other in concluding FTAs with ASEAN. Japan agreed with Singapore to initiate a tripartite (government, industry, and academic) study on an FTA in December 1999, officially started negotiating over an FTA with Singapore in January 2001, and signed it in January 2002. Then Prime Minister Junichiro Koizumi simultaneously proposed a Comprehensive Economic

62

Source:

Note:

–

– –

99.9

12.0

3.4

5.0

30.1

100.0

Peru 2008

3.8

5.0

99.7

9.9

10.0

100.0

China 2006

0.0

0.0

44.1

7.7

37.2

93.9

6.7

7.0

98.8

6.2

25.4

64.2

11.0

n.a.

n.a.

0.0

5.3

94.3

5.7

5.7

100.0

12.8

17.2

91.5

11.8

29.0

70.1

4.8

4.8

100.0

United States 2006

7.7

14.5

62.0

11.8

n.a.

n.a.

Vietnam 2007

10.9

36.0

100.0

Korea Malaysia Mexico 2006 2008 2007

Thailand 2008

Hong Indonesia Japan Kong 2008 2008 2006

Philippines Russia Singapore Taiwan 2008 2007 2003 2005

6.0

25.1

100.0

Chile 2006

APEC Electronic Individual Action Plan (e-IAP) (http://www.apec-iap.org/).

*PNG = Papua New Guinea.

Bound tariff lines as a percentage of all lines Simple average bound tariff rate Simple average applied tariff rate

3.6

3.5

PNG* 2004

27.8

10.3

New Zealand 2006

92.9

96.5

Australia Brunei Canada 2006 2008 2006

Tariff bindings overhang in APEC participants (%)

Bound tariff lines as a percentage of all lines Simple average bound tariff rate Simple average applied tariff rate

Table 3.7

Economic integration in extended East Asia

63

Partnership between Japan and ASEAN, also in January 2002. On the other hand, China was largely tied up with the WTO accession issue up to the middle of 2000 and felt somewhat left behind for involving itself in rising regionalism.9 Chinese Prime Minister Zhu Rongji took a top-down approach and suddenly proposed the formation of a study group with ASEAN in November 2000, which was followed up by the China–ASEAN FTA proposal in November 2001 and the conclusion of the framework agreement in November 2002. Such a quick move by China stimulated Japan, which initiated a set of negotiations over bilateral and plurilateral FTAs (see Appendix 3A, Table 3A.1). In parallel, other countries including Korea and India also negotiated over FTAs with ASEAN. Unlike European integration, ASEAN is not a huge market or a dominating economy for neighboring countries. China’s motivation for FTA networking was mainly political. At the time, China was looking for friendly neighbors, and ASEAN was a natural choice for deeper commitments. Japan was more motivated by economic self-interest. The FDI stock of Japan in ASEAN was, and is even now, larger than that in China, and Japan had a strong incentive for taking opportunities to improve the investment climate in ASEAN. From different incentive schemes, the two countries competed with each other in concluding FTAs with ASEAN. Then ASEAN successfully became a focal point in the regionalism in East Asia, also attracting countries other than Japan and China. Juggernaut Effect The juggernaut effect focuses on the politico-economic interactions between exporting industries and import-competing industries. In order for exporting industries to enjoy a trade opening in partner countries of an FTA, they first need to persuade import-competing industries to remove trade barriers for exporting industries of the partner country. Internal negotiations between exporting industries and import-competing industries can end up with freer trade under the FTA scheme. Furthermore, in a dynamic setting, FTA conclusions may strengthen the political power of exporting industries vis-à-vis import-competing industries, and thus even freer trade can be promoted over time. Such a situation is also convincing in a number of episodes in East Asian FTA networking with some twists. In the case of ASEAN, export industries are basically multinationals or local firms with strong links with them, and thus multinationals and their local chambers of commerce are major actors in the promotion of freer trade and a better investment climate. On the other hand, import-competing industries consist of three groups: (i) small local firms without much political voice, for example, in food processing and light

64

Globalization and economic integration

industries; (ii) large local firms, typically state owned, with certain political power such as in iron and steel and petrochemicals; and (iii) importsubstituting multinationals in electric machinery and automobiles, for example. Types (ii) and (iii) can present some effective resistance against freer trade. Their interaction with freer trade supporters constitutes the politico-economic background of FTA negotiations. Note that freer trade supporters become stronger over time, which accelerates FTA conclusions. AFTA was a sort of practice round in order to adjust to freer trade, preparing for more thorough trade liberalization in FTAs with outsiders. Japan’s case presents harsh domestic interactions between globalizing manufacturing firms and an import-competing agricultural sector. Compared with the concession at the level of the Uruguay Round, FTA negotiations, other than the one with Singapore where no liberalization was conducted in agriculture, forced the agricultural sector to make additional, though modest, trade openings. However, the liberalization coverage is not of a high standard. Such coverage tends to increase over time, and politico-economic power is visibly weakened as the number of concluded FTAs increases, which is consistent with the dynamic juggernaut effect. Kuno and Kimura (2008) examine the liberalization coverage of the agricultural sector in JapanSingapore, Japan-Mexico, and Japan-Malaysia FTAs. In these three FTAs, among 196 agricultural items at the HS 4–digit level, 139 have tariff removals within 10 years for 80–100 percent of sub-items. On the other hand, only 17 items are untouched in these negotiations. All of these products, with rice as the only exception, present a high concentration in production sites. In other words, trade protection is supported by local pressure groups on a narrow geographical basis. These subsectors are certain to be highlighted and isolated in the Japanese political scene. The existence of side-payments in FTA negotiations between Japan and ASEAN allows Japan not to proceed with the trade liberalization of import-competing sectors, namely agriculture, and some “dirty” elements are left over in these FTAs. However, Japan cannot use the same kind of side-payments when negotiating FTAs with developed countries such as Australia, New Zealand, and the United States. Agricultural protection has been well recognized as a major obstacle for Japan to pursue aggressive FTA strategies. Some scholars point to impenetrable difficulties in reforming the notorious agricultural sector and pursuing FTAs with a high liberalization coverage.10 However, the issue is the power balance between free trade supporters and protectionists. Whether the former can claim large benefits or not may be the key to the juggernaut effect.

Economic integration in extended East Asia

65

WHERE NOW? Unlike the European integration, economic integration in East Asia has not been driven by a unified political will of governments in the region. Unlike economic integration in North America, there is not a single dominant leader in East Asia, either. Decentralized forces of political economy have pushed forward FTA networking in East Asia and an openend FTA system has been formulated. Functional deepening of economic integration is likely to continue in order to further activate international production networks. The mechanics of such networks would work for narrowing development gaps across countries and regions in East Asia, which would present a successful case of inclusive or pro-poor growth. If we review the accomplishment of economic integration so far, East Asianwide consolidation of FTAs does not seem to be impossible, at least for trade in goods and some elements of functional WTO+. For the ministerial meetings in August 2009, study groups of EAFTA (ASEAN+3) and CEPEA (ASEAN+6) proposed possible paths of FTA consolidation in East Asia. However, due to the lack of FTAs among Japan, Korea, and China, an East Asian-wide consolidated FTA is unlikely to be realized in the next few years. Rather, Asia-Pacific FTA networking is likely to proceed earlier. FTAs in Asia-Pacific, possibly led by the TPP initiative, would have characteristics different from East Asian FTAs: they tend to have higher coverage of trade liberalization and to be more rule oriented. Singapore, Australia, New Zealand, and possibly Korea seem to be ready to be on board. If such an initiative proceeds, how will Japan, China, and ASEAN respond? New forces of political economy will certainly emerge. All in all, FTA networking has developed in an open setting in East Asia and the Asia-Pacific. The development has been backed up by the logic of political economy. With economic dynamism, East Asia and Asia-Pacific are likely to become a focal point for multilateralizing regionalism.

NOTES *

1.

2.

The author would like to thank Noel Gaston and other participants in the Globalisation and Development Centre Conference “How Globalization Is Shaping the Asia-Pacific: Multi-disciplinary Perspectives”, held at Bond University in September 2008, for helpful comments and encouragement. In this chapter, “East Asia” primarily indicates ASEAN10 (Brunei, Indonesia, Malaysia, the Philippines, Singapore, Thailand, Cambodia, Laos, Myanmar, and Vietnam) plus 3 (China, Japan, and Korea) while “extended East Asia” means ASEAN10 plus 6 (China, Japan, Korea, Australia, New Zealand, and India). The recent cancellation of a number of ministerial and summit meetings in Pattaya,

66

3. 4. 5. 6.

7.

8.

9.

10.

Globalization and economic integration Thailand in April 2009 due to anti-government demonstrations is generating some doubt about ASEAN’s ability to keep providing convenient venues for international meetings. On the other hand, China, Japan, and Korea had the “first” summit meeting, held independently of other international gatherings, in Fukuoka, Japan in December 2008, and announced further annual meetings. With regard to the characteristics of East Asian production networks and the background policy environment, see Ando and Kimura (2005) and Kimura (2006). 1947 GATT XXIV5(c) stated “reasonable length of time”, which is further specified as “10 years” by the Understanding on the Interpretation of Article XXIV of the General Agreement on Tariffs and Trade 1994. Latecomers of ASEAN, that is, Vietnam, Laos, Myanmar, and Cambodia, are supposed to eliminate tariffs for almost all commodities by 2015 or 2018. As of August 2008, the percentage of tariff lines with zero tariffs is 85.4 percent in Brunei, 80.0 percent in Indonesia, 82.6 percent in Malaysia, 82.9 percent in the Philippines, 100 percent in Singapore, and 80.0 percent in Thailand, which clear the interim target of 80 percent. The average tariff rates are 1.95 percent for ASEAN10 and 0.97 percent for ASEAN6 in 2008. See JETRO (2009, p. 24). These figures are obtained from the homepage of the Ministry of Foreign Affairs, Government of Japan (http://www.mofa.go.jp/). Note that the measurement of liberalization coverage in terms of trade values is sensitive to the trade pattern in the base year, which may not properly reflect high spikes of protection. Kuno and Kimura (2008) show that the liberalization coverage of some bilateral FTAs concluded by Japan in terms of the number of tariff lines is substantially lower than the announced figures based on trade values. With regard to agricultural protection in FTA negotiations by Japan, see Ando and Kimura (2008) and Mulgan (2008a, 2008b). Kuno and Kimura (2008) analyze the nature of heavily protected agricultural products, focusing on their geographical concentration of production in Japan. Low coverage of liberalization for agricultural products becomes an obvious obstacle to Japan’s further extending FTA strategies. The official accession was granted in December 2001. Bilateral negotiations over China’s accession with Japan, the United States, and the European Union were agreed in July 1999, November 1999, and May 2000, respectively. The United States endorsed permanent MFN status for China in October 2000. See Mulgan (2008a), for example. She points to the decentralized nature of Japanese negotiating team for FTAs, which reflects the multi-polar bureaucratic system in Kasumigaseki.

REFERENCES Ando, M. (2007), “Impacts of Japanese FTAs/EPAs: post evaluation from the initial data”, RIETI Discussion Paper Series 07–E-041, Research Institute of the Economy, Trade and Industry, Tokyo, available at: http://www.rieti.go.jp/jp/ publications/act_dp.html (accessed August 20, 2009). Ando, M. and F. Kimura (2005), “The formation of international production and distribution networks in East Asia”, in T. Ito and A.K. Rose (eds), International Trade in East Asia, Chicago, IL: University of Chicago Press: 177–213. Ando, M. and F. Kimura (2008), “Japanese FTA/EPA strategies and agricultural protection”, Keio Business Review, 44 (1): 1–25. Association of Southeast Asian Nations (ASEAN) (2008), ASEAN Economic Community Blueprint, Jakarta: ASEAN Secretariat, available at: http://www. aseansec.org/ (accessed August 20, 2009).

Economic integration in extended East Asia

67

Baldwin, R.E. (2006), “Multilateralizing regionalism: spaghetti bowls as building blocs on the path to global free trade”, The World Economy, 29 (11): 1451–518. Calvo-Pardo, H., C. Freund and E. Ornelas (2009), “The ASEAN Free Trade Agreement: impact on trade flows and external trade barriers”, Centre for Economic Performance Discussion Paper. Estevadeordal, A., C. Freund and E. Ornelas (2008a), “Regionalism worries?”, July 25, available at: http://www.voxeu.org/index.php?q=node/1460 (accessed August 20, 2009). Estevadeordal, A., C. Freund and E. Ornelas (2008b), “Does regionalism affect trade liberalization toward nonmembers?”, Quarterly Journal of Economics, 123 (4): 1531–75. Estevadeordal, A., J. Harris and K. Suominen (2007), “Multilateralizing preferential rules of origin around the world”, paper presented at WTO/HEI/NCCR Trade/CEPR Conference “Multilateralizing Regionalism”, Geneva, September 10–12, available at: http://www.wto.org/english/tratop_e/region_e/conference_ sept07_e.htm (accessed August 20, 2009). Hayakawa, K., D. Hiratsuka, K. Shiino and S. Sukegawa (2008), “Maximizing benefits from FTAs in ASEAN”, in Jenny Corbett and Su Umezazi (eds), Deepening East Asian Economic Integration, ERIA Research Project Report 2008, No. 1, pp. 407–55. Japan External Trade Organization (JETRO) (2008), 2008 White Paper on International Trade and Foreign Direct Investment, Tokyo: JETRO (in Japanese). Also see http://www.jetro.go.jp/en/news/releases/20080807699–news. Japan External Trade Organization (JETRO) (2009), Zai Azia Oseania Nikkei Kigyo Katsudo Jittai Chosa 2008 Nendo Chosa (2008FY Survey on Foreign Affiliates of Japanese Firms in Asia and Oceania), 08–ORF 70F-213FB 11, Tokyo: JETRO (in Japanese). Kawai, M. and G. Wignaraja (2009), “The Asian ‘noodle bowl’: is it serious for business?”, ADBI Working Paper Series No. 136 (April), Asian Development Bank Institute, Tokyo. Kimura, F. (2006), “International production and distribution networks in East Asia: eighteen facts, mechanics, and policy implications”, Asian Economic Policy Review, 1 (2): 326–44. Kuno, A. and F. Kimura (2008), “Northeast Asia and FTAs: issues and perspectives”, ERINA Report, 82, July: 3–14. Medalla, E.M. and J. Balboa (2009), “ASEAN roles of origin: lessons and recommendations for best practice”, ERIA Discussion Paper ERIA-DP-2009–17, Economic Research Institute for ASEAN and East Asia, Jakarta. Mulgan, A.G. (2008a), “Japan’s FTA politics and the problem of agricultural trade liberalization”, Australian Journal of International Affairs, 62 (2): 164–78. Mulgan, A.G. (2008b), “Where Japan’s foreign policy meets agricultural trade policy: the Australia–Japan Free Trade Agreement”, Japanese Studies, 28 (1): 31–44. Viner, J. (1950), The Customs Union Issue, New York: Carnegie Endowment for International Peace.

68

Globalization and economic integration

APPENDIX 3A Table 3A.1

Japan’s FTA negotiations (as of March 2009)

Counterpart

Negotiation started

Agreement signed

Singapore Mexico Malaysia Chile Thailand Indonesia Brunei ASEAN Philippines

01/2001 11/2002 01/2004 02/2006 02/2004 07/2005 06/2006 04/2005 02/2004

01/2002 09/2004 12/2005 03/2007 04/2007 08/2007 06/2007 04/2008 09/2006

Vietnam Switzerland

01/2007 05/2007

12/2008 02/2009

GCC** India Australia (Korea)

09/2006 01/2007 04/2007 12/2003

Entry into force 11/2002 04/2005 07/2006 09/2007 11/2007 07/2008 07/2008 12/2008–* 12/2008

(11/2004: negotiation suspended)

Note: *Effective between Japan and Singapore/Laos/Vietnam/Myanmar in December 2008. Other countries are expected to follow soon. ** GCC = Gulf Cooperation Council (Bahrain, Kuwait, Oman, Qatar, Saudi Arabia and the United Arab Emirates). Source:

Ministry of Foreign Affairs, Government of Japan (http://www.mofa.go.jp).

4.

The politics of (anti-)globalization: what do we learn from simple models? David Greenaway and Douglas Nelson*

INTRODUCTION Contrary to some of the more overheated rhetoric on globalization, this process is, in fact, quite reversible. Sometimes lost in the attempts to determine whether current levels of globalization are higher or lower than those in the late 19th century is the fact that globalization came to a screaming halt in the 1930s.1 Simple technological determinism misses the essential role of politics in supporting, or undermining, globalization.2 Such determinism also distracts from at least one fundamental difference between the late 19th-century globalization and the late 20th-century version: the former was characterized by far more restricted democratic politics in the core countries than the latter. As we observe an increasingly confident and aggressive anti-globalization movement, proponents of a liberal international order, to say nothing of stable liberal domestic political economies, need to think hard about both the roots of anti-globalism and the nature of its politics. In this chapter we focus on the latter. Our focus is a preliminary investigation of the link between democratic politics and the stability of globalization in three steps. First, we briefly develop two key distinctions that will provide an analytical framework for our discussion. Specifically, we shall argue that most of the literature on political economy of trade and immigration fails to distinguish between the average level of a policy (say, a tariff) and the variance of that policy (for example, the dispersion of the tariff across sectors), and we shall distinguish between two very broad classes of political economy model (Weberian models and interest-group models). Second, we shall consider how well these models account for policy outcomes, both mean and variance, in trade and immigration policies. We conclude that the pattern of successes and failures is difficult to account for within any of the standard political economy frameworks. This will lead us to the third part of the 69

70

Globalization and economic integration

chapter in which we propose what, for want of a better label, we call the social values extension of both the Weberian and interest-group models.

MODELING THE DOMESTIC POLITICS OF GLOBALIZATION: TRADE AND IMMIGRATION At least among economists, there are two broad approaches to the systematic explanation of policy – the Weberian model and the interestgroup model.3 In both cases, analysis proceeds by constructing a model of the underlying economy of an essentially neoclassical sort. That is, we assume a given set of households and firms: the former are characterized by preferences over final consumption goods as well as endowments of goods and factors; and the latter by technologies for transforming inputs into outputs. Most research on trade and immigration further simplifies by assuming that households are endowed only with factors of production and firms produce only final consumption goods, technologies are constant returns to scale, markets for all goods and factors exist and are perfectly competitive, and there are no externalities in production or consumption.4 Finally, it is quite often assumed that all consumers share the same, generally homothetic, preferences. With a well-specified model of the economy in place, we can complete the political economy model by identifying the politically relevant agent(s), the policy space, and the institutions that constrain policy choice. In both the Weberian and the interest-group models, individual household preferences play a fundamental role. In the former an ideal bureaucrat seeks to choose the policy which is, in some sense, best for society. For the economist, this is an invitation to transform the Samuelsonian social planner of welfare economics into Weber’s ideal bureaucrat, thus transforming normative into positive theory. If we are willing to endow the ideal bureaucrat with a utilitarian objective function and assume that preferences are identical and homothetic (thus aggregable), the analysis becomes trivially easy. The group politics (class of) model is inherently more complex. Where the ideal bureaucrat operates directly on individual welfares and selects an optimal policy based on his or her own objective function, the analysis of group politics proceeds from individual preferences over policy. Each of these must be derived relative to household preferences over final consumption, for each household, and then be mapped somehow into a final policy choice. It is well known, at least since Arrow’s (1951) pioneering work, that this final selection will not generally satisfy a small set of normative axioms intended to reflect minimally democratic commitments.

The politics of (anti-)globalization

71

The positive version of Arrow’s theorem is that we cannot generally expect to predict an equilibrium policy based on detailed knowledge of preferences in a minimally complex institutional environment (that is, we cannot generally expect a preference-induced equilibrium to exist).5 As a result, all analyses make very restrictive assumptions on preferences, economic structure, and political institutions. Standard referendum models, such as Mayer’s (1984) classic paper, generally assume identical, homothetic preferences with a key, but little noticed, assumption of single-peakedness over the one-dimensional policy space, and policy determination by direct referendum. Similarly, the currently popular model of Grossman and Helpman (1994) rests on an exceptionally restrictive model of preferences (identical quasi-linear), economic structure (perfect competition in all markets, specific factors with a freely traded Ricardian numeraire), and political institution (direct sale of clearly delimited policy by a unitary, rational policy maker with a very simple objective function).6 While these restrictions render the models highly dubious as frameworks for structural estimation, the clarity they bring makes them extremely useful as loose guides to both research and thinking about the future. It is this latter purpose for which we use them in this chapter. Specifically, we want to use the Weberian and group politics models to look at the ways that domestic politics respond to changes in international trade and immigration. Because we are particularly interested in the potential for transformation in support for globalization, we argue that it is essential to distinguish between change in the average level of policy and in the dispersion of policy around that average. For example, in the case of international trade policy, considered as level of protection, it is well known that there was a break in the average level of US protection occurring around the time of the Reciprocal Trade Agreements Act of 1934 (RTAA). As Figure 4.1 shows, the US went from being a country characterized by rather highly variable tariffs around a high average, to quite stable low tariffs. While widely commented upon, this systemic transformation has received very little systematic research. At the same time, the variance of the tariff across sectors has continued to be substantial.7 Similarly, overall levels of immigration have varied over time, while dispersion across sources of immigrants as well as a wide variety of other immigrant characteristics (skill/education, gender, age, family status) is also substantial. The distinction between mean and variance in the dependent variable has not generally been made in systematic analysis of the political economy of globalization, so the next two sections consider this issue for trade and immigration policy in both the Weberian and the group politics frameworks.

72

Globalization and economic integration

70 60

Average tariff

50 40 30 20 10

1986-1

1975-1

1964-1

1953-1

1942-1

1931-1

1920-1

1909-1

1898-1

1887-1

1876-1

1865-1

1854-1

1843-1

1832-1

1821-1

0

Year Note:

The average tariff here is defined as (tariff revenues)/(total value dutiable imports).

Source: Historical Statistics of the United States: From Colonial Times to 1970, Washington, DC: US Bureau of the Census; updated from Census Bureau data.

Figure 4.1

US average tariff

UNDERSTANDING THE POLITICS OF GLOBALIZATION: THE WEBERIAN MODEL For our purposes, Max Weber’s theory of the modern state characterizes the ideal state as autonomous from group pressure, unitary, and legitimate.8 It is the first attribute that distinguishes the Weberian model from the group politics model. The second attribute, along with the rationality that is so central to all of Weber’s analysis, allows us to treat state decision making as if it were done by an individual. We shall return later to legitimacy; at this point we simply assume that the objective function characterizing the policy preferences of the state is widely accepted as legitimate. In fact, we shall assume that the state’s objective is to maximize social welfare. That is, we now conceive of the state as an ideal Samuelsonian social planner. It is well known that, under quite general conditions, free trade can be shown to dominate protection for a small economy. From a comparative

The politics of (anti-)globalization

73

static perspective, again for a small country, a liberalizing change is welfare improving.9 There are, of course, a virtually infinite number of exceptions. The theory of economic policy, as applied to international trade policy, deals with the major cases of such exceptions.10 For a country like the US it might seem that the optimal tariff argument (application of monopoly power in trade) would be relevant, but there is very little evidence that politicians in the US have ever considered this a credible argument for protection. As a practical matter, the optimal tariff structure of the US would be as complex as the economy itself. Thus, at least when thinking about the average tariff, free trade is probably as good a baseline as any for thinking about the optimal policy. The striking thing about US policy (in common with virtually all major trading countries), as illustrated in Figure 4.1, is that current policy is strikingly close to this optimum. With an average tariff of less than 4 percent in all the main trading countries of the industrial world, the Weberian model would seem to do an excellent job of accounting for current trade policy in the average sense.11 Furthermore, again as illustrated for the US in the figure, the direction of change in the average has been consistently in the direction of the optimum. When we turn to dispersion of tariff rates, the story would appear to be quite different. While a very small number of countries have a uniform tariff (for example, Chile), not a single major trading nation has adopted such a policy. Not only are statutory tariff rates highly varied, but administered protection mechanisms generate rates that are sizable multiples of bound rates for very specific imports. While any one of these rates might be justified in terms of the theory of economic policy, the structures would appear to be incoherent from an overall perspective. Thus, while the Weberian model seems to provide a coherent account of the average tariff, it appears to fail completely to account for the variance. This, in fact, is the opening wedge for the group politics model of trade policy making that we consider in the next section. When we turn to immigration policy, the content of the state’s objective function is considerably less certain. Immigrants carry many traits that enter only very indirectly in economic welfare, usually proxied by income, but may be highly relevant to social welfare more broadly construed. However, if we apply the same objective function that we used for the case of trade policy, in the context of the same sort of underlying economy, the implication is fairly clear. Even in this case, there are tricky issues about where to count the welfare of the immigrants, but if we use the same utilitarian framework and perfectly competitive baseline, something like free migration would seem to be the central policy prediction of the Weberian model.12 However, where most of the industrial countries that make up the core of the liberal international economy are committed to something

74

Globalization and economic integration

approximating free trade, none of them is committed to anything like free immigration.13 With considerably less confidence than for the case of trade, we conclude that the Weberian model fails to account for the average immigration policy. As with trade, we suppose that the “average” immigration flow (for example, the annual total) is fixed, and consider the allocation of that number across categories of potential immigrants. Sticking with national income as shorthand for aggregate welfare, Borjas (1999) argues that policy should seek to admit “high quality” immigrants – that is, immigrants with a high value of marginal product given the existing technology, tastes and endowment in the host country. Borjas further suggests that this can be implemented by focusing on labor market properties (education or other measures of relevant skills) as well as any directly productive capital that might be brought.14 While most industrial countries are constrained by noneconomic objectives, such as family unification, to give some weight to these goals, virtually all countries give some considerable weight to such economic considerations. By comparison with trade policy, and again with less confidence than for the case of trade policy, we conclude that the Weberian model does a better job of accounting for variance in immigration policy.15

UNDERSTANDING THE POLITICS OF GLOBALIZATION: THE GROUP POLITICS MODEL16 As we noted above, it is the failure of the Weberian model to account for cross-section variance in protection that is the usual opening wedge for the group politics (class of) model. We are using the phrase “group politics model” to refer to all models in which policy is determined fundamentally by more-or-less organized citizen preferences, where those are determined by selfish preferences defined over bundles of final consumption goods. Based on those, it is straightforward to derive preferences over policy. Under the standard assumptions that fundamental preferences are identical across households and factor markets are perfectly competitive so all factors of a given type earn the same rental in equilibrium, preferences over policy are primarily driven by the effect of policy on factor rentals and any government transfers. That is, representing preferences with the indirect utility function mh 5 nh ( p; gh) ; where gh 5 a rizhi 1 Th,

(4.1)

i[I

where zi is factor i ∈ I, ri is the return to factor i, Th is the transfer to h, and h ∈ H refers to a given household. Since the direct effect of increases in

The politics of (anti-)globalization

75

the elements of p on nh(·) is negative for all households, through the effect on cost of consumption, heterogeneity in policy preference comes though endowment and transfer heterogeneity (that is, zh and Th vary among households). While the literature on the political economy of trade policy does not make the distinction between average and variance of policy, one approach is to think of low-dimensionality models (that is, models with one importable and one exportable good) as being about the average level of policy and high-dimensionality models (many goods, many factors) as being about its variance. For the low-dimensionality case, the comparative static effects of a change in trade policy are clear: the Stolper–Samuelson theorem yields factor-based preferences for the case of two inter-sectorally mobile factors and two goods; while the equivalent result for the case of two goods, two sector-specific factors and one mobile factor yields sectorbased preferences. In terms of the equation above, these results tell us about the rj, not about household income (or welfare). The way the models are usually used is to assume that each household is endowed with some quantity of a single factor, so that household income can be tied to returns to that factor. When we turn to the high-dimensionality case, things get considerably trickier. Only analytically local effects for single-factor households can be determined, and these are unlikely to be very useful in thinking about the sorts of non-marginal changes that characterize the sorts of major change in policy that induce political economy analysis.17 It is analytically straightforward to derive the effects of a change in the policy vector on household portfolios, but virtually impossible to identify these portfolios in the data.18 It is probably not surprising that, even in papers that use data with many factors and many goods, the fundamental intuition for interpreting the results runs off low-dimensionality results like Stolper–Samuelson or restrictive economic structures like the specificfactors model. For the case of the average level of protection, we can draw on the sizable literature on citizen preferences over trade policy based on public opinion surveys. As columns two and three in Table 4.1 (from Mayda and Rodrik, 2005) illustrate, with the exceptions of the Netherlands and Japan, majorities of respondents to questions about support for trade restrictions show majorities in favor.19 Loosely speaking, in the context of the referendum model, the median voter in nearly all countries supports trade restriction. To the extent that we have historical data on such opinions, it would appear that in most industrial countries, for most historical periods, and for the period of dramatic liberalization beginning some time in the 1930s in particular, the median voter has been a supporter of protection, or at least an opponent of further liberalization.20 Without reference

76

Globalization and economic integration

Table 4.1

Average opinion on trade and immigration Pro-trade

Anti-trade

Pro-immigration

Germany West Germany East Great Britain USA Austria Hungary Italy Ireland Netherlands Norway Sweden Czech Republic Slovenia Poland Bulgaria Russia New Zealand Canada Philippines Japan Spain Latvia Slovak Republic

0.36 0.22 0.14 0.13 0.16 0.09 0.23 0.22 0.37 0.28 0.24 0.27 0.24 0.14 0.08 0.22 0.25 0.28 0.16 0.34 0.10 0.13 0.25

0.39 0.56 0.63 0.64 0.70 0.71 0.60 0.66 0.29 0.38 0.41 0.52 0.51 0.65 0.77 0.60 0.52 0.46 0.66 0.31 0.71 0.71 0.55

0.03 0.02 0.04 0.08 0.04 0.01 0.04 0.20 0.05 0.07 0.07 0.02 0.02 0.09 0.06 0.08 0.11 0.20 0.11 0.16 0.08 0.00 0.03

Mean Standard deviation

0.22 0.41

0.54 0.50

0.07 0.26

Sources: (2006).

For trade question: Mayda and Rodrik (2005); for immigration question: Mayda

to the roots of preference, we can already see that the group politics model has trouble accounting for average policy. This is what we have elsewhere called “the mystery of missing protection” (Nelson, 2003; Greenaway and Nelson, 2005).21 Attempts to account for cross-sectional variation in protection (or liberalization) constitutes the core of empirical research on the political economy of trade. This large body of research ranges from essentially ad hoc search for correlates of sectoral protection, through work that is loosely motivated by one or another model of group pressure, to putative structural estimates that take a given model very seriously indeed. The results of this large literature, which examines tariff and non-tariff barriers,

The politics of (anti-)globalization

77

voting on trade policy by legislators, and implementation of the administered protection mechanism, are consistent with both informed knowledge about the actual politics involved in the making and enforcement of trade policy and loosely consistent with standard group pressure models.22 In particular, variables intended to capture the return to political activity, the ability to organize, access (for example, representation in key committees, chairmanship of key committees, and so on), and resources invested (for example, Political Action Committee (PAC) contributions and testimony in hearings on trade legislation) consistently and significantly appear with the signs predicted by the theory. Overall, our judgment is that these models do a rather good job of accounting for cross-sectional variance in levels of protection. The inference in group pressure models of immigration is the same as that for trade: we identify the predicted effect of immigration flows on citizen-agents and then identify an equilibrium policy (or comparative static change in policy) based on the distribution of those preferences (or change in the distribution of those preferences). As with trade, for a first cut at explaining the average level of policy, we can examine the distribution of preferences over immigration policy without reference to the underlying determinants of preferences. Like trade (in fact, even more strongly so) the median voter would seem to reject liberalization of the immigration regime. Column four of Table 4.1, drawn from Mayda (2006), shows that in no country is there more than a small proportion of the population willing to support liberalization of a national immigration regime.23 That is, in all countries polled, the median voter would seem to support a regime no less restrictive than the existing regime. Unlike trade, most of the governments involved have adopted restrictive regimes and, at least in recent years, have tightened up those regimes. Thus, at least with respect to average policy, the group pressure model performs reasonably well. The situation is different when we turn to cross-section variance in policy. We have already noted that there are a number of possible ways in which immigrants could be differentiated, but we shall focus on the one most commonly considered in the political economy literature: labor market characteristics and, specifically, skill. Most political economy research that focuses on direct labor market effects as the basis of policy preferences follows the literature in labor economics in assuming a single final output (GDP) produced by a variety of inputs, with a particular focus on various types of labor. This framework naturally drives all adjustment through the wage and makes the attachment of standard trade-theoretic political economy models particularly easy.24 That is, for the small open economy (that is, a country for which the p vector is fixed) an increase

78

Globalization and economic integration

in the endowment of a factor (say, unskilled labor) reduces the return to that factor and raises the return to all others, while raising aggregate income. In the context of a lobbying model, all citizen-agents should resist immigration by factors that compete with their factor and support immigration by others.25 The problem with this research is that, while there is a certain prima facie plausibility to this prediction, the existence of generally restrictive immigration regimes, and occasional outbreaks of aggressive anti-immigrant politics, seems far out of scale relative to estimated labor market effects of immigration.26 Furthermore, public opinion data suggest that it is precisely those groups for whom there are statistically significant wage effects that are less opposed to immigration. Thus, where average trade policy is characterized by a “mystery of missing protection”, variance of immigration policy is characterized by a “mystery of too much protection”. An alternative channel via which immigration might produce a political response based on material interest recognizes the presence of a redistributive state as an essential part of the political economy of immigration. In terms of equation (4.1), we need to focus on the Th terms. In addition to a number of insightful, if informal, analyses of the effect of immigration on the welfare state (for example, Freeman, 1986) a sizable literature has developed seeking to model interactions between immigration and various aspects of the welfare state.27 While this work helps organize thinking about the topic, the results tend to be very model specific and cover a very wide range, with the varying results depending on which redistributive programs are considered (pensions, unemployment, and so on) and what is assumed about the properties of the immigrants and the size and timing of shocks. It is not surprising that there is very little in the way of systematic empirical research based on this theory, and that what there is must be seen as very preliminary. There is a body of empirical research based only loosely on this sort of theory, but seeking to link institutional detail to underlying economics.28 Most of this work is by political scientists and tends to assume the existence of the sorts of labor market effects that we have already seen are hard to establish. Thus, it is hard to see this work as representing a successful account of existing patterns of policy across categories of immigrant. In the context of the federal political system in the US, there are two components of the redistributive system: a federal welfare state and a state welfare state. Simplifying, considerably, the former deals with pensions while the latter deals with health and education. Immigrants tend to be relatively expensive in the state welfare because they tend to: have more school-age children; be poorer and thus receive more state-funded aid; and have lower incomes and, thus, pay less in property and other state taxes.

The politics of (anti-)globalization

79

The National Academy of Sciences study of immigration in the US estimated that, as a result of state welfare expenditures, immigrants in major immigrant gateways resulted in net negative effects, while the effects in the rest of the country were net positive – with an aggregate effect near zero (the members of the commission that produced the report differed on the aggregate effect). In only one gateway was this effect large: in California the central estimate was that the average Californian household paid an additional $1,178 (for the 1994/95 tax year) in taxes as a result of immigration (Smith and Edmonston, 1997, Chapter 6).29 This makes it relatively easy to understand the highly politicized nature of immigration politics in California, but it is not very useful in understanding the overwhelming rejection of a liberal immigration regime in parts of the US that seem to gain via the welfare state channel. On the other hand, if the Borjas/ national labor market model is correct, it is hard to understand why the aggressive public politics of unskilled immigration are so locally focused in California. Overall, and again in contrast to trade policy, it is hard to see that either the direct labor effect models or the indirect redistributive state models provide much explanatory power of variance across categories of immigrant. The central mystery this chapter sets out to identify is illustrated in Table 4.2. It is not clear at all why, for either the Weberian or the group politics model: the politics of average and variance should differ within either policy domain; or why the pattern of success and failure in accounting for these should differ across policy domains. The core of both the standard Weberian and group pressure models are that citizen preferences, derivable from observable, self-regarding, material conditions, fundamentally determine policies. While we can construct a consistent account for each separate case, looking at the average and variance of policy, for two fundamental components of globalization leads us to the puzzle identified in Table 4.2. In the next section we offer the beginnings of an approach to this puzzle.

Table 4.2

Summary of argument Policy domain is:

Dependent variable is

Trade

Immigration

Average

Fails

Works (?)

Dispersion

Works

Fails

80

Globalization and economic integration

UNDERSTANDING THE POLITICS OF GLOBALIZATION: PUBLIC POLITICS AND SOCIAL VALUES In this section, we argue that the difference between the politics of international trade and immigration is not due to any underlying material difference between the issues, but rather to the fact that one, international trade, is treated as a technical issue, while the other, immigration, is treated, when it is treated at all, as a public political issue. We shall argue that the more private the issue, the more it can approach the pure grouptheoretic ideal modeled in the endogenous policy framework, while the more public the issue, the more it becomes attached to broad considerations of social values and the less predictable its outcomes. The first step in developing this argument is to provide greater clarity by developing a distinction due to Schattschneider (1960) between “group politics” and “democratic politics”. For Schattschneider, and as we have used the expression above, group politics is about the pursuit of relatively narrowly defined private interests. In US parlance, group politics is “inside the Beltway” politics – the politics of lobbying. Because group politics are solidly rooted in relatively stable interests, they are predictable and change in predictable ways in response to the, generally marginal, changes in the environment embedding these interests. Not surprisingly, group politics is the focus of virtually all endogenous policy analysis. Early research in the group-theoretic tradition (often called “pluralist theory” by political scientists) saw group politics as a natural mechanism for aggregating preferences.30 Where voting could not convey much information about intensity of preference and, except in the relatively rare case of single-issue referenda, could not convey much specific information about policy, lobbying does both.31 While accepting the positive analysis of group politics, critical pluralists rejected group politics as a normative basis for democratic theory. Much research by critical pluralists involved detailed case studies of particular policy areas which demonstrated the presence of severe asymmetries in representation, resulting in biased outcomes.32 One of the earliest, and most influential, of these critical pluralist analyses was Schattschneider’s (1935) classic study of the making of the Hawley–Smoot tariff. Because of the link between group politics and democracy, many saw critical pluralists as making the argument that democracy was a sham.33 It was in this context that Schattschneider produced his “realist’s view of democracy in America” as a response to this line of argument. Specifically, he argued that while democracies, in common with every other form of political organization known to man, were characterized by a group politics system

The politics of (anti-)globalization

81

possessing all of the biases identified by the critical pluralists, what distinguished democracies from other systems was the presence of a democratic political system that acted as a check on the group politics system. For Schattschneider, democratic politics revolves around the public attempt to identify collectively satisfactory policies. That is, democratic politics is seen as the public politics through which a democratic civil society constitutes itself and through which it is linked to the policymaking apparatus. It is about the legitimation of policies and the governments that formulate them. While elections are the final defense of democratic politics, as well as the key stimulus to the public discourse, as stressed by theorists of deliberative democracy, the core of democratic politics is the public discourse itself. For our purposes, one of the essential attributes of this discourse is that its terms emphasize public interests/ values and downplay private/individual interests. Note that the claim is not that private interests are unimportant in defining one’s interpretation of the public interest, or one’s position in the public discourse, but only that widely held notions of the public interest (as well as attendant notions like “fairness”) constrain that discourse. One of the main sources for the constraining power of these widely held notions is that they affect the willingness of unengaged citizens to take sides in the public discourse. This is one of the keys to the link between democratic and group politics. Stable group politics depends on the participants being generally satisfied with the outcomes. This does not mean that there are not winners and losers, but that both prefer the outcomes under the group politics regime to their expected outcome from public politics. When this condition fails, the loser(s) in the interaction may seek to change the structure by turning to public politics. Similarly, the emergence of new groups may produce a dynamic in which those groups seek to use democratic politics as a resource in their bid to enter the group politics system, or even to overturn the existing group politics in the interest of more radical goals. In either case, success in such strategies involves recruiting citizens who have not taken strong positions, and that involves explicit attempts to link the issue to broadly held normative commitments. Given the relatively unstable nature of such commitment, these strategies tend to be risky.34 At the founding of the Republic, trade was sufficiently central to the definition of the state to be written into the Constitution as a defined responsibility of Congress. In the first century of its existence the tariff was primarily about revenue, however with the end of Reconstruction (1877) both parties cast about for an issue to replace the “Bloody Shirt” and settled on “The Tariff”.35 The capitalization is appropriate here since the issue was not any specific tariff, but the system of high tariffs, often

82

Globalization and economic integration

referred to in the period as “the American system”. With the emergence of the tariff as a, if not the, major basis of continuing electoral contestation between Republicans and Democrats, both parties sought to attach a tariff to a wide range of national goals. Broadly speaking, to the Republicans “The Tariff” was a symbol of national strength, independence, and a strong central government; to the Democrats it was “the mother of trusts”, a symbol of the corruption of national government. However, as a number of ethno-cultural studies of 19th-century voting suggest, the meaning of “The Tariff” varied greatly across elections and regions.36 This is the characteristic of democratic politics in which we are particularly interested. For reasons that are still far from clear, “The Tariff” disappears as a public issue sometime between the writing of the Hawley–Smoot tariff (when classic tariff politics were very much on display) and the Trade Expansion Act of 1962 (when the structure of new trade politics was firmly established).37 Given the suspicion with which the general public treats trade liberalization (as reflected in the poll data) it seems clear that this transition was more-or-less independent of public preference and fundamental to the long period of general trade liberalization. However, while the democratic politics of “The Tariff” disappear, individual tariffs continued to be determined by group politics.38 Even while the Executive branch negotiated steady reductions in the average level of protection, the lobbying system surrounding both the legislation setting the rules under which protection is given (administered protection) and the quasijudicial process actually granting that protection has become even better established. Part of the reason why standard political economy models do such a good job of accounting for cross-sector dispersion of protection is the isolation of the group politics of trade from democratic politics. On the other hand, the failure of such models to account for the average is a result of the decoupling of the setting of the average from democratic politics.39 Unlike trade policy, the early Republic of the US was essentially unconcerned with immigration. The basic immigration law of the US simply asserted that free, white males were free to enter and become citizens. This remained the basic law until the late 19th century when immigration of Chinese and then Japanese became a major public issue in California – resulting in the first major change in the immigration law (the Chinese Exclusion Act of 1882). Even before this, however, nativism (that is, antiimmigrant political activity) had occasionally been a feature of public politics. Like the public politics of trade, and like contemporary antiglobalization politics, there appears to have been a strong link between poor macroeconomic conditions and the attractiveness of restrictive

The politics of (anti-)globalization

83

policies. Also like the public politics of trade, the presence of direct economic foundations (that is, factor-market foundations) of anti-immigrant politics is hard to identify.40 As with contemporary anti-immigrant politics, these public politics tended to be local, episodic, and intense. That is, unlike the politics of trade, immigration tended not to be a continuing issue of national political competition. Rather, anti-immigrant politics tended to emerge in what are now called “gateway” communities, and to emerge primarily in times of economic and/or political stress. Because there was an obvious target, these moments of anti-immigrant politics were often characterized by violence. Perhaps most strikingly, to the extent that group politics grew up around the immigration issue, it did not bear nearly the strong relationship to underlying material interests that characterize the group politics of trade. One of the most telling facts is that there is no equivalent, long-lived, group-based politics surrounding immigration. Following the establishment of the national origin quotas in the Johnson–Reed Act (1924), immigration more or less disappears as a political issue (democratic or group) for forty years – not because it is taken off the table, as with trade, but because the public seem to have no particular interest in the issue. Interestingly, the Immigration and Nationality Act of 1965, which ended the quota system, reflected neither the emergence of new public pressure nor the operation of group politics, but rather derived from its attachment to civil rights issues and, to some extent, to a liberal framing of US international obligations (Gimpel and Edwards, 1999). By the time of the landmark Immigration Reform and Control Act of 1986, while there was a more established set of groups in play: these groups do not have the long history that groups on trade do (that is, most of the established groups go back no further than the politics surrounding the 1965 Act); and, more importantly, there is not the same straightforward material foundation, or broad base, in immigration-related groups.41 It is interesting that, although there is interest-based organization on the immigration issue, this organization does not cover the wide range of economic interests that organization on trade does, and, as we have just noted, much of it focuses on issues that are essentially orthogonal to economic issues in general, and distributive issues in particular. Comparing the lack of both broad interest-based organization and sustained interestbased politics on immigration, to the presence of both on trade would seem to provide strong evidence in favor of our central claim.42 Thus, to the extent that standard political economy models account well for average immigration policy it is because these politics are public politics; while their failure to account for dispersion reflects the lack of a clear material basis for those politics.

84

Globalization and economic integration

CONCLUSIONS Overall, then, the peculiar pattern of success and failure of the grouptheoretic model across dependent variables (average and variance) and issue domains (trade and immigration) does not seem to rely on any obvious material basis. Thus, it seems unlikely that it will be compellingly accounted for by reference to standard political economy models. So, what do we learn from simple political economy models about the coming politics of globalization? It seems likely that, to the extent that globalization becomes a public issue, group-theoretic models (especially those with a strong analytical link to lobbying) will be of very limited use. Globalization politics seem much more likely to be like the politics of immigration than the politics of trade. General public attitudes will play an important role in setting the terms of the democratic politics of globalization, but those terms will be highly contestable. The terms of the public discourse will not be set by economists, and will not likely be identifiable in any simple way from economic self-interest of identifiable groups. At least as important will be how globalization is related to widely, but loosely, held notions like “fairness”. Some steps in the direction of a more systematic understanding of such notions have been taken by scholars working in behavioral economics, but we need much more systematic research on how these work in aggregate in the political economy.43 This is an area where new work on public opinion would be useful – but the emphasis needs to be less on the material foundations of policy attitudes (since these seem weakly held in any event), and more on how citizens see globalization attaching to broader social values. As we noted above, there have been a number of studies of the ways in which the public discourse of democratic politics works, but similar studies on the evolution of trade policy would be very useful. Specifically, the transition in the political economy of trade that we mentioned above is an ideal laboratory for understanding the interaction among insiders, outsiders, institutions, and policy equilibrium. We know too little about how elite attitudes on trade changed so dramatically at a time when citizen attitudes appear not to have changed to the same degree. We still know too little about which institutional changes were essential to the transition, and which less so. And we know too little about how immigration remained a public issue, when trade did not. There is clearly a major agenda for research on the domestic political foundations of a Liberal international political economy, but we shall make little headway if we continue to focus exclusively on the political economy of protection.

The politics of (anti-)globalization

85

NOTES *

1.

2. 3.

4.

5.

6.

7.

8.

9.

10. 11.

This chapter was presented at the 2005 CESifo workshop on Global Economic Negotiations, held in Venice on 20–21 July 2005. The authors are grateful to the conference participants for very useful discussions. Financial support from the Leverhulme Trust under Programme Grant F114/BF is gratefully acknowledged. James (2001) is an excellent treatment of the reversal of globalization during the Great Depression. Bordo et al. (1999) present a useful overview of the data on globalization in the late-19th and late-20th centuries. On these latter issues, also see the papers collected in Bordo et al. (2003). While much political economy research on globalization focuses on the ways in which globalization constrains democratic politics, in this chapter we emphasize the ways in which democratic politics constrains globalization. We abstract from the “stuff happens” approach, according to which all policy acts are specific and require specific explanation involving shifting mixes of ignorance and chance along with the sort of systematic effects on which the Weberian and interestgroup models focus. There are, of course, a number of more technical assumptions that yield sufficient structure to carry out standard comparative static analysis. Virtually any trade theory text develops these assumptions and the fundamental results derivable from such models. Particularly useful texts are: Dixit and Norman (1980); Woodland (1982); and Wong (1995). The logic of this result has been developed in most detail for the case of majority rule in the context of more than one issue, but the point is quite general (as Arrow’s theorem suggests). Among the many excellent presentations of this literature, see Riker (1982) for a sophisticated development of the relationship between results of this sort and the theory of democracy. For an admirably clear development of the formal theory, see Austen-Smith and Banks (1999). This should not be taken as a criticism of the theory. Both the Mayer and the Grossman–Helpman models are paragons of political economy theory. Rather, we simply note that the attempt to capture the structure and dynamics of political economy for complex policy areas, like globalization, require radical simplification to get any results at all. As a result of formula cuts in the GATT/WTO, the absolute dispersion has almost certainly fallen. However, I know of no research suggesting whether this dispersion has fallen relative to the average (that is, change in the variance). Furthermore, the domestic factors affecting the dispersion have certainly changed far less than whatever are the factors that determine the average. See Weber (1978) for the classic treatment and Weber (1921) for a short, but admirably clear, development of key aspects of this theory. Recall that this is an ideal type. Actually existing states will possess these properties only to a greater or lesser extent. This is directly parallel to the treatment of individuals in standard microeconomic theory. For the analysis of policy change, one of the essential assumptions, especially from a positive perspective, is the existence of an ideal redistributive mechanism (lump-sum transfers). While such a mechanism does not exist, the presence of a sizable welfare state surely goes a long way toward moving a potential welfare improvement in the direction of an actual welfare improvement. Bordo et al. (1999), among others, argue that the existence of a welfare state has played a role in supporting trade liberalization for precisely this reason. See Bhagwati et al. (1998) for an admirably clear presentation of all the issues in this paragraph, and the theory of economic policy in particular. The Weberian model does not, at least without some additional work, provide an account for why the tariff was so high prior to the change that occurred in the 1930s. It is in some sense correct, but trivial, to say that the content of the state’s objective

86

12.

13.

14. 15.

16. 17. 18.

19.

20. 21.

22.

Globalization and economic integration function changed. What is clearly needed is a more general account of the content of state preferences and sources of change in those preferences. We return to this question following our discussion of group politics. One of the earliest treatments of this issue is by Henry Sidgwick (1891). Sidgwick concludes that free immigration is the implication of a thoroughgoing utilitarian position, but ultimately rejects that position on practical political grounds. Borjas (1999) simply asserts that maximizing national income is the obvious objective function for immigration policy. In fact, one is far less able to predict a scholar or an activist’s position on immigration based on knowledge of the usual predictors (political commitments: left versus right; normative position: (utilitarian versus Rawlsian versus libertarian versus communitarian; and so on) than it is to predict their position on trade policy. Most economists are perfectly willing to contemplate a smoothly operating redistributive system when advocating free trade, but see only barriers to such redistribution when considering immigration. Contemporary levels of immigration are quite high – approaching the extraordinary levels of the 19th century. These levels, however, only very imperfectly reflect policy. In fact, all countries maintain highly restrictive immigration regimes that tend to be overwhelmed by large illegal flows as well as refugee flows that are only very tangentially related to policies of the sort considered by models of immigration policy. On the evolution of immigration policy, which includes some comparison with trade policy, see the important work reported in Williamson (2004), Hatton (2005), and Hatton and Williamson (2006). Borjas (1999) also suggests that, as a result of the close correlation between country of origin and labor market properties, country of origin could be used as a proxy for policy orientation. From a comparative perspective, countries vary considerably in the degree to which such economic objectives figure in the construction of immigration policy. For example, Australia and Canada have explicit point systems with sizable weight assigned to economic factors, while the US system of preferences assigns much higher weight to noneconomic factors. As with trade, explaining a shift in preferences, such as that implied by the US Immigration Reform and Control Act of 1986, requires an explicit account of the preference shift from a Weberian perspective. This section provides only representative references. For a more detailed survey of the relevant research, see Greenaway and Nelson (2005). See Jones and Scheinkman’s (1977) analysis of friends and enemies for the classic presentation. Ethier (1984) is the industry standard presentation of the results in the high-dimensionality case. See Cassing (1981) for the first paper that showed how using such portfolios allows identification of determinate household income effects, and thus preferences over policy, in the high-dimensional case. For useful generalizations, see Lloyd and Schweinberger (1997) and Lloyd (2000). Based on ISSP question: “(Respondent’s country) should limit the import of foreign products in order to protect the national economy”. The pro- and anti- categories come from aggregating the “agree” and “strongly agree”, and “disagree” and “strongly disagree” responses. Scheve and Slaughter (2001a) provides some of the historical data for the US. This is revealed most clearly in papers of the Goldberg and Maggi (1999) sort, in which a state “preference” parameter accounts for the overwhelming majority of (lack of) protection in cross-section. Given that this parameter has no analytical foundation, these results are essentially the same as saying that the average level of protection is unaccounted for by the model. For surveys with good coverage of the empirical literature, see: Baldwin (1984); Anderson and Baldwin (1987); Rodrik (1995); Magee (1997); and Gawande and Krishna (2003). On administered protection, see Blonigen and Prusa (2003) and Nelson (2006).

The politics of (anti-)globalization 23.

24.

25. 26.

27. 28. 29. 30.

31.

87

The question Mayda (2006) focuses on asks: “Do you think the number of immigrants to the country should be a) reduced a lot; b) reduced a little; c) remain the same as it is; d) increased a little; or e) increased a lot”. The number reported in the table is the share of the population expressing a “pro-immigration” attitude – that is, a response of (d) or (e). See Bilal et al. (2003) for a derivation of policy preferences over immigration policy, and for examples of papers that derive equilibrium immigration policy in such an environment, see Grether et al. (2001) who use a median voter framework or Facchini and Willmann (2005) who use a menu auction framework. The theoretical extension to domestic and immigrant households endowed with multiple factors is as straightforward as its empirical implementation is difficult. With the exception of recent work by George Borjas (for example, 2003), the overwhelming majority of research on labor market effects agrees that these are small and concentrated on very narrowly defined groups made up primarily of earlier cohorts of immigrants with essentially identical labor market traits. For a recent evaluation of this literature, see Card (2005). Borjas’s work proceeds from a fundamental critique of earlier work based on comparison of local labor markets. The key claim is that the US labor market is essentially national, since highly mobile workers will adjust their internal migration decisions to avoid labor markets faced with large immigration shocks. There are two problems with this analysis: first, Card (2001) finds little evidence of such effects on the pattern of migration by native workers; and, more importantly, from a political economy perspective, it is precisely the highly localized nature of public response to immigration that is most striking. Gaston and Nelson (2000, 2002) argue that the standard trade-theoretic model is identical to the standard labor theoretic model in all details but dimensionality. If there are at least as many productive sectors as inter-sectorally mobile factors of production, then Leamer’s (1995) factor-price insensitivity result holds. Intuitively, and without taking factor-price insensitivity as a perfect description of reality, the trade-theoretic model emphasizes adjustment on the output margin as an alternative to adjustment on the wage margin. Given that most estimates of wage adjustment are small, the prima facie plausibility of the trade model as a basic framework for intuition in this case seems established. As a matter of fact, technological change seems to have played a major role in this case. But that just pushes the analysis back to the issue of how to conceive of technological change. If such change were a random phenomenon, unrelated to the immigration shock, then we might still retain the labor model as our intuition driver. That is, citizen-agents unable to count on an appropriate technological shock, should still be expected to resist immigration based on an expectation of negative wage effects. However, if the technological change were a rational response to the immigration shock, then from the perspective of political economy modeling, that is essentially the same as adjustment on the output margin. That is, rational agents should expect adjustments in input mix that will tend to protect the existing wage structure. Recent work by Ethan Lewis (2004a, b, 2005; Card and Lewis, 2007) presents strong evidence that the technological response is endogenous in this latter sense. For representative work of this sort, see Razin and Sadka (2001, 2005), Razin et al. (2002), and Facchini et al. (2005). For representative work of this sort, see: Gimpel and Edwards (1999) on the US; Lahav (2004) on Europe; and Money (1999) for a very interesting comparative study. The next largest negative impact was in New Jersey, where the estimated effect was $232. Greenstone (1975) is still an excellent overview of the classic work in the group-theoretic tradition, with particular reference to the link between group theory and democratic theory more broadly, both the early work that emphasized the democratic virtues of lobbying and the later critical pluralism emphasizing asymmetries and democratic problems. We are referring here to actual lobbying on an issue. The data most commonly used to

88

32. 33.

34.

35. 36. 37.

38. 39.

40.

41.

Globalization and economic integration represent this variable in empirical studies, total lobbying expenditure (on any issue) by organization, are really little more informative than voting data. Note that exactly the same data are used in congressional voting studies on every other issue, so interpretation is bedeviled by exactly the same problem as giving meaning to a vote in a multiple issue referendum or election. Critical pluralism was given the beginnings of a solid theoretical foundation by Mancur Olson’s (1965) The Logic of Collective Action, which provided a systematic account of asymmetric organization among groups seeking private outcomes from government. Many Marxists at the time made this claim quite explicitly. More generally, this research was seen as related to results from the early voting studies suggesting that large numbers of citizens had very little detailed knowledge of the candidates or issues on which they were casting votes. Together with the detailed case studies of critical pluralists, this led to something of a crisis in normative democratic theory. Schattschneider (1960) likens this to the process by which a fight is transformed into a brawl as participants in the fight seek allies from the crowd. Riker (1986) develops a more formal analysis of such strategies, which he calls “heresthetic”. For case studies, see Baumgartner and Jones (1993, 2002) and Rochefort and Cobb (1994). “The Bloody Shirt” refers to the wounds suffered by the Union soldiers in the Civil War. The transition to the use of trade as the most important ongoing issue between Republicans and Democrats is well described in Reitano (1994). McCormick (1974) is an excellent overview of this research. The literature on the transformation of US trade politics in this period is large and has produced no compelling account of the transformation. Among such accounts are those stressing: elite learning; domestic institutional change (specifically adoption of an income tax and the Reciprocal Trade Agreements Act of 1934); international institutional change (mainly the General Agreement on Tariffs and Trade); female franchise; and change in the mobility of factors of production. Reviews of these accounts can be found in Hiscox (1999) and Nelson (2003). In addition to Schattschneider’s (1935) classic, which we have already mentioned, the importance of group politics to the determination of early tariff politics is made clear in Taussig (1931). This is also why poll data on general public preferences over trade data are of very little use in understanding the politics of trade policy: they have essentially nothing to do with the average level of protection because of this decoupling; and they tell us very little about the politics of dispersion because dispersion is set by lobbying, not by public politics. That is, the great majority of citizens, whose preferences may be well measured by the polls, are simply unrepresented in the politics of trade. The literature on public opinion on immigration is large and suggests strongly that broad social values have a major impact on preferences for immigration policy. A number of recent studies have identified a significant element of material interest (Scheve and Slaughter, 2001b; O’Rourke and Sinnott, 2002; Mayda, 2006; Hatton, 2005). However, recent work by Hainmueller and Hiscox (2007) successfully shows that these results, which use education to identify labor market position, are more likely to be identifying general values. The best treatment of the politics of this period is Schuck (1992). A couple of exceptions require careful consideration. On the one hand, there are a small number of groups with clear material interests that have been involved in immigration politics on more or less the same terms as trade-related groups. Southwestern farmers, orchard owners, and ranchers have been actively involved in immigration politics. More recently are employers in the computer industry that have aggressively sought liberalization of entry for skilled labor. However, the narrowness of these interests relative to the wide base of economic interests makes the immigration groups exceptions that prove the rule. On the other hand, immigration lawyers have played an important role in the politics of immigration policy. In understanding their role, however, it is useful to compare the immigration bar with the trade bar. Both have an obvious interest in the details of the

The politics of (anti-)globalization

42.

43.

89

law regulating their areas of practice, but these two groups of lawyers do very different things: the trade bar is essentially in the lobbying business, they represent broad parts of American industry and labor; the immigration bar represents a much less obviously material interest and what they do seems different. In addition, a range of humanitarian, religious, and other groups play large roles that they do not play in the trade context. It may be that part of the reason why the group politics of immigration appear so different from those of trade is that the opportunities to engage in group-based politics are so few. In addition to fairly regular legislation on trade issues, there are anti-dumping, countervailing duty, escape clause, unfair trade practices (301: that is, Section 301 of the 1974 (US) Trade Act), (a few) national security cases, and so on. In all of these the plaintiff is an industry. This is also, indirectly, true in the Court of International Trade cases. And we should not forget that there is virtually always Geneva-based action of one kind or another. All of these induce broad sector-based, and, since unions are actively involved, factor-based organization on the issue. There do not appear to be nearly the range of opportunities for group-based politics on immigration. For a very preliminary effort in this direction, with some related references, see Davidson et al. (2006).

REFERENCES Anderson, K. and R. Baldwin (1987), “The political market for protection in industrial countries”, in A.M. El-Agraa (ed.), Protection, Cooperation, Integration and Development, New York: Macmillan. Arrow, K. (1951 [1968]), Social Choice and Individual Values, New Haven, CT: Yale. Austen-Smith, D. and J.S. Banks (1999), Positive Political Theory, I: Collective Preference, Ann Arbor, MI: University of Michigan Press. Baldwin, R. (1984), “Trade policies in developed countries”, in R. Jones and P. Kenen (eds), Handbook of International Economics, vol. 1, Amsterdam: North-Holland. Baumgartner, F. and B. Jones (eds) (1993), Agendas and Instability in American Politics, Chicago, IL: University of Chicago Press. Baumgartner, F. and B. Jones (eds) (2002), Policy Dynamics, Chicago, IL: University of Chicago Press. Bhagwati, J., A. Panagariya and T.N. Srinivasan (1998), Lectures in International Trade, 2nd edn, Cambridge, MA: MIT Press. Bilal, S., J.-M. Grether and J. de Melo (2003), “Attitudes towards immigration: a trade-theoretic approach”, Review of International Economics, 11 (2), 253–67. Blonigen, B. and T. Prusa (2003), “Antidumping”, in E.K. Choi and J. Harrigan (eds), Handbook of International Trade, Oxford: Blackwell, pp. 251–84. Bordo, M., B. Eichengreen and D. Irwin (1999), “Is globalization today really different from globalization a hundred years ago?”, in S. Collins and R. Lawrence (eds), Brookings Trade Forum, 1999, Washington, DC: Brookings Institution, pp. 1–72. Bordo, M., A. Taylor and J. Williamson (2003), Globalization in Historical Perspective, Chicago, IL: University of Chicago Press/NBER. Borjas, G. (1999), Heaven’s Door: Immigration Policy and the American Economy, Princeton, NJ: Princeton University Press.

90

Globalization and economic integration

Borjas, G. (2003), “The labor demand curve is downward sloping: reexamining the impact of immigration on the labor market”, Quarterly Journal of Economics, 118 (4), 1335–74. Card, D. (2001), “Immigrant inflows, native outflows, and the local labor market impacts of higher immigration”, Journal of Labor Economics, 19 (1), 22–64. Card, D. (2005), “Is the new immigration really so bad?”, Economic Journal, 115 (507), F300–F23. Card, D. and E. Lewis (2007), “The diffusion of Mexican immigrants during the 1990s: explanations and impacts”, in G. Borjas (ed.), Mexican Immigration to the United States, Chicago: University of Chicago Press/NBER, pp. 193–277. Cassing, J. (1981), “On the relationship between commodity price changes and factor-owners real positions”, Journal of Political Economy, 89 (3), 593–5. Davidson, C., S. Matusz and D. Nelson (2006), “Fairness and the politial economy of trade”, World Economy, 29 (8), 989–1004. Dixit, A. and V. Norman (1980), Theory of International Trade, Cambridge: Cambridge University Press. Ethier, W. (1984), “Higher dimensional issues in trade theory”, in R. Jones and P. Kenen (eds), Handbook of International Economics, vol. 1, Amsterdam: North-Holland, pp. 131–84. Facchini, G., A. Razin and G. Willmann (2005), “Welfare leakage and immigration policy”, CESifo Economic Studies, 50 (4), 627–45. Facchini, G. and G. Willmann (2005), “The political economy of international factor mobility”, Journal of International Economics, 67 (1), 201–19. Freeman, G. (1986), “Migration and the political economy of the welfare state”, Annals of the American Academy of Political and Social Science, no. 485, 51–63. Gaston, N. and D. Nelson (2000), “Immigration and labour-market outcomes in the United States: a political-economy puzzle”, Oxford Review of Economic Policy, 16 (3), 104–14. Gaston, N. and D. Nelson (2002), “The wage and employment effects of immigration: trade and labour economics perspectives”, in D. Greenaway, R. Upward and K. Wakelin (eds), Trade, Investment, Migration and Labour Market Adjustment, Basingstoke: Palgrave-Macmillan, pp. 201–35. Gawande, K. and P. Krishna (2003), “The political economy of trade policy: empirical approaches”, in E. K. Choi and J. Harrigan (eds), Handbook of International Trade, Oxford: Blackwell, pp. 213–50. Gimpel, J. and J. Edwards (1999), The Congressional Politics of Immigration Reform, Boston, MA: Allyn & Bacon. Goldberg, P. and G. Maggi (1999), “Protection for sale: an empirical investigation”, American Economic Review, 89 (5), 1135–55. Greenaway, D. and D. Nelson (2005), “The distinct political economies of trade and immigration policies”, in F. Foders and R.J. Langhammer (eds), Labor Mobility and the World Economy, New York: Springer-Verlag, pp. 295–327. Greenstone, J.D. (1975), “Group theories”, in F. Greenstein and N. Polsby (eds), Micropolitical Theory, Reading, MA: Addison-Wesley, pp. 243–318. Grether, J.-M., J. de Melo and T. Muller (2001), “The political economy of migration in a Ricardo–Viner model”, in S. Djajic (ed.), International Migration: Trends, Policy, Impact, London: Routledge, pp. 42–68. Grossman, G. and E. Helpman (1994), “Protection for Sale”, American Economic Review, 84 (4), 833–50.

The politics of (anti-)globalization

91

Hainmueller, J. and M. Hiscox (2007), “Educated preferences: explaining attitudes toward immigration in Europe”, International Organization, 61 (2), 399–442. Hatton, T. (2005), “Trade policy and migration policy: why the difference?”, Manuscript, Australian National University, Canberra. Hatton, T. and J. Williamson (2006), Global Migration in the World Economy: Two Centuries of Policy and Performance, Cambridge, MA: MIT Press. Hiscox, M. (1999), “The magic bullet? The RTAA, institutional reform and trade liberalization”, International Organization, 53 (4), 669–98. James, H. (2001), The End of Globalization: Lessons from the Great Depression, Cambridge, MA: Harvard University Press. Jones, R. and J. Scheinkman (1977), “The relevance of the two-sector production model in trade theory”, Journal of Political Economy, 85 (5), 909–35. Lahav, G. (2004), Immigration and Politics in the New Europe: Reinventing Borders, Cambridge: Cambridge University Press. Leamer, E. (1995), “The Heckscher–Ohlin model in theory and practice”, Princeton Studies in International Finance, no. 77. Lewis, E. (2004a), “Local open economies within the US: how do industries respond to immigration?”, Federal Reserve Bank of Philadelphia Working Paper 04–1. Lewis, E. (2004b), “How did the Miami labor market absorb the Mariel Immigrants?”, Federal Reserve Bank of Philadelphia Working Paper 04–3. Lewis, E. (2005), “Immigration, skill mix, and the choice of technique”, Federal Reserve Bank of Philadelphia Working Paper 05–8. Lloyd, P. (2000), “Generalizing the Stolper–Samuelson theorem: a tale of two matrices”, Review of International Economics, 8 (4), 597–613. Lloyd, P. and A. Schweinberger (1997), “Conflict generating product price changes: the imputed output approach”, European Economic Review, 41 (8), 1569–87. Magee, S. (1997), “Endogenous protection: the empirical evidence”, in D.C. Mueller (ed.), Perspectives On Public Choice: A Handbook, New York: Cambridge University Press, pp. 526–61. Mayda, A.M. (2006), “Who is against immigration? A cross country investigation of individual attitudes toward immigrants”, Review of Economics and Statistics, 88 (3), 510–30. Mayda, A.M. and D. Rodrik (2005), “Why are some people (and countries) more protectionist than others?”, European Economic Review, 49 (6), 1393–430. Mayer, W. (1984), “Endogenous tariff formation”, American Economic Review, 74 (5), 970–85. McCormick, R. (1974), “Ethno-cultural interpretations of nineteenth century voting behavior”, Political Science Quarterly, 89 (2), 351–77. Money, J. (1999), Fences and Neighbors: The Political Geography of Immigration Control, Ithaca, NY: Cornell University Press. Nelson, D. (2003), “Political economy problems in the analysis of trade policy”, Inaugural Lecture, University of Nottingham. Nelson, D. (2006), “The political economy of antidumping: a survey”, European Journal of Political Economy, 22 (3), 554–90. O’Rourke, K. and R. Sinnott (2002), “The determinants of individual trade policy preferences: international survey evidence”, Brookings Trade Policy Forum 2001, pp. 157–96.

92

Globalization and economic integration

Olson, M. (1965), The Logic of Collective Action, Boston, MA: Harvard University Press. Razin, A. and E. Sadka (2001), “Interactions between international migration and the welfare state”, in S. Djajic (ed.), International Migration: Trends, Policy, Impact, London: Routledge, pp. 69–88. Razin, A. and E. Sadka (2005), The Decline of the Welfare State: Demography and Globalization, Cambridge, MA: MIT Press/CESifo. Razin, A., E. Sadka and P. Swagel (2002), “Tax burden and migration: a politicaleconomy theory and evidence”, Journal of Public Economics, 85 (2), 167–90. Reitano, J. (1994), The Tariff Question in the Gilded Age, University Park, PA: Pennsylvania State University Press. Riker, W. (1982), Liberalism against Populism: A Confrontation between the Theory of Democracy and the Theory of Social Choice, San Francisco, CA: W.H. Freeman. Riker, W. (1986), The Art of Political Manipulation, New Haven, CT: Yale University Press. Rochefort, D. and R. Cobb (eds) (1994), The Politics of Problem Definition: Shaping the Policy Agenda, Lawrence, KS: University Press of Kansas. Rodrik, D. (1995), “Political economy of trade policy”, in G. Grossman and K. Rogoff (eds), Handbook of International Economics, vol. 3, Amsterdam: North-Holland, pp. 1457–94. Schattschneider, E.E. (1935), Politics, Pressure, and the Tariff, Englewood Cliffs, NJ: Prentice-Hall. Schattschneider, E.E. (1960), The Semi-Sovereign People: A Realist’s View of Democracy in America, New York: Holt, Rinehart. Scheve, K. and M. Slaughter (2001a), Globalization and the Perceptions of American Workers, Washington, DC: Institute for International Economics. Scheve, K. and M. Slaughter (2001b), “Labor market competition and individual preferences over immigration policy”, Review of Economics and Statistics, 83 (1), 133–45. Schuck, P. (1992), “The politics of rapid legal change: immigration policy in the 1980s”, Studies in American Political Development, 6, 37–92. Sidgwick, H. (1891), “Principles of external policy”, Chapter XVIII of The Elements of Politics, London: Macmillan, pp. 285–315. Smith, J. and B. Edmonston (eds) (1997), The New Americans: Economic, Demographic, and Fiscal Effects of Immigration, Washington, DC: National Academies Press. Taussig, F. (1931), The Tariff History of the US, New York: G.P. Putnam’s Sons. Weber, M. (1921), “Politics as vocation”, Translation in H.H. Gerth and C. Wright Mills (eds) (1946), From Max Weber: Essays in Sociology, New York: Oxford University Press, pp. 77–128. Weber, M. (1978), Economy and Society, Berkeley, CA: University of California Press. Williamson, J. (2004), The Political Economy of World Mass Migration, Washington, DC: AEI Press. Woodland, A. (1982), International Trade and Resource Allocation, Amsterdam: North-Holland. Wong, K. (1995), International Trade in Goods and Factor Mobility, Cambridge, MA: MIT Press.

PART III

Macroeconomic coordination and financial market integration

5.

Asian financial integration Jennifer Corbett

INTRODUCTION Many studies of the causes of the Asian financial crisis of 1997–98, and of the changes since those years, focused on the way the region’s financial systems linked with other financial systems and on flows of financial resources between them. The crisis was marked by the double problem of currency and maturity mismatches when firms borrowed short term, in foreign currencies, to fund long-term projects. In the years since the crisis the region has generated large savings surpluses and concern is now sometimes expressed that these savings are not available within the region to fund the obvious investment needs in infrastructure and other development areas. That concern also grows out of the experience leading to the crisis, when dependence on outside sources of financing exposed financial systems to instability. But it would be the wrong lesson to draw the conclusion that regional cooperation should be aimed at directing regional savings to regional investment needs. The patterns of financial flows within the region and between it and the rest of the world, are a reflection of the fundamental macroeconomic patterns of savings and investment and of the risk–return characteristics of financial instruments available regionally and globally. As long as the regions’ excess savings result in accumulated holdings of foreign reserves or in financial claims held outside the region it will be the case that the regions’ finances are intermediated outside the region. The financial resources do not necessarily stay outside the region (the data make it impossible to identify the ultimate owners of many forms of financial asset) and they can be presumed to be flowing to the location of highest risk-adjusted return or, if not, to be reflecting other calculations of risk (such as expropriation risk) or institutional and regulatory barriers to the desired allocation of capital. The policy lessons to be drawn from these characteristics of regional finance need to be carefully considered since they may not be those that appear to leap most readily to mind. This chapter is intended to raise some issues about the meaning of regional financial integration and to the policy questions that arise from a

95

96

Globalization and economic integration

careful consideration of what can, and cannot, be said about the present state of, and benefits from, integration. The next section discusses definitions of financial integration and draws a distinction between measures of integration that capture features of markets for particular financial assets and measures that capture behavior in geographically defined areas which directly relate to the welfare of residents in those areas. The third section extends this to highlight the costs and benefits of financial integration as a prelude to a discussion of whether there is a theoretical case to be made for closer regional financial integration. The fourth section presents some data on regional integration and surveys a number of studies of aspects of financial integration in the region. This section makes a distinction, often overlooked in the literature, between measures of “openness”, measures of financial “engagement”, and measures of “integration”. Openness can be affected by policy actions since it reflects direct policy measures such as the presence or absence of barriers but neither engagement nor integration, in the measurement sense used here, can be directly affected by policy. These two measures reflect the outcomes of financial flows. Furthermore it does not necessarily follow that an increase in openness will lead either to greater financial engagement or to more integrated markets. The fifth section addresses policy implications of the preceding discussion. Policy makers should recall that welfare results come from the effects of integration, not from the integration itself, that is either from the growth enhancing effect of capital inflow or from the consumption risk sharing achieved by integration. There is no basis for piecemeal policy interventions that seek sporadically to build institutional or market structures merely because they appear to be missing rather than because they can be shown to be welfare enhancing. The final section concludes.

WHAT IS FINANCIAL INTEGRATION? Despite the frequency of discussion of financial integration (even before the global financial crisis of 2008) there is no single definition of what precisely it means, no clear view of which markets are integrated with which others, and a considerable debate about how it should be appropriately measured. Considering the amount of policy energy devoted to either promoting or restricting it, this constitutes a puzzle in itself. There are other puzzles that will be described in this chapter, but the main focus of the discussion here is to consider what regional financial integration within Asia might mean and how the policy debate about it should be interpreted. The chapter does not discuss the various institutional developments that have taken place as part of the efforts to build a new

Asian financial integration

97

financial architecture in the region – the purpose is to enquire about the economic rationale underlying those actions. Definitions can be divided into broadly two types: those dealing with a geographical notion of financial markets and those dealing with markets defined by asset classes. A fundamental definition that would command fairly wide agreement is that a set of financial markets would be integrated when all investors have access to all markets, and all financial instruments within them, without regard to national borders or other regulatory impediments. Flood and Rose (2004) provide a technical definition that applies naturally to particular asset markets and would be recognized by financial economists and practitioners: “financial markets are integrated when asset prices are priced by the same stochastic discount rate” (p. 2). This definition can also (as in their work) be extended to geographically defined financial markets. Another definition, explored below, which applies more naturally to geographic markets, is that they are integrated when there is perfect consumption correlation between them. Such a definition makes sense when thinking of groups of consumers who “live” within the scope of the particular market and whose consumption can be measured. The idea of financial integration is also closely linked to the idea of openness of financial markets in the sense of an absence of barriers to transactions. This is also related to the idea of financial market liberalization but is not the same. Liberalization refers to removal of regulatory barriers and would be expected to increase openness and could lead to greater integration if barriers both at the border and behind the border are removed. As will be shown below, there is an intuitive link between these notions but it is not always easy to bring the existing empirical literature into line with them. By creating the possibility of arbitrage, liberalization and openness have implications for prices, and this is the sense in which the three terms above are linked. Asset prices are unlikely to be priced in the same way (that is, using the same discount rate) unless there are possibilities for arbitrage. These come about when investors have the open market access across borders described above. All this seems very obvious and is similar to the meaning usually attached to integration of markets for goods. Where difficulties enter is in attempts to measure financial integration as discussed further below. There are few (possibly no) complete examples of financial integration beyond national borders and even within borders. There is surprising evidence that markets that appear to exist virtually side by side are not integrated in the Flood and Rose sense (see Flood and Rose, 2004 on the evidence against integration within the US NASDAQ (National

98

Globalization and economic integration

Association of Securities Dealers Automated Quotations), the Toronto Stock Exchange, and three different US bond markets and between these markets). Should we expect that the apparent globalization of financial markets implies a trend towards integration in some segments of finance, or in some regions? It would be possible to go beyond the technical definitions of integration and look for a deeper meaning of financial integration. The definition could extend to looking at markets for financial services as well as to the markets for assets. For example, whether banking services markets are integrated is not the same as whether the money market, in which deposits are traded, is. Similarly, insurance services market integration might be an element of financial integration separately from any trade in insurance contracts. Additionally, integration could relate to the extent of the similarity and interoperability of financial systems (for example, whether they share clearance and settlement systems or coordinated regulatory structures), or their convergence to such similarity. At present there is very little discussion of these issues of “deep integration” in the literature on financial integration.

BENEFITS AND COSTS OF FINANCIAL INTEGRATION Policy makers in many regions claim to see benefits from integration. Europe, the Asia-Pacific Economic Cooperation Conference (APEC), and the Association of South East Asian Nations (ASEAN) have statements about the desirability of closer financial integration that either explicitly or implicitly rely on the idea of these benefits. A recent European Central Bank report (ECB, 2008), for example, states “Financial integration is a key component of the general economic policy of the EU, as it promotes the development of the financial system, thereby raising the potential for stronger non-inflationary economic growth” (p. 6). Similarly, the Asian Development Bank (ADB, 2008) says, “The case for Asia’s financial development and integration is clear-cut” (p. 109). “In short Asia’s financial integration could both bolster the region’s economic growth and reduce its vulnerability to global shocks. Strengthening financial stability regionally would also bolster it globally” (p. 112). There is less economic theory to support these propositions than might be supposed. The propositions that are clear in the existing theory mainly relate to gains from financial development. Larger and deeper financial systems have been associated with consumers’ improved ability to achieve consumption smoothing (between time periods), with investment increases

Asian financial integration

99

and growth and with increases in the efficiency of capital allocation. They may also be associated with increases in savings (though the effect is contested in the literature). The theoretical contributions on the welfare improvements from financial integration come from a variety of sources. An early contribution was MacDougall (1960), who noted that while the inflow of foreign capital reduces the marginal rate of return to existing capital some of the losses are to the incumbent foreign owners of capital while the gain from increased marginal return to labor accrues to domestic labor.1 More important are the likely tax gains for the domestic government (these are losses for the government of the originating country) and the possibilities of economies of scale and of what are now called spillovers (MacDougall’s “know how”). The subsequent literature on the effects of foreign direct investment (FDI) is not usually focused on the discussion of financial integration, but does in fact speak to the same issue. Gourinchas and Jeanne (2006), for example, examine the effect of integration through the additional capital accumulation that comes from encouraging capital inflow. Lee and Shin develop this idea in a manner reminiscent of MacDougall, and estimate the additional welfare gains from the technology diffusion that comes with foreign capital. The alternative approach to the welfare effects of financial integration (see, for example, Cole and Obstfeld 1991; and Backus et al., 1992; van Wincoop, 1994, 1999; Tesar, 1995; Imbs, 2006; Lee and Shin, 2008) derives from the idea that “under complete markets, the social planner equates the marginal utilities of consumption across countries . . . isoelastic preferences then imply that consumption plans be perfectly correlated” (Imbs, 2006, p. 299). Put more loosely “welfare gains are measured by the degree of consumption risk shared through financial integration” (Lee and Shin, 2008, p. 2). The same question has also been posed the other way around to argue that those countries with low levels of consumption risk sharing have most to gain from greater integration. Empirical results based on this approach have varied enormously in their estimates of the welfare gain but are beginning to converge. Van Wincoop (1999) is currently the definitive study of welfare gains from risk sharing among OECD (Organization for Economic Cooperation and Development) countries and gives the range of gains from 1.1 to 3.3 percent of consumption over a 50–year horizon. While he claims that these are large gains, the gains need to be compared to some earlier studies. Using the same method (and some of the same parameter values), Kim et al. (2006) derive estimates for East Asia as a group at 1.5 percent over a 10–year horizon, rising to 8.48 percent over 50 years, while Liu and Zheng (2007) get slightly lower gains for all countries in the region and an

100

Globalization and economic integration

average of 1.2 percent over 10 years, rising to 7.61 percent over 50 years. Consensus numbers would seem to be around 3 percent for within-OECD risk sharing, around 6 percent for developing countries integrating with the world (developed plus developing) and perhaps 7–8 percent for East Asia over a 50–year horizon. For a group of countries with average growth rates of 5–7 percent per annum, this does not seem a large gain. Lee and Shin (2008) also examine the supply-side effects of financial integration to estimate welfare gains, focusing on the growth impact of higher levels of capital accumulation, noting that while the welfare effects of integration from pure capital accumulation are fairly small, they rise with technological diffusion.2 It is difficult to compare the scale of results with those reported above3 and Lee and Shin’s percentage increases in consumption vary widely for individual countries and between different technology diffusion scenarios. There is clearly more work to be done in this vein but they conclude rather modestly that there is “some room for welfare improvement in Asian countries from enhancing further integration” as a result of the gains from technology diffusion. Integration in their study seems to mean the extent to which countries attract capital inflows and there is no attention paid to the source of the inflow. We return to this point below. Though rarely discussed in the context of financial integration, there are also macroeconomic implications of internationally open and integrated financial markets. By embracing unimpeded capital flows, countries accept a discipline on the operation of monetary and fiscal policy and will find the operation of exchange rate regimes subject to much greater scrutiny. In some cases this may introduce a desirable degree of discipline on macro policy. Arguably the presence of foreign financial institutions, usually another outcome of an integrated market, may enable economies of scale in services and give depth and liquidity to markets. It may also enable “leapfrogging” of stages of financial development. There are few studies of these aspects of financial integration. There are also costs of financial integration. Greater integration may expose markets to external shocks and financial contagion. There is now evidence (Imbs, 2006) that more financially integrated economies have more synchronized GDP fluctuations despite the theoretical presumption that the result should be neutral or should actually be negatively correlated. The ambivalence and resistance to openness among some governments in the region comes to some extent from concerns about this possibility, reinforced by the experience of the Asian crisis of 1997. The ambivalence is demonstrated in the ASEAN Blueprint for closer market integration which seeks free markets in goods, services and labor but only freer markets for capital. Resistance also comes, as always, from financial

Asian financial integration

101

sector incumbents wanting protection – some of whom may be foreign firms that have already established themselves behind protective barriers.

HOW OPEN, ENGAGED AND INTEGRATED ARE ASIAN FINANCIAL MARKETS? There is an emerging conventional wisdom that Asian financial markets are more closely integrated with financial markets outside the region than within it, and that financial integration has progressed much less than trade integration (see ADB, 2008 and Kim and Lee, 2008). This section of the chapter reviews the most recent studies. Despite the theoretical discussion (reviewed above) implying that either similarity of the underlying discount rates or correlation of consumption flows are the appropriate indicators of integration, the empirical literature has continued to use a range of measures of three broad types: (i) regulatory or institutional measures, (ii) quantity measures of financial flows or holdings of stocks, and (iii) price measures. Taking account of the theoretical background to the notion of financial integration it is more accurate to consider the regulatory (de jure) measures as capturing the extent of openness, while the quantity measures describe the extent of financial engagement within the region or with the world. Openness Regulatory or institutional measures that directly capture “openness” will normally apply equally to all potential sources of financial flows. In the absence of discriminatory (non-MFN – most favored nation) barriers (rare in financial services and financial markets4) a financial market is not usually “more open” to some outside markets than to others. The information derived from these measures can therefore reveal whether the Asian region is more or less closed than other regions, but cannot show whether it is more regionally or globally integrated. However, these measures can be used as variables to explain the extent of integration (as in Imbs, 2006). In the face of apparent low levels of integration in Asia there is some interest in what measures of openness reveal about the level of barriers in the regional economies. Chinn and Ito provide a data-rich and useful index based on the International Monetary Fund (IMF) Annual Report on Exchange Arrangements and Exchange Restrictions, which suggests that Asian economies do not have particularly high levels of measured barriers (including exchange controls).5 Figure 5.1 shows that the Asian region has had relatively high levels

102

Globalization and economic integration

KA openness

60

40

20

0

1970–1979

1980–1989

Less developed Latin Eastern & Central Europe

1990–1999

2000–2005

Asia & Pacific South Asia, Middle East & Africa

Note: The index is normalized with the highest degree of financial openness captured by the value of 100 and the lowest by the value of zero. Source:

Chinn and Ito (2008).

Figure 5.1

Development of capital account openness by region

of financial openness since the 1970s compared with other less-developed countries (Latin America, the Middle East, Africa, and Eastern and Central Europe). Because of the reversal of openness in some countries in Asia after the Asian crisis of 1997, the position had reversed in the 2000–05 period so that Asia was no longer the most open region among those shown here. Engagement One consequence of having more (or less) open markets might be actual flows of capital and accumulation of foreign assets and liabilities (though this will not necessarily follow automatically from openness). Quantity measures of this kind are purely descriptive, have no theoretical content and cannot be regarded as revealing about the extent of integration. What they reflect is the actual state of transactions. Their size may be interpreted in a variety of ways, but they do allow a comparison of the extent to which

Asian financial integration

103

2.5 2.0

Ratio of foreign liabilities to GDP Ratio of foreign assets to GDP

1.5 1.0 0.5

ia Ze al an d Ph ilip pi ne s Si ng ap or e Th ai la nd Vi et Eu na ro m pe an U ni on

N ew

of lic

al ay s

M

ub

ep

Ja pa n R Ko

re

a,

In di a In do ne si a

Au st ra lia C hi na ,P .R .

0

Source: IMF, CIPS Data, and World Bank Key Development Data and Statistics at http://web.worldbank.org/WBSITE/EXTERNAL/DATASTATISTICS/.

Figure 5.2

Foreign assets and liabilities relative to GDP, 2007

the region engages in global financial activity as against activity within a region.6 Measures of the size of the holdings of foreign financial assets relative to GDP give information about the extent of past interactions with foreign financial markets. To give a sense of perspective we need to compare the size of financial holdings relative to GDP across countries and regions (Lane and Milesi-Ferretti, 2003; Obstfeld and Rogoff, 1996). Asian economies hold, on average, assets valued at only about half the European level relative to GDP (Poonpatpibul et al., 2006 using IFS: International Financial Statistics and Lee, 2008, using CIPS) but there are important caveats. Compared to the United States, the average of Asian economies shows a quite high foreign asset ratio. The United States has a much lower value of foreign assets to GDP than either Asia or Europe. Within the Asian region there is very large variation. As Figure 5.2 shows, Singapore, as a regional financial center, holds stocks of assets larger than the average European economy while most of the other countries of the region still hold very modest foreign assets in relation to their GDP. Data problems inhibit a more detailed picture of the intra-regional and inter-regional holdings of assets. For developing countries the accuracy and coverage of cross-border capital flow data are particularly problematic, and for some asset classes almost non-existent. Data from the IMF’s Consolidated International Portfolio Survey (CIPS) give the basic patterns

104

Globalization and economic integration

of cross-border activity. Both assets held abroad by the region and the stock of foreign-owned assets inside the region have grown over the last 10 years (Ghosh, 2006). Intra-regional holdings of both equities and bonds have increased, driven mainly by the industrializing economies (Hong Kong, Korea, and Singapore) investing both into each other’s economies and into the developing economies (China, Indonesia, Malaysia, the Philippines, and Thailand). The developing East Asian economies have become considerable owners of equities in both global markets and in the industrializing East Asian economies, but do not have significant bond holdings within the region. Using CIPS data, Lee (2008) shows that East Asian economies held about 4.9 percent of their foreign portfolio assets within East Asia in 2003, while they owned 8.6 percent of the total foreign-held assets in the region. These compared with 57 and 62 percent for Europe’s holdings in Europe (see Lee, Tables 1 and 2). In terms of scale, Lee and Poonpatpibul et al., conclude that Asia is much less engaged with the world’s financial system than Europe. In cross-border bank flows, Eichengreen and Park (2004) conclude the same. The same data can be used to draw attention to the important role played by Japan. Because Japan mainly invests outside the region, the percentage of regional holdings is much higher when Japan is excluded. In 2005, the 10 emerging East Asian economies (Hong Kong, Korea, Singapore, Taiwan, Indonesia, Malaysia, the Philippines, Thailand, Vietnam, and China) held 24 percent of their foreign assets within the same group, while 16 percent of the total foreign-owned assets in the group were held by the group.7 In some types of assets the proportions are much higher and in 2003, in short-term debt, Asia’s holdings of the regions’ liabilities, at 86 percent excluding Japan, were very high. On these data, the markets in which the Asian region has accumulated most assets within the region are first the equity markets, then shortterm debt and finally long-term debt (ranked excluding Japan), but the international holdings of short-term debt are much smaller than either of the other two and are almost entirely accounted for by Hong Kong and Singapore. In terms of the ownership of liabilities in the region, the order is reversed. The regional share is largest in short-term debt, next in longterm debt, and least in equities. The reason is that extra-regional investors account for very large proportions of the inward portfolio investments in equities. These interpretations beg the question of whether the Asian region would be expected to look like Europe. Two approaches provide a benchmark for the size of asset holdings. Intensity indices can indicate whether bilateral holdings of assets are large or small in terms of the investing

Asian financial integration

105

country’s total holdings of foreign assets and the recipient country’s size in the world. Table 5.1 shows the geographic pattern of portfolio asset holdings while Table 5.2 shows intensity indices for the region. The intensity index formula is given by (Aij/Ai) / (Lj/Lw) , where Aij is the asset holding by country i in country j, Ai is the total foreign asset holding by country i, Lj is the total of assets held by foreigners in country j, and Lw is the global total of foreign-held assets (= liabilities). The indices parallel the idea of intensity indices used in international trade and indicate whether the asset holding by a particular country in another country is a large share of the investor country’s holdings relative to the size of the host country. For example, if country i holds 10 percent of its total overseas assets in country j and country j accounts for 10 percent of the world’s foreign-held assets, then the intensity of i’s holdings in j would be 1. Numbers larger than 1 indicate that holdings by j in i are disproportionately large relative to i’s role as a host to foreign portfolio investment. Table 5.2 indicates the variation among Asian countries and the importance of Singapore as both an investor and a recipient of regional investment. Since the ultimate owner of assets cannot be distinguished in the data, Singapore’s role as a financial center and conduit of portfolio investment will be part of the explanation. The patterns also suggest that physical proximity may matter. Japan’s lack of portfolio investment in the developing economies of the region is also clear. The contrast between Japan and South Korea is notable. Unfortunately, there are no data for China as a source of portfolio investment. A “gravity” model, based on norms for holdings of assets and liabilities for pairs of countries in a larger international sample, is another way to assess this. Taking several characteristics of country pairs into account, this approach considers whether bilateral financial relations are lower or higher than might be expected, given the range of factors that generally explain bilateral cross-border holdings of assets (size of the two economies, distance between the two, common language, common borders, and common colonial heritage). The gravity approach suggests that the extra effect of being within the East Asian region is significant (compare Eichengreen and Park, 2004 and Lee, 2008). This means that country pairs within East Asia hold assets in each other that are somewhat larger than would be held by a random pair of countries, although the regional effect is smaller than the “Europe effect” which is on the order of nine times the size for a random pair. For cross-border bank claims (Eichengreen and Park, 2004) the Asian regional effect is even greater than the effect for Europe. Most of the regional effect, however, comes from the link between trade and financial flows. Once the

106

.... (c) 3,265 306 25,058 4,024 534 4,467 (c) 2,781 (c) (c) 404,666

2 35 .... 7 .... 54 3 16 4 11 8 .... 1104

16 – 504 .... 20 5 4 .... 2 188 23 .... 2,129

Australia India Indonesia 65,931 15,501 6,584 1,433 – 13,763 3,229 5,390 1,952 11,314 1,771 69 2,523,566

Japan 2,949 11,055 5,752 742 4,881 – 934 13 103 1,579 430 1,213 158,651

South Korea 363 106 105 323 220 961 .... 4 81 2,251 129 1 12,935

Malaysia 7,903 (c) (c) (c) 2,353 (c) (c) .... (c) 68 (c) .... 35,602

New Zealand

Source:

IMF, Coordinated Portfolio Investment Survey.

360 53 49 49 83 170 272 11 .... 501 35 – 6,520

Philippines 24,119 10,554 14,697 9,332 8,417 20,277 25,538 571 1,871 .... 4,495 1,126 325,063

4,157 49 21 20 63 526 48 1 3 566 – 33 14,692

Singapore Thailand

Total 714,339 411,329 332,765 57,630 1,463,796 366,346 100,922 37,890 39,820 171,430 50,833 5,028 3,752,128

Geographic breakdown of total portfolio investment assets, year-end 2007 (in millions of US dollars)

Note: The data are derived from the creditor side for both assets and liabilities. Countries in the top row are the investor, countries in the first column are the recipient. – Indicates a zero value or a value less than US$500,000. . . . . Indicates an unavailable datum. (c) Indicates that a non-zero datum was not disclosed for reasons of confidentiality.

Australia China India Indonesia Japan South Korea Malaysia New Zealand Philippines Singapore Thailand Vietnam Total

Table 5.1

107

Source:

1.57

0.51 11.41

0.95 0.51 1.66 1.06

1.01 14.57 3.22 2.29 5.27

5.23

4.12

2.98

0.1

1.05 20.17 8.46

0.73

0.26 0.27

27.86

0.41

Indonesia

0.5 2.21 0.76 1.02 0.54 0.21

0.58

0.58 0.31 0.39

1.43

Japan

IMF, Coordinated Portfolio Investment Survey.

Australia Cambodia China India Indonesia Japan Korea Laos Malaysia New Zealand Philippines Singapore Thailand Vietnam

Recipient:

India

2.28 0.08 0.64 2.27 2.09 59.52

6.62 4.27 3.18 0.82

1.02

South Korea

0.29 6.14 39.74 7.67 0.36

0.78 0.96 16.96 0.45 7.94

1.54

Malaysia

Intensity index of total portfolio investment assets, 2007

Investor: Australia

Table 5.2

0.44

1.77

12.17

New Zealand

17.54 4.15

16.18 1.68

0.78 0.88 5.06 0.34 2.78

3.02

Philippines

10.65 26.97

30.48 1.82 5.66

3.09 5.32 19.5 0.69 6.67

4.07

Singapore

17.61

15.51 674.14 0.31 0.17 0.94 0.11 3.82 103.82 1.27 0.08 0.23 8.79

Thailand

108

Globalization and economic integration

extent of bilateral trade flows is accounted for, the Asian regional effects in both portfolio and bank claims become negative. This means that larger bilateral trade flows have such a large positive effect on financial flows that, once that effect is taken out, the additional effect of being within the Asian region is negative – financial flows in both Asia and Europe are lower than they would be between random pairs of countries with similar bilateral trade flows.8 The conclusion from these studies is that there is evidence of above-average financial engagement among Asian economies, even though it is still less than between European economies, but that it may be largely explained by the degree of bilateral trade between countries in the region. Once trade is accounted for there is less engagement in the region than normal and this pattern is replicated across most types of portfolio assets (equities, long-term debt and bank claims; short-term debt is less affected).9 These are important indicators of what drives or impedes financial integration, to which we shall return. Integration The best price-based measure of integration would be the Flood and Rose approach to extracting the common discount rate across markets, but so far no studies have used that method. The conventional price measures, either measuring covered interest parity or co-movements between returns on various types of assets, have a variety of imperfections, so interpreting them is difficult. Typically these measures fail to demonstrate high degrees of integration even between well-developed financial systems, so it is perhaps not surprising that the Asian regional data also fail to find significant, consistent evidence for high degrees of integration within the region. Kim and Lee (2008) give a good account of the results of the usual range of price measures: covered interest parity, bond yield spreads over a benchmark bond, overnight bank rates, and equity prices. While each of these measures shows a reduction in dispersion over time (bond yield spreads and bank rates), or an increase in correlation (equity prices), they still display higher distances from the integrated values shown in other integrated areas such as the European Union (EU). Furthermore, their movements are generally consistent with a hypothesis that integration is being driven by the closer integration of most countries in the region with the US market.10 Thus the appearance of regional integration may be an artifact of the growing integration of the parts with an outside player. By market sector, Asian equity markets show most co-movement though still low, while bonds, money markets, and banking still exhibit interest differentials. The other theoretically justified measure of integration is the correlation

Asian financial integration

109

of consumption flows. The rationale, as noted above, is that integrated financial markets allow either, or both, risk sharing and consumption smoothing.11 One element of this is the global movement of savings and investment. This should mean that investors in one country are not limited by access to only domestic savings.12 Fully open capital markets and fully integrated financial systems should imply no correlation between domestic savings and investment for any individual country. Furthermore, the desire to smooth consumption can be met by access to capital markets either at home or abroad, so a related approach is to look for convergence of consumption paths between countries. If countries are able to smooth the path of their consumption, this implies that they have access to capital markets and are able to reduce or share the risks of consumption volatility. Under perfect risk sharing the consumption growth rate of one country would equal that of the world consumption growth (see Obstfeld and Rogoff, 1996 for the standard analysis). For the region as a whole there is evidence that consumption growth is closely linked to income growth (Poonpatpibul et al., 2006) and that that there have always been quite high correlations between domestic investment and savings, indicating low degrees of openness before and after the crisis years (Montiel, 1994). There is also evidence that GDP shocks for countries within East Asia are smoothed to a smaller degree than in the OECD and the EU (Kim et al., 2006) and that Asian economies have a lower degree of risk sharing within the region but a higher degree of risk sharing globally than Europe. Not all countries are achieving risk sharing. China, Hong Kong, South Korea, and Taiwan achieved lower levels than Japan, whereas the Philippines, Singapore, and Thailand have significant global risk sharing. On average, Asian economies have lower overall levels of risk sharing than Europe (Kim et al., 2008).13

POLICY IMPLICATIONS Given the evidence that the region is more financially integrated with global markets than with other economies in the region and is incompletely integrated with both, a number of questions arise. A key question for current policy debate is whether there is a clear need for greater regional financial integration. The answer depends on whether there are welfare gains from specifically regional integration. If there are, a further question is how to identify the barriers holding back regional integration and what can and should be done about them.14 If the main gains actually come from further global integration, is there a role for regional integration as a stepping-stone towards the final goal?

110

Globalization and economic integration

As noted in the introduction there is a capital outflow, associated with trade surpluses, from the region to the rest of the world. It is not clear whether that capital remains outside the region or finds its way back into it. What is clear is that the intermediation of those flows is taking place outside the region. What causes the external intermediation of financing, whether it is a problem for the region and whether it requires policy intervention are important questions for both analysis and policy advice. These questions go to the heart of where the benefits of closer integration lie and how much is to be achieved by greater country-level liberalization and how much by regional policy. Discussion of them is still at an early stage in the region but should be an explicit part of the future regional agenda. The fact that finance flows out of the region towards the developed financial centers of Europe and the United States, and that intermediation functions lie outside the region, may have some benefits, and may persist even after closer regional integration progresses. Even so, having inefficient regional markets that could not perform intermediation functions within the region will be costly and may be part of the explanation of lowered levels of investment. Policy can legitimately focus on improving the quality of domestic and regional markets and institutions in the hope that more financial flows will occur within the region, but it cannot cause that pattern of financial allocation to happen. Welfare Gains from Regional Integration As noted in the section on the benefits and costs of financial integration, calibrated models estimating welfare gains from closer integration, whether in terms of smoother consumption flows or from capital accumulation and technology transfer, suggest that the gains are not very large. Furthermore, few (arguably no) studies address the question of whether regional integration would give greater benefits than more global integration. Some implications can be deduced from the existing literature but more research is needed. From the consumption-smoothing welfare literature, potential welfare gains from closer integration positively depend on the degree of risk aversion, and the persistence and volatility of output shocks (Obstfeld and Rogoff, 1996 and van Wincoop, 1999). Countries with risk-averse consumers who highly value smoother consumption but face persistent or volatile output shocks will gain most from increased opportunities for consumption smoothing. Welfare gains are larger over long time horizons and will depend negatively on the cross-country correlation of output shocks. Thus integrating with similar countries, subject to similar shocks, would not give the gains. The theory does not otherwise identify which

Asian financial integration

111

countries should seek to integrate with which others, and there is a strong theoretical presumption that global integration will dominate any smaller group, although it might turn out that most of the gains can be achieved by integrating with a few well-chosen partners.15 Indeed Kim et al. (2008) note that under a capital asset pricing model (CAPM) view of the world, investors should diversify their portfolios to the greatest possible extent, choosing securities with low correlation with each other and with the home portfolio. If this argument carries across to countries (and it is much more complex to make this transition) then [C]ountries with different structures, subject to different economic shocks, with low business cycle correlation, will find it more advantageous to develop closer financial links with one another. In this regard, extensive portfolio diversification within East Asia may not be necessarily an optimal strategy, considering the homogeneity of East Asian economies (p. 192).

By homogeneity, the authors mean close correlation of output growth rates within the region. This argument seems to imply that “the welfare gains for regional financial integration are lower in East Asia than in Europe” (ibid., p. 193). Some evidence is provided by the empirical results in Kim et al. (2006). Calibrating the gains from complete consumption smoothing among a group of 10 East Asian countries16 they find that the gains are greatest for integration of the whole region.17 Tellingly, the gains increase when the United States is included in the grouping. However there is no systematic comparison of the gains just from integration with the United States alone or with other groups of developed countries outside the region, so these results cannot be used to provide strong evidence in favor of closer regional integration over global. As noted earlier, some studies add additional elements to the gains from consumption smoothing by appealing to the capital accumulation literature. Lee and Shin (2008) add to the conventional calibrated welfare estimates by assuming that there is FDI from advanced countries to less advanced and that there is technology diffusion associated with that. Their results on the gains from integration depend crucially on these characteristics of the capital inflow, but they acknowledge that capital inflows could come from different sources and that “regional financial integration would have different welfare implications compared to global integration” (p. 20). While they are not explicit about this point, it seems to follow from their method that flows from more advanced countries in the region to less advanced, with accompanying technology transfers, would be beneficial but other regional flows would not have those characteristics and, in any case, to the extent that the most advanced technologies are outside

112

Globalization and economic integration

the region, the benefits will be larger from opening to flows from those sources. Barriers to Regional Integration So far a picture emerges of a slow, and sometimes sporadic, increase in financial integration within the region that is yet to reach levels observed among other groups of closely linked economies. What can be said about the main drivers of the process? In particular, do remaining legal and policy (de jure) barriers explain a low level of financial integration or are other factors more important? The major explanation of degrees of financial integration appears to be the level of regional trade. Countries with large trade flows will likely also have large capital flows relative to GDP. In so far as trade integration has been a market-driven, bottom-up process in Asia, it is probable that the same will be true of financial integration. In general terms, Imbs (2006) suggests that the failure of capital to flow from high to low capital–labor ratio countries must reflect some unobserved risk, and he suggests that a likely candidate is the fear of expropriation. This is consistent with research suggesting that removing de jure capital controls does not automatically result in increased capital flows or greater integration. Lee and Shin note also the point by Caselli and Feyrer (2007) that countries with low levels of capital and low rates of return have this because of high costs of “implementing” capital. As a result, actual flows will be minimal even if the capital market is completely open – that is, openness will result in much less welfare improvement. From the policy perspective there is very little empirically established link between increasing financial integration and the policy and financial structure variables that regularly turn up in lists of desirable reforms to promote greater integration. The World Bank (Ghosh, 2006) lists the impediments to greater cross-border transactions as including withholding taxes, a lack of hedging instruments, differences in market practices and infrastructure such as trading platforms and conventions, procedures for clearance and settlement and custodian systems, differences in rating standards, national legal and regulatory frameworks, and accounting and auditing practices. To this list others have added the “underdevelopment of financial markets”, inadequate financial and legal structure, low auditing and accounting standards, low transparency, and weak corporate governance (Lee, 2008). Parreñas (2007) surveys several papers and picks out recommendations to improve transparency, legal systems, insolvency systems and workout procedures, competition and free entry, risk pricing undistorted by subsidies or interest rate controls, clearing and settlement

Asian financial integration

113

systems, government bond markets and benchmark yields and the promotion of securitization. There is only limited consistent statistical evidence on which of these desiderata actually affect the extent of financial integration. Lane and Milesi-Ferretti (2007) find that financial depth and the size of stock markets have an effect but, tellingly, corporate income tax rates and the introduction of insider trading laws have no effect. Ostry et al. (2007) suggest that domestic policies towards the financial system do have an effect. Capital controls, institutional quality, trade openness and the level of economic development affect the overall extent of “openness” (measured by total foreign liabilities as a share of GDP, so closer to what this chapter calls “engagement”). The institutional quality index is an average of indicators covering voice and accountability, political stability and absence of violence, government effectiveness, regulatory quality, rule of law, and control of corruption. The measurement of all of these is contentious and they are much bigger policy issues than simply improving financial systems, difficult as that may be. Much of the policy advice and discussion on regional integration thus seems to be based on intuition and anecdotal evidence about what would improve financial systems, rather than on hard evidence. There is a clear need for more research on these issues. Political Economy? If there is no strong presumption that closer regional financial integration is directly welfare enhancing in itself is it the case that regional financial integration could pave the way for globally integrated markets? Does it help to make the case for greater financial liberalization generally that is, is it a building block for globally freer capital markets in an analogy with the argument that is sometimes made for preferential trade agreements? Such arguments are implied by a recent ADB (2008) study which argues that a lack of regional integration is “damaging in itself” and that “National capital markets need to be developed and connected to improve liquidity. Such measures may be easier . . . regionally. . . . Regional institutions could also foster dialogue, information sharing and peer pressure that promote financial development and integration as well as best practices in financial regulation . . .” (p. 140). So far the empirical evidence for Asia suggests that the reverse may be the case. If closer regional financial integration is to take place it will follow, not lead closer global integration. A further element of the political economy story is provided by Kim et al. (2006) whose research question is predicated on the idea that a higher degree of risk sharing makes a group of countries more likely to fill the

114

Globalization and economic integration

requirements of an optimal currency area than otherwise. If business cycles and output shocks are not highly correlated, then the ability to risk share within a group of countries can offset the high cost of forming a currency union. While the discussion of a currency union in the region is in its infancy it is not unrelated to the discussion of regional financial integration.

CONCLUSIONS This chapter has argued that care is needed in the discussion of the idea of closer regional financial integration. Some of the existing literature describes financial phenomena that are only indirectly linked to the issue of integration. These show that the region engages in financial transactions, and accumulates claims and liabilities, more with markets outside the region than with those inside it. This is consistent with the picture of a region with less-developed intermediaries and with excess savings that need to be invested in assets with a preferred risk–return profile. Asian regional financial engagement currently takes place at a level that may be slightly below that observed in other geographically defined regions such as the EU or the United States. More careful studies do suggest that the regions’ markets are more closely integrated with global markets than with markets within the region (though there is much more research needed to establish this result in a theoretically sound way). Such research also shows that there are potential gains in welfare from increased consumption risk sharing for many countries in the region (that is, that the present level of consumption smoothing and risk sharing is quite low for most countries in the region). These could be achieved by greater financial integration (though that is not the only way to achieve them. Fiscal transfers are another alternative). At present there is no clearly articulated theoretical basis for an argument that closer regional financial integration would deliver greater welfare gains than would more global integration. There are also no empirical studies that demonstrate the advantages of regional integration over global. There is much further research needed, both theoretical and empirical, to understand which countries would make natural partners in a financially integrated region. It is possible that the extensions to existing theory would show that the best outcome is global integration and anything less is second best. This naturally leads to a consideration of the political economy arguments for whether a regional approach would be preferred to a more global approach. The policy lesson that comes from this is that we do not yet have good information on what characteristics of the region’s financial structure

Asian financial integration

115

explain the current pattern of financial engagement. Some of the methods reviewed in this chapter have the potential to elucidate these issues. There may be lessons from gravity models that are detailed enough to identify the elements of “distance” that could respond to policy (for example, institutional variability, transparency and so on). What we do know is that there are unexploited potential welfare gains in the Asian region from improving consumption smoothing, from enabling better mechanisms for intertemporal transfers of consumption streams (which would help to reduce both household and national anxieties that lead to the accumulation of excess savings) and from increasing capital accumulation with technology transfers. What we do not yet know are the most effective methods for achieving these.

NOTES 1. 2. 3. 4. 5.

6.

7.

I am indebted to Vijay Joshi for reminding me of this paper. Technological diffusion rises with the proportion of capital inflows that comes as FDI. Since the timeframe over which the welfare gains are achieved is not clearly stated (though it looks to be somewhere between 20 and 50 years). It would be possible to develop an index of discriminatory barriers to financial flows by studying the finance and investment provisions of regional and bilateral preferential trade agreements, but I am not aware of studies of that type. Chinn and Ito (2008) derive the principal component of four variables that include both the conventional capital account controls and also measures of the presence of multiple exchange rates, controls on current account transactions and the need to surrender export proceeds. The index is expressed in terms of degrees of openness rather than whether restrictions exist, and data are available for 181 countries back to 1983. This measure has the advantage of incorporating a more extensive range of restrictions that would impact on cross-border capital movements. The data are available at: http:// www.ssc.wisc.edu/~mchinn/research.html. There are other measures of institutional restrictions, including the Edison and Warnock (2003) measures of the proportion of the stock market available for foreign investment and Takagi and Hirose (2004) using principal components on a set of de facto indicators (exchange rate volatility, deviations of purchasing power parity (PPP), deviations of uncovered interest parity condition (UIP), trade intensity and interest rate correlations. Imbs (2006) uses other indices composed of proxies for financial openness including institutional variables covering legal traditions, shareholders’ and creditors’ rights, and enforcement variables. For example, it could be argued that if markets are perfectly integrated there will be no opportunities for arbitrage and there would be no incentive for any flows to take place at all. So the usual commentary, looking at the size of flows as a sign of greater integration with external markets, would be quite wrong. In 2003 in equities the proportion is 20 percent compared with Europe’s 53 percent, while in long-term debt the comparison is 15 percent against 46 percent and in short-term debt 18 percent against 59 percent. There is still, certainly, a large difference between Asia and Europe, but excluding Japan increases the figures by several multiples. This reflects the fact that Japan, a large, post-industrial economy is heavily invested in equivalent economies elsewhere, rather than near neighbors. It also highlights the growth in Asia’s integration in all directions since the first detailed data became available in 1997.

116 8.

9.

10.

11.

12. 13.

14.

15. 16. 17.

Globalization and economic integration There is evidence that this “finance follows trade” effect occurs in other countries too (Lane and Milesi-Ferreti, 2003 and Rose and Spiegel, 2002 cited in Eichengreen and Park 2004; and Portes and Rey, 2005). Imbs (2004) also indirectly tests whether financial integration is driven by trade integration. Interestingly similar patterns are found in Europe – once trade is accounted for, the positive regional effect becomes negative (in bank claims), or very small (shortterm debt) or at least much smaller than before-trade effects (equities and long-term debt). The ADB regularly uses the standard deviation of the absolute average cross-market long-term government bond yield spread over benchmark US Treasury bonds as an indicator of integration and claims that it “has been falling since 1999, although it remained substantial until 2005. Since March 2007, it has fallen to a new low, with an average standard deviation of about 2 basis points. While the dispersion of interest rates in the region has declined over the past decade, it has remained substantial until very recently”. This measure in fact shows that regional prices vary increasingly closely with the United States over time. The same ADB study notes “In terms of comovement, the bilateral correlations of equity price indexes across markets have risen over the past decade. This is not necessarily proof of greater regional financial integration; it may simply reflect growing links among most Asian bourses via the US or Europe” (2008, p. 122). Looking at bilateral correlations, only three countries are more closely correlated with the region than with the United States. Other studies have shown that regional equity markets still respond more strongly to global news shocks than to regional ones. As explained by Obstfeld and Rogoff (1996), risk sharing and consumption smoothing are conceptually different. Risk sharing implies contingent claims (insurance) with other countries to insure idiosyncratic shocks while consumption smoothing can be achieved by borrowing and lending (non-contingent claims such as bonds) to achieve intertemporal smoothing. The famous result of Feldstein and Horioka (1980) showing high degrees of correlation has been challenged but is largely replicated for many countries and time periods. Kim et al. (2006) also claim to show that regional capital and credit markets provide very small opportunities for consumption smoothing for the countries of East Asia but in fact their method shows only that each country in their set of 10 achieves very low consumption smoothing by international factor income and by intertemporal shifting of consumption via savings. There is nothing in the method that can identify the role played by specifically regional capital and credit markets. There are a number of ways to address these issues. One is to look for direct evidence on what explains the remaining interest differentials or the low level of consumption correlation in order to get a clearer sense of where the barriers lie. This is the approach of Imbs (2006) but has not yet been applied specifically to the region. Another is to observe that, since arbitrage is clearly incomplete, as implied by the results, then the actual levels of financial flows, and the ownership patterns of financial assets within the region, reflect the existing barriers (whatever their source may be). If these quantitative flows differ from what would be expected it should be possible to analyze what factors explain the gap between the actual and expected patterns. This is also a research opportunity that has not been developed in the region. The theory is built on assumptions of identical consumers with identical preferences in all countries, so there is further scope to consider how gains are impacted by integration between countries with different preferences. Indonesia, Malaysia, the Philippines, Singapore and Thailand (ASEAN5), and China, Korea and Japan (North East Asia), plus Hong Kong and Taiwan. Integration within the ASEAN5 countries alone gives very slightly higher average gains for the members than does integration with wider East Asia. This may be an artifact of the calculation method since it does not seem consistent with several of the other results.

Asian financial integration

117

REFERENCES Asian Development Bank (ADB) (2008), Emerging Asian Regionalism, Manila: ADB. Backus, D.K., P.J. Kehoe and F.E. Kydland (1992), “International real business cycles”, Journal of Political Economy, 100 (4), 745–75. Caselli, F. and J. Feyrer (2007), “The marginal product of capital”, Quarterly Journal of Economics, 122 (2), 535–68. Chinn, M.D. and H. Ito (2008), “A new measure of financial openness”, Journal of Comparative Policy Analysis: Research and Practice, 10 (3), 309–22. Cole, H.L. and M. Obstfeld (1991), “Commodity trade and international risk sharing: how much do financial markets matter?”, NBER Working Paper 3027, Cambridge, MA. Edison, H.J. and F.E. Warnock (2003), “A simple measure of the intensity of capital controls”, Journal of Empirical Finance, 10 (1–2), 81–103. Eichengreen, B. and Y.C. Park (2004), “Why has there been less financial integration in Asia than in Europe?”, Monetary Authority of Singapore Staff Paper, Singapore. European Central Bank (ECB) (2008), Financial Integration in Europe, Frankfurt: European Central Bank. Feldstein, M. and C. Horioka (1980), “Domestic saving and international capital flows”, Economic Journal, 90 (358), 314–29. Flood, R.P. and A.K. Rose (2004), “Financial integration: a new methodology and an illustration”, IMF Working Paper, Washington, DC, available at: http:// ssrn.com/abstract=878934 (accessed August 20, 2009). Ghosh, S. (2006), East Asian Finance: The Road to Robust Markets, Washington, DC: World Bank. Gourinchas, P.-O. and O. Jeanne (2006), “The elusive gains from international financial integration”, Review of Economic Studies, 73 (3), 715–41. Imbs, J. (2004), “Trade, finance, specialization and synchronization”, Review of Economics and Statistics, 8 (3), 723–34. Imbs, J. (2006), “The real effects of financial integration”, Journal of International Economics, 68 (2), 296–324. Kim, S. and J.W. Lee (2008) “Real and financial integration in East Asia”, ADB Working Paper Series on Regional Economic Integration 17, Manila. Kim, S., S.H. Kim and Y. Wang (2006), “Financial integration and consumption risk sharing in East Asia”, Japan and the World Economy, 18, 143–57. Kim, S., J.-W. Lee and K. Shin (2008), “Regional and global financial integration in East Asia”, in B. Eichengreen, C. Wyplosz and Y.C. Park (eds), China, Asia and the New World Economy, Oxford: Oxford University Press, pp. 168–99. Lane, P.R. and G.M. Milesi-Ferretti (2003), “International financial integration”, Trinity Economics Papers 20031, Department of Economics, Trinity College Dublin. Lane, P.R. and G.M. Milesi-Ferretti (2007), “The external wealth of nations mark II: revised and extended estimates of foreign assets and liabilities, 1970–2004”, Journal of International Economics, 73 (2), 223–50. Lee, J.W. (2008), “Patterns and determinants for cross-border financial asset holdings in East Asia”, ADB Working Paper Series on Regional Economic Integration 13, Manila.

118

Globalization and economic integration

Lee, J.W. and K. Shin (2008), “Welfare implications of international financial integration”, ADB Working Paper Series on Regional Economic Integration 20, Manila. Liu, H. and H. Zheng (2007), “Regional cooperation on the integration of Asian financial markets”, background paper for ADB, Emerging Asian Regionalism, Manila. MacDougall, G.D.A. (1960), “The benefits and costs of private investment from abroad: a theoretical approach”, Economic Record, 36, 13–35. Montiel, P.J. (1994), “Capital mobility in developing countries: some measurement issues and empirical estimates”, Policy Research Working Paper Series (WPS 1103), 1–57, World Bank, Washington, DC. Obstfeld, M. and K. Rogoff (1996), Foundations of International Macroeconomics, Cambridge, MA: MIT Press. Ostry, J.D., P. Mauro, G. Dell’Ariccia, J. di Giovanni, A. Faria, A. Kose, M. Schindler and M. Terrones (2007), “Reaping the benefits of financial globalization”, Research Department Discussion Paper, IMF, Washington, DC. Parreñas, J.C. (2007), “Financial liberalization and integration in East Asia: challenges, prospects and cooperation opportunities”, ABA Journal, 22 (2), 1–31. Poonpatpibul, C., S. Tanboon and P. Leelapornchai (2006), “The role of financial integration in East Asia in promoting regional growth and stability”, Bank of Thailand, mimeo. Portes, R. and H. Rey (2005), “The determinants of cross-border equity flows”, Journal of International Economics, 65 (2), 269–96. Rose, A.K. and M.M. Spiegel (2002), “A gravity model of international lending: trade, default and credit”, CEPR Discussion Paper 3539, Centre for Economic Policy Research, London. Takagi, S. and K. Hirose (2004), “A multivariate approach to grouping financially integrated economies”, in M. Kawai and G. de Brouwer (eds), Exchange Rate Regimes in East Asia, New York: Routledge, pp. 131–52. Tesar, L.L. (1995), “Evaluating the gains from international risk sharing”, Carnegie-Rochester Conference Series on Public Policy, 42, 95–143. van Wincoop, E. (1994), “Welfare gains from international risk sharing”, Journal of Monetary Economics, 34, 175–200. van Wincoop, E. (1999), “How big are potential welfare gains from international risk sharing?” Journal of International Economics, 47, 109–35.

6.

Forecasting international financial prices with fundamentals: how do stocks and exchange rates compare? Robert P. Flood and Andrew K. Rose*

MOTIVATION We are now in the midst of an era of financial integration, which is leading to convergence of international financial prices. Indeed, international financial flows are largely driven by discrepancies in asset prices across borders. This would be widely seen as a welcome development if asset prices were well grounded in economic fundamentals. Part of what our analysis seeks to do is question this latter assumption, and thus the presumption that financial integration can be automatically viewed as good for economic welfare. The integration of international financial markets is a key part of globalization. Understanding how these markets operate in practice is thus important. In this chapter, we study price determination in international markets for two important assets: stock prices and foreign exchange rates. We ask how our knowledge of foreign exchange markets for developed countries compares with our knowledge of comparable equity markets. More specifically, we are interested in our ability to model and forecast international equity prices and exchange rates. In their now-classic papers (1983a, b), Richard Meese and Kenneth Rogoff (hereafter “MR”) examined the forecasting performance of a number of then-popular exchange rate models. They found that a random walk “model” of the exchange rate consistently outforecast the structural models, despite the latter’s being given the advantage of using actual future values of market fundamentals. The full reaction to the MR message took years to process, but was eventually devastating for the field of international finance. Academic modeling of exchange rate determination basically ceased. The area fell into disrepute; indeed, the area is not even represented on many first-rate academic faculties. By academic standards the MR paper had a huge impact and its fallout is still felt whenever 119

120

Globalization and economic integration

exchange rates are intelligently discussed. The negative conclusions from the original MR exercise still seem to characterize the data today; see for example, Chinn et al. (2005) and Rogoff and Stavrakeva (2008). In the current chapter we ask a simple question: what happens when the MR method is applied to assets other than foreign exchange? We consider aggregate stock market indices in Germany, Japan, the United Kingdom and the United States, countries that correspond to the bilateral exchange rates considered by MR. We carry out the same forecasting analysis as MR, initially over the same sample period they considered, 1973m3 through 1981m6. Just as MR did for foreign exchange rates, we consider a number of time-series and structural models, and use a number of metrics to compare them out of sample with a random walk. Crucially, we follow MR in allowing structural models to “forecast” asset prices with actual future values of fundamentals; since these fundamentals would ordinarily be unknown, ours is most certainly not a standard ex ante forecasting exercise. Where MR forecast exchange rates with money, income, and the like, we provide the forecaster with information about the actual future levels and growth rates of earnings, dividends, and interest rates. It turns out that not only is our methodology similar to that of MR; so is our conclusion. Just as MR found with foreign exchange, we find that none of our models with fundamentals performs consistently and substantially better than a simple random walk model of the aggregate stock market. Our intention is to focus on the stock market to demonstrate that foreign exchange prices are broadly as difficult to model as other asset prices of relevance like the stock market. Accordingly, we stick close to the original MR setup. Consistent with MR, we use each of our competing models to generate forecasts at a number of different forecasting horizons. We examine four large stock markets, and estimate the coefficients of each model with the most up-to-date information available at the time of a given forecast. We follow MR in doing this by adding an observation and re-estimating each of our forecasting models for each forecast period. We use a number of different models of stock prices, primarily to specify the set of future fundamentals that should drive current prices. The models we use rely on both the level and growth rate of two key fundamental determinants of stock prices: dividends and earnings. In particular, we focus on the well-known “Gordon growth” model of stock values, using either dividends or earnings as the appropriate fundamental. We also consider a composite model that incorporates the levels and growth rates of both dividends and earnings, as well as the short-term interest rate. We estimate these models with a variety of different estimators, and also consider both forecasts from univariate models and vector autoregressions

Forecasting international financial prices with fundamentals

121

(VARs). Yet despite our broad-ranging sensitivity checks, we are unable to find a model that consistently outperforms the simple-minded random walk “model”. In the next section, we present the theoretical framework for our analysis; we then discuss the empirical methodology and data in the following section. The results follow, and the chapter ends with a brief conclusion.

THEORY The standard way we think about stock prices is that a firm’s value at a point in time (Pt) is based entirely on either the present value of dividends (D) or earnings (N); the two must be equal by the firm’s budget constraint. We follow the profession in assuming that stocks give no non-pecuniary returns. A firm’s stock price is given by: Pt 5 PV (Dt11, Dt 12, . . .) 5 PV (Nt 11, Nt12, . . .) ,

(6.1)

where PV(·) is the present-value operator, which can take many forms. For our purposes, we start by assuming: `

PV (Xt11, Xt12, . . .) 5 Et a Xt1i ri,

(6.2)

i51

where E is the expectation operator, and 0 < r < 1 is the non-stochastic discount rate. (The assumption of a constant discount rate is not as restrictive as it might seem, since we never estimate r and indeed, we do not assume that it is constant in practice; more on this below.)1 Combining equations (6.1) and (6.2), we can write: `

`

Pt 5 Et a Dt1i ri 5 Et a Nt1i ri. i51

(6.3)

i51

Since price is equal to the present value of either dividends or earnings, it must equal a weighted average of the two. So for any q, `

`

Pt 5 qEt a Dt1i ri 1 (1 2 q) Et a Nt1i ri. i51

(6.4)

i51

To take equation (6.4) to the data, we need to make an assumption about how to form expectations of future values of dividends and earnings. We assume that growth is proportional: Xt1i11 5 (1 1 gxt) Xt1i 1 et1i11,

(6.5)

122

Globalization and economic integration

where et is white noise orthogonal to X, and gxt is the growth rate of X estimated with data through time t, which is assumed to be constant from t onward. This assumption seems to be broadly consistent with evidence from the asset-pricing literature.2 We use these assumptions and equation (6.5) to get: `

`

Pt 5 qDt a (1 1 gDt) iri 1 (1 2 q) Nt a (1 1 gNt) iri. i51

(6.6)

i51

With a little algebra we collapse the infinite sums to: Pt 5

(1 2 q) Nt (1 1 gNt) r qDt (1 1 gDt) r 1 . 1 2 (1 1 gDt) r 1 2 (1 1 gNt) r

(6.7)

Equation (6.7) is a simple formula that gives stock price as a weighted function of earnings and dividends and their respective growth rates.3 Versions of equation (6.7) are commonly known as the “Gordon growth model”, after Gordon (1962). We follow MR, and consider a log-linearized version of (6.7), with the interest rate added: pt 5 b0 1 bd dt 1 bdg ln (1 1 gDt) 1 bnnt 1 bng ln (1 1 gNt) 1 bi ln (1 1 it) 1 ut,

(6.8)

where i represents the interest rate, {b} are a set of coefficients of interest, and lower-case letters are natural logs of their upper-case counterparts.4 In our forecasting analysis, we consider three “structural” models of stock prices; all are contained within equation (6.8). The Gordon growth model for dividends sets bn = bng = bi = 0, while the Gordon growth model for earnings symmetrically sets bd = bdg = bi = 0. We also consider a composite model without any parametric constraints. We estimate our models with ordinary least squares (OLS).

METHODOLOGY AND DATA Estimating and Forecasting with the Fundamentals-based Models We try to stick as closely as possible to MR in estimating our models and using them to produce forecasts. Thus, we estimate our models using our monthly dataset, beginning in March 1973. We initially estimate our models with OLS using data through the start of the first forecasting period, November 1976 (we choose these dates to match those of MR).

Forecasting international financial prices with fundamentals

123

Using the estimated coefficients, we then generate forecasts for stock prices at one-, three-, six-, and 12–month horizons. We then add data for December 1976, re-estimate our models, generate new forecasts, and continue in this fashion through the end of the sample period of June 1981.5 Our most interesting models rely on “fundamentals” such as earnings and/or dividends. Forecasting future stock prices with these models requires future values of these explanatory variables, and we follow MR in using actual realized values in place of forecasts of explanatory variables. We compare our models using out-of-sample statistical techniques, and follow MR in using three measures of forecast accuracy. These are: (i) the root mean square forecast error, (ii) the mean forecast error, and the (iii) mean absolute forecast error. These are defined as follows: Nk 21

MeanError 5 a [ F (t 1 s 1 k) 2 A (t 1 s 1 k) ] /Nk, s50

(6.9a)

MeanAbsoluteError 5 a [ 0 F (t 1 s 1 k) 2 A (t 1 s 1 k) 0 ] /Nk, s50 (6.9b) Nk 21

Nk 21

RootMeanSquareError 5 e a [ F (t 1 s 1 k) 2 A (t 1 s 1 k) ] 2/Nk f , s50 (6.9c) 1/2

where: k = 1, 3, 6, 12, and 24 denotes the forecast horizon; Nk is the total number of forecasts in the projection period; A(t) is the actual known value; F(t) is the forecast value; and forecasting begins in period t. Since we follow MR in examining natural logarithms, our statistics are unit-less, approximately percentages, and comparable across markets.6 The root mean square error is the most important of these statistical criteria, though we consider the mean absolute error to be of almost equal importance. We are less interested in the mean error, since a very noisy forecast which wildly over- and underforecasts can have a low average forecast error. However, we follow MR in using it as an auxiliary measure. Moreover, as MR note, one can easily compare mean and mean absolute errors to determine whether a particular model habitually over- or underforecasts. Alternative Models We follow MR in comparing models with fundamentals to atheoretical univariate and multivariate time-series alternatives. As a univariate alternative to the random walk prediction of no change, we use the “long AR” technique favored by MR. This is an unconstrained autoregression, where the longest lag included (M) is a function of the

124

Globalization and economic integration

sample size (T), M = T/ln(T). Other univariate techniques tend to deliver comparable results.7 For our atheoretical multivariate alternative, we follow MR in using an unconstrained VAR. This consists of a VAR in three variables: the natural logarithms of stock prices, earnings, and dividends. We chose the lag lengths for the VARs taking into account three different criteria. the final prediction error (FPE), the Hannan and Quinn information criterion (HQIC) and the Schwarz information criterion (SIC).8 Since more parsimonious VARs typically forecast better in short samples, it is unsurprising to us that we end up with relatively short lag lengths: Germany (2); Japan (2); the UK (3); and the US (4).9 The Dataset Our dataset relies on conventional measures of popular stock price indices that cover most of the national stock market. We use the following indices: the CDAX (Germany); the Nikkei 225 (Japan); the FTSE All-Share (UK), and the S&P 500 (US). We think of these broad measures of national stock markets as being roughly analogous to the necessarily aggregate bilateral exchange rates considered by MR.10 We combine these prices with price/ earnings (P/E) and dividend/yield (D/P) ratios to “back out” measures of dividends and earnings. This may well induce measurement error, since we only have indirect measurements of earnings and dividends. This problem is potentially more serious at short horizons, since P/E and D/P ratios may be updated with a lag. Accordingly, we tend to place more credence on our long-horizon forecasting results, and try to be cautious in our conclusions.11 We follow MR and use closing month-end prices rather than temporally averaged prices. We also follow MR by taking natural logarithms of our key variables. In particular, we transform stock price indices, earnings, and dividends by logs, and consider interest rates in the form of ln{1 + [i(t)/100]} where i(t) is the annualized interest rate in percentage points at time t.12 Our data have not been seasonally adjusted or adjusted for inflation. Our data sources are described more fully in Appendix 6A1. We have checked all our series with a number of different statistical and ocular filters; we have also checked them with comparable and similar (for example, monthly average) series from other sources (such as Datastream and Bloomberg), and performed spot checks with the Financial Times. Our key raw variables are plotted in three figures: stock price indices are graphed in Figure 6.1, while dividends and earnings are plotted in Figures 6.2 and 6.3, respectively. The trends over this period of time are similar

Forecasting international financial prices with fundamentals 332.18

47.98

66.66

26.04

1973m3

UK

1981m6

7867.4

Japan

1981m6

1973m3

USA

1981m6

2.6059

6.6017

0.989728 UK

1981m6

114.077

1973m3

Germany

1981m6

6.39084

76.6122

Figure 6.2

1981m6

Raw stock price indices

18.5024

1973m3

Germany

63.54

1973m3

1973m3

1973m3

140.52

3594.55

Figure 6.1

125

3.16717 Japan

1981m6

1973m3

USA

Dividends extracted from dividend/price ratios

1981m6

126

Globalization and economic integration

46.8918

6.00411

10.4786 1973m3

2.14259 UK

1981m6

367.636

Figure 6.3

Germany

1981m6

15.2974

175.587 1973m3

1973m3

6.8 Japan

1981m6

1973m3

USA

1981m6

Earnings extracted from price/earnings ratios

across countries for prices and both fundamentals. The depressing effect of the mid-1970s recession is also quite apparent in all the series. The natural logarithm transformation is quite consistent with the upward drift of the series.13 Standard univariate tests are usually quite consistent with the hypothesis that our key variables – (natural logarithms of) stock prices, earnings, and dividends – have a unit root. Further, standard tests typically also show that these variables are co-integrated with a single common trend. Under these circumstances, OLS provides “super consistent” estimates of the {b} coefficients in (6.8). Nevertheless, we stress that we have no intrinsic interest in our estimates of {b}, since our focus is on forecasting precision. Our fundamentals-based models rely on growth rates of both earnings and dividends. We have experimented with a number of different ways of measuring rates of change. In particular, we have looked at annual growth rates over both one- and three-year periods; we have also measured growth rates as both forward and backward looking.14 Since no one measure of growth seems to perform consistently better than the others, we err on the side of giving our forecasters more rather than less information, and use a three-year forward-looking growth rate as our default. In Appendix 6A2, we present results for all four combinations that we have considered (estimating growth rates over both one- and three-year horizons, looking either forwards or backwards).

Forecasting international financial prices with fundamentals

127

RESULTS Benchmark Results Our key results are presented in Tables 6.1 to 6.3. These present root mean square forecast errors, mean absolute forecast errors, and mean forecast errors from November 1976 through June 1981. In each table, we present results for the random walk benchmark, and five alternative models. These are, respectively, (a) the univariate long autoregression; (b) the vector autoregression; (c) the structural Gordon growth model with earnings; (d) the structural Gordon growth model with dividends; and (e) the composite model. Each of these models is compared at three different forecast horizons (one, three, and 12 months) for the key stock market indices of four different countries (Germany, Japan, the UK, and the US). Tables 6.1 and 6.3 are exact analogs to Tables 1 and 2 of MR (1983a) with one gap; we have no equivalent of forward rates for stock indices.15,16 The most striking feature of the tables is how poorly all five alternative models forecast future stock prices, compared with the simple random walk prediction of no change. In terms of our preferred root mean square error (RMSE) metric, the random walk model outforecasts all four of the models that take advantage of fundamentals, for all four countries, at all horizons. This is true despite the fact that the forecasts of the Gordon growth and composite models are based on realized future values of earnings, dividends, and interest rates. Moreover, the differences in forecast performance are not small; using fundamentals often results in a substantive deterioration in forecast performance. For instance, at the 12–month horizon, the random walk has an RMSE that is almost 50 percent lower than that of the Gordon growth earning models for Germany; results are even more dramatic for other countries. In this sample, the only model to beat the random walk model in terms of RMSE is the univariate long autoregression. Even then, it only beats the random walk decisively at the 12–month horizon for Japan. At other horizons, the much-simpler random walk performs slightly better for Japan. The long AR also outforecasts the random walk at the one-month horizon for the US, though only marginally.17 The generally dismal performance of fundamentals-based models is also apparent when one looks at mean absolute forecast errors instead of root mean square forecast errors. The long autoregression still performs better than the random walk at the 12–month horizon for Japan, but it also forecasts better at the six-month horizon for Japan, and at 12 months for Germany. The VAR also beats the random walk at the one-month horizon for the UK in terms of mean absolute error (MAE).18

128

Globalization and economic integration

Table 6.1

Root mean square forecast errors

Stock market

Horizon Random Univ. walk long AR

VAR

Germany

1 mon 6 mon 12 mon 1 mon 6 mon 12 mon 1 mon 6 mon 12 mon 1 mon 6 mon 12 mon

3.09 10.86 19.31 3.53 13.18 24.44 5.59 12.96 19.31 4.64 10.25 14.80

Japan

UK

USA

Note:

2.68 6.13 8.65 2.70 6.03 10.93 5.51 12.95 18.09 4.14 8.68 12.59

2.86 8.41 9.11 2.75 6.09 8.19 5.85 22.20 37.02 4.06 10.40 15.32

Horizon Random Univ. walk Long AR

VAR

Germany

1 mon 6 mon 12 mon 1 mon 6 mon 12 mon 1 mon 6 mon 12 mon 1 mon 6 mon 12 mon

2.45 9.14 15.82 2.84 11.99 22.78 4.22 11.16 16.81 3.69 8.78 12.41

USA

Note:

6.59 13.65 24.24 11.27 14.98 20.03 15.16 20.05 21.94 10.43 14.01 18.71

5.30 8.36 12.41 9.99 14.23 19.03 13.49 19.13 24.68 13.26 25.11 40.73

Mean absolute forecast errors

Stock market

UK

8.05 10.45 12.97 16.93 21.84 27.85 27.44 41.60 58.12 12.53 17.18 24.32

Percentage terms. Three-year forward-looking growth rates for structural models.

Table 6.2

Japan

Gordon Gordon Composite earnings dividends

2.13 5.24 7.61 2.02 5.22 10.19 4.51 10.64 15.36 3.25 6.96 10.38

2.42 6.55 7.22 2.16 4.84 6.05 4.61 20.28 36.27 3.25 8.24 13.09

Gordon Gordon Composite earnings dividends 6.81 9.96 10.88 16.34 21.52 27.67 23.42 34.90 48.69 10.94 15.57 22.51

5.24 9.60 16.57 10.00 13.89 19.13 12.26 15.82 18.39 8.19 10.99 14.84

4.20 6.54 9.63 8.65 12.45 16.95 10.81 14.94 19.82 10.53 20.51 35.06

Percentage terms. Three-year forward-looking growth rates for structural models.

It seems remarkable that forecasts which take advantage of unknowable future values of fundamentals do not seem to perform systematically better than the mindless prediction of no change. Still, this echoes the celebrated negative results of MR. We emphasize that we are not claiming that stock price indices actually follow a random walk. We are simply following MR

Forecasting international financial prices with fundamentals

Table 6.3

Mean forecast errors

Stock market

Horizon Random Univ. walk long AR

Germany

1 mon 6 mon 12 mon 1 mon 6 mon 12 mon 1 mon 6 mon 12 mon 1 mon 6 mon 12 mon

Japan

UK

USA

Note:

129

0.52 2.33 4.18 1.01 4.96 10.08 1.66 7.88 14.05 0.46 3.19 7.51

0.75 2.74 2.10 0.93 2.88 4.07 2.62 20.28 36.27 1.52 8.00 13.09

VAR −1.20 −7.64 −13.84 2.55 11.99 22.78 −0.05 −5.19 −14.16 0.20 2.45 5.75

Gordon Gordon Composite earnings dividends 4.24 5.31 6.52 16.34 21.52 27.67 22.70 34.31 48.69 10.73 15.46 22.51

2.48 6.70 13.39 10.00 13.89 19.13 6.67 7.22 6.00 7.27 10.22 14.78

1.67 2.77 4.26 7.99 11.91 16.89 7.56 12.89 18.82 9.46 20.42 35.06

Percentage terms. Three-year forward-looking growth rates for structural models.

in showing that the forecasts from the simplest forecasting model we can imagine – that of no change in an asset price – works approximately as well as more-sophisticated forecasting models using unknowable future information on “fundamentals”. Sensitivity Checks A more complete set of results is tabulated in Appendix 6A2. These include a number of different perturbations of the models that rely on fundamentals, primarily using backward- (instead of just forward-) looking growth rates of fundamentals for one- (as well as three-) year horizons. We provide tabulations of: (a) root mean square forecast error; (b) mean absolute forecast error; and (c) mean error forecast statistics, at five different forecast horizons: (a) one; (b) three; (c) six; (d) 12; and (e) 24 months. We also tabulate the comparable figures for the random walk models to facilitate comparisons. In Tables 6.4a through 6.4c, we provide statistics for the random walk model inclusive of a freely estimated drift term. In Tables 6.5a and 6.5b, we show robustness with respect to the sample period. In Table 6.5a, we show the results of changing the start of the forecast period from November 1976 to November 1978, while in Table 6.5b we change the end of the forecast period from June 1981 to November 1980.19 We choose these dates to match those of MR. We also add over 20 years of data and extend the sample into the present in Tables 6.5c

130

Table 6.4a

Globalization and economic integration

Composite “model”: dividends, earnings, their growth rates, and the interest rate: root mean square forecast errors (percentages)

Stock market

Horizon Random Random Growth: Growth: Growth: Growth: walk walk with 1-yr bkd 3-yr bkd 1-yr fwd 3-yr fwd drift

Germany

1 mon 3 mon 6 mon 12 mon 24 mon 1 mon 3 mon 6 mon 12 mon 24 mon 1 mon 3 mon 6 mon 12 mon 24 mon 1 mon 3 mon 6 mon 12 mon 24 mon

Japan

UK

USA

Table 6.4b

2.68 4.59 6.13 8.65 10.19 2.70 4.37 6.03 10.93 21.74 5.51 9.84 12.95 18.09 26.82 4.14 6.40 8.68 12.59 19.73

2.65 4.49 5.93 8.35 12.27 2.62 3.95 4.72 8.08 16.65 5.58 10.01 13.22 18.24 27.82 4.16 6.46 8.81 12.94 22.30

5.71 7.17 8.80 11.18 10.82 14.94 17.88 22.11 28.47 38.02 19.98 24.24 30.60 40.03 62.11 11.47 14.35 18.53 25.48 38.44

7.47 10.22 13.24 12.52 21.22 17.17 19.82 23.39 29.34 37.61 20.28 25.15 32.83 49.05 87.40 8.93 11.22 14.02 19.63 25.11

6.11 7.20 8.53 10.89 13.30 10.61 13.13 16.89 23.59 31.48 19.47 24.23 31.90 45.14 72.77 10.82 13.22 17.06 24.84 37.01

5.30 6.70 8.36 12.41 18.33 9.99 11.79 14.23 19.03 28.51 13.49 16.03 19.13 24.68 40.61 13.26 17.92 25.11 40.73 67.87

Composite: dividends, earnings, their growth rates, and the interest rate: mean absolute forecast errors (percentages)

Stock market

Horizon Random Random Growth: Growth: Growth: Growth: walk walk with 1-yr bkd 3-yr bkd 1-yr fwd 3-yr fwd drift

Germany

1 mon 3 mon 6 mon 12 mon 24 mon 1 mon 3 mon 6 mon 12 mon 24 mon

Japan

2.13 3.72 5.24 7.61 7.32 2.02 3.46 5.22 10.19 21.50

2.09 3.70 5.10 6.71 10.79 1.94 2.94 3.80 6.69 15.56

4.46 5.62 7.00 9.55 10.02 12.11 14.97 19.19 26.34 37.54

5.79 7.80 10.01 8.75 15.46 13.33 15.93 19.60 26.79 37.02

4.43 5.33 6.32 7.56 10.08 9.14 11.19 14.19 20.49 30.37

4.20 5.35 6.54 9.63 13.66 8.65 10.27 12.45 16.95 27.24

Forecasting international financial prices with fundamentals

Table 6.4b

131

(continued)

Stock market

Horizon Random Random Growth: Growth: Growth: Growth: walk walk with 1-yr bkd 3-yr bkd 1-yr fwd 3-yr fwd drift

UK

1 mon 3 mon 6 mon 12 mon 24 mon 1 mon 3 mon 6 mon 12 mon 24 mon

USA

Table 6.4c

4.51 7.97 10.64 15.36 24.68 3.25 5.35 6.96 10.38 16.64

4.54 8.04 10.31 14.51 22.39 3.28 5.40 7.01 11.01 19.29

17.90 21.49 26.73 35.45 56.50 9.13 11.90 16.16 22.77 32.99

16.08 19.63 25.31 40.47 79.76 7.25 9.10 11.41 16.91 20.53

17.18 21.45 28.63 42.70 71.14 8.54 10.87 14.38 20.99 28.32

10.81 12.85 14.94 19.82 35.11 10.53 14.35 20.51 35.06 64.09

Composite: dividends, earnings, their growth rates, and the interest rate: mean forecast errors (percentages)

Stock market

Horizon Random Random Growth: Growth: Growth: Growth: walk walk with 1-yr bkd 3-yr bkd 1-yr fwd 3-yr fwd drift

Germany

1 mon 3 mon 6 mon 12 mon 24 mon 1 mon 3 mon 6 mon 12 mon 24 mon 1 mon 3 mon 6 mon 12 mon 24 mon 1 mon 3 mon 6 mon 12 mon 24 mon

Japan

UK

USA

0.52 1.38 2.33 4.18 5.09 1.01 2.66 4.96 10.08 21.50 1.66 4.40 7.88 14.05 24.68 0.46 1.37 3.19 7.51 15.12

0.98 0.10 −0.26 −1.08 −6.07 0.68 1.67 3.05 6.48 15.56 1.43 3.76 6.71 12.19 22.39 0.47 1.42 3.41 8.39 17.96

1.70 2.00 2.09 0.79 −6.22 11.82 14.71 19.19 26.34 37.54 14.41 17.45 22.46 31.46 3.95 6.96 8.55 10.80 14.72 25.71

3.31 4.50 6.09 6.97 6.15 12.79 15.28 19.19 26.55 37.02 12.51 15.74 22.36 37.85 78.48 2.62 3.44 5.18 8.66 19.08

3.21 4.11 5.30 6.95 9.81 8.71 10.75 13.92 20.36 30.37 17.18 21.45 28.63 42.70 71.36 6.19 7.63 10.05 15.38 28.00

1.67 2.08 2.77 4.26 6.14 7.99 9.58 11.91 16.89 27.24 7.56 9.69 12.89 18.82 34.86 9.46 13.34 20.42 35.06 64.09

132

Table 6.5a

Globalization and economic integration

Changing forecast start date from 1976m11 to 1978m11: root mean square forecast errors (in percentage terms; 3-yr fwd growth rates)

Stock market

Horizon

Random walk

Gordon earnings

Gordon dividends

Composite

Germany

1 mon 3 mon 6 mon 12 mon 24 mon 1 mon 3 mon 6 mon 12 mon 24 mon 1 mon 3 mon 6 mon 12 mon 24 mon 1 mon 3 mon 6 mon 12 mon 24 mon

2.90 4.68 5.59 6.70 6.92 2.14 3.21 5.19 9.21 18.63 5.31 9.79 11.05 17.52 24.11 4.29 6.68 9.81 16.66 30.38

5.73 5.34 5.19 6.31 5.94 19.55 21.40 24.34 29.89 41.64 17.86 19.39 18.56 19.60 39.78 15.12 17.38 21.04 28.25 42.53

3.79 4.01 4.36 5.00 3.90 13.82 15.61 18.50 24.36 38.80 9.98 11.79 13.69 16.13 9.26 9.87 11.39 13.82 18.68 31.11

4.98 5.76 6.60 7.81 6.76 12.54 14.85 18.28 24.84 33.14 9.22 10.46 10.76 11.88 11.19 16.33 21.64 29.40 42.83 42.31

Japan

UK

USA

Table 6.5b

Changing forecast end date from 1981m6 to 1980m11: root mean square forecast errors (in percentage terms; 3-yr fwd growth rates)

Stock market

Horizon

Random walk

Gordon earnings

Gordon dividends

Composite

Germany

1 mon 3 mon 6 mon 12 mon 24 mon 1 mon 3 mon 6 mon 12 mon 24 mon

2.59 4.47 6.28 8.86 10.85 2.68 4.40 6.02 11.13 22.53

8.37 9.37 10.97 13.74 16.61 16.82 18.80 21.72 27.58 39.04

7.00 9.62 14.70 26.08 40.14 9.74 11.17 13.38 18.45 29.49

4.75 6.26 8.16 12.78 20.40 9.06 10.73 13.07 18.13 27.19

Japan

Forecasting international financial prices with fundamentals

Table 6.5b

133

(continued)

Stock market

Horizon

Random walk

Gordon earnings

Gordon dividends

Composite

UK

1 mon 3 mon 6 mon 12 mon 24 mon 1 mon 3 mon 6 mon 12 mon 24 mon

5.66 10.26 13.69 17.24 27.86 4.30 6.65 8.70 10.74 16.28

28.44 34.46 44.01 62.63 107.07 11.10 12.73 15.96 23.89 43.37

16.08 18.55 21.48 23.75 27.57 10.10 11.04 13.41 18.09 30.20

14.21 17.00 20.46 26.78 45.63 13.77 18.30 25.69 42.60 74.30

USA

Table 6.5c

Forecasting from 1989m3 to 2001m12: root mean square forecast errors (in percentage terms; 3-yr fwd growth rates)

Stock market

Horizon

Random walk

Gordon earnings

Gordon dividends

Composite

Germany

1 mon 3 mon 6 mon 12 mon 24 mon 1 mon 3 mon 6 mon 12 mon 24 mon 1 mon 3 mon 6 mon 12 mon 24 mon 1 mon 3 mon 6 mon 12 mon 24 mon

5.52 10.18 14.68 21.26 29.96 6.84 11.42 16.08 22.84 27.46 4.18 7.28 9.69 13.86 21.94 4.23 7.27 10.50 17.28 30.99

46.15 49.24 53.35 59.08 65.91 69.91 71.74 74.03 78.07 84.92 27.28 27.97 28.97 30.83 34.23 32.77 34.32 36.43 39.93 44.82

27.01 28.62 30.97 34.98 40.17 35.01 36.47 38.44 41.99 48.28 26.18 27.00 28.23 30.71 35.84 39.59 41.08 43.25 47.42 55.58

22.14 24.45 27.81 34.16 43.45 30.38 33.34 36.82 41.52 41.65 23.17 24.69 27.03 31.69 39.23 22.86 24.56 26.99 31.72 41.76

Japan

UK

USA

Note:

Estimation begins in 1973m3.

134

Globalization and economic integration

Table 6.5d

Forecasting from 1989m3 to 2001m12: root mean square forecast errors (in percentage terms; 3-yr fwd growth rates)

Stock market

Horizon

Random walk

Gordon earnings

Gordon dividends

Composite

Germany

1 mon 3 mon 6 mon 12 mon 24 mon 1 mon 3 mon 6 mon 12 mon 24 mon 1 mon 3 mon 6 mon 12 mon 24 mon 1 mon 3 mon 6 mon 12 mon 24 mon

5.52 10.18 14.68 21.26 29.96 6.84 11.42 16.08 22.84 27.46 4.18 7.28 9.69 13.86 21.94 4.23 7.27 10.50 17.28 30.99

46.67 50.83 56.22 62.97 68.81 52.33 55.69 60.13 68.32 80.41 20.54 21.18 22.08 23.57 25.22 30.39 32.54 35.64 41.17 50.04

17.36 18.60 20.38 23.28 26.20 27.64 28.88 30.22 33.06 37.97 26.02 26.94 28.32 31.19 36.66 25.05 26.43 28.45 32.34 39.93

16.45 19.20 23.33 34.33 50.77 29.38 33.21 37.57 44.35 45.72 17.09 18.40 20.24 23.40 28.17 21.22 23.21 26.06 31.25 40.34

Japan

UK

USA

Note:

Estimation begins in 1980m1.

and 6.5d, with similar results; the similarity of our results with MR is not a result of a special sample. Finally, in Table 6.6, we use three different estimators. In Table 6.6a we use a generalized least squares (GLS) with a first-order (Prais–Winsten) correction for serial correlation. In Table 6.6b, we use one-year lags for instrumental variables (IVs).20 In Table 6.6c, we use least-absolute deviations (LADs) instead of OLS as our estimator. In Table 6.6d, we add a comprehensive set of seasonal dummies. In their second paper, MR (1983b) eschewed coefficient estimation altogether and checked the out of sample forecasting ability of their models against that of a random walk model using a grid-search technique. We have found essentially the same results ourselves with a comparable gridsearch technique; the results are in Table 6.7. MR also performed a number of experiments that we consider to be unimportant in our context. These include: (a) the Granger–Newbold technique for combining forecasts; (b) estimation in first differences; and

Forecasting international financial prices with fundamentals

Table 6.6a

135

GLS instead of OLS: root mean square forecast errors (in percentage terms; 3-yr fwd growth rates)

Stock market

Horizon

Random walk

Gordon earnings

Gordon dividends

Composite

Germany

1 mon 3 mon 6 mon 12 mon 24 mon 1 mon 3 mon 6 mon 12 mon 24 mon 1 mon 3 mon 6 mon 12 mon 24 mon 1 mon 3 mon 6 mon 12 mon 24 mon

2.68 4.59 6.13 8.65 10.19 2.70 4.37 6.03 10.93 21.74 5.51 9.84 12.95 18.09 26.82 4.14 6.40 8.68 12.59 19.73

20.76 21.97 23.60 25.90 27.52 17.01 18.52 20.70 25.36 35.77 29.14 31.93 35.13 38.68 46.12 13.29 14.74 17.30 24.37 38.73

3.64 4.77 5.60 7.59 13.93 15.35 16.36 17.95 21.42 26.53 12.49 14.29 16.13 17.65 19.97 12.66 14.47 17.44 24.35 40.10

3.99 5.23 6.42 7.71 11.62 16.37 17.67 19.60 24.04 33.05 20.84 23.27 25.90 28.02 32.45 16.46 19.58 24.53 35.51 56.81

Japan

UK

USA

Table 6.6b

IVs instead of OLS: root mean square forecast errors (in percentage terms; 1-yr fwd growth rates)

Stock market

Horizon

Random walk

Gordon earnings

Gordon dividends

Composite

Germany

1 mon 3 mon 6 mon 12 mon 24 mon 1 mon 3 mon 6 mon 12 mon 24 mon

2.68 4.59 6.13 8.65 10.19 2.70 4.37 6.03 10.93 21.74

13.98 16.15 17.90 18.59 21.95 18.43 21.62 23.88 −30.83 −25.84

11.22 13.60 17.25 24.07 38.07 45.20 45.30 67.10 72.50 20.30

31.13 43.93 58.36 96.93 112 410 308 299 761 629

Japan

136

Globalization and economic integration

Table 6.6b

(continued)

Stock market

Horizon

Random walk

Gordon earnings

Gordon dividends

UK

1 mon 3 mon 6 mon 12 mon 24 mon 1 mon 3 mon 6 mon 12 mon 24 mon

5.51 9.84 12.95 18.09 26.82 4.14 6.40 8.68 12.59 19.73

16.82 18.91 22.23 27.60 33.27 20.66 22.49 25.94 32.83 55.39

11.76 13.59 15.59 17.80 22.80 12.94 14.68 17.48 22.99 34.64

USA

Note:

Composite 990 1153 1513 2128 1535 120 610 143 916 8435

12-month lags of (log) earnings/dividends/interest rates used for current values.

Table 6.6c

LAD instead of OLS: root mean square forecast errors (in percentage terms; 3-yr fwd growth rates)

Stock market

Horizon

Random walk

Gordon earnings

Gordon dividends

Composite

Germany

1 mon 3 mon 6 mon 12 mon 24 mon 1 mon 3 mon 6 mon 12 mon 24 mon 1 mon 3 mon 6 mon 12 mon 24 mon 1 mon 3 mon 6 mon 12 mon 24 mon

2.68 4.59 6.13 8.65 10.19 2.70 4.37 6.03 10.93 21.74 5.51 9.84 12.95 18.09 26.82 4.14 6.40 8.68 12.59 19.73

8.05 8.99 10.45 12.97 14.97 16.93 18.89 21.84 27.85 39.66 27.44 32.93 41.60 58.12 96.26 12.53 14.13 17.18 24.32 42.68

5.18 7.05 11.52 23.75 36.69 10.07 10.94 12.40 16.26 26.66 12.03 14.94 18.17 21.01 27.44 12.31 13.87 17.71 24.81 41.40

5.77 7.27 9.09 12.68 22.19 10.13 11.36 13.08 17.68 26.03 10.92 12.75 15.73 20.46 34.50 13.41 17.83 25.00 39.35 65.00

Japan

UK

USA

Forecasting international financial prices with fundamentals

Table 6.6d

137

OLS with seasonal dummies: root mean square forecast errors (in percentage terms; 3-yr fwd growth rates)

Stock market

Horizon

Random walk

Gordon earnings

Gordon dividends

Composite

Germany

1 mon 3 mon 6 mon 12 mon 24 mon 1 mon 3 mon 6 mon 12 mon 24 mon 1 mon 3 mon 6 mon 12 mon 24 mon 1 mon 3 mon 6 mon 12 mon 24 mon

2.68 4.59 6.13 8.65 10.19 2.70 4.37 6.03 10.93 21.74 5.51 9.84 12.95 18.09 26.82 4.14 6.40 8.68 12.59 19.73

8.91 9.54 10.77 12.12 13.86 18.03 19.53 21.91 26.16 37.83 30.56 35.48 43.40 56.87 96.06 13.64 14.82 17.50 23.58 41.17

7.87 10.05 13.99 23.62 35.19 12.27 13.56 15.67 19.86 32.22 16.90 18.14 20.72 20.69 23.10 11.82 12.07 14.30 18.47 29.56

5.89 7.26 8.65 12.36 20.32 10.61 12.20 14.50 18.85 28.13 15.38 17.95 19.90 18.84 19.62 14.59 19.33 26.08 39.83 66.00

Japan

UK

USA

(c) more univariate models. One could also add more sophisticated estimation strategies (for example, using generalized method of moments: GMM) or forecasting metrics (for example, Diebold–Mariano), both of which have advanced greatly since the time of the original MR research. We see no reason why any of these perturbations should deliver better results than MR, who uniformly found that considerations like these typically did little to affect the key conclusions. We note in passing that our analysis is linked to two ongoing research programs. Engel and West (2005) consider the case where discount rates are high and fundamentals are I(1); see also Engel et al. (2007). They show that in the limit, asset prices will move like random walks in the sense that (p. 486) “the change in the time t asset price will be uncorrelated with information known at time t − 1”. We agree with all of this. However, such ex ante forecasting ability is not the issue of concern to us, since we follow the MR methodology in giving the forecaster information on actual future fundamentals, dated t + 1 or greater. Further, we do not

138

Table 6.7

Globalization and economic integration

Shortest forecast horizon (in months) for which at least x percent of each model’s parameter grid improves on the random walk model in MAE/RMSE when realized values of the explanatory variables are used (3-yr fwd growth rates)

Model

Stock market Threshold

Germany

Japan

UK

US

MAE RMSE MAE RMSE MAE RMSE MAE RMSE Gordon earnings

Gordon dividends

0–1% 10 25 50 0–1% 10 25 50

–

– – – – –

– – – – – – – –

24 24 24 24 36 – – –

24 24 24 30 – – –

12 18 24 – 6 6 12 12

12 18 30 – 6 6 6 12

36 – – – 18 18 30 –

36 – – – 18 18 24 –

Note: Grid-size = 49. Parameters for grid were 0.7–1.3 in increments of 0.1, for each coefficient. Forecasting horizons used: 1, 3, 6, 12, 18, 24, 30, 36.

make assumptions above regarding either the persistence in fundamentals or the discount rate. Rossi (2005) shows that persistence in disturbances may drive MR results, because of asymptotic estimation (Hurwicz) bias. However, we assume away this issue above; adding this very real estimation problem back into our setup is likely to make the MR methodology even more troublesome.

SUMMARY AND CONCLUSION Naive “random walk” models of no change seem to forecast future stock prices just as well as more-sophisticated models based on a variety of fundamentals such as earnings and dividends. Crucially, this is true even giving a hypothetical “forecaster” the advantage of actual future information on fundamentals, which are unobservable in practice. In this sense, domestic financial prices of great interest (stock market indices) are just as difficult to forecast as international financial prices (exchange rates). Succinctly, for a number of crucial international financial markets, models of aggregate prices of equity seem to work about as well as aggregate models of the price of foreign exchange. It is tempting to speculate that international finance might have progressed differently if MR had considered other prices above and beyond

Forecasting international financial prices with fundamentals

139

the exchange rate when they first conducted their research. Still, such an exercise would be just that: speculation. Instead, we conclude that any feelings that researchers have about our (in-)ability to model foreign exchange rates should also be felt about international equity markets.

NOTES *

1. 2.

3. 4.

5. 6.

7. 8. 9.

10.

We thank Jelena Kmezic for tenacious research assistance. For comments, we thank Yu-chin Chen, Levis Kochin, Assaf Razin, Barbara Rossi, Steve Turnovsky, and seminar participants at Cornell, Crete, the IMF, Luxembourg, National University of Singapore, Oregon, Syracuse, University of California Santa Cruz, and Washington. Rose also thanks the Federal Reserve Bank of New York, the International Monetary Fund, the National University of Singapore and the Monetary Authority of Singapore for hospitality during the course of this work. This chapter is a revision of “Why So Glum?”, CEPR DP 6714. The dataset and sample output are available at http://faculty. haas.berkeley.edu/arose. We follow MR in developing a model of the nominal price of stock. With suitable assumptions, this can be derived from a model of real asset prices. Fama and French (2002, p. 640) state “[c]onfirming Campbell (1991), Cochrane (1994), and Campbell and Shiller (1998), we find that dividend and earnings growth rates for 1950–2000 are largely unpredictable”. Also, see Cochrane (2001, p. 404). We have not directly attempted to test (6.5) ourselves; this might be a topic for future research. The growth rates for dividends and earnings must be similar over the long run, which may raise estimation issues. Equation (6.8) does not follow strictly from equation (6.7), which is a risk-neutral pricing equation that does not involve interest rates. Equation (6.8) is a generalization that allows for the possibility that there is some portfolio substitution in stock pricing, so that when interest rates change, the stock market can react. Note that this estimation strategy means that we do not impose the assumption of a constant discount rate on the model; indeed, our discount rates vary arbitrarily across time and forecast horizons. To compare the out-of-sample statistics formally, one would need to take account of the temporal dependence of the forecast errors, since forecasts of over one month overlap. As it turns out, this is not a major issue in our sample, since the random walk model often dominates the forecasting models we consider. In Tables 6.4a–4c, we tabulate results for one of the extensions we considered, the random walk “model” with an estimated intercept. We choose our lag length after we estimate the VARs over the sample, 1973m3 through 1981m6. After we choose the VAR lag length for a particular country, we then keep it constant for all forecast periods. For Germany, the FPE is minimized at two lags, while both the HQIC and the SIC indicate one lag. We are wary of including only a single lag, and thus we err on the side of (minor) potential overparameterization. In the case of Japan, the FPE, HQIC, and SIC all point to two lags. Both HQIC and SIC indicate three lags for the UK, while the FPE indicates four lags. Finally, four lags are indicated for the US data by both HQIC and SIC, while the FPE indicates 19 lags. One could also use the Akaike information criteria, though this is well known to indicate much longer lag lengths. One could also use standard Bayesian techniques to reduce the overparameterized VARs, even without resorting to more structural techniques. They also match the P/E and D/P ratios we use closely. This match is exact in the case of Japan, the UK and the US. In the case of Germany, the GFD description is “[m]onthly data from the Bundesamt begin . . . in January 1956. The current index is an aggregate

140

11. 12. 13. 14. 15. 16.

17. 18.

19. 20.

Globalization and economic integration of dividends (including stock dividends) divided by the aggregate market value (end of period) of all quoted shares and do not include the effect of the withholding tax. The data from 1953 through 1995 were calculated by the Statistisches Bundesamt”. In future work we also plan to tackle this problem by examining data from individual firms. Thus a 5 percent interest rate is measured as ln(1 + 0.05) ≈ 0.05. This is above and beyond ensuring positivity, and, more importantly, our goal of keeping our technique as compatible as possible with that of MR. To be precise, we measure a one-year backward-looking growth rate at time t as [ln(xt) − ln(xt-12)], while an annualized three-year forward-looking growth rate at time t is measured as [ln(xt+36) − ln(xt)]/3. For comparison, Appendix 6A3, Tables 6A3.1a and 1b are transcribed versions of Meese and Rogoff (1983a), Tables 1 and 2, respectively. These contain RMSE and ME for the data and models that MR consider, formatted analogously. Stock index futures do exist for some of our indices. For instance, the Chicago Mercantile Exchange trades a futures contract for the S&P 500, though trading only began in 1982, see http://www.cme.com/trading/prd/equity/index.html. Likewise, the Osaka Securities Exchange has traded a Nikkei 225 stock index future since 1988, see http://en.wikipedia.org/wiki/Osaka_Securities_Exchange. We have been unable to find evidence that either the FTSE All-Share or CDAX is traded as a futures contract. As MR found with exchange rates, the random walk may perform as well as other forecasting models, but it does not perform well in any absolute sense; the RMSE at even the one-month horizons exceed 2 percent for all the countries we consider. Unsurprisingly, these positive results are also reflected in superior mean forecast errors; these also show that there are some cases where the British and American VARs – and the Gordon growth for the United Kingdom – outforecast the random walk. We are less concerned with these outcomes for the reasons spelled out above, given that they perform worse in terms of root mean square and MAE. Note that the number of forecasts in Table 6.5a is small, especially at the longer horizons; for instance, there are only eight 24–month forecasts available. We use one-year lags of the (log) levels of dividends, earnings, and interest rates; we do not instrument for the growth rates at all. We realize that this is lame, and are open to suggestions for superior instrumental variables, mostly to account for the potential measurement error.

REFERENCES Chinn, M., Y.-W. Cheung and A. Garcia (2005), “Empirical exchange rate models of the nineties: are any fit to survive?”, Journal of International Money and Finance, 24 (7), 1150–75. Cochrane, J.H. (2001), Asset Pricing, Princeton, NJ: Princeton University Press. Engel, C.M. and K. D. West (2005), “Exchange rates and fundamentals”, Journal of Political Economy, 113 (3), 485–517. Engel, C.M., N.C. Mark and K.D. West (2007), “Exchange rate models are not as bad as you think”, NBER Working Paper W13318, with comments from K. Rogoff and B. Rossi. Fama, E.F. and K.R. French (2002), “The equity premium”, Journal of Finance, 57 (2), 637–59. Gordon, M.J. (1962), Investment, Financing and Valuation of the Corporation, Homewood, IL: R.D. Irwin. Meese, R.A. and K. Rogoff (1983a), “Empirical exchange rate models of the

Forecasting international financial prices with fundamentals

141

seventies: do they fit out of sample?”, Journal of International Economics, 14 (1), 3–24. Meese, R.A. and K. Rogoff (1983b), “The out-of-sample failure of empirical exchange rate models: sampling error or misspecification?”, in Exchange Rates and International Macroeconomics, edited by J.A. Frenkel, Chicago, IL: University of Chicago Press, pp. 67–105. Rogoff, K. and V. Stavrakeva (2008), “The continuing puzzle of short horizon exchange rate forecasting”, NBER Working Paper 14071, Cambridge, MA. Rossi, B. (2005), “Testing long-horizon predictive ability with high persistence, and the Meese–Rogoff puzzle”, International Economic Review, 46 (1), 61–92.

142

Globalization and economic integration

APPENDIX 6A1

DATA SOURCES

Stock Price Indices for Germany, Japan, and the United Kingdom are taken from Datastream (URL: http://www.datastream.net) and Global Financial Dataset (GFD, URL: http://www.globalfinancialdata.com/). ●

●

●

●

“The CDAX includes the shares of all domestic companies listed in Prime Standard and General Standard. The index represents the German equity market in its entirety, that is, all companies listed on FWB Frankfurter Wertpapierbörse (Frankfurt Stock Exchange)”.1 “The Nikkei Stock Average is Japan’s most widely watched index of stock market activity and has been calculated continuously since September 7, 1950 . . . The 225 components of the Nikkei Stock Average are among the most actively traded issues on the first section of the TSE”.2 “FTSE All-Share Index: Representing 98–99% of the UK market capitalisation, FTSE All-Share is the aggregation of the FTSE 100, FTSE 250 and FTSE Small Cap Indices”.3 “S&P 500: Widely regarded as the best single gauge of the U.S. equities market, this world-renowned index includes 500 leading companies in leading industries of the U.S. economy”.4

US data for the S&P 500 stock price index, earnings and dividends are taken from Robert Shiller’s website (URL: http://www.econ.yale. edu/~shiller/data/ie_data.htm). Price/earnings and dividend/price ratios for Germany, Japan, and UK are taken from the GFD.5 Interest rates for Germany, the United Kingdom, and the United States are 1–month euro-interest rates taken from the Bank for International Settlements data set, mnemonic “JDBA”. The Japanese interest rate starts too late to be useable, and has been replaced by the call-money interest rate taken from the IFS data set, IFS line 60b. Notes 1. See http://deutsche-boerse.com/dbag/dispatch/en/isg/gdb_navigation/market_data_analytics/20_indices/24_all_share_indices/20_CDAX L?module=InOverview_Index&folder type=_Index&lang=de&wp=DE0008469602&wplist=DE0008469602&active= overview. 2. See http://www.nni.nikkei.co.jp/FR/SERV/nikkei_indexes/nifaq225.html. 3. See http://www.ftse.com/Indices/UK_Indices/index.jsp.

Forecasting international financial prices with fundamentals

143

4. See http://www2.standardandpoors.com/portal/site/sp/en/us/page.topic/indices_500/2,3, 2,2,0,0,0,0,0,0,0,0,0,0,0,0.html. 5. The GFD notes on these ratios: “Dividend yields and P/E ratios are generally based upon large cap stocks in each country representing about 75% of the capitalization of that country. Since dividends and earnings are reported quarterly or annually, and at a lag, while prices change daily, the yields are based upon historical data. Dividend data are based upon the dividends reported for the trailing twelve months and do not include any forecast dividends. Fourth quarter dividends and earnings, for example, are generally not reported until February and only at this point are fourth quarter dividends and earnings included in the calculations. Hence, January dividend yields are based upon dividends through the third quarter of the previous year, but prices of the stocks in January. Earnings data are generally based upon trailing twelve-month as reported earnings”.

144

Globalization and economic integration

APPENDIX 6A2 Table 6A2.1a

STOCK MARKET RESULTS

Root mean square forecast errors (percentages)

Country

Horizon

Random walk

Univ AR

VAR

Germany

1 mon 3 mon 6 mon 12 mon 24 mon 1 mon 3 mon 6 mon 12 mon 24 mon 1 mon 3 mon 6 mon 12 mon 24 mon 1 mon 3 mon 6 mon 12 mon 24 mon

2.68 4.59 6.13 8.65 10.19 2.70 4.37 6.03 10.93 21.74 5.51 9.84 12.95 18.09 26.82 4.14 6.40 8.68 12.59 19.73

2.86 5.67 8.41 9.11 5.28 2.75 4.61 6.09 8.19 11.59 5.85 13.31 22.20 37.02 25.91 4.06 6.75 10.40 15.32 16.22

3.09 6.40 10.86 19.31 39.76 3.53 7.72 13.18 24.44 37.51 5.59 10.09 12.96 19.31 35.73 4.64 7.25 10.25 14.80 18.31

Japan

UK

USA

Table 6A2.1b

Mean absolute forecast errors (percentages)

Stock market

Horizon

Random walk

Univ AR

VAR

Germany

1 mon 3 mon 6 mon 12 mon 24 mon 1 mon 3 mon 6 mon 12 mon 24 mon

2.13 3.72 5.24 7.61 7.32 2.02 3.46 5.22 10.19 21.50

2.42 4.61 6.55 7.22 4.24 2.16 3.61 4.84 6.05 9.42

2.45 5.44 9.14 15.82 35.03 2.84 6.76 11.99 22.78 36.33

Japan

Forecasting international financial prices with fundamentals

Table 6A2.1b

145

(continued)

Stock market

Horizon

Random walk

Univ AR

VAR

UK

1 mon 3 mon 6 mon 12 mon 24 mon 1 mon 3 mon 6 mon 12 mon 24 mon

4.51 7.97 10.64 15.36 24.68 3.25 5.35 6.96 10.38 16.64

4.61 11.25 20.28 36.27 24.63 3.25 5.32 8.24 13.09 13.65

4.22 8.22 11.16 16.81 30.78 3.69 5.96 8.78 12.41 14.40

USA

Table 6A2.1c

Mean forecast errors (percentages)

Stock market

Horizon

Random walk

Univ AR

VAR

Germany

1 mon 3 mon 6 mon 12 mon 24 mon 1 mon 3 mon 6 mon 12 mon 24 mon 1 mon 3 mon 6 mon 12 mon 24 mon 1 mon 3 mon 6 mon 12 mon 24 mon

0.52 1.38 2.33 4.18 5.09 1.01 2.66 4.96 10.08 21.50 1.66 4.40 7.88 14.05 24.68 0.46 1.37 3.19 7.51 15.12

0.75 1.86 2.74 2.10 −0.79 0.93 2.41 2.88 4.07 6.43 2.62 9.75 20.28 36.27 24.63 1.52 3.90 8.00 13.09 13.65

−1.20 −3.82 −7.64 −13.84 −30.39 2.55 6.74 11.99 22.78 36.33 −0.05 −1.10 −5.19 −14.16 −28.33 0.20 0.88 2.45 5.75 7.49

Japan

UK

USA

146

Globalization and economic integration

Table 6A2.2a

Gordon earnings: root mean square forecast errors (percentages)

Stock market

Horizon

Random walk

Growth: 1-yr bkd

Growth: 3-yr bkd

Growth: 1-yr fwd

Growth: 3-yr fwd

Germany

1 mon 3 mon 6 mon 12 mon 24 mon 1 mon 3 mon 6 mon 12 mon 24 mon 1 mon 3 mon 6 mon 12 mon 24 mon 1 mon 3 mon 6 mon 12 mon 24 mon

2.68 4.59 6.13 8.65 10.19 2.70 4.37 6.03 10.93 21.74 5.51 9.84 12.95 18.09 26.82 4.14 6.40 8.68 12.59 19.73

7.64 7.73 6.48 8.36 9.52 21.02 22.93 25.87 31.99 44.70 24.61 28.40 33.29 37.59 47.85 12.53 13.78 15.91 19.59 26.97

9.61 11.26 13.61 17.46 18.61 20.31 22.14 24.95 30.41 39.03 24.14 28.63 35.31 48.59 80.06 11.14 12.43 14.65 19.40 30.03

7.80 9.04 10.79 12.71 11.70 12.06 13.24 14.98 18.53 28.02 22.58 26.91 32.88 40.29 52.67 13.26 14.83 17.30 21.79 26.56

8.05 8.99 10.45 12.97 14.97 16.93 18.89 21.84 27.85 39.66 27.44 32.93 41.60 58.12 96.26 12.53 14.13 17.18 24.32 42.68

Japan

UK

USA

Table 6A2.2b

Gordon earnings: mean absolute forecast errors (percentages)

Stock market

Horizon

Random walk

Growth: 1-yr bkd

Growth: 3-yr bkd

Growth: 1-yr fwd

Growth: 3-yr fwd

Germany

1 mon 3 mon 6 mon 12 mon 24 mon 1 mon 3 mon 6 mon 12 mon 24 mon

2.13 3.72 5.24 7.61 7.32 2.02 3.46 5.22 10.19 21.50

5.65 5.78 5.94 7.53 8.56 18.91 20.94 24.08 30.71 44.18

8.04 9.33 10.84 12.99 16.28 17.76 19.79 22.98 29.11 38.74

6.78 7.93 9.44 10.66 8.29 11.07 12.32 14.17 17.97 27.90

6.81 7.61 9.96 10.88 11.51 16.34 18.41 21.52 27.67 39.59

Japan

Forecasting international financial prices with fundamentals

Table 6A2.2b

147

(continued)

Stock market

Horizon

Random walk

Growth: 1-yr bkd

Growth: 3-yr bkd

Growth: 1-yr fwd

Growth: 3-yr fwd

UK

1 mon 3 mon 6 mon 12 mon 24 mon 1 mon 3 mon 6 mon 12 mon 24 mon

4.51 7.97 10.64 15.36 24.68 3.25 5.35 6.96 10.38 16.64

17.75 20.21 23.26 26.27 38.05 10.05 11.00 12.68 16.15 25.02

17.84 21.10 26.47 40.14 74.25 9.24 10.48 12.78 18.00 29.50

19.40 22.83 27.44 33.72 47.59 10.84 12.35 14.99 20.13 22.56

23.42 27.94 34.90 48.69 89.65 10.94 12.57 15.57 22.51 41.93

USA

Table 6A2.2c

Gordon earnings: mean forecast errors (percentages)

Stock market

Horizon

Germany

1 mon 3 mon 6 mon 12 mon 24 mon 1 mon 3 mon 6 mon 12 mon 24 mon 1 mon 3 mon 6 mon 12 mon 24 mon 1 mon 3 mon 6 mon 12 mon 24 mon

Japan

UK

USA

Random Growth: walk 1-yr bkd 0.52 1.38 2.33 4.18 5.09 1.01 2.66 4.96 10.08 21.50 1.66 4.40 7.88 14.05 24.68 0.46 1.37 3.19 7.51 15.12

−0.96 −1.73 −3.12 −4.64 −8.50 18.91 20.94 24.08 30.71 44.18 15.41 17.63 20.55 24.07 36.06 9.67 10.70 12.52 16.15 25.02

Growth: 3-yr bkd

Growth: 1-yr fwd

Growth: 3-yr fwd

7.31 8.53 10.17 12.68 16.28 17.71 19.79 22.98 29.11 38.74 16.25 19.62 26.04 40.04 74.25 8.46 9.61 11.87 17.39 29.50

2.54 3.16 3.95 4.32 3.75 11.07 12.32 14.17 17.97 27.90 17.07 20.14 24.41 31.04 47.59 5.84 6.21 6.97 8.62 16.12

4.24 4.65 5.31 6.52 8.92 16.34 18.41 21.52 27.67 39.59 22.70 27.21 34.31 48.69 89.65 10.73 12.38 15.46 22.51 41.93

148

Globalization and economic integration

Table 6A2.3a

Gordon dividends: root mean square forecast errors (percentages)

Stock market

Horizon

Random walk

Growth: 1-yr bkd

Growth: 3-yr bkd

Growth: 1-yr fwd

Growth: 3-yr fwd

Germany

1 mon 3 mon 6 mon 12 mon 24 mon 1 mon 3 mon 6 mon 12 mon 24 mon 1 mon 3 mon 6 mon 12 mon 24 mon 1 mon 3 mon 6 mon 12 mon 24 mon

2.68 4.59 6.13 8.65 10.19 2.70 4.37 6.03 10.93 21.74 5.51 9.84 12.95 18.09 26.82 4.14 6.40 8.68 12.59 19.73

10.65 13.01 16.44 23.48 35.39 24.08 26.37 29.62 35.05 42.27 12.6 14.41 16.43 17.96 24.55 10.55 11.31 12.87 14.52 19.03

10.22 12.8 15.92 19.02 26.05 24.54 26.56 29.52 35.09 43.1 14.1 16.23 18.76 22.17 28.81 8.05 9.69 11.78 15.01 13.79

6.73 7.66 8.92 11.45 17.59 18.17 20.15 23.13 29.12 40.18 14.14 16.21 18.56 21.37 29.21 12.96 14.58 17.23 22.5 33.08

6.59 9.01 13.65 24.24 35.53 11.27 12.73 14.98 20.03 31.82 15.16 17.42 20.05 21.94 24.78 10.43 11.51 14.01 18.71 30.17

Japan

UK

USA

Table 6A2.3b

Gordon dividends: mean absolute forecast errors (percentages)

Stock market

Horizon

Random walk

Growth: 1-yr bkd

Growth: 3-yr bkd

Growth: 1-yr fwd

Growth: 3-yr fwd

Germany

1 mon 3 mon 6 mon 12 mon 24 mon 1 mon 3 mon 6 mon 12 mon 24 mon

2.13 3.72 5.24 7.61 7.32 2.02 3.46 5.22 10.19 21.50

9.06 10.75 13.16 19.77 32.48 19.76 21.80 24.70 30.24 39.62

8.00 9.81 11.78 14.45 19.36 21.38 23.44 26.44 32.26 41.62

5.50 6.33 7.69 10.54 16.16 16.48 18.51 21.50 27.33 38.62

5.24 6.84 9.60 16.57 26.72 10.00 11.54 13.89 19.13 31.37

Japan

Forecasting international financial prices with fundamentals

Table 6A2.3b

149

(continued)

Stock market

Horizon

Random walk

Growth: 1-yr bkd

Growth: 3-yr bkd

Growth: 1-yr fwd

Growth: 3-yr fwd

UK

1 mon 3 mon 6 mon 12 mon 24 mon 1 mon 3 mon 6 mon 12 mon 24 mon

4.51 7.97 10.64 15.36 24.68 3.25 5.35 6.96 10.38 16.64

9.89 11.11 12.49 13.95 16.38 8.25 8.92 9.96 11.52 16.84

11.46 13.08 14.92 17.15 19.26 7.00 8.46 10.13 12.10 11.33

11.88 13.49 15.23 17.21 19.27 11.02 12.55 15.21 20.59 32.22

12.26 13.96 15.82 18.39 18.96 8.19 9.12 10.99 14.84 25.95

USA

Table 6A2.3c Gordon dividends: mean forecast errors (percentages) Stock market

Horizon

Germany

1 mon 3 mon 6 mon 12 mon 24 mon 1 mon 3 mon 6 mon 12 mon 24 mon 1 mon 3 mon 6 mon 12 mon 24 mon 1 mon 3 mon 6 mon 12 mon 24 mon

Japan

UK

USA

Random Growth: walk 1-yr bkd 0.52 1.38 2.33 4.18 5.09 1.01 2.66 4.96 10.08 21.50 1.66 4.40 7.88 14.05 24.68 0.46 1.37 3.19 7.51 15.12

−5.68 −6.75 −8.61 −14.14 −31.70 19.56 21.62 24.60 30.23 39.62 6.26 7.04 7.77 8.45 13.94 6.98 7.62 8.82 11.07 16.84

Growth: 3-yr bkd

Growth: 1-yr fwd

2.31 3.21 4.35 5.15 8.11 21.38 23.44 26.44 32.26 41.62 8.01 9.08 10.24 12.00 18.90 −3.23 −4.18 −5.33 −7.51 −10.66

−3.64 −4.51 −5.93 −9.45 −16.16 16.48 18.51 21.50 27.33 38.62 8.61 9.70 10.81 12.27 18.72 10.87 12.48 15.21 20.59 32.33

Growth: 3-yr fwd 2.48 3.92 6.70 13.39 25.06 10.00 11.54 13.89 19.13 31.37 6.67 7.11 7.22 6.00 7.99 7.27 8.25 10.22 14.78 25.95

150

Globalization and economic integration

APPENDIX 6A3 Table 6A3.1a

THE MEESE–ROGOFF RESULTS

Meese–Rogoff (1983a) Table 1: Root mean square forecast errors (in percentage terms)

Horizon Random Forward Univ VAR Frenkel walk rate AR –Bilson $/mark

1 mon 6 mon 12 mon $/yen 1 mon 6 mon 12 mon $/pound 1 mon 6 mon 12 mon EER 1 mon 6 mon 12 mon

Table 6A3.1b

3.72 8.71 12.98 3.68 11.58 18.31 2.56 6.45 9.96 1.99 6.09 8.65

3.20 9.03 12.60 3.72 11.93 18.95 2.67 7.23 11.62 N.A. N.A. 14.24

3.51 12.40 22.53 4.46 22.04 52.18 2.79 7.27 13.35 2.72 6.82 11.14

5.40 11.83 15.06 7.76 18.90 22.98 5.56 12.97 21.28 4.10 8.91 10.96

3.17 9.64 16.12 4.11 13.38 18.55 2.82 8.90 14.62 2.40 7.07 11.40

1 mon 6 mon 12 mon $/yen 1 mon 6 mon 12 mon $/pound 1 mon 6 mon 12 mon EER 1 mon 6 mon 12 mon

3.65 12.03 18.87 4.40 13.94 20.41 2.90 8.88 13.66 2.50 6.49 9.80

3.50 9.95 15.69 4.20 11.94 19.20 3.03 9.08 14.57 2.74 7.11 10.35

Meese–Rogoff (1983a) Table 2: Mean forecast errors (in percentage terms)

Horizon Random Forward Univ VAR Frenkel walk rate AR –Bilson $/mark

Dorn- Hooper busch –Morton –Frankel

0.04 −0.92 −3.93 −0.46 −3.32 −6.48 −0.31 −3.09 −7.75 −0.03 0.77 3.18

0.35 1.31 0.29 −0.06 −1.26 −2.62 −0.38 −4.05 −9.55 N.A. N.A. 7.66

0.26 −1.12 0.37 1.99 −3.31 1.23 5.20 −5.22 0.55 −0.15 −2.64 1.36 −3.17 −7.51 −8.00 −8.91 −10.45 −14.05 −0.12 −3.72 −0.48 −1.32 −9.45 −5.55 −4.17 −18.54 −13.21 0.06 0.89 0.63 1.61 3.91 3.86 6.44 7.11 7.69

Dorn- Hooper busch –Morton –Frankel −0.17 0.07 −0.59 0.17 −3.06 −1.52 −1.46 −0.18 −8.53 −1.81 −14.82 −2.38 −0.37 −0.52 −4.53 −5.30 −12.07 −11.69 0.54 0.68 2.79 3.52 5.30 5.78

Forecasting international financial prices with fundamentals

151

Table 6A3.2a Meese–Rogoff (1983b) Table 3.7: Shortest forecast horizon (in months) for which at least x percent of each model’s parameter grid improves on the random walk model in MAE/RMSE when realized values of the explanatory variables are used Model

FX rate

$/DM

Threshold/ MAE metric Frenkel – Bilson

Dornbusch – Frankel

Hooper – Morton

0–1% 10 25 50 0–1% 10 25 50 0–1% 10 25 50

24 30 30 36 12 18 30 – 12 18 30 –

$/Pound

$/yen

RMSE

MAE

RMSE

MAE

RMSE

30 30 30 36 18 18 30 – 18 18 30 –

18 18 24 30 18 24 30 – 18 24 30 –

24 24 30 36 18 24 36 – 18 24 36 –

12 18 24 36 12 12 12 24 12 12 12 24

12 18 24 30 12 12 12 18 12 12 18 18

Table 6A3.2b Meese–Rogoff (1983b) Table 3.8: Comparing the random walk and the structural models (with their best representative parameter configuration) when realized values of the explanatory variables are used (in percentage terms) Model

Random walk

Frenkel – Bilson

FX rate

$/DM

$/Pound

$/yen

Threshold/ metric

MAE

RMSE

MAE

RMSE

MAE

RMSE

1 3 12 36 1 3 12 36

2.4 4.8 9.4 18.1 9.1 11.5 12.2 12.6

3.2 6.2 10.9 21.0 11.4 14.2 15.2 17.0

2.0 3.2 9.8 23.4 4.2 8.7 13.5 15.5

2.5 5.1 11.5 25.4 6.1 11.1 16.6 18.8

2.1 4.2 10.6 19.4 4.5 7.9 9.7 10.2

3.0 5.7 13.8 23.3 6.4 11.5 13.3 14.5

152

Globalization and economic integration

Table 6A3.2b

(continued)

Model

FX rate Threshold/ metric

Dornbusch – Frankel

Hooper – Morton

1 3 12 36 1 3 12 36

$/DM MAE 5.5 8.1 8.8 8.2 8.3 8.8 9.2 9.3

$/Pound

RMSE

MAE

RMSE

6.9 9.7 10.8 10.5 10.4 11.0 11.6 11.6

8.1 8.6 10.4 8.3 4.0 8.5 10.0 10.5

10.0 10.5 12.3 10.0 10.0 10.5 12.0 12.5

$/yen MAE 4.4 7.4 7.0 8.8 4.9 8.3 8.8 9.3

RMSE 8.4 9.4 8.5 10.2 6.7 10.9 11.7 12.2

7.

Trade policy, exchange rate adjustment and unemployment Yoshiyasu Ono*

INTRODUCTION When a country has unemployment, the government often attributes it to massive imports from foreign countries. It then imposes trade restrictions so that domestic firms can expand their market share and increase employment. A famous example is the tariff war during the Great Depression that was triggered by the Smoot–Hawley Tariff Act.1 Although most economists agree that such policies would harm the world economy, they are still occasionally adopted.2 In the 1980s, for example, the United States was annoyed by a huge trade deficit with Japan. Various US industries were threatened by massive Japanese exports and many employees were laid off. Consequently, trade frictions arose in various industries, such as steel, color TV, automobiles, and microchips. The US government imposed trade restrictions to protect threatened industries, believing that they would provide a larger market to import-competing firms, thereby increasing employment. In reality, however, employment did not recover as expected. The US dollar continued to rise against the Japanese yen, causing the price competitiveness of US products to further deteriorate, and new trade frictions occurred. This chapter examines this process of unemployment and currency appreciation due to trade restrictions. It introduces an import tariff into a two-country two-commodity continuous-time competitive model presented by Ono (2006) in which people behave rationally and wages and prices continue to adjust, albeit sluggishly. However, a persistent demand shortage arises.3 In this framework it is found that an import tariff improves the current account and raises the value of the home currency, which raises the relative prices of home products.4 Moreover, the price rises are so high that home employment eventually decreases, whereas the relative prices of foreign products decline and foreign employment expands. In the conventional analysis of the optimal tariff with full employment and perfect competition, it is well known that a zero tariff is optimal for a 153

154

Globalization and economic integration

small country. In a two-country setting a positive tariff benefits the tariffimposing country, and harms the other, through an improvement in the terms of trade. Moreover, in the presence of unemployment a positive tariff seems even more beneficial since it protects an import-competing industry and raises employment. In fact, Choi and Beladi (1993, 1998) assume wage and interest rate rigidities and show that a positive tariff is beneficial even for a small economy. However, in a small-country setting with Harris–Todaro-type urban unemployment caused by fixed urban and flexible rural wages, Chen and Choi (1994) find that a negative tariff benefits the country. This occurs because they assume that the country imports a capital-intensive commodity and thus import promotion lowers the capital rent and raises the rural wage, which causes urban unemployment to decrease. Chao and Yip (2000) introduce a cash-in-advance constraint into a monetary economy with Harris–Todaro-type unemployment, and find that a negative tariff can benefit the country.5 All of these studies use a static framework. There are also some approaches that analyze the effect of a tariff on the dynamic profile of current account in a continuous-time dynamic optimization setting. For example, Roldos (1991) with capital accumulation, Mansoorian (1993) with habit formation, and Ikeda (2003, 2006) with weakly non-separable intertemporal preferences.6 They consider a small country and ignore the possibility of unemployment.7 Van Wijnbergen (1987) uses a two-period model with wage rigidity in the first period and analyzes the effect of a tariff on the current account and employment. Unlike the present analysis it regards unemployment as a short-run phenomenon that disappears in the second period. When analyzing unemployment in an open-economy dynamic framework, the new open-economy macroeconomic models, for example, Obstfeld and Rogoff (1995), Christiano et al. (1997), Betts and Devereux (2000a, 2000b), and Hau (2000), are often used.8 They analyze monopolistic competition and nominal price/wage stickiness and assume that prices and wages can be modified only at the beginning of each period. In this setup, if a government imposes a trade restriction in the middle of a period, disequilibrium arises in that period only; full employment is achieved in the next period and thereafter. Hence, they cannot deal with persistent unemployment. Since the present chapter focuses on persistent unemployment, it uses the dynamic stagnation model of Ono (2006). The next section starts with an intuitive discussion about the effect of an import tariff on the exchange rate, employment and consumption in the presence of persistent unemployment. The third section then models firm and household behavior, presents market equilibrium conditions, and formulates an autonomous dynamic system. The fourth section derives

Trade policy, exchange rate adjustment and unemployment

155

the steady state with full employment and with unemployment, and formally analyzes the effect of an import tariff in each case. It is shown that an import tariff that is aimed at raising employment actually “lowers” it, since the exchange rate appreciates and the price competitiveness of home products declines. The final section summarizes and concludes.

OUTLINE OF THE ANALYSIS Before starting the analysis, this section discusses the intuition of the model. Suppose that there are two countries h and f, the home and foreign countries, and that country h (or country f ) produces commodity 1 (or commodity 2) and exports it to the other country. Country h# imposes an import tariff t while country f does not. The current account b j ( j = h, f ) is represented as a function of relative price w and tariff t, where b j is country j’s foreign asset holdings and a dot denotes a time derivative. The relative price is the international relative price of the foreign commodity: w=

eP2 , P1

(7.1)

where e is the exchange rate of the home currency against the foreign currency and Pi (i = 1, 2) is the nominal price of commodity i measured in terms of the currency of the country in which commodity i is produced. #h The tariff t improves country h’s current account b and worsens country # f’s current account b f , that is, # # 0bh (w, t) 0b f (w, t) > 0, < 0. (7.2) 0t 0t Further, the Marshall–Lerner condition implies #that an increase # in the relative price of the foreign commodity improves bh and worsens bf : # # 0bh (w, t) 0bf (w, t) > 0, < 0. (7.3) 0w 0w With flexible exchange # rates the# exchange rate adjusts current account imbalances, that is, bh (w, t) = bf (w, t) = 0. Equations (7.2) and (7.3) yield: dw < 0, dt

(7.4)

that is, an import tariff lowers the relative price of the foreign commodity. In a trade-theoretic setting the improvement in the terms of trade,

156

Globalization and economic integration

represented by (7.4), benefits the home country and harms the foreign country. This property is valid under full employment since the improvement in the terms of trade directly increases real income of the home country and decreases that of the foreign country. However, a tariff is more often imposed when the world economy faces unemployment and each country wants to expand demand for its own product so that employment increases. It is shown below that in this case an improvement in the terms of trade due to the imposition of an import tariff harms the tariff-imposing country since it lowers international competitiveness of the home commodity and therefore decreases employment and consumption.

THE MODEL By introducing the home country’s import tariff t in the model of Ono (2006), this section presents the behavior of firms, households and governments. It also describes market adjustments and formulates an autonomous dynamic system. Firms The firm sector of country h produces commodity 1 while that of country f produces commodity 2. They have linear technology and each output is q1xh, q2xf,

(7.5)

where xj is labor input in country j and qi is the labor productivity of commodity i (= 1, 2), which is assumed to be constant. Given nominal wages Wh and Wf, under this technology each firm maximizes profits: (q1P1 − Wh) xh and (q2P2 − Wf )xf. In other words, q1 =

Wh Wf , q2 = . P1 P2

(7.6)

Thus, the firm values are zero. Households Each household’s lifetime utility depends on the path of real consumption, cj, and of real balances, mj (j = h, f ): `

Uj = 3 [ ln (cj) 1 v (mj) ] exp (2rs) ds, where v9(m) > 0, v0(m) < 0. (7.7) 0

Trade policy, exchange rate adjustment and unemployment

157

For simplicity, it is assumed that the two countries have the same utility function and the same subjective discount rate r. Real consumption cj (for j = h, f ) is cj = p1jc1j + p2jc2j, where cij is country j’s consumption of commodity i, pi j is the real consumer price of it (for i = 1, 2 and j = h, f ): p1h =

P1 h e (1 1 t) P2 f P1 P2 , p2 = , p1 = f , p2f = f , h h P P eP P

(7.8)

and Pj (j = h, f) is the consumer price index in country j, which is also used when calculating real balances mj. As shown in Appendix 7A1, if the utility function is a constant elasticity of substitution (CES) with respect to c1j and c2j, that is, (1/s) ln[k1(c1j)s + k2(c2j)s], where 1 > s > 0, then the following properties obtain. First, real prices pi j’s are functions of the relative consumer price wj in country j and satisfy: pi j = pi(wj), p19(wj) < 0, p29(wj) > 0 for i = 1, 2; j = h, f,

(7.9)

where wh and wf are: wh = (1 + t)w, wf = w,

(7.10)

since only country h imposes an import tariff. Second, utility of consumption equals ln(cj), as assumed in (7.7), after optimally allocating real consumption expenditure cj between the two commodities. Third, demand for each commodity is given by: p1(wj)c1j = d(wj)cj, p2(wj)c2j = [1 − d(wj)]cj,

(7.11)

where d(.) is the ratio of consumption expenditure on commodity 1 and satisfies: d(wj) = 1 +

wjpr 1 (wj) wjpr 2 (wj) = , 1 > d(.) > 0, d9(.) > 0. p1 (wj) p2 (wj)

(7.12)

Each household’s total asset aj is the sum of real balances mj and real international assets bj: aj = bj + mj.

(7.13)

Owing to international asset market adjustment, home and foreign nominal interest rates Rh and Rf satisfy: # e (7.14) Rh = + Rf. e

158

Globalization and economic integration

Thus, households are indifferent to holding home and foreign interestbearing assets. The flow budget equation of each household measured in each real term is: # a j = rjaj + wjxj − cj − Rjmj + zj,

(7.15)

where zj is the real lump-sum transfer, rj is the real interest rate: rj = Rj − pj,

(7.16)

and pj is the inflation rate of Pj. Each household maximizes (7.7) subject to (7.13) and (7.15) and consequently satisfies the following intertemporal optimal conditions: # cj r + j + pj = Rj = v9(mj)cj, for j = h and f, (7.17) c and the transversality condition: lims ➝ ∞lj(s)aj(s)exp(− rs) = 0,

(7.18)

where lj is the co-state variable for aj. The first equality in (7.17) shows the Ramsey equation, which implies equality between the time preference rate and the real interest rate. The second equality gives money demand which is determined so that the asset rate of interest equals the liquidity premium of money. Thus, the three values given in (7.17) respectively exhibit the desire for consumption, asset holding and money holding. Governments The home government imposes import tariff t and transfers tariff revenues to households as lump sums, whereas the foreign government neither imposes a tariff nor gives any transfer to households. Since country h’s nominal imports measured in terms of the home currency is eP2c2h, from the second equation of (7.8), zh =

tp2 (wh) ch2 f , z = 0. (1 1 t)

(7.19)

Market Adjustments The equilibrium conditions in each country’s money market and the international asset market are:

Trade policy, exchange rate adjustment and unemployment

the money markets:

Mh Mf = mh, f = mf, h P P

the asset market: Phbh + ePfbf = Ph cbh 1

p1 (wh) f b d = 0, p1 (wf)

159

(7.20)

where Mj is the nominal money supply in country j. Commodity market adjustments are perfect; from (7.5) and (7.11), q1xh =

d (wh) h d (wf) f 1 2 d (wh) h 1 2 d (wf) f c + c , q2xf = c + c. h f p1 (w ) p1 (w ) p2 (wh) p2 (wf)

(7.21)

Nominal-wage adjustments in each country are assumed to be sluggish,9 # Wj = a(xj − 1), (7.22) Wj where a is the adjustment speed of nominal wage Wj and the labor input xj equals the employment rate since each country’s population is normalized to unity. Dynamics The money market equilibrium conditions given in (7.20) yield: m# j = − pj for j = h and f. mj

(7.23)

Since (7.6) and (7.22) generate movements of commodity prices P1 and P2, # # P1 P2 = a(xh − 1), = a(xf − 1), (7.24) P1 P2 from (7.10), the time differentiation of the first and fourth equations in (7.8) and (7.12), one obtains the inflation rates of consumer price indices Ph and Pf: w# w# ph = a(xh − 1) + [1 − d(wh)] , pf = a(xf − 1) − d(wf) . (7.25) w w Substituting (7.17) into (7.14) gives the dynamics of exchange rate e: # e = v9(mh)ch − v9(mf)cf. (7.26) e From (7.6) and (7.8), real wages wh and wf are wh = Wh /Ph= p1(wh)q1, wf = Wf/Pf = p2(wf)q2. Applying (7.13), (7.16), and (7.19) to which c2h in (7.11)

160

Globalization and economic integration

is substituted, (7.23) and wh and wf given above into # # (7.15) and using (7.20) yields the two countries’ current accounts bh and bf : # 1 1 td (wh) h # f bh = rhbh + p1(wh)q1xh − c , b = rfbf + p2(wf)q2xf − cf, 11t (7.27) where bh = − [ p1 (wh) /p1 (wf) ] bf. They show that each country’s current account equals the sum of its interest earnings and total product minus consumption expenditure. In the following analysis it is assumed that international assets are held in the form of real bonds measured in the international real price. Thus, bf is unaffected by a change in tariff t whereas bh is affected, since tariff t varies the home country’s consumer price index. Since the time differentiation of (7.1), (7.24) and (7.26) generate the dynamics of relative price w, w# = v9(mh)ch − v9(mf)cf + a(xf − 1) − a(xh − 1), (7.28) w from (7.10), (7.21) and (7.25) one obtains ph and pf as functions of w, ch, cf, mh and mf. Thus, seven dynamic equations given by (7.17), (7.23), (7.27) and (7.28) formulate an autonomous dynamic system with respect to ch , cf, mh, mf, bh, bf and w.10

EFFECTS OF TARIFFS This section obtains the steady state with full employment and that with persistent unemployment and discusses the effects of an import tariff in each case. Before starting the analysis the steady-state conditions that hold in common are summarized. In the steady state of the present dynamics, where w, ch and cf are constant, (7.16), (7.17) and (7.25) yield: rh = rf = r, pj = a(xj − 1), Rj = v9(mj)cj = r + pj for j = h, f.

(7.29)

# Each country’s real current account bj given by (7.27) is zero and hence from (7.29), p1 (wh) f # 1 1 td (wh) h h h bh = 2 rb + p (w )q x − c =0 1 1 p1 (wf) 11t and # bf = rbf + p2(w)q2xf − cf = 0.

(7.30)

Trade policy, exchange rate adjustment and unemployment

161

Full Employment If full employment is reached in the steady state (xh = xf = 1), (7.10) and (7.30) yield: ch = e

(1 1 t) p1 [ (1 1 t) w ] rbf 1 q1 b and cf = rbf + p2(w)q2. f a2 1 1 td [ (1 1 t) w ] p1 (w) (7.31)

Totally differentiating (7.31) and using (7.9) and (7.12) one finds that in the neighborhood where t is zero ch and cf satisfy: 0ch 0ch 0cf 0cf = 0, < 0; = 0, > 0. 0t 0w 0t 0w

(7.32)

Substituting wh and wf in (7.10) and ch and cf in (7.31) into the first equation of (7.21) in which xh = xf = 1 leads to the solution of w as a function of t, which yields the solutions of ch and cf. Applying the properties given in (7.9), (7.12) and (7.32), this yields a natural result: tariff t lowers w, the relative price of the foreign commodity. Thus, from the properties of ch and cf given by (7.32) it is found that tariff t increases ch and decreases cf, as is consistent with the standard result of the optimal tariff. Note that the steady-state values of w, ch and cf obtained above are independent of mj (for j = h, f ). Given such w, ch and cf, from (7.29) in which xh = xf = 1 one obtains mj that satisfies: v9(mj)cj = r.

(7.33)

Stagnation The full-employment steady state represented by (7.31) and (7.33) may not exist, and then persistent stagnation arises. As proved by Ono (2006), if v9(m) has a positive lower bound, that is,11 limm ➝ ∞v9(m) = b > 0,

(7.34)

a liquidity trap arises and persistent stagnation may occur. The money demand function is the relationship between Rj and mj given by the second equality of (7.17), Rj = v9(mj)cj, for given cj. Since v0(mj) < 0, Rj is negatively related to mj. In the presence of a positive lower bound of v9(mj), as mj increases, Rj decreasingly approaches bcj but never becomes lower than this, implying the existence of a liquidity trap. In the presence of the liquidity trap one finds the following property:

162

Globalization and economic integration

bcj ≥ r 3 mj that satisfies (7.33) does not exist,

(7.35)

where cj takes the full-employment level given by (7.31). Appendix 7A2 exhibits the range of the foreign asset position that makes (7.35) for both countries valid.12 Under (7.35) the left-hand side of (7.33) exceeds the right-hand side for any mj, implying that a desire for saving exceeds that for consumption as long as the volume of consumption is so high as to attain full employment. Therefore, a demand shortage arises. In this case nominal prices and wages continue to decline and v9(mj) converges to b. Therefore, from (7.29), Rj = bcj = r + a(xj − 1).

(7.36)

Since (7.21) gives xh and xf as functions of w, t, ch and cf, from (7.10) and (7.36) ch and cf are: ch = ␾h(cf, w; t), cf = ␾f(ch, w; t),

(7.37)

which naturally satisfy: 0␾ h (cf, w; t) 0␾ h (cf, w; t) 0␾ h (cf, w; t) > 0, > 0, > 0; 0cf 0w 0t 0␾ f (ch, w; t) 0␾ f (ch, w; t) 0␾ f (ch, w; t) > 0, < 0, < 0. h 0c 0w 0t

(7.38)

The proof of (7.38) is set out in Appendix 7A3. Since (7.36) generates: dcj a = > 0, dxj b

(7.39)

the implications of (7.38) are the following. A country’s employment and consumption are stimulated by an increase in the other country’s consumption. A rise in the relative price of the foreign commodity expands the home country’s employment and consumption. Import tariff t increases the home country’s employment and consumption but decreases the other country’s. Note that these properties are valid only in a partial sense. The total effect of an import tariff is quite different, as discussed in the following analysis. The two curves of (7.37) are shown in Figure 7.1, where the following properties hold. For given w a country’s consumption is positive even if the other country’s consumption is zero, an increase in a country’s consumption raises the other’s, and the intersection exists. Point E denotes the steady-state levels of ch and cf as functions of w and t:

Trade policy, exchange rate adjustment and unemployment

ch

163

␾f(c h, ␻; t)

␾h(c f, ␻; t)

E ␾h(0, ␻; t)

0 Figure 7.1

␾f(0, ␻; t)

cf

Interdependence of consumption ch = ch(w, t), cf = cf(w, t).

(7.40)

The steady-state level of w is determined so as to adjust current account imbalances, as shown by (7.30), into which ch and cf in (7.40) are substituted. In this state, xh and xf are smaller than 1 and thus all nominal prices and wages continue to decline and mh and m f continue to expand. Nevertheless, the transversality condition in (7.18) holds since aj satisfies (7.13), bj stays finite, and (7.23), (7.29) and (7.34) yield m# j/mj = 2pj = r 2 bcj < r. In this state the effect of an import tariff is examined in the neighborj hood where t = 0. Since (7.36) shows xj as a function# of cj and # f c is repreh sented as a function of # # w #and t in (7.40), # (7.30) gives b and b as functions of w and t, that is, bh= bh (w, t) , bf = bf (w, t) . Using these functions one finds that (7.4) is valid under the Marshall–Lerner condition represented by (7.3), whose proof is set out in Appendix 7A3. Thus, an increase in tariff t decreases w, the relative price of the foreign commodity. It harms home firms and benefits foreign firms, which reduces home employment xh and consumption ch (see (7.39)) and expands foreign employment xf and consumption cf:

164

Globalization and economic integration

t c 1 wT; xhT, chT; xf c, cf c,

(7.41)

as is formally proved in Appendix 7A3. Equations (7.26), (7.29), (7.34) and (7.41) yield: # d (e/e) dch dcf = ba 2 (7.42) b < 0, t c 1 RhT, phT, Rf c, pf c. dt dt dt Since neither Wh nor Wf can discontinuously jump and, from (7.6), P1 and P2 are always set equal to Wh/q1 and Wf/q2, respectively, neither P1 nor P2 can discontinuously jump. Therefore, the first property of (7.41) implies that the exchange rate e (= wP1/P2) discontinuously decreases when tariff t is imposed. Thereafter, exchange rate e gradually moves along with P1 and P2 so that w stays constant over time. The first property of (7.42) means that the appreciation speed of the home currency also increases. Intuitively, a rise in t potentially improves the home current account, causing the value of the home currency to appreciate. It causes the home commodity to lose (and the foreign commodity to gain) price competitiveness in the international market. Therefore, demand for the foreign commodity increases, causing foreign employment and consumption to improve, while demand for the home commodity decreases so much that home employment and consumption decrease. The abovementioned mechanism seems opposite to what policy makers usually have in mind. It may be because they ignore adjustments of commodity prices and the exchange rate. In fact, from the properties of ␾h and ␾f given in (7.38), if the exchange rate and prices are unchanged and thus relative price w is constant, a rise in the tariff rate shifts the ␾h curve upward and the ␾f curve leftward, as illustrated in Figure 7.2. Consequently, the intersection point of the two curves moves in the north-west direction from E to E9, causing the tariff-imposing country’s consumption to increase and the other’s consumption to decrease, as is usually expected. Appendix 7A3 formally proves this property: 0ch (w, t) 0cf (w, t) > 0, < 0. 0t 0t

(7.43)

However, this process accompanies an improvement in the home current account. Thus, under the flexible exchange rate regime the exchange rate appreciates, which harms competitiveness of the home commodity. The present analysis implies that the harmful effect is so strong that home employment and consumption decrease while foreign employment and consumption increase.

Trade policy, exchange rate adjustment and unemployment

165

ch ␾f(c h, ␻; t) t ␾h(c f, ␻; t) t E' E

cf

0 Figure 7.2

The effect of a tariff under fixed prices

CONCLUSIONS An import tariff benefits the country that imposes it if full employment holds since the well-known mechanism of the optimal tariff operates. In the presence of unemployment, it is believed to benefit the country since it is thought to expand demand for the home commodity and raise employment and income in the home country. In this argument, however, the exchange rate adjustment is ignored. If it is taken into account, the result is quite different from the usual belief. An import tariff reduces home employment and consumption and increases foreign employment and consumption as a result of the exchange rate adjustment. An import tariff reduces imports and thus improves the current account. It makes the home currency appreciate against the foreign currency so as to adjust current account imbalances, causing the international relative price of the home commodity to rise. If the economy faces a demand shortage, the appreciation of the home currency is so high that home production decreases, which worsens home employment and consumption. In

166

Globalization and economic integration

turn this reduces the price of the foreign commodity so much that world demand for the foreign commodity expands and hence foreign employment and consumption increase, even though the home country imposed a tariff. Such a consequence may make the home country regard the tariff as being not high enough to expand home employment and lead to successive impositions of tariffs, with the obvious consequences.

NOTES * 1. 2. 3.

4. 5.

6. 7. 8. 9.

10. 11.

12.

This research is financially supported by the Grants-in-Aid for Scientific Research, Japan Society for the Promotion of Science (JSPS). See Archibald and Feldman (1998) for a history of the Act. There are many empirical papers on the effect of the Act, for example, Crucini (1994), Crucini and Kahn (1996), Archibald and Feldman (1998) and Madsen (2001). A closed-economy version of this model is presented by Ono (1994, 2001). This type of model is widely used in various analyses that examine persistent stagnation in a dynamic-optimization framework of a monetary economy. For example, Matsuzaki (2003) finds the effect of a consumption tax on effective demand in the presence of poor and rich people. Hashimoto (2004) examines the intergenerational redistribution effects of the public pensions system in an overlapping generations framework with the present type of stagnation. Johdo (2006) considers the relationship between R&D subsidies and unemployment in the present stagnation setting. Rodriguez-Arana (2007) examines the dynamic path with a public deficit in the stagnation case and compares it with that in the neoclassical case. Johdo and Hashimoto (2009) introduce FDI into a two-country model with the present stagnation mechanism and analyze the effect of the corporation tax on employment in each country. In fact, recent empirical research by Santos-Paulino and Thirlwall (2004) finds that a trade liberalization worsens the current account. There are also other types of distortions that cause a negative tariff to be beneficial. For example, under an oligopolistic vertical relationship, Lahiri and Ono (1999) show a negative tariff to be beneficial. See also Laure and Gervais (2002) for the same possibility with oligopolistic distortions. Inoue (2000) adopts a two-period trade model and finds that a negative tariff can benefit the country. Ikeda (2006) also analyzes a two-country case. See Lane (2001) for an extensive survey on the new open-economy macroeconomics. This assumption is imposed in order to allow disequilibrium to occur in the labor market; otherwise the possibility of unemployment is not possible by definition. Note that even under this assumption the possibility of the full-employment steady state is not eliminated. In fact, the full-employment steady state is presented in the next section. See Ono (2006) for the stability and determinacy of the present dynamics when t = 0 in the case where full employment is reached and in the case where persistent stagnation occurs. Keynes (1936, ch.17) mentions that this is an essential property of money. Under this property a shortage of effective demand obtains. See Ono (1994, pp. 31–3; 2001) for the implication of this property on the effective-demand shortage caused by a liquidity trap. This property is empirically shown by Ono (1994, ch. 3) using generalized method of moments, and more extensively by Ono et al. (2004) using both a parametric and a non-parametric approach. See Ono (2007) for the foreign asset position under which a country faces stagnation and the other country has full employment in the absence of a tariff.

Trade policy, exchange rate adjustment and unemployment

167

REFERENCES Archibald, R.B. and D.H. Feldman (1998), “Investment during the Great Depression: uncertainty and the role of the Smoot–Hawley tariff”, Southern Economic Journal, 64, 857–79. Betts, C. and M.B. Devereux (2000a), “Exchange rate dynamics in a model of pricing-to-market”, Journal of International Economics, 50, 215–44. Betts, C. and M.B. Devereux (2000b), “International monetary policy coordination and competitive depreciation: a reevaluation”, Journal of Money, Credit, and Banking, 32, 722–45. Chao, C.-C. and C.K. Yip (2000), “Urban unemployment and optimal trade policy in a cash-in-advance economy”, Journal of Economics, 71, 59–77. Chen, J. and E.K. Choi (1994), “Trade policies and welfare in a Harris–Todaro Economy”, Southern Economic Journal, 61, 426–34. Choi, E.K. and H. Beladi (1993), “Optimal trade policies for a small open economy”, Economica, 60, 475–86. Choi, E.K. and H. Beladi (1998), “Welfare reducing trade and optimal trade policy”, Japan and the World Economy, 10, 187–98. Christiano, L.J., M. Eichenbaum and C.L. Evance (1997), “Sticky prices and limited participation models of money: a comparison”, European Economic Review, 41, 1201–49. Crucini, M.J. (1994), “Sources of variation in real tariff rates: the United States 1900 to 1940”, American Economic Review, 84, 732–43. Crucini, M.J. and J. Kahn (1996), “Tariffs and aggregate economic activity: lessons from the Great Depression”, Journal of Monetary Economics, 38, 427–67. Hashimoto, K. (2004), “Intergenerational transfer and effective demand”, Economics Bulletin, 5, 1–13. Hau, H. (2000), “Exchange rate determination: the role of factor price rigidities and nontradeables”, Journal of International Economics, 50, 421–47. Ikeda, S. (2003), “Tariffs, time preference, and the current account under weakly nonseparable preferences”, Review of International Economics, 11, 101–13. Ikeda, S. (2006), “Luxury and wealth”, International Economic Review, 47, 495–526. Inoue, T. (2000), “The optimal tariff formula in a two-period economy”, Japanese Economic Review, 51, 596–604. Johdo, W. (2006), “Geographical space and effective demand under stagnation”, Australian Economic Papers, 45, 286–98. Johdo, W. and K. Hashimoto (2009), “International relocation, the real exchange rate and effective demand”, Japan and the World Economy, 21, 39–54. Keynes, J.M. (1936), The General Theory of Employment, Interest and Money, London: Macmillan. Lahiri, S. and Y. Ono (1999), “Optimal tariffs in the presence of middlemen”, Canadian Journal of Economics, 32, 55–70. Lane, P.R. (2001), “The new open economy macroeconomics: a survey”, Journal of International Economics, 54, 235–66. Laure, B. and J.-P. Gervais (2002), “Welfare-maximizing and revenue-maximizing tariffs with a few domestic firms”, Canadian Journal of Economics, 35, 786–804. Madsen, J.B. (2001), “Trade barriers and the collapse of world trade during the Great Depression”, Southern Economic Journal, 67, 848–68.

168

Globalization and economic integration

Mansoorian, A. (1993), “Tariffs, habit persistence, and the current account”, Canadian Journal of Economics, 26, 194–207. Matsuzaki, D. (2003), “The effects of a consumption tax on effective demand under stagnation”, Japanese Economic Review, 54, 101–18. Obstfeld, M. and K. Rogoff (1995), “Exchange rate dynamics redux”, Journal of Political Economy, 103, 624–60. Ono, Y. (1994), Money, Interest, and Stagnation, Oxford: Oxford University Press. Ono, Y. (2001), “A Reinterpretation of Chapter 17 of Keynes’s General Theory: effective demand shortage under dynamic optimization”, International Economic Review, 42, 207–36. Ono, Y. (2006), “International asymmetry in business activity and appreciation of a stagnant country’s currency”, Japanese Economic Review, 57, 101–20. Ono, Y. (2007), “International transfer under stagnation”, in Theory and Practice of Foreign Aid, edited by S. Lahiri, Amsterdam: Elsevier, pp. 155–71. Ono, Y., K. Ogawa and A. Yoshida (2004), “Liquidity Preference and persistent unemployment with dynamic optimizing agents”, Japanese Economic Review, 55, 355–71. Rodriguez-Arana, A. (2007), “Inflation and the public deficit when the utility of money is insatiable”, Japanese Economic Review, 58, 238–54. Roldos, J.E. (1991), “Tariffs, investment and the current account”, International Economic Review, 32, 175–94. Santos-Paulino, A. and A.P. Thirlwall (2004), “The impact of trade liberalization on exports, imports and the balance of payments of developing countries”, Economic Journal, 114, F50–F72. van Wijnbergen, S. (1987), “Tariffs, employment and the current account: real wage resistance and the macroeconomics of protection”, International Economic Review, 28, 691–706.

Trade policy, exchange rate adjustment and unemployment

169

APPENDIX 7A1 For the CES utility function of consumption: 1 u^ (cj1,cj2) = a bln[k1(c1j)s + k2(c2j)s], where 1 > s > 0, s

(7A1.1)

p1j and p2j, namely real consumer prices of the two commodities in country j, are respectively: 1

1

p1j = p1(wj) ; ck 11 2s 1 k 12 2s (wj) p2 = p2(w ) = w p1(w ) ; ck j

j

j

j

1 1 2s 1

2

s 1 2s

s

d

1 2s s

, p19(wj) > 0, 1

(wj) 1 2s 1 k21 2s d

1 2s s

, p29(wj) < 0, (7A1.2)

where wj is given in (7.10). These equations give (7.9). Each household maximizes lifetime utility Uj subject to (7.13) and (7.15) and consequently the intra-temporal optimal conditions are: p1(wj)c1j = d(wj)cj, p2(wj)c2j = [1 − d(wj)]cj,

(7A1.3)

where d(.) is the expenditure ratio of commodity 1, which satisfies: 1

k11 2s d(wj) = k

1 1 2s 1

1 1 2s 2

1k

(wj)

(7A1.2) and (7A1.4) yield (7.12).

2

s 1 2s

, 1 > d(.) > 0, d9(.) > 0.

(7A1.4)

170

Globalization and economic integration

APPENDIX 7A2 The region of bh in which (7.35) is valid for both countries is examined when t = 0. From (7A1.2), (7A1.4), (7.21) where xh = xf = 1, and (7.31), ch and cf under full employment are: ch =k1qs1 (k1qs1 1 k2qs2 ) cf = k2qs2 (k1qs1 1 k2qs2 )

1 2s s 1 2s s

+ rbh, 2 rbh.

(7A2.1)

Therefore, (7.35) reduces to: 1 2s 1 2s r r 2 k1qs1 (k1qs1 1 k2qs2 ) s ≤ rbh (= 2rbf) ≤ k2qs2 (k1qs1 1 k2qs2 ) s 2 , b b

(7A2.2) where the first inequality is the condition under which the full-employment level of mh does not exist, while the second one is the condition under which the full-employment level of mf does not exist. Note that the region given by1 (7A2.2) exists if b, q1 or q2 is large enough to satisfy r/b ≤ 2 (k1qs1 1 k2qs2 ) s . This condition is essentially the same as that for the stagnation steady state to obtain in Ono (2001, 2006).

Trade policy, exchange rate adjustment and unemployment

171

APPENDIX 7A3 Since (7.21) gives xh and xf as functions of w, t, ch and cf, from (7.10) and (7.36), ch and cf are: ch = ␾h(cf, w; t) = c

f1 [ (1 1 t) w ] a cff1 (w) a r2a 1 d^e1 2 f, b bq1 bq1

cf = ␾f(ch, w; t) = e

chf2 [ (1 1 t) w ] a f2 (w) a r2a 1 f ^ c1 2 d, b bq2 bq2 (7A3.1)

where: f1(wj) =

d (wj) 1 2 d (wj) j , f (w ) = , 2 p1 (wj) p2 (wj)

f19(wj) = 2wjf29(wj) > 0.

(7A3.2)

In order for the two curves to have the shapes depicted in Figure 7.1 it must be that: r > a and f1 [ (1 1 t) w ] a f2 (w) a f c1 2 d bq1 bq2 f1 (w) a f2 [ (1 1 t) w ] a −c de f > 0. bq1 bq2

A ; e1 2

(7A3.3)

From (7A1.2), (7A1.4), (7A3.1), (7A3.2) and (7A3.3), the properties in (7.38) are derived. Totally differentiating (7.30) in the neighborhood where t = 0 and using (7.12) and (7.36) gives: bp1q1 # dw 2 1bdch − (1 − d)p1q1xh a b = 0, dbh = a a w bp2q2 # dw 2 1bdcf + dp2q2xf a b = 0. dbf = a a w From (7A3.1) and (7A3.2) one finds: a afr1 a1 2 b bp 2q2 dch = £ § [(ch + cf)dw + chwdt], Abq1

(7A3.4)

172

Globalization and economic integration

dcf = £

afr2 a1 2

a b bp1q1 § [(ch + cf)dw + chwdt], Abq2

(7A3.5)

where A is given by (7A3.3). Substituting them into (7A3.4) and using (7A1.4) and (7A3.2) yields: a a # a1 2 b a1 2 b h f bp1q1 bp2q2 0b (w,t) d (1 2 d) (c 1 c ) = c 2 (1 2 s) § , w d£ 0w 12s A a a # a1 2 b a1 2 b h f bp1q1 bp2q2 0b (w,t) d (1 2 d) (c 1 c ) = c . (7A3.6) d 0t 12s A # Therefore, under the Marshall–Lerner condition (∂b (w,t) /∂w > 0), it must be that:1 a a a1 2 b a1 2 b bp1q1 bp2q2 2 (1 2 s) > 0. (7A3.7) A Since if t = 0 then A in (7A3.3) reduces to: A ; da1 2

a a b + (1 2 d)a1 2 b > 0, bp1q1 bp2q2

(7A3.7) implies: 12

a a > 0, 1 2 > 0. bp1q1 bp2q2

(7A3.8)

(7A3.3), (7A3.5) and (7A3.8) generate (7.43). From (7A3.3), (7A3.6) and (7A3.8), dw/dt satisfies: # 0b (w, t) /0t dw < 0, (7A3.9) =2 # dt 0b (w, t) /0w which implies (7.4). From (7.39), (7A3.4), (7A3.8) and (7A3.9), properties in (7.41) obtain. Note 1. It is valid when s is close to 1 and hence the two commodities are close substitutes to each other.

8.

Trade and wage inequality in a world of incomplete diversification Christis G. Tombazos*

INTRODUCTION The significant increase in wage inequality that has been observed over the last few decades in the United States coupled with a concurrent and equally significant expansion of trade has attracted considerable attention in the literature.1 Early research in this area relied heavily on the well-known Heckscher– Ohlin (HO) channels. However, in light of little evidence of a substantial increase in the relative price of skilled to unskilled-intensive goods it was soon concluded that the HO mechanism was an unlikely culprit (see Slaughter, 1998, for example). The potential relevance of trade in the wage inequality debate was re-introduced in the literature by the pioneering work of Feenstra and Hanson (1996a, 1996b, 1999). These authors argue that preoccupation with HO dynamics that emphasize trade in final goods obfuscates the full range of channels through which import competition impacts labor markets. They explain that imports of intermediate goods have the potential to have a significant impact on wages by fragmenting the set of production processes that typically take place within individual manufacturing industries into distinct sub-activities which are then re-allocated across countries (Feenstra and Hanson, 1996b, p. 240). In this context, Feenstra and Hanson show that to the extent that intermediate imports in the US are low-skill intensive they shift employment away from low-skill labor and contribute to wage inequality. In an effort to shed light on how trade can facilitate inequality in developing countries, Zhu and Trefler (2005) extend the general framework of Feenstra and Hanson by introducing Ricardian sources of comparative advantage. Yet, despite the evolving importance of how trade in intermediate goods promotes specialization, relevant research continues to rely on the regularities of a globally diversified production. For example, research in this area typically reconciles 4–digit SITC (Standard 173

174

Globalization and economic integration

International Trade Classification) classifications of imports into the US with 4–digit SIC (Standard Industrial Classification) figures irrespectively of the source of these imports (see ibid., p. 33, for example). Of course, the potential for aggregation bias in such studies that may derive from the extent of specialized production is well known. However, it was not until the recent findings of Davis and Weinstein (2001) and Schott (2003) that the true scope of this bias was revealed. These authors show that traditional product categorizations hide fundamental and profound cross-country specific product differences. They further demonstrate that cross-country specialization in largely exclusive subsets of goods is significant. Aggregation bias aside, the prevalence of specialization can have profound implications for the study of trade and wages. For example, an increase in non-competing intermediate imports can potentially stimulate demand for all domestic labor with uncertain effects on wage inequality. In this chapter, we extend Tombazos (2003, 2007) in an effort to investigate the role of imports disaggregated by kind in wage inequality using an economy-wide production theory approach that relies on a flexible functional representation of aggregate production. This approach has a number of advantages over competing methodologies. By not requiring “matching” imports with domestic output, it is not susceptible to the aggregation bias. In addition, by assuming an economy-wide perspective that does not depend on, or require, fully diversified production, this approach can investigate potential substitution between different types of labor that may ensue from a process of trade-induced fragmentation of domestic production. Moreover, the flexible representation of production used in this approach differs from the Cobb–Douglas and the constant elasticity of substitution (CES) in that it does not restrict, a priori, the signs or sizes of estimated coefficients. Hence, our estimations capture potential complementarities between non-competing imports and domestic labor. The empirical model proposed in this article disaggregates imports by kind on the basis of their degree of “intermediateness”. This facilitates a study of the special role of traded inputs in promoting fragmentation that has been highlighted in the “outsourcing” literature. In addition, disaggregation of imports on this basis sheds light on the results of a host of studies, such as Aw and Roberts (1985) and Tombazos (1998, 1999b), that have identified substantial differences in the role of intermediate and final imports in domestic production. In particular, these authors find that imports of final goods, or more precisely goods not subject to extensive domestic handling, typically lead to net substitutions with aggregate

Trade and wage inequality

175

domestic labor, while imports of ordinary intermediate products generally complement aggregate labor demand. While such studies suggest that the degree of import “intermediateness” is likely to differentially impact the demand for different types of labor, they do not employ frameworks that can relate such effects directly to wages. The remainder of this chapter is organized as follows. The econometric model is examined in the next section. The discussion of the data, estimation details, and results are in the third section. Concluding remarks are reserved for the fourth section.

THE MODEL The production theory approach to modeling international trade was originally proposed by Burgess (1974, 1976) and further developed by Kohli (1978, 1991). A derivative of this framework was applied to the study of trade and wages by Tombazos (1999a), and was subsequently extended in a number of relevant applications by Falk and Koebel (2002), Tombazos (2003), Hijzen et al. (2005), and others. This approach requires the specification of a model that treats imports as an input in the domestic economy on the premise that all imports, including those of so-called “final” goods, are subject to extensive domestic “downstream handling” before reaching the consumer. This is reflected in the well-known observation that a significant portion of the “shelf price” of imported commodities reflects value added domestically.2 Given our objective, labor is disaggregated into skilled (S) and unskilled (U) categories and imports in a total of four groups: Industrial Supplies (D), Capital Equipment (F), Consumer Goods (G) and Services (M). Following Krugman (1995) and Kohli (1991), the prices of capital (K), skilled labor (S), and unskilled labor (U), together with the quantities of imports (D, F, G, and M) and aggregate output (Y) are treated as variable. Hence, the production technology is represented using the symmetric normalized quadratic (SNQ) variable profit function developed by Kohli (1993). This functional form allows the required non-uniform statistical treatment of inputs, and is given by: p 5 12 (grx) prAp/ (vrp) 1 12 (vrp) xrBx/ (grx) 1 prCx 1 prDxt 1 12 (vrp) (grx) xt2,

(8.1)

where x 5 (xU, xS, xK)r and p 5 (pY, pD, pF, pG, pM) collect the quantities of fixed inputs and the prices of “outputs”, respectively; A ; [ aih ] ,

176

Globalization and economic integration

B ; [ bjk ] , C ; [ cij ] , and D ; [ dij ] denote unknown symmetric parameter matrices, g ; [ gj ] and v ; [ vi ] represent vectors of pre-selected parameters 4i, h [ (Y, D, F, G, M) and j, k [ (U, S, K) ; x denotes an unknown scalar parameter; Sh(Y,D,F,G,M)aih 5 0 4i, h [ (Y, D, F, G, M) ; Sk(U,S,K)bjk 5 0 4j [ (U, S, K) ; Si(Y,D,F,G,M)vi 5 1 and Sj(U,S,K)gj 5 1. Equation (8.1) is a fully flexible functional form and it is neither necessarily concave in fixed input quantities, nor necessarily convex in the prices of the variable quantities. Using the Gorman–Diewert adaptation of Hotelling’s lemma, in the context of the variable profit function, differentiation of p(.) with respect to input prices yields the supply of output and the demand for variable inputs (that is, imports) given by (see Diewert, 1974, p. 137): y 5 (grx) Ap/ (vrp) 2 12 v (grx) prAp/ (vrp) 2 1 12vxrBx/ (grx) 1 Cx 1 Dxt 1 12v (grx) xt2,

(8.2)

where y collects the quantities of “outputs”.3 If, under competitive conditions, producers optimize with respect to fixed inputs, the inverse fixed input demand functions for capital and the two types of labor can be derived from the “marginal product” conditions and are given by: w 5 12gprAp/ (vrp) 1 (vrp) Bx/ (grx) 2 12g (vrp) xrBx/ (grx) 2 1 Crp 1 Drpt 1 12 (vrp) gxt2,

(8.3)

where w represents the prices of the fixed inputs. Substitution possibilities between inputs and outputs can be identified from the elasticity matrix: E=

c

({diag [ⵜpp( # ) ] } 21 # [ⵜ2pp p( # ) ] # [diag (p) ]) ({diag [ⵜxp( # ) ] } 21 # [ⵜ2xp p( # ) ] # [diag (p) ])

({diag [ⵜpp( # ) ] } 21 # [ⵜ 2px p( # ) ] # [diag (x) ]) d ({diag [ⵜxp( # ) ] } 21 # [ⵜ 2xx p( # ) ] # [diag (x) ])

(8.4) where, given m, n [ (p, x) , ⵜ mp ( # ) , represents the gradient of p(.) with respect to m and ⵜ 2mmp ( # ) denotes the sub-Hessian of p(.) with respect to m and n. Similarly, the impact of technical change, made possible by the passage of time, on the prices of the fixed inputs can be captured by the time semi-elasticity: Ext 5 [ diag (ⵜ x) p ( # ) ] 21 # [ ⵜ 2xtp ( # ) ] .

(8.5)

Trade and wage inequality

177

DATA, ESTIMATION, AND RESULTS Estimation of the model requires data on prices and quantities for skilled and unskilled labor, capital, intermediate and semi-final imports, and aggregate output. The data used to construct these variables are from the US Department of Commerce, Bureau of Economic Analysis National Income and Product Accounts of the United States (NIPA), the Survey of Current Business, and the Census of Population 1970. Subject Reports: Occupations by Industry. To generate price and quantity indices corresponding to economy-wide output, capital, skilled labor, and unskilled labor, in accordance with the requirements of the model discussed in the previous section, we follow the approach used by Tombazos (2003). Specifically, construction of economy-wide output entails a Tornqvist aggregation of all categories of income listed in the NIPA.4 Given that the model treats imports as inputs into domestic production, nominal capital expenditure for any given year corresponds to the nominal value of economy-wide output net of the wage bill and expenditures on imports. To determine the implicit rental rate of capital for each year we divide nominal capital expenditures by the capital stock which is defined as the sum of the net stock of fixed non-residential equipment and structures and the net stock of residential capital. Economy-wide data on employment and wages for skilled and unskilled labor are not currently available for the United States. To produce representative indices we rely on the definition for skilled labor used by Tombazos (ibid.) who classifies skilled occupations to include professionals and managers as well as sales, clerical, and precision production labor. Using this definition and the occupational breakdown across industries that appear in the Census of Population 1970. Subject Reports: Occupations by Industry, we calculate the percentage of skilled workers in total employment in each of the 54 industries identified in the NIPA. We then classify industries as either high-skill intensive or low-skill intensive. An industry is considered to be high-skill intensive if it employs a higher percentage of skilled labor than the average of all industries we examine (this approximates the economy’s skilled labor–aggregate labor ratio). Representative wages and employment levels for the clusters of high- and low-skill intensive industries were derived using Tornqvist aggregations. Disaggregation of imports into the various categories outlined in the previous section was based on NIPA’s end-use classification of imports. The data cover the 1968–94 period. The econometric model consists of eight equations: the aggregate output supply function, the four variable input demand functions, and the three inverse demand functions of capital and the two types of

178

Table 8.1

Globalization and economic integration

Average elasticities of the symmetric normalized quadratic variable profit function

Price elasticities of inverse factor demands (Stolper–Samuelson Elasticities) eji 5 0 ln wj /0 ln pi eUD 5 20.103 (22.99) < eSD 5 20.063 (24.08) eUF 5 20.177 (22.26) < eSF 5 20.101 (23.39) eUG 5 0.007 (0.128) > eSG 5 20.055 (23.77) eUM 5 20.029 (20.061) < eSM 5 20.016 (21.32) Quantity elasticities of inverse factor demands ejk 5 0 ln wj /0 ln xk eUK 5 0.0199 (2.32) < eSK 5 0.424 (4.84) Time semi-elasticities of inverse factor demands ejt 5 0 ln wj/0t eUt 5 20.01 (22.96) < eSt 5 0.066 (3.41) Note: t-statistics in parentheses. Superscripts “a”, “b” and “c” denote significance at the 1%, 5% and 10% level with a two-tailed test, respectively.

labor. Simultaneous estimation of these equations is performed using an autocorrelation-adjusted nonlinear three-stage-least-squares (AN3SLS) method that accounts for the likely endogeneity of import prices.5 Table 8.1 reports the estimated average elasticities that were calculated using equations (8.4) and (8.5). As may be noted, all inverse labor demand elasticities are small and, in all but one case, negative signifying a net complementary relationship between imports and domestic labor. Perhaps more curiously, the relative magnitudes of all but one set of these elasticities suggest that an increase in imports decreases relative wage inequality. In fact, 75 percent of all US imports (all but consumer goods: category G) decrease rather than increase wage inequality. This is a relatively novel and perhaps unintuitive result, yet it appears quite robust with respect to alternative definitions of skill, set of instruments and so on. Contrary to the role of the majority of imports, capital accumulation and technical change are found to drive the observed trend in wage inequality The results are consistent with the findings of Appelbaum and Kohli (1997) as well as Tombazos (1998, 1999b, 2007) who account for the potentially positive downstream effects of non-competing intermediate imports as well as the effects of output substitution. However, they contradict the findings of studies that concentrate on the latter effect, either in the context of the HO mechanism, such as Wood (1998), or through a process of outsourcing, such as Feenstra and Hanson (1999). Hence, contrary to the results of mainstream studies in this area, trade liberalization schemes which decrease the relative price of imports, such as uniform tariff

Trade and wage inequality

179

reductions, are found to compress, rather than augment, the prevailing wage inequality.

CONCLUDING REMARKS The recent findings of Davis and Weinstein (2001) and Schott (2003) suggest that cross-country specialization in the production of different goods is significant. Yet specialization has not received much attention in the literature on trade and wage inequality. This is despite the fact that the prevalence of extensive specialization can have profound implications for the study of trade and wages. An increase in non-competing intermediate imports has the potential to stimulate demand for all domestic labor with uncertain effects on wage inequality. Furthermore, to the extent that increases in imports facilitate further fragmentation of production and lead to new structures of specialization, they can set in motion substitution between skilled and unskilled labor that can have significant, and not easily predictable, effects on relative wages. To investigate the potential role of imports in stimulating domestic demand for labor we used an economy-wide production theory approach that relies on an SNQ flexible functional representation of the US economy. Unlike similar studies that utilize the translog, the SNQ retains its flexibility in the estimation process and can therefore correctly measure potential complementarities between imports and domestic labor. Deviating from other studies in this area, our results suggest that the preponderance of imports stimulate the relative wage of unskilled labor. Interest in the finding that imports can decrease wage inequality does not derive from the size of the effect. Instead, it derives from the theoretical plausibility that, in the context of specialized production and trade in intermediate goods, the impact of imports on the relative wages of unskilled labor may be considerably smaller than the –20 percent consensus that has emerged in the literature (for example, Zhu and Trefler, 2005). In fact, it may well even be positive.

NOTES * I thank Noel Gaston for useful comments on earlier drafts. 1. Surveys include Richardson (1995), Burtless (1995), Slaughter (2000), and Feenstra and Hanson (2003). See also Sachs and Shatz (1994). 2. According to Rousslang and To (1993, p. 214), domestic value-added increases the final price of US imports by a greater margin than the combined effect of import tariffs and

180

Globalization and economic integration

international transportation costs in the case of about half of all import categories, and for the overall average of all sectors examined by these authors. 3. Note that variable inputs, such as imports, are treated as negative outputs in the model. 4. These include 14 categories of private consumption, nine categories of private investment, six categories of consumption by the state and federal governments, six categories of investment by the state and federal governments, three categories of exports (including services), and the changes in durable and nondurable business inventories. 5. The instrumental variables employed are: excise taxes, sales taxes, and personal savings as percentages of personal disposable income; the budget deficit, net foreign investment, and the government wage bill as percentages of GDP; the discount rate; the producer price indices of Canada, Japan, the UK and Germany; the population of the US, Canada, Japan, Germany, and the UK; the time trend and the time trend squared; and a constant.

REFERENCES Appelbaum, E. and U. Kohli (1997), “Import price uncertainty and the distribution of income”, Review of Economics and Statistics, 79, 620–30. Aw, B.Y. and M.J. Roberts (1985), “The role of imports for the newly industrializing countries in US production”, Review of Economics and Statistics, 67 (1), 108–17. Burgess, D.F. (1974), “Production theory and the derived demand for imports”, Journal of International Economics, 4, 103–18. Burgess, D.F. (1976), “Tariffs and income distribution: some empirical evidence for the United States”, Journal of Political Economy, 84, 17–43. Burtless, G. (1995), “International trade and the rise in earnings inequality”, Journal of Economic Literature, 33, 800–816. Davis, D.R. and D.E. Weinstein (2001), “An account of global factor trade”, American Economic Review, 91, 1423–53. Diewert, W.E. (1974), “Applications of duality theory,” in M.D. Intriligator and D.A. Kendrick (eds), Frontiers of Quantitative Economics, Amsterdam: NorthHolland, pp. 106–206. Falk, M. and B.M. Koebel (2002), “Outsourcing, imports and labor demand”, Scandinavian Journal of Economics, 103, 567–86. Feenstra, R.C. and G.H. Hanson (1996a), “Foreign investment, outsourcing and relative wages,” in R.C. Feenstra, G.M. Grossman and D.A. Irwin (eds), The Political Economy of Trade Policy: Papers in Honor of Jagdish Bhagwati, Cambridge, MA and London: MIT Press, pp. 89–127. Feenstra, R.C. and G.H. Hanson (1996b), “Globalisation, outsourcing and wage inequality”, American Economic Review, 86, 240–45. Feenstra, R.C. and G.H. Hanson (1999), “The impact of outsourcing and hightechnology capital on wages: estimates for the United States, 1979–1990”, Quarterly Journal of Economics, 114, 907–40. Feenstra, R.C. and G.H. Hanson (2003), “Global production sharing and rising inequality: a survey of trade and wages”, in E.K. Choi and J. Harrigan (eds), Handbook of International Trade, Oxford: Blackwell, pp. 146–85. Hijzen, A., H. Görg and R.C. Hine (2005), “International outsourcing and the skill structure of labor demand in the United Kingdom”, Economic Journal, 116, 860–78.

Trade and wage inequality

181

Kohli, U.R. (1978), “A gross national product function and the derived demand for imports and supply of exports”, Canadian Journal of Economics, 11, 167–82. Kohli, U.R. (1991), Technology, Duality, and Foreign Trade: The GNP Function Approach to Modeling Imports and Exports, Ann Arbor, MI: University of Michigan Press. Kohli, U.R. (1993), “A symmetric normalized quadratic GNP function and the US demand for imports and supply of exports”, International Economic Review, 34, 243–55. Krugman, P. (1995), “Technology, trade and factor prices”, NBER Working Paper 5355, National Bureau of Economic Research, Cambridge, MA. Richardson, J.D. (1995), “Income inequality and trade: how to think, what to conclude”, Journal of Economic Perspectives, 9, 33–55. Rousslang, D.J. and T. To (1993), “Domestic trade and transportation costs as barriers to international trade”, Canadian Journal of Economics, 26, 208–21. Sachs, J.D. and H.J. Shatz (1994), “Trade and jobs in US manufacturing”, Brookings Papers on Economic Activity, 1994 (1), 1–69. Schott, P.K. (2003), “One size fits all? Heckscher–Ohlin specialization in global production”, American Economic Review, 93, 686–708. Slaughter, M.J. (1998), “What are the results of product price studies and what can we learn from their differences”, NBER Working Paper 6591, Cambridge, MA. Slaughter, M.J. (2000), “What are the results of product-price studies and what can we learn from their differences?”, in R.C. Feenstra (ed.), The Impact of International Trade on Wages, Chicago, IL: University of Chicago Press, pp. 129–65. Tombazos, C.G. (1998), “US production technology and the effects of imports on the demand for primary factors”, Review of Economics and Statistics, 80, 480–83. Tombazos, C.G. (1999a), “The role of imports in expanding the demand gap between skilled and unskilled labor in the US”, Applied Economics, 31, 509–17. Tombazos, C.G. (1999b), “The impact of imports on the demand for labor in Australia”, Economics Letters, 62, 351–56. Tombazos, C.G. (2003), “A production theory approach to the imports and wage inequality nexus”, Economic Inquiry, 41, 42–61. Tombazos, C.G. (2007), “Specialisation, the intermediate nature of traded commodities, and the myth of import driven wage inequality in the United States”, Pacific Economic Review, 12 (1), 117–28. United States Department of Commerce, Bureau of Economic Analysis (various annual issues), The National Income and Product Accounts of the United States – Statistical Tables, Washington, DC: US Government Printing Office. United States Department of Commerce, Bureau of Economic Analysis (various monthly issues), Survey of Current Business, Washington, DC: US Government Printing Office. United States Department of Commerce, Bureau of the Census (1973), Census of Population 1970, Subject Reports: Occupations by Industry, Washington, DC: US Government Printing Office. Wood, A. (1998), “Globalisation and the rise in labor market inequalities”, Economic Journal, 108, 1463–82. Zhu, S.C. and D. Trefler (2005), “Trade and inequality in developing countries: a general equilibrium analysis”, Journal of International Economics, 65 (1), 21–48.

PART IV

Globalization, immigration and labor market issues

9.

Globalization and labor markets: developments in the Asia-Pacific Ian Harper

INTRODUCTION The Asia-Pacific region is very diverse. Countries are at different stages of development; their social features are distinct; and economic regulations vary widely. Each country has a different story to tell concerning its experience of globalization. History tells us that globalization is not a new phenomenon. People (as well as trade, finance and technology) have naturally flowed between locations. However, the size, pace and nature of these flows have varied over time, as have their effects on the labor market. This chapter takes a medium-term perspective on labor market developments in the Asia-Pacific region, through a period in which these forces of globalization have continued to change and, arguably, intensify. The next section surveys how labor markets in a selection of countries have performed over the past 20 years. This is followed by a review of the effects of globalization on low-skilled workers. The fourth section discusses the role of minimum wages in the context of globalization. The final section concludes.

LABOR MARKET PERFORMANCE IN THE ASIAPACIFIC The Asia-Pacific is one of the world’s most interesting regions for research on labor market issues: ● ● ● ● ●

the region includes countries at very different stages of development; some countries are changing at a faster pace than others; the composition of economic activity varies widely across countries; there is a high rate of labor mobility; and the region’s engagement with the rest of the world is increasing. 185

186 300

Globalization and economic integration Percentage change GDP per capita Employment

250 200 150 100 50

tn am ie V

sia

pp in es Ph ili

M

al

ay

a K or e

sia In do ne

In di a

ji Fi

hi na C

A us tr al ia

0

Note: Change in GDP per capita is the change between 1989 and 2004 except for Fiji, India, Malaysia, and Vietnam, for whom the change is between 1989 and 2003. Change in employment is between 1989 and 2006 except for India, for whom the change is between 1993 and 2005. Employment for Australia is the average over each year. Sources: ADB (2007); Australian Bureau of Statistics (ABS), Labour Force Spreadsheets, Catalogue No. 6202.0.55.001; Center for International Comparisons at the University of Pennsylvania (CICUP), Penn World Tables, .

Figure 9.1

Growth in real GDP per capita and employment in selected countries since 1989

While the availability of reliable and comparable data on labor markets in developing countries is limited, a broad indication of the trends in labor markets can be discerned. The level of economic activity is a major factor determining the level of employment in a country. Over the past two decades, real gross domestic product (GDP) per capita increased significantly in most of the selected countries in the Asia-Pacific region (Figure 9.1). Average economic growth slowed a little following the regional financial crisis in the late 1990s, compared with average growth over the previous 10 years. The strength of growth varies between countries. This reflects many factors, including the stage of economic development already attained by a country and each country’s ability to harness its economic potential. The Asia-Pacific economies have been growing at a generally faster rate than other parts of the world, reflecting their relative stage of development. Between 1990 and 2002, GDP per capita in the region spanning East Asia, South Asia and the Pacific increased by 3 to 5 percent per annum

Globalization and labor markets

187

on average, whereas GDP per capita declined by 1 percent on average in Europe and Central Asia (UNDP, 2006, p. 29). Certainly, there are cases where economic progress has been limited. For example, in isolated parts of the region such as some of the Pacific island countries (ILO, 2008, p. 29). However, geographic isolation has not always been a significant barrier to further development, with countries such as Fiji demonstrating good economic growth over the medium term. The extent to which increased economic activity has translated into increased employment is also illustrated in Figure 9.1. The “employment dividend”, as measured by the growth in employment relative to GDP per capita, appears to have been relatively low in China over the past 20 years, and relatively high in Malaysia. While most countries have seen a significant increase in employment, there has been no clear trend in unemployment. Some countries have seen a reduction in unemployment over the past 20 years, while in others unemployment has increased dramatically. For example, in Indonesia, the unemployment rate increased from 2 percent in 1991 to 11 percent in 2005. In contrast, the unemployment rate in Malaysia decreased from 4 percent to 3.3 percent over the same period (Figure 9.2). For the broader region between 1993 and 2003, the Asian Development Bank (ADB) has found that unemployment rates in Asia generally increased or remained steady, with the unemployment rate in South-East Asia increasing by 2.4 percentage points (UNDP, 2006, p. 43). While employment may have increased significantly in some countries, in others, including Indonesia, the increase has not been sufficient to match growth in the labor force (Figure 9.3), resulting in higher unemployment. There can be many explanations for these trends but we do not explore the issues further in this chapter. However, these examples do illustrate the diversity of labor market experience in individual countries over the period. The data also indicate that very large shifts are occurring in labor markets, both in demand for and supply of labor, making the level of unemployment difficult to predict. A further feature of Asia-Pacific labor markets has been the changing composition of employment (Figure 9.4). Since 1990, in most of the selected countries: ● ● ●

the share of employment in agriculture has decreased; the share of employment in secondary industries has been steady; and the share of employment in services has increased.

There may be a number of so-called “push” and “pull” factors relating to these changes in employment patterns. For example, the trend towards

188

Globalization and economic integration

Percent 12 1991 1998 2005

10 8 6 4 2

am

es

tn ie V

pp ili Ph

al

ay

in

sia

or ea M

do In

K

ne sia

a di

ji Fi

In

C

hi

na

A us tr al ia

0

Note: The years shown for India are 1994, 2000, and 2005. The unemployment rate for Australia is the average over the year. Beginning April 2005, for China and the Philippines, the unemployed definition was revised to include all persons who are 15 years old and over and are reported as: (1) without work; (2) currently available for work; and (3) seeking work or not seeking work due to valid reasons. Sources:

ADB (2007); ABS, Labour Force Spreadsheets, Catalogue No. 6202.0.55.001.

Figure 9.2

Unemployment rates in selected countries since 1991

less employment in the agricultural sector could be due to displacement reflecting higher labor productivity levels. It might also reflect the greater availability of service sector jobs in urban areas reflecting higher average incomes and higher average skill levels. Employment patterns have also varied by skill level. In most countries, employment has strongly favored people with medium- and high-level skills. However, growth in employment for people with low levels of skill has generally continued at moderate though still significant rates (Figure 9.5). This may reflect, in part, an improvement in the general level of educational attainment in all countries, with higher literacy rates and higher secondary school enrolments (ADB, 2007). Whatever the explanation, the potential to raise average living standards through employment growth at all skill levels has improved, as a result of greater levels of economic activity. There are only limited data on longer-term trends in average wages in

Globalization and labor markets

189

Percentage change 70 Employment Labor force

60 50 40 30 20 10

tn am ie V

pp in es ili Ph

a

al ay sia M

or e K

ia do ne s In

In di a

ji Fi

C hi na

A us tr al ia

0

Note: Change in employment is between 1989 and 2006 except for India where the change is between 1993 and 2005. Change in the labor force is between 1989 and 2006 except for India (1994 to 2005) and Vietnam (1998 to 2005). Figures for Australia are the average over the year. Sources:

ADB (2007); ABS, Labour Force Spreadsheets, Catalogue No. 6202.0.55.001.

Figure 9.3

Growth in employment and the labor force in selected countries since 1989

many of the developing Asia-Pacific countries. However, some data are available for a small selection of countries at different stages of economic development. Average real wages have increased over the past decade in each of the countries shown (Australia, China and Korea), with a particularly large increase in China (Figure 9.6). Research by the International Labour Organization (ILO) and the World Bank suggest that the incidence of poverty has decreased in most parts of the Asia-Pacific (ILO, 2008, pp. 26, 29, 31). For example, the World Bank has estimated that the number of people in the Asia-Pacific region who live on US$1 a day or less has declined since the early 1980s. However, the incidence of poverty remains significant, with an estimated 530 million people in the Asia-Pacific region living on US$1 a day or less (Chen and Ravallion, 2008, p. 31). The evidence on the dispersion of income across the population is mixed. Over the past 10 years, the ratio of income for the highest 20 percent of the population compared with the lowest 20 percent of the population

190

Globalization and economic integration Agriculture

Industry

Services

100%

80%

60%

40%

20%

0% 1990 2003 1990 1999 1990 2005 1990 2006 1990 2006 1990 2006 China

Fiji

India

Korea

Malaysia

Vietnam

Note: Indicates the percentage of total employment in each country by industry, 1990 and 2006, except for China (1990 and 2003), Fiji (1990 and 1999) and India (1990 and 2005). Source:

ADB (2007).

Figure 9.4

Shares of employment by economic sector in selected countries, 1990 compared with 2006

has increased in some countries, while in others it has declined, with most seeing little change in wage relativities (see ADB, 2007). In summary, the Asia-Pacific economies have expanded significantly over the past 20 years, with an associated increase in employment. While unemployment has declined in some countries, it remains high in others. At the same time, there has been considerable change in the industrial composition of activity and of employment. While recorded levels of employment are subject to a high degree of uncertainty, one may nevertheless conclude that labor markets in most countries in the Asia-Pacific have adjusted flexibly to economic developments over the past 20 years.

THE EFFECTS OF GLOBALIZATION ON LOWSKILLED WORKERS The effects of globalization on labor markets can be highly varied. For the lowest-paid workers, these effects depend on factors such as their location,

Globalization and labor markets

600 500 400

191

Percentage change Total Low Medium High

300 200 100

am

s V

ie

tn

in e pp ili Ph

al ay sia M

ea K or

sia In

do

ne

di a In

C

A us tr al ia

hi na

–100

Fi ji

0

Note: The percentage changes in the high-skilled labor force in Indonesia and Malaysia were nearly 1000 percent. Source: World Bank, Brain Drain database (Docquier 1975–2000), available at: http:// econ.worldbank.org/research.

Figure 9.5

Growth in the labor force in selected countries by skill level, between 1980 and 2000

skill level, mobility, sector of employment, gender and the local economy’s stage of development. This section surveys the relevance of some of these factors for the effects of globalization on low-paid workers in the AsiaPacific. Employment Globalization may be expected to have a different effect on countries depending upon their level of economic development. Less open and developed economies are more likely to experience major changes in their labor markets as a result of globalization, both in terms of risks and in terms of opportunities. The Organization for Economic Cooperation and Development (OECD) found that in the more highly developed countries, employment and unemployment rates have generally improved over the medium term and the rate of change in sectoral employment patterns has been fairly stable

192

Globalization and economic integration

Percentage change 250

200

150

100

50

0 Australia Note:

China

Korea

Data for Australia and South Korea are for non-farm wages.

Source: Australian Fair Pay Commission Secretariat calculations based on International Labour Organization (ILO) data, available at: http://laborsta.ilo.org/.

Figure 9.6

Growth in real average wages, 1996 to 2006

(see OECD, 2007a). By contrast, in the lesser-developed countries of the Asia-Pacific, there has been a significant increase in the level of employment and major change in the composition of employment towards the services sector, as data in the preceding section illustrated. This is consistent with Williamson’s conclusion that over the longer term ‘history offers an unambiguous positive correlation between globalization and convergence’ (of poor countries catching up with wealthier countries) (Aghion and Williamson, 1998, p. 170). However, there are ebbs and flows in economic history, and these can be reflected in employment patterns over time. In South Korea, for example, unemployment increased significantly over the course of the 1990s, before returning to earlier low levels of unemployment by 2005 (Figure 9.2). Significant change in employment patterns has also taken place within countries, with an uneven balance of employment growth between urban and rural areas. Part of this shift may be due to the economic development process itself, while another part may be due to globalization. There are high levels of underemployment in rural areas, including

Globalization and labor markets

193

in China and India (see OECD, 2007b). However, the data presented earlier indicate that growth in the demand for labor has strongly favored urban-based industries. As a result, labor supply has needed to adjust to the new patterns of demand, both in terms of the skills required and their location. Most countries have seen a shift in the share of employment from low-skill industries such as agriculture into higher-skill activities. This development has been supported by an increase in average skill levels in most countries, demonstrated in higher levels of educational attainment and literacy rates. This suggests that low-skilled people have been, and are becoming, better placed to take advantage of higher-skilled and betterpaying job opportunities. Accurately measuring employment can be difficult in some countries. One reason for this is the significance of the informal sector, in which regulation is generally absent and there are no formal employment arrangements. Many countries have a dual labor market, with employment in the informal sector, including self-employment, acting as a substitute for employment in the formal sector, with some interaction between the two. Rapid growth in one of the formal or informal sectors attracts labor from the other sector. As a result, employment may fall in one sector but remain steady if a broader measure of employment is used. For example, greater foreign investment in the formal sector of the economy can induce workers out of the informal sector into more secure jobs. Equally, less foreign investment may result in workers being displaced from the formal sector and returning to the informal sector (see Ghose et al., 2008). In many developing countries, most of the jobs can be in the informal sector. The ILO estimates that six to eight out of every 10 workers in Asia in 2007 were in the informal sector (see ADB, 2008a). Some researchers have estimated that the informal sector accounts for around 85 percent of all employment in India (see OECD, 2007b). In these environments, unemployment may be disguised in a high incidence of underemployment, and heightened levels of self-employment and casual employment (see Ghose et al., 2008). Average wages also tend to be lower in the informal sector than in the formal sector (see Carr and Alter Chen, 2002). Recent estimates suggest that the proportion of workers considered to be in vulnerable forms of employment has declined over the past decade (ILO, 2008, pp. 25, 28, 31). A further aspect of the labor market effects of globalization relates to increasing trade exposure. Over the past 15 years, total trade as a proportion of GDP nearly doubled in China and more than doubled in India (see OECD, 2007c). While offering more job opportunities, this development

194

Globalization and economic integration

can lead to more precarious employment as well as downward pressure on wages for low-skilled workers in both developed and less-developed countries. One effect of opening an economy to greater trade and investment is that production can be off-shored. This shift can affect workers at all skill levels. However, arguably, it may place low-skilled workers at greater risk in some countries compared to others where there is a surplus of low-skilled labor and demand for low-skilled labor may increase. Hence, off-shoring has different implications for lesser and more developed economies, respectively. Following an assessment of the available literature, Coe (2007) found several major labor market effects of off-shoring for developed countries: ●

●

● ●

they were disproportionately felt by low-skilled workers, although some low-skilled service sector jobs continue to be demanded by the local population; the impact on relative wages depends on a variety of structural features of economies, including factor endowments and factor intensities of production; a likely increase in uncertainty about job security resulting in more elastic labor demand; and aggregate employment and unemployment in the long run is determined by macroeconomic policies and structural aspects of labor markets.

Wages Globalization can also have different effects on the wages of the low skilled in the short term compared with the longer term. Recent wage growth has been found to favor higher-skilled occupations, with less change in the wages of the unskilled (see Corley et al., 2005). While some studies have found an initial negative effect from globalization on wages for low-skilled workers, this effect appears to dissipate over time and eventually turns positive. Data on wages in the manufacturing sector suggest that real wages in less-developed economies, particularly in Asia, have been catching up with those in the United States (IMF, 2007, pp. 169, 172). More generally, researchers have found evidence of wage growth across all occupations and particularly in those countries with a high trade exposure (see Corley et al., 2005). There were clear differences in wage growth across occupational groupings, with high-skilled occupations experiencing stronger wage increases in the 1990s than low-skilled occupations. Part of the reason may be the surplus of labor in developing economies, whereby

Globalization and labor markets

195

the initial impact of globalization and growth more generally may be to bring previously underemployed or unemployed people into the formal labor market, with no direct effect on wages or employment. Openness to trade has been found to contribute to an increase in occupational wage inequality within developed countries, although that effect diminishes with an increased level of development as general skill levels improve. However, in developing countries, the effect of openness to trade on wage inequality has been found to be insignificant and does not vary with the level of development (see Munshi, 2008). Labor Mobility A further defining feature of Asia-Pacific labor markets is the high rate of growth in the labor force, with greater rates of natural increase and net migration for some countries. With labor force growth being so high in some cases, employment growth may not have kept pace, resulting in increased unemployment and underemployment in some countries. International labor mobility has been an important feature of the economic development of the Asia-Pacific region, with mobility patterns changing in response to the relative availability of employment opportunities in different countries. Until the late 1980s, most of the migration flows were between Asia and countries outside the region. However, since the late 1980s, the rate of migration within the region has increased (see Ahmad, 2003), with migration from those countries with surplus labor, particularly of an unskilled nature (see Athukorala, 2006). Increased trade between countries in the Asia-Pacific has also seen increased flows of labor between countries, particularly temporary flows of labor. Globalization and the ability to migrate is a positive development for many people, as it creates more job opportunities for those who are unemployed. Most emigration is concentrated among young people with reasonable levels of education and some skills, but facing poor domestic job prospects (see Jones, 2006). Some researchers consider that the higher-performing economies are increasingly drawing on migrants to meet needs along the whole continuum of skills (see ADB, 2008b). Skill shortages in some of the moredeveloped economies and the higher wages that are offered raise the incentive for people living in countries with surplus labor to migrate. The share of migrant labor in the total labor force varies widely between countries in the Asia-Pacific. For example, the proportion of migrants in the labor force amounts to around 30 percent in Singapore and 12 percent in Malaysia (Jones, 2006, p. 4). Migrants on short-term labor contracts

196

Globalization and economic integration

dominate labor flows within the Asian region, accounting for over 90 percent of the total (see Athukorala, 2006). As a result, there is a considerable degree of turnover in the labor force in some countries. The temporary nature of migration has a positive aspect for many countries. Migration can help to address some of the economic challenges facing isolated nations in the short term, while potentially enabling the home country to benefit from the income and skills in the medium term. For some traditionally labor-sending countries, globalization has reduced emigration rates as a result of greater job opportunities being available in the home country (see ADB, 2008c). Summary The effect of globalization on employment opportunities has varied with factors such as location within the Asia-Pacific region and, within country, skill levels and sectors of the economy. Most countries in the Asia-Pacific have experienced an improvement in living standards, particularly those countries that are less isolated from others. The past 20 years has seen an improvement in employment and wage outcomes, and a reduction in both absolute and relative poverty. Over a period in which indicators of globalization have increased, GDP per capita and employment have increased. However, in some countries, unemployment remains a significant concern. The increase in the demand for skilled labor in the Asia-Pacific has benefited many people, with many who are unskilled also benefiting from increased job opportunities and greater opportunities to increase their income. Globalization has resulted in a higher rate of structural change in both economic activity and labor markets in many developing countries. The labor markets in some countries have adjusted more effectively to those developments, with consequent reductions in unemployment. A more open labor market presents both risks and opportunities for countries: on the one hand, employment may become more sensitive to changes in economic variables; while, on the other hand, it opens more economic opportunities for domestic as well as migrant labor.

THE RELEVANCE OF MINIMUM WAGES TO THE ASIA-PACIFIC Globalization is a multidimensional process, with many interdependent parts. This complicates the analysis of globalization and associated policy implications. Particularly in the Asia-Pacific, where there is wide diversity

Globalization and labor markets

197

in the stages of development of individual countries, examining the effects of globalization is challenging, as is analyzing the potential role for labor market regulation. Minimum Wage Arrangements in the Asia-Pacific ILO records show that minimum wage legislation exists in more than 90 percent of countries throughout the world (ILO, 2006, p. 1). Minimum wage arrangements vary considerably across countries, and especially between developing and developed countries. The starting point for an assessment of the role of minimum wages is to compare their levels. Table 9.1 compares the monthly minimum wage in selected countries to GDP per capita in purchasing power parity (PPP) terms. The table illustrates that: ●

minimum wages in many developing countries in the Asia-Pacific are very low in absolute terms compared with those in many of the

Table 9.1

Standardized minimum wages and GDP per capita (PPP dollars 2002–04)

Country Australia China Fiji India Indonesia Japan Korea, Republic of Malaysia Papua New Guinea Philippines Thailand Vietnam Note:

Monthly minimum wage

GDP per capita per month

Ratio

1475 228 302 177 184 816 519 103 122 322 260 153

2428 416 460 242 280 1347 1492 808 209 360 632 207

0.61 0.55 0.66 0.73 0.66 0.35 0.35 0.13 0.59 0.90 0.41 0.74

Minimum wage in PPP dollars.

Sources: ILO database of Work and Employment Laws and United States Department of Labor. Monthly minimum wage rates have been standardized on the basis of 40 hours of work a week, using information on legal maximum hours of work. GDP per capita at purchaser price in 2003 were taken from the World Development Indicators (World Bank). In the third column, ‘ratio’ is the ratio of one of the standardized monthly wage to GDP per capita per month. Reproduced from Saget (2008).

198

●

Globalization and economic integration

more-developed countries, from very low rates (where the ratio of minimum wage to GDP per capita is below 30 percent) to very high rates (with a ratio above 60 percent); and the ratio of a minimum wage to average wages in a given country can be high or low, regardless of their level of economic development.

As a result, there is a need to consider what effects, if any, differences in minimum wages between countries may have on the mobility of labor between countries. Another matter to consider is the effect of the particular minimum wage arrangements within a given country in relation to its particular economic and labor market circumstances. A key aspect of this analysis is the coverage of minimum wages. In most countries this is incomplete, resulting in a complex set of arrangements within, as well as between, countries. Table 9.2 provides a survey of minimum wage arrangements in AsiaPacific countries, particularly the bodies that determine minimum wages and also which elements of the labor market are covered by minimum wages. Generally, minimum wages are determined by government agencies or tripartite wage councils. In most developing countries, minimum wages have been decentralized, while there are national rates in others. In India, for example, minimum wages vary by industry, occupation and region (Sugiyarto and Endriga, 2008, p. 3). The coverage of minimum wages is not uniform in any country, with variations by region, industry, occupation and citizenship status. The level of the minimum wage combined with its coverage determines its overall effectiveness. Saget (2008) identified two distinct groups of developing countries based on their minimum wage arrangements: one group with very low minimum wages and the other with very high minimum wages. She concluded that some developing countries have set their minimum wage levels too high or too low “to constitute a meaningful constraint on employers” (p. 25). On the other hand, Saget found that high minimum wages, such as those in Indonesia, the Philippines and Thailand, were often a result of the minimum wage being fixed at a level closer to the actual wage, acting as a substitute for wage bargaining at the local level. Countries with very low minimum wages may not provide an adequate “floor” in the labor market; while very high minimum wages can result in non-compliance. Saget gives the example of Indonesia, where an estimated 30 percent of all full-time workers earn below the minimum wage.

Globalization and labor markets

Table 9.2

199

Institutional arrangements for minimum wages in different countries, 2008

Country

Minimum wage decision body

Coverage

China

No national minimum wage, but the labor law requires local governments to set their own minimum wage according to standards promulgated by the Ministry of Labor and Social Security. Wage rates are set by provinces, municipalities and regions Wages councils provide regulations for Wage Regulation Orders. A council consists of independent members as well as workers’ and employers’ representatives Minimum wages set by state and regional governments

Different minimum wage rates apply in different regions

Fiji

India

Indonesia

Japan

Wages set by district and provincial tripartite wage councils comprising workers, employers, and government representatives. Local districts set district minimum wages using the provincial levels as references. Districts also set minimum wages in some industrial sectors on an ad hoc basis The minimum wage is set and frequently adjusted by the Ministry of Health, Labor,

No single, minimum wage; industry-based minimum wages prescribed as Wage Regulation Orders – currently in place for 10 industries

Currently 1,230 occupational and sectoral minimum wage rates; some industries such as the apparel and footwear industries do not have a minimum wage; state governments set a separate minimum wage for agricultural workers Minimum wages vary across authorities based on proposals from different regions

Minimum wage rates vary according to prefecture

200

Table 9.2

Globalization and economic integration

(continued)

Country

Minimum wage decision body

Korea, Republic of

and Welfare in consultation with minimum wage councils. Councils comprise delegations from workers, employers and the public Determined annually by the Ministry of Labor

Malaysia

Papua New Guinea

Philippines

Thailand

Wage councils composed of labor, management and government representatives, provide a recommended minimum wage for sectors where the market rate is deemed to be insufficient Minimum Wage Board, a quasi-governmental body with labor and employer representatives sets minimum wages for the private sector Minimum wages set by tripartite regional wage boards, comprising labor, employer, and government representatives

Minimum wages are set by provincial wage committees that sometimes include only employer representatives

Coverage

Single minimum wage rate. Domestic workers are not entitled to minimum wages. Some workers may be exempted from the minimum wage upon permission from the Ministry of Labor No national minimum wage in effect; in practice, wage council decisions have little effect in any sector

Minimum wage set for private sector employees

The regional wage board orders cover all private sector workers except domestic workers and other employees in the service of another person. Certain employers may be exempted from minimum wage laws because of factors such as business size, industry sector, export intensity, financial distress and level of capitalization Minimum wages vary between different provinces

Globalization and labor markets

Table 9.2

201

(continued)

Country

Minimum wage decision body

Coverage

Vietnam

Government required by law to set minimum wage

Single minimum wage rate. Government may temporarily exempt certain joint ventures from paying the minimum wage

Sources: US Bureau of Democracy, Human Rights, and Labor, 2007 Country Reports on Human Rights Practices, available at: http://www.state.gov/g/drl/rls/hrrpt/2007; ILO, Working Conditions Laws: a global review 2006–2007, 2008, p. 3.; ILO, Minimum wages policy, Information Sheet No. W-1, 2006, available at: http://www.ilo.org/public/english/ protection/condtrav/pdf/infosheets/w-1.pdf; ILO, National Labour Law Profiles, 2007, available at: http://www.ilo.org/public/english/dialogue/ifpdial/info/national/index.htm.

Policy Issues A number of studies have investigated whether the introduction or increase of minimum wages in developing countries can mitigate poverty. Opinions on the effectiveness of minimum wages vary widely. Most studies conclude that a minimum wage has a negative effect on employment (see Neumark and Wascher, 2006). However, such studies generally focus on minimum wages in developed countries, perhaps in part because the availability of quality data in many developing nations is scant. Minimum wages may not be applicable to the majority of the labor force in developing countries. In many developing countries, this group includes those working in the informal, unregulated sector and who are generally the lowest paid. Thus, for many workers, minimum wages are unlikely to have a significant effect on their wages. Countries vary in the alternative forms of social protection they offer to mitigate social disadvantage. In some countries, where a governmentfunded income safety net is not well developed, the relevance of minimum wages may be greater. A key question is whether the implementation or increase of a minimum wage will adversely affect the employment prospects of low-skilled or lowwage workers. When looking at disemployment effects in developing countries, Saget found that the minimum wage has an insignificant effect on employment levels but a significant ameliorating effect on poverty levels. This suggests that minimum wages can be an effective tool for poverty reduction in developing countries.

202

Globalization and economic integration

Labor markets in many countries in the Asia-Pacific are experiencing rapid economic development and are moving quickly through the traditional stages of development. There remains a significant level of underemployment in many countries. Therefore on the one hand raising minimum wages or improving the coverage of minimum wages in such circumstances carries risks, but on the other may have little if any effect. Minimum wages may be a useful tool for providing a safety-net income level and help to limit exploitation of labor by some employers. However, depending on the compliance arrangements that exist, the effectiveness of raising a minimum wage to achieve these outcomes may be minimal. In many countries, the minimum wage is limited to those in some elements of the formal sector or to other specific elements of the labor market. As labor markets in the Asia-Pacific become more integrated, the ability to sustain partial minimum wage structures will be challenged.

CONCLUSION Globalization involves changes in the economic potential of a country, its industrial structure, relative prices and the composition of consumption. As a result, different countries and different people within those countries are affected in different ways, according to their labor force status, income and where they live. As the ILO has indicated, ‘invariably some have been adversely affected while others have gained from this often intense process of change’ (World Commission on the Social Dimension of Globalization, 2004, p. 45). Overall, it is difficult to predict with certainty whether domestic minimum wage legislation can play an effective role in mitigating some of the negative effects of globalization. It appears that, in the right circumstances, minimum wages do have a role in reducing poverty, but this depends very much on the specific country context. Returning to the macroeconomic scene, globalization has expanded the availability of job opportunities, to local as well as migrant labor. Employment offers the most important means for people to emerge from poverty. While wage rates may not have improved significantly for some lowskilled workers in the short term, over the longer term, globalization offers significant economic benefits for each country. It is an important role for governments to implement policies that spread the benefits of globalization more evenly within their communities and enhance the ability of people to adapt to the changes that globalization brings.

Globalization and labor markets

203

REFERENCES Aghion, P. and J. Williamson (1998), Growth, Inequality and Globalization, Cambridge: Cambridge University Press. Ahmad, S. (2003), “The constant flux, the mobile reserve and the limits of control: Malaysia and the legal dimensions of international migration”, in R. Iredale, C. Hawksley and S. Castles (eds), Migration in the Asia Pacific: Population, Settlement and Citizenship Issues, Cheltenham, UK and Northampton, MA, USA: Edward Elgar, pp. 141–58. Asian Development Bank (ADB) (2007), Key Indicators 2007: Inequality in Asia, 2007, available at: http://www.adb.org/documents/books/key_indicators/2007/ default.asp. Asian Development Bank (ADB) (2008a), “Structure and quality of employment”, Asian Development Outlook 2008, available at: http://www.adb.org/Documents/ Books/ADO/2008/part020104.asp. Asian Development Bank (ADB) (2008b), “Migration trends and directions”, Asian Development Outlook 2008, available at: http://www.adb.org/documents/ books/ado/2008/part020302.asp. Asian Development Bank (ADB) (2008c), “Asia’s labor migration dynamics”, Asian Development Outlook 2008, available at: http://www.adb.org/documents/ books/ado/2008/part020303.asp. Athukorala, P. (2006), “International labor migration in East Asia: trends, patterns and policy issues”, Asian-Pacific Economic Literature, 20 (1), pp.18–39. Carr, M. and M. Alter Chen (2002), ‘Globalization and the informal economy: how trade and investment impact on the working poor’, International Labour Organization, available at: http://www.ilo.org/public/english/employment/ infeco/download/wp1.pdf. Chen, S. and M. Ravallion (2008), “The developing world is poorer than we thought, but no less successful in the fight against poverty”, World Bank Policy Research Working Paper 4703, available at: http://wwwwds.worldbank.org/ external/default/WDSContentServer/IW3P/IB/2008/08/26/000158349_2008082 6113239/Rendered/PD F/WPS4703.pdf. Coe, D. (2007), “Globalization and labor markets: policy issues arising from the emergence of China and India”, OECD Social, Employment and Migration Working Papers 63, Organization for Economic Cooperation and Development, available at: http://www.oecd.org/dataoecd/10/44/39608656.pdf. Corley, M., Y. Perardel and K. Popova (2005), “Wage inequality by gender and occupation: a cross-country analysis”, Employment Strategy Papers 2005/20, International Labour Organization, available at: http://www.ilo.org/public/ english/employment/strat/download/esp2005–20.pdf. Ghose, A., N. Majid and C. Ernst (2008), “The Global Employment Challenge”, executive summary, International Labour Organization, available at: http:// www.ilo.org/wcmsp5/groups/public/---dgreports/---dcomm/---publ/documents/ article/wcms_092210.pdf. International Labour Organization (ILO) (2006), “Minimum wages policy”, Information Sheet no.W-1, available at: http://www.ilo.org/public/english/protection/condtrav/pdf/infosheets/w-1.pdf. International Labour Organization (ILO) (2008), Global Employment Trends, January 2008, Geneva: ILO.

204

Globalization and economic integration

International Monetary Fund (IMF) (2007), World Economic Outlook, April 2007, available at: http://www.imf.org/external/pubs/ft/weo/2007/01/index.htm. Jones, G. (2006), “Cross-border labor flows: policy issues, options and implications for Australia–Asia Relations”, Melbourne Asia Policy Papers 8, University of Melbourne, available at: http://www.asialink.unimelb.edu.au/__data/assets/ pdf_file/0006/4200/mapp8.pdf. Munshi, F. (2008), “Globalization and inter-occupational inequality in a panel of countries: 1983–2003”, Working Papers in Economics 302, University Of Gothenburg, available at: http://gupea.ub.gu.se/dspace/bitstream/2077/10025/1/ gunwpe0302.pdf. Neumark, D. and W.L. Wascher (2006), “Minimum wages and employment: a review of evidence from the new minimum wage research”, Working Paper 12663, National Bureau of Economic Research, Cambridge, MA. Organization for Economic Cooperation and Development (OECD) (2007a), OECD Economic Outlook No. 81, Special Chapter: Making the Most of Globalization, available at: http://www.oecd.org/dataoecd/21/32/38628438.pdf. Organization for Economic Cooperation and Development (OECD) (2007b), OECD Employment Outlook – 2007, summary, available at: http://www.oecd. org/dataoecd/28/32/38798341.pdf. Organization for Economic Cooperation and Development (OECD) (2007c), “Globalization, Jobs and Wages”, OECD Policy Brief, available at: http://www. oecd.org/dataoecd/27/1/38796126.pdf. Saget, C. (2008), “Fixing minimum wage levels in developing countries: common failures and remedies”, International Labor Review, 147 (1), 25–42. Sugiyarto, G. and B. Endriga (2008), “Do minimum wages reduce employment and training”, ERD Working Paper 113, Economic Research Department, Asian Development Bank, Manila, available at: http://www.adb.org/Documents/ ERD/Working_Papers/WP113.pdf. United Nations Development Programme (UNDP) (2006), “Asia-Pacific Human Development Report 2006: Trade on Human Terms”, available at: http:// www.undprcc.lk/rdhr2006/G2235H835352H/P24314143234344343242_ PDF_214335/Chap%2002.pdf. World Commission on the Social Dimension of Globalization (2004), “A Fair Globalization: Creating Opportunities for All”, International Labour Organization, available at: http://www.ilo.org/public/english/wcsdg/docs/report. pdf.

10.

Aging and migration in Japan Junichi Goto

INTRODUCTION Due to the low and declining fertility rate, the Japanese population is rapidly aging. As a result, Japan is facing a potentially serious labor shortage in the near future. Many people argue that Japan should admit more migrant workers in order to compensate for the decline in working population due to aging. However, there are various alternatives to international migration, such as the increased labor participation of women and older persons. Moreover, instead of importing workers from abroad, Japan can utilize foreign labor indirectly through imports of labor-intensive products made abroad. In view of the above, this chapter will examine merits of the international movement of labor and various alternatives to it, using a rigorous framework. In the next section, the impact of migration on the host county is analyzed theoretically, using a rigorous, but general, equilibrium model. Although traditional economic theorists are generally in favor of migration, at least as far as the economic effects are concerned, it is demonstrated that such a rosy scenario depends on fairly restrictive assumptions, such as perfect competition, and that when we relax some assumptions of the traditional theory, the admission of migrant workers can create adverse economic effects to the host country. In the third section, the social effects of migration are discussed because migration involves the movement of workers as human beings as a whole, rather than the movement of labor as a factor of production. The fourth section, discusses some empirical issues about the relative benefit of various measures to cope with a possible labor shortage due to Japan’s aging population. Based on a simple simulation result, it will be argued that, at least in aggregate, the decline in the working-age population and the resulting increase in the burden on the working population can to be compensated if various measures, such as the utilization of female labor, are taken into consideration. The final section summarizes all the major findings of the chapter.

205

206

Globalization and economic integration

ECONOMIC IMPACT OF MIGRATION Traditional View of Migration Traditional economic theorists usually consider that the overall effect of international migration is favorable to both the home and host countries because it involves a movement of labor from a labor-abundant (and capital-scarce) country to a labor-scarce (and capital-abundant) country. This increases productivity (and economic welfare) in both countries. For example, when some workers move from Indonesia (a labor-abundant country) to Japan (a labor-scarce country), the Japanese employers who have unfilled vacancies can gain from hiring these workers and the Indonesian workers can earn more than what they could earn in their home country. If these workers remit some part of their income earned in Japan to their home country, people left behind in Indonesia also benefit from the migration of their fellow Indonesians. Of course, there could be some conflicts of interest among various economic agents in each country. For example, an inflow of Indonesian workers may dampen Japanese wages, thereby reducing the income of Japanese workers, while the income of employers in Japan increases. However, the overall effect is argued to be positive in both countries. Thus, the movement of workers (or unemployed persons) from the home to the host country would increase the national income (and economic welfare) in both countries. The economic reasoning behind their argument is summarized in Figure 10.1. In this figure, the horizontal axis plots the amount of labor, where the amount of labor supply in country 1 (home country) and that in country 2 (host country) are measured from O1 and O2, respectively. The vertical axis plots the marginal productivity of labor (MPL), which is equal to the wage rate in the competitive equilibrium. The MPL in country 1 (country 2) is expressed by line NE (by line AT). Suppose that at the initial stage before migration the labor endowment in country 1 (home country) is O1H and that in country 2 (host country) is O2H, and therefore labor supply in the two countries as a whole is O1O2. At this stage, the value of total production (that is, national income) of the sending home country is the area of trapezoid NGHO1, and the value of the national income of the receiving host country is the area AFHO2. In this pre-migration situation, the wage rate in the host country is BO2, which is higher than that in the home country (SO1). Such a wage gap between the two countries constitutes an incentive for the workers in country 1 to migrate to country 2. Now, suppose that some workers in country 1 (the number of workers expressed by HK) migrate to country 2 in search of higher wages. In this

Aging and migration in Japan

207

(MPL,W)

N

A F B I

D

C

L Q J

R

G

S

E

T

O2

H

K

M

L2

Figure 10.1

O1 L1

Economic effect of migration

post-migration situation, the amount of labor in country 2 is increased to O2K, and that in country 1 is decreased to O1K, because the labor HK moves from country 1 to country 2. Now, the value of goods produced in country 2 (gross domestic product, GDP) increases to the area AIKO2. But, the area of rectangular DIKH is paid as wages to the workers from country 1, and the net gain of income of host country’s citizens is equal to the area of triangle FDI. While the GDP in country 1 is decreased to area NJKO1, the national income of the citizens of country 1 (gross national product, GNP), which includes the income earned by the workers who migrated to country 2, is increased to the area NJIDHO1. So the net gain to country 1 is equal to the area DIJG. Therefore, according to traditional economic theory, international migration increases the national income (and economic welfare) of both the sending and receiving countries. As shown in Figure 10.1, when HK of workers move from country 1 to country 2, the economic welfare of

208

Globalization and economic integration

country 1 is increased by the area of triangle FDI and that of country 2 is increased by the area of trapezoid DIJG. Based on such reasoning, traditional economists often argue that international migration gives economic gains to both countries, although both countries may incur some social costs, as discussed below. The above rosy picture of international migration, however, hinges upon various restrictive assumptions, such as perfect competition and constant returns to scale. There are several studies demonstrating that the conclusions of the traditional framework break down when more realistic assumptions are introduced. For example, Schiff (1999) introduced the role of social capital and Goto (1998) analyzed the impact of migration when trade barriers and non-traded goods exist. Both authors cast some doubt on the orthodox conclusions based on Figure 10.1 and point out that migration can (and is likely to) have adverse economic effects for receiving countries. In the next subsection, I summarize Goto’s (1998) argument on the economic impact of migration of the host country: if we incorporate other realities, such as the existence of trade restrictions and non-traded goods in the economy, the simple argument in this subsection collapses. The effect of migration is not as simple as the conventional argument implies. Economic Effects under the New Framework Basic characteristics of the model The formal model developed in Goto (1998) has three main characteristics, which are somewhat different from the traditional theory. The following three features have been added in order to capture more reality for the analysis of the migrant workers in Japan: (i) a possible change in factor prices (for example, a decline in wage rate) by admission of migrant workers, even when the prices of tradable goods are constant; (ii) the existence of non-traded goods; and (iii) the existence of trade barriers. Possible change in factor prices Under the standard Heckscher–Ohlin– Samuelson (HOS) framework, factor prices are completely determined by the price of goods without regard to relative factor endowments. Thus, for a small open economy where the price of goods is given, the factor prices are unchanged by the admission of foreign workers. However, many empirical studies, including Morgan and Gardner (1982), found that the influx of foreign unskilled labor results in a depressing effect on the domestic wage rate. In order to overcome the discrepancies between the theoretical prediction and the empirical findings, a Jones-type specific factors model is used.

Aging and migration in Japan

209

More specifically, capital is assumed to be fixed to each sector while labor is mobile between sectors. Under the specific factors model, the change in relative factor endowments due to an influx of foreign labor plays an important role in determining factor prices. Non-traded goods In the real world, the share of the non-traded goods sector, which includes industries such as construction and services, in the total economy is high. However, less attention has been paid to this by traditional trade theories. Also, many migrant workers in Japan are employed in the non-traded goods sector. While most formal studies of international labor mobility have been based on the two-sector model (that is, exportables and importables), the economy in the model here is assumed to consist of three sectors producing three kinds of goods: exportables, importables, and non-tradables. Combined with the specific factors assumption, the model in this chapter is a 3 × 4 (three goods and four factors) model rather than the traditional 2 × 2 model. As seen below, the inclusion of non-traded goods yields additional insights into the economic effect of migration. Trade barriers While most studies of international factor mobility have assumed free trade, the model below shows that international trade is assumed to be restricted by tariffs and/or non-tariff barriers (NTBs), as is often the case for many other countries. This is an application of the framework developed by Brecher and Diaz-Alejandro (1977), which was used to analyze international capital mobility. As shown in their study, the economic impact of factor inflows under trade barriers is very different from that under free trade. Specification of the model In the model, consumers are characterized by the following Cobb–Douglas social utility function. U = C1aC2bC3g, a + b + g = 1,

(10.1)

where C1, C2, and C3 are the amount of consumption of exportables (good 1), importables (good 2), and non-tradables (good 3), respectively. U is social utility. Consumers maximize (10.1) subject to the budget constraint (10.2). P1C1 + (1 + t)C2 + P3C3 = Y,

(10.2)

where P1 and P3 are the prices of exportables and non-tradables, respectively, and Y is national income. The world price of importables, which

210

Globalization and economic integration

is considered to be the numeraire good, is set to unity; t is the rate of domestic price markup of importables due to tariffs and NTBs. In order to avoid further complications, it is assumed that the world price of tradables is given to the economy (that is, the “small country” assumption). From the maximization problem, the following three demand functions are obtained: C1 = aY/P1

(10.3)

C2 = bY/(1 + t)

(10.4)

C3 = gY/P3.

(10.5)

The producers in the three sectors are characterized by the following Cobb–Douglas production functions: Q1 = K1aL11–a

(10.6)

Q2 = K2bL21–b

(10.7)

Q3 = K3cL31–c,

(10.8)

where a > b > c. Qi, Li, and Ki are, respectively, production, labor and capital in the production sector of the ith good (i = 1, 2, 3). Note that capital is assumed to be fixed in each sector in the equilibrium after foreign workers are admitted, even though it was mobile before the initial longrun equilibrium was reached. Producers maximize the following profit function: pi = PiQi – (riKi + wLi),

(10.9)

where pi and ri are, respectively, the profit and rental rate of the ith production sector, and w is the wage rate. Solving the profit-maximization problem, the following equilibrium conditions are obtained: aK1a–1L11–aP1 = r1,

(10.10)

(1 – a)K1aL1–aP1 = w,

(10.11)

bK2b–1L21–b(1 + t) = r2,

(10.12)

(1 – b)K2bL2–b(1 + t) = w,

(10.13)

Aging and migration in Japan

211

cK3c–1L31–cP3 = r3,

(10.14)

(1 – c)K3cL3–cP3 = w.

(10.15)

Note that equations (10.10) to (10.15) show that factor prices are equal to their marginal value products. Domestic labor supply is assumed to be given, that is, there is no wage– leisure trade-off. Therefore, the sum of labor inputs in the three sectors is equal to the sum of the domestic labor endowment (Lh) plus the number of admitted foreign workers (Lf): L1 + L2 + L3 = Lh + Lf.

(10.16)

The domestic supply of non-tradables must equal the domestic demand because, by definition, no international trade is allowed. Therefore, equation (10.17) holds in equilibrium: C3 = Q3.

(10.17)

Since the tariff revenue accrued to the government is assumed to be distributed to domestic consumers in a lump-sum fashion, and since there is no profit in equilibrium, the national income (GNP, rather than GDP), which does not include the income accrued to migrant workers, consists of factor payments and tariff revenue: r1K1 + r2K2 + r3K3 + wLh + t(C2 – Q2) = Y.

(10.18)

By substitution, equation (10.18) can be expressed by (10.19): P1Q1 + (1 + t)Q2 + P3Q3 – WLf + t(C2 – Q2) = Y.

(10.19)

Note that temporary guest workers are assumed to be paid in terms of exportables. In other words, the consumption pattern of guest workers is different from that of domestic workers, in that the former are assumed to spend their entire income on the consumption of exportable goods of the host country. This assumption seems reasonable because the temporary guest workers stay in Japan for a short period of time and remit a large portion of their income to the home country. Triple effects of immigration on the welfare of the host country Using the above framework, we can analyze the model in order to obtain some insight into the welfare effect of labor inflow under the three realistic

212

Globalization and economic integration

assumptions included in this model. The following is an examination of the impact of the admission of migrant workers on welfare which is represented by the social utility (U) in equation (10.1), above. Substituting (10.3), (10.4) and (10.5) into (10.1), (10.20) is obtained: U = (a/P1)a[b/(1 + t)]bggY/P3g.

(10.20)

Taking natural logarithms of both sides of (10.20), we obtain: lnU = ln(a/P1)a[b/(1 + t)]bgg + lnY – glnP3.

(10.21)

Differentiating equation (10.21) with respect to Lf, (10.22) is obtained: (lnU)9 = (ln Y)9 – g(lnP3)9,

(10.22)

where the variables with primes are defined as the derivative of those variables with respect to Lf. Similar short-hand notation is used throughout this chapter. From (10.22), (10.23) is obtained: (lnU)9 = Y9/Y – gP39/P3.

(10.23)

Equation (10.23) shows that the total welfare effect of admission of migrant workers can be broken down into the effect of the change in income and the effect of the change in the price of non-traded goods. Tedious but straightforward substitution using equilibrium conditions yields equation (10.24), which divides the total welfare effect into four sub-effects: (ln U)9Y = B(–Lfw9) Effect 1 (Cheaper foreign labor effect) + B(–tQ29) Effect 2 (Trade barrier effect) + (B/Y)(Q3P39) – C3P39 Effect 3 (Non-traded goods effect), (10.24) where B ; (1 + t)/(1 + t–bt), note that B > 1. The following is the underlying logic behind the above three sub-effects. Cheaper foreign labor effect It is often the case that the wage rate in the host country becomes lower as more and more foreign workers are admitted. In other words, as the number of admitted migrant workers increases, the cost of hiring them becomes cheaper because the increase in the number of foreign workers gives a dampening effect on the level of the

Aging and migration in Japan

213

MVPL, W

A B

W0 W1

0 Figure 10.2

E C

D

F

G

Labor

Cheaper foreign labor effect

prevailing wage rate in the host country. So, the host country as a whole can benefit from the cheaper foreign labor. Needless to say, there would be conflicts of interest between employers and workers in the host country, because workers, including native workers, incur a loss from the decline in the wage. Figure 10.2 demonstrates an intuitive reasoning for this effect, that is, cheaper foreign labor effect. In the figure, ABEG shows the marginal value product of labor (MVPL) curve. Since the wage rate is equal to the MVPL, the equilibrium before the admission of foreign labor is B, where domestic labor (0D) is employed with a wage rate of W 0. In this case, total labor income is W 00DB and total capital income is AW 0B. If foreign labor of DF is admitted to the country, the new equilibrium point moves to E, and the wage rate decreases to W 1. In this case, capital income increases to AW 1E; total labor income accrued to the native workers and the income accrued to migrant workers become W 10DC and CDFE, respectively. Thus, total income of domestic factors (capital and labor) is increased by the hatched area BCE. Note that the magnitude of the (positive) cheaper foreign labor effect increases, ceteris paribus, as the scale of the admission of migrant workers becomes larger.

214

Globalization and economic integration

Import restriction

Price of importable goods (labor-intensive) higher than international price

Inflated factor price of labor (=wage)

Overpayment to imported factor

Welfare loss to host country Figure 10.3

Mechanism of the Brecher–Diaz-Alejandro effect

Trade barrier effect Brecher and Diaz-Alejandro (1977) note this effect in the context of international capital movement, but similar reasoning also holds for the international movement of labor. Although the mechanism of this effect is a little complicated, intuitive reasoning is as follows. Suppose that country 2 is imposing tariffs on labor-intensive importable goods, such as textiles and clothing. In this situation, the domestic price of the labor-intensive goods is higher than that of the goods in the international market due to the tariff, and therefore, the price of the factor used intensively for the production of labor-intensive goods (that is, wage rate) is inflated and higher than that under free trade. When the admitted migrant workers are paid by this inflated wage rate, they are effectively overpaid, and therefore the host country incurs an economic loss. Figure 10.3 summarizes the mechanism of this effect.1 Non-tradable-good effect When we take account of the existence of the non-traded goods sector, additional insights into the economic effects of immigration can be obtained. In fact, the number of immigrants working in the non-traded goods sector is large in many countries, such as Japan, the United States and Europe. Due to the employment of immigrants, the price of non-traded goods is generally lower than otherwise. In other words, thanks to immigrant workers, consumers can enjoy less-expensive non-traded goods, for example, cheaper maid service or street cleaning (positive consumption effect). On the other hand, the income of native

Aging and migration in Japan

215

U Curve II

Curve I

0 Figure 10.4

Lf1,II

Lf1,I

Lf2,II

Lf2,I

Lf

Migrant workers and host-country welfare

workers in the non-tradable goods sector would be lowered by hiring cheaper immigrants in that sector (negative income effect). Overall effect – trickle versus flood The next important question is whether the net effect of the above three sub-effects has any systematic relationship to the level of admitted migrant workers (Lf) and the magnitude of trade barriers (t). The answer to this question is “yes”. After some algebra, it can be shown that the following two propositions hold under reasonable parameter values: 1. 2.

Welfare declines due to the initial inflow of migrant workers, but after a certain number of admitted foreign workers (Lf1), welfare increases; The smaller the value of t, the smaller the value of Lf1. In other words, the less severe the trade barriers are, the more likely it is that the admission of a certain number of migrant workers is welfare improving.

Figure 10.4 summarizes the above two propositions. In the figure, the welfare level of host country (U) is plotted on the vertical axis, while the number of admitted foreign workers is plotted on the horizontal axis.

216

Globalization and economic integration

Curve I plots the welfare level as a function of admitted migrant workers when the magnitude of trade barriers is t1. The admission of migrant workers decreases the welfare level of host country first, but when the number of admitted foreign workers reaches Lf1,I, the welfare level begins to increase, and exceeds the initial level when the number of admitted foreign workers exceeds Lf2,I. In other words, the admission of a small number (or trickle) of migrant workers produces a negative effect on the host country while a large number (or flood) produces a positive impact on the host country. This finding implies that when migrant workers are admitted, the admission quota should be large if it is to produce a positive welfare impact. Curve II plots the welfare level when the magnitude of trade barriers decreases to t2 due to, for example, a successful conclusion to the Doha Round. The curve shifts upward and leftward, and therefore the trough of the curve also shifts leftward. In other words, a smaller number of migrant workers can be welfare improving. Although the proof of the above propositions requires a cumbersome manipulation of the equilibrium conditions of the model, the intuition behind these propositions is obtained from the following argument: first, note that equation (10.24) can be rearranged as equation (10.25): (ln U)9 = (1/Y)(–Lfw9 + tM9),

(10.25)

where M ; C2 – Q2 (that is, the amount of imports). We know that –Lfw9 (the cheaper foreign labor effect) is positive. It can be shown that M9 is negative, that is, the amount of labor-intensive goods imports declines as the number of migrant workers (or the import of labor) increases, and therefore the second effect in equation (10.25) is negative. In equation (10.25), first note that, when Lf is zero, the first effect (–Lfw) is also zero. Thus, the economic effect of the influx of migrant workers is always welfare worsening. Second, note that the cheaper foreign labor effect is the change in the wage rate (w9) multiplied by the number of foreign workers (Lf), and therefore the magnitude of the first effect tends to increase more rapidly than the second effect, as the number of admitted foreign workers increases. Therefore, at a certain level of admitted foreign workers (Lf1,I in Figure 10.4), the first effect begins to dominate the second effect, which means that an additional admission of migrant workers is welfare improving. Third, the second negative effect (tM9) seems to decrease as t decreases because it is M9 multiplied by t. Therefore, the admission of the same number of migrant workers can produce a positive impact on the welfare of the host country when the magnitude of trade barriers is smaller. The above analysis has the following policy implications for Japan: (i) while small-scale admission of foreign workers has a negative impact, a

Aging and migration in Japan

217

large-scale admission is beneficial; and (ii) the liberalization of trade barriers increases Japan’s chance of benefiting from the admission of foreign labor.

THE SOCIAL EFFECT OF MIGRATION Internationalization – a Positive Externality Since international migration involves the international movement of human beings as a whole, it brings about various social effects in addition to the economic effects discussed above. For example, suppose that a Japanese university in Tokyo hires a Vietnamese medical professor. Although the job description of the Vietnamese professor is probably to teach and carry out research in the field of medical science, his/her contribution to the university, and perhaps to society, is much more than that. His/her colleagues in the university in Tokyo can learn from him/her about the economic and social situation in Vietnam and ASEAN (Association of South East Asian Nations) countries as well as medical science, and listening to the Vietnamese fellow professor will widen their perspective. These interactions between the Japanese and Vietnamese would enhance mutual understanding between the two countries. In economic jargon, such an effect is a positive externality. Fiscal Burden on Governments Since migrant workers pay taxes and receive various social services from the government of the host country, they have public finance effects for the host country. On the one hand, the existence of these migrant workers increases the revenue of the government, because they pay income taxes, consumption tax, and property taxes and so on. If they are enrolled in the social security system in the host country, they also contribute to the social security system. On the other hand, it increases the expenditure of the government, because they receive various social services from the government of the host country, for example, education for their children, medical services, and pensions if they enrol. The Japanese government has published an estimate of the fiscal cost and benefit to the host government (both central and local government) for three different stages of admission of migrant workers (see Table 10.1). While the host government benefits from migrant workers in Stage 1 (only single youths are admitted) because their tax payment exceeds social services they receive, in Stage 2 (with spouse), and Stage 3 (with spouse and

218

Globalization and economic integration

Table 10.1

Social cost of migrant workers Stage 1

Stage 2

Stage 3

Receipt Central government Local government Social security Total

181.1 0.0 145.5 326.6

93.1 38.3 180.2 311.6

77.5 28.2 193.2 298.9

Expenditure Central government Local government Social security Total

12.8 15.5 52.3 80.6

77.8 486.1 89.1 653.0

353.8 901.5 158.1 1,413.4

Net Central government Local government Social security Total

168.3 −15.5 93.2 246.0

15.3 −447.8 91.1 −341.4

−276.3 −873.3 35.1 −1,114.5

Note: Stage 1: Single; Stage 2: Married; Stage 3: Married, two children. Figure in billions of yen. Source:

Japanese Ministry of Labor.

two children), the fiscal cost for the social expenditure far exceeds tax revenues. When half a million migrant workers are admitted, the net cost to the government in Stage 3 exceeds ¥1 trillion (or about US$8 billion). Possible Increase in Crime It is often argued that the admission of migrant workers may increase criminal activity in the host country. Since Japan is a relatively homogeneous society and enjoys a very low crime rate, some Japanese argue that the migrant workers may increase crime in the safe country. The logic behind such a fear is very simplistic and in Japan goes something like: most large US cities like Los Angeles and New York are filled with migrant workers and the crime situation there is extremely serious, and therefore, there must be some correlation (and causality) between migration and crime. Therefore, the increase in migrant workers in Japan would make it a more dangerous country. However, in my view, such a claim is not substantiated. There is no scientific evidence that connects high crime rate with migrant workers.

Aging and migration in Japan

219

On the contrary, some studies suggest that, as far as the first generation is concerned, migrant workers commit fewer crimes than natives in the host country. That is probably because newly migrated workers have a higher motivation to achieve success in their adopted country, and because the effective penalty for committing a crime tends to be far more severe for migrant workers than for the native workers (for example, only migrant workers face possible deportation!). Continuation of “3–D” Jobs In many cases, migrant workers are employed in a job where working conditions are less favorable than in other jobs. In Japan, such jobs are often referred to as “3–D jobs”, that is, “dangerous”, “dirty”, and “demanding” jobs. Since nobody prefers 3–D jobs to other more congenial ones, employers of 3–D jobs often have unfilled vacancies. The existence of such vacancies would encourage employers to make working conditions there more favorable (for example, better wages, safer workplace, and so on). However, if such 3–D vacancies are easily filled by migrant workers, the incentive for employers to achieve better working conditions would disappear. In other words, due to the hiring of migrant workers, who are willing to do 3–D jobs because such jobs may be better than regular jobs in their home country, the 3–D jobs in the host country may persist in the long run.

ALTERNATIVE MEASURES TO MIGRATION There are various alternative measures to migration as a means to cope with labor shortages that result from an aging population. First, it should be noted that the decline in the number of working-age people in a population does not necessarily cause a labor shortage or a burden on the working population. The labor market balance is determined by the demand side (that is, how much labor is demanded by the national economy) and by the supply side (that is, what percentage of the working-age population actually works). Therefore, if the Japanese economy makes a structural change to a labor-saving economy, and if the female labor participation rate increases, the decline in the working-age population does not necessarily result in labor shortages. One of the most important measures to avoid a labor shortage or an increase in the burden on the working population is to increase labor productivity, because, even if the number of workers decreases in the future, the productivity increase can compensate for the decline. The working-age

220

Globalization and economic integration

population in Japan is expected to decline by 17 million in 25 years, or an annual decline of 0.9 percent. On the other hand, labor productivity (at least until the recent recession) used to increase by about 3 percent a year mainly due to the active investment by Japanese firms. So, if the Japanese economy is put on track again, in the age of the IT revolution, a 0.9 percent annual decline of the working-age population will easily be offset by the increase in labor productivity. In addition to the increase in labor productivity, various measures can be taken to prevent labor shortages in the future. In what follows, I shall discuss some alternative measures: (i) measures to utilize Japanese labor effectively (for example, female participation and rationalization of agriculture); and (ii) measures to utilize foreign labor indirectly (for example, trade liberalization and outflow of foreign direct investment: FDI). First, increasing female participation seems promising. Although more women have entered the labor market in recent years, there are still a large number who want to work but cannot do so due to family responsibilities such as housework, small children, or sick parents. So, if a better working environment for women, such as an adequate number of good-quality daycare centers, is realized, the supply of female labor will increase. This would mitigate the expected labor shortage in Japan. Second, the rationalization of inefficient industries, such as agriculture and the distribution sectors, is important. While the number of agricultural workers has been steadily declining, labor productivity in agriculture is still much lower compared to many countries, if Japanese agricultural products are evaluated at international prices. So, if a further rationalization of Japanese agriculture makes it possible to reallocate agricultural workers to other sectors, the labor shortage in the economy as a whole will be relieved. Third, trade liberalization is also promising. If (instead of producing goods in Japan using Japanese workers) Japan imports more foreignmade products, which are produced by foreign labor in foreign countries, it can save domestic labor. This is particularly the case for the import of labor-intensive products, such as textiles and clothing. Fourth, closely related to the third point, a curb on export promotion should be helpful. Ever since the Meiji Restoration, Japan has strongly encouraged exports. In recent years the accumulation of a trade surplus led to trade conflicts with many countries. If Japan curbs its strong export drive and reduces the export of labor-intensive products, then the labor demand by the Japanese economy will be decreased substantially. Fifth, FDI is also important. Instead of building factories in Japan and hiring Japanese workers there, Japanese firms can move production sites to foreign countries, where they can produce goods by hiring foreign

Aging and migration in Japan

Table 10.2

Labor saving and labor creation (a simulation) (thousands)

Female participation Efficient agriculture Trade liberalization Curb on the export drive FDI Total of the above Note: Source:

221

Normal

High speed

379 223 153 96 158 1009

783 325 304 96 158 1666

Number of expected decline in working-age population in 2000–25 is 17 million. Goto (1994).

workers. The outflow of Japanese FDI will create jobs in foreign countries, and, at the same time, save labor in Japan. So, how big is the effect of these five measures? I have done a simple simulation of the magnitude of the labor-creation and labor-saving effects of the measures (see Goto, 1994), the results of which are shown in Table 10.2. The results show that the total size of the labor-saving (and laborcreating) effect ranges from 10 million to 17 million, depending on the underlining assumptions. These figures are close to the expected decline in the working-age population in the early 21st century (17 million). Moreover, in addition to the measures listed here, there are many other ways to cope with future labor shortages, such as mobilization of older workers, and a restructuring of the currently inefficient distribution system in Japan. Thus, at least in aggregate, the decline in the working-age population and the resulting increase in the burden on the working population seem to be compensated if various measures are implemented.

CONCLUDING REMARKS In the above, I have discussed the merits of international migration as a measure to cope with the labor shortage due to aging in Japan. In the second section, I examined the economic impact of migration. While the effect is complex, it can be divided into sub-effects, some of which are welfare improving and others are welfare worsening. The negative trade barrier effect (the Brecher–Diaz-Alejandro effect) seems to dominate other positive effects in many cases. This implies that, at least in the long run, the admission of guest workers may not be a desirable policy, unless the above negative effect is compensated by positive externalities. The migration of

222

Globalization and economic integration

professional and skilled workers is more likely to have positive externalities than that of unskilled workers. Unfortunately, the majority of the current intra-Asian migration involves the movement of unskilled workers, such as maids and construction workers. In the third section, some social effects of migration were examined. Since migration is an international movement of human beings rather than that of labor as a production factor alone, it has various social as well as economic impacts. In the fourth section, I discussed some empirical issues on the merits of various measures to cope with possible labor shortages due to the aging population in Japan. In order to cope with a future labor shortage, migration might be an alternative, but, at least in aggregate, the decline in the working-age population and the resulting increase in the burden on the working population would seem to be compensated as long as various measures, such as the utilization of female labor, are taken effectively. Thus, it seems important to take measures to increase the domestic labor supply and/or utilize foreign labor indirectly through further imports of labor-intensive goods from abroad.

NOTE 1. Note that the argument here assumes that both native and migrant workers are paid according to their labor productivity, that is, there is no wage discrimination. In reality, however, it is often reported that migrant workers are paid substantially less than native workers. In the event that wage discrimination exists, the magnitude of the negative trade barrier effect becomes smaller (if the wage discrimination is severe, the effect can be positive to the host country).

REFERENCES Brecher, R. and C.F. Diaz-Alejandro (1977), “Tariffs, foreign capital, and immiserizing growth”, Journal of International Economics, 7, 317–22. Goto, J. (1994), Gaikokujin Rodosha to Nihon Keizai (Migrant workers and the Japanese economy), Tokyo: Yuhikaku Ltd. Goto, J. (1998), “The impact of migrant workers on the Japanese economy: trickle versus flood”, Japan and the World Economy, 10, 63–83. Morgan, L. and B. Gardner (1982), “Potential for a US guest-worker program in agriculture: lessons from the Braceros”, in B.R. Chiswick (ed.), The Gateway: US Immigration Issues and Policy, Washington, DC: American Enterprise Institute. Schiff, M. (1999), “Trade, migration, and welfare”, Policy Research Working Paper 2044, World Bank, Washington, DC.

11.

Labor market transitions for female workers in Japan: the role of global competition Tomoko Kishi and Noel Gaston*

INTRODUCTION Japan has only recently turned the corner on a dismally long decade of stagnant economic growth and unprecedented high levels of postwar unemployment. The labor market woes were coincident with the contraction of manufacturing industries. Manufacturing employment declined due to a variety of factors, for example, the protracted economic slump, financial and bad loan-related problems as well as deindustrialization and the structural shift to service sector industries. Another commonly perceived culprit has been globalization, more generally; and outsourcing, more specifically. In Japan, globalization is commonly thought to be driving the “hollowing-out” of manufacturing industry. Associated with the industrial shifts has been a significant increase in part-time and casual forms of employment as opposed to permanent, fulltime work. Some commentators regard that the growth of non-standard work arrangements is the single most important change taking place in the Japanese labor market (for example, Rebick, 2005). According to the Ministry of Health, Labor and Welfare, for the 10 years to 2003, a period during which total employment in Japan stagnated, the number of parttime workers increased by 52.5 percent (MHLW, 2004). As always, there are concerns about the quality of part-time jobs compared with that of full-time work. A central plank of the Japanese government’s labor market policy has been to increase the flexibility of the labor market. Gaston and Kishi (2005) argue that certain characteristics of Japan’s unique industrial relations system may hamper rather than ease the plight of the unemployed. The need for greater flexibility may also explain the increasing preference by Japanese firms for part-time employees (Houseman and Osawa, 223

224

Globalization and economic integration

1995). Gaston and Kishi (2007) explain that the demand for more flexible working arrangements has also increased. While the reasons for the growth in the part-time workforce are still not completely understood, changes in the industrial structure are likely to have been important. According to the Ministry of Internal Affairs and Communications in its Labor Force Survey (MIAC, 2005), from 1992 to 2002 the share of total employment in manufacturing industries fell from 24 to 19 percent; while service sector industries expanded their share from 22 to 28 percent. Due to falling labor productivity and the instability of labor demand, firms in the services sector have come increasingly to rely on “non-standard” (that is, those other than permanent, full-time employees) workers (Kishi, 2003). In the services sector, the proportion of full-time workers fell from 77 to 68 percent from 1992 to 2002. Nearly all OECD countries have experienced rapid growth in part-time employment. Between 1987 and 1997 nearly 70 percent of all new jobs created in Japan were parttime jobs – the OECD average was about 50 percent (OECD, 1999). While the number of male non-standard workers has been increasing in recent years (Rebick, 2005), such workers are still overwhelmingly female. In a flat labor market it is not obvious that workers are necessarily disadvantaged by part-time work arrangements, vis-à-vis equivalent types of standard work. However, a common (and seemingly, reasonable) presumption in the literature is that in a declining economy, part-time and temporary workers are both vulnerable and exploitable (for example, Shinozaki et al., 2003). This finding is not confirmed by studies for other countries. For example, Stratton (1996), Farber (1999), and Gaston and Timcke (1999) all find that casual and part-time workers are more likely to make successful transitions to full-time work than their unemployed counterparts. With this background in mind, this chapter seeks to uncover the most important determinants of labor market transitions for female workers in Japan. The time period we consider is 1993 to 2004 – a period of unprecedented stagnation for Japan. It provides an excellent, albeit salutary, setting to investigate how workers are affected and how firms adjusted their workforces during such trying economic times. To achieve this purpose we use data from the Japanese Panel Data on Consumers (JPSC). In the next section, we take the slightly unorthodox approach of describing our data before adumbrating the list of hypotheses we test. We do so for a number of reasons. First, we need to make clear at the outset the precise definitions of full-time, part-time, and temporary workers, which are somewhat idiosyncratic to Japan as well as specific to the data we use in this chapter. Second, in order to economize on our justification for some of our hypotheses, we describe previous research done using the same data. Some of the findings are unsurprising and do not bear detailed

Labor market transitions for female workers in Japan

225

discussion. Another advantage of doing so is that it enables us to spend a little more time describing what we consider to be the most novel aspect of our study: specifically, how globalization may have affected labor market transitions. The third section contains a brief discussion of our econometric methodology and presents the results. In particular, we estimate a panel generalized linear latent and mixed model for workers making transitions between three labor market states: full-time jobs, part-time jobs, and temporary jobs (which are combined with those not in the labor force). The final section contains our concluding observations.

THE JPSC AND OUR HYPOTHESES The JPSC The data used for our study combine data from the JPSC and industry-level data obtained from various official sources. The JPSC collects information about economic and subjective well-being, labor market dynamics, and family dynamics over time. All the respondents are women from a specific year (for example, Wave 1 respondents were women aged between 25 and 34 in 1993). Fortunately, for our purposes, most of the women are neither students nor retirees. The characteristics of each wave of the JPSC samples given in Table 11A.1. We use data from all 12 waves, that is, 1993 to 2004. The first stage of our empirical analysis uses industry fixed effects to capture industry-specific effects affecting the transitions between labor market states for female workers. Subsequently, since (rather unsurprisingly) we find that these industry effects are statistically significant, we then replace the industry dummies with industry-specific variables. Industry Data Most previous studies of female labor supply in Japan have concentrated on the effects of personal characteristics – marriage, childbirth, availability of childcare services, and spousal incomes – on employment status. These factors are undeniably important; after all they are well-grounded theoretically and empirically. Any analysis needs to control for individual-level factors, such as marriage and the existence of children, in an investigation of transitions in female employment status. In addition, we also investigate the factors specific to the industries to which the female workers’ employers belong. In particular, given the concerns voiced about microeconomic reforms, deregulation, changes

226

Globalization and economic integration

in corporate governance, immigration, outsourcing, hollowing-out, and globalization more generally, we pay specific attention to each industry’s degree of exposure to international competition and the consequent impact on the employment status of female workers. Previous research indicates that industries exposed to global competition tend to have more part-time employees than do purely domestic industries (Toyoda, 2005). Thus, those seeking jobs in more competitive tradable goods sectors have a higher probability of getting part-time jobs compared to jobs in less-competitive, non-tradable goods sectors. In addition, as we discuss below, full-time employees in more open industries could have a higher probability of losing their job, a lower probability of re-employment or a higher probability of becoming a part-time employee or leaving the workforce altogether than counterpart workers in more competitively sheltered sectors. The degree of “outward orientation” for each industry is measured by import penetration (that is, imports/total demand), export orientation (that is, exports/total demand), and foreign direct investment, FDI (that is, nominal inward and outward FDI/nominal GDP). These data are contained in Tables 11A.2–4. Obviously, the pattern of the import and export ratios varies across industries. In the manufacturing, mining, and wholesale and retail trade industries, the import and export ratios have been increasing. In construction and utilities, both import and export ratios are almost zero, as they are for the government sector. In the transportation and communication sector, import penetration and export orientation have been fluctuating and display no obvious trend. The FDI and import data indicate Japan’s dependence on natural resources. We also include data on industry growth rates. As we discuss further below, there is a widespread view that the low growth rates of the 1990s were in large measure responsible for some of the structural changes that took place in the Japanese labor market. These data appear in Table 11A.5. What is immediately obvious is the severity of the recession in the second half of the 1990s as well as the marked differences across industries in growth rates. One might anticipate that the combination of both of these factors would result in significant structural shifts and changing employment relationships. As we shall see below, such an expectation is not borne out statistically. Macroeconomic Data and Measuring Globalization To control for the overall macroeconomic environment, we include data for the female unemployment rate to capture the effects of the economywide slowdown on female workers. To assess the extent to which Japan has globalized we use the KOF index

Labor market transitions for female workers in Japan

227

65 60 55 50 45 40 35 30 1970

1975

Figure 11.1

1980

1985

1990

1995

2000

2005

The KOF index of globalization for Japan, 1970–2005

(Dreher et al., 2008); in the following, we simply label it as KOF. As discussed in Chapter 1, the index is derived from 25 variables grouped into six ‘sub-indices’: actual flows of trade and investment, restrictions, variables measuring the degree of political integration, data quantifying the extent of personal contact with people living in foreign countries, data measuring transborder flows of information and a proxy for cultural integration. Table 11A.6 reports the weights of the individual components.1 Overall, since 1970 Japan has globalized. As Figure 11.1 illustrates, globalization is not inexorable, there have been reverses, some of which occur in the sample period.2 Part-time Employees in Japan Before proceeding to our hypotheses and econometric analysis, we need to define precisely what it means to be a part-time employee in Japan, since it is somewhat different from other countries (see OECD, 2007). In Japan, there are essentially three definitions of part-time workers: 1.

In accordance with international convention, MIAC conducts the Labor Force Survey, which defines part-time workers as those working less than 35 hours a week.

228

2.

3.

Globalization and economic integration

MIAC also conducts the Employment Status Survey. Part-time workers are defined as employees in a particular workplace for whom the wage tables applied are substantially different to those for regular employees. In the Survey, part-time employees with the same hours of work, or possibly longer hours of work, compared to full-time employees are identified. MHLW conducts the General Survey on Diversified Types of Employment. Part-time employees are defined as those working fewer hours per day or fewer days per week than regular employees.

The definition of part-time employees used in the JPSC is 2).3 In Japanese firms, human resource management for part-time employees is completely different from that for full-time employees. First, the ports of entry for part-time employees are different. Usually, regular employees are carefully screened and chosen from new graduates during a designated recruitment period. Such workers are implicitly assumed to be long-term employees. According to the Japan Institute for Labor Policy and Training (JILPT, 2006), approximately 60 percent of companies assign tests for general knowledge and academic proficiency and approximately 50 percent perform vocational aptitude tests. On the other hand, part-time employees are generally selected from a pool of housewives and students without such careful scrutiny. In addition, the average career paths for regular employees and part-time employees are quite different. Newly hired regular employees undergo onthe-job training and gain a wide range of skills (Koike and Inoki, 2003). Unlike regular employees, part-time employees receive limited formal training, as they are not expected to climb the career ladder within the firm’s internal labor market. As mentioned, for regular employees, wages are mainly based on job grades and age. However, part-time employees’ wage rates are often anchored to the outside labor market and may have little connection to a worker’s previous experience. Overall the Japanese labor market displays a dualism: a primary market characterized by relatively high wages, good working conditions, employment stability and the prospects of career advancement and a secondary market which tends to have low wages, poor working conditions, high turnover rates, and little chance of advancement (Kishi, 2003). While this perspective seems to be useful in understanding the coexistence of regular and part-time employee labor markets, there are still transitions from part- to full-time and what is perceived in the current global environment to be increasingly common full- to part-time work transitions. This is our main concern in this chapter.

Labor market transitions for female workers in Japan

229

Preceding Analyses Based on the JPSC In the long period of stagnation following the collapse of the “bubble” economy, the proportion of non-standard employees to total employment trended sharply upwards. At the same time, wage inequality and household income inequality became significant. This led to a large number of research papers examining the effects of stagnation on labor market outcomes and wage distributions. As for research papers using the JPSC data, there have been relatively few analyses of the changes in employment status and incomes as opposed to studies of the relationship between female labor force participation, marriage, and childbirth.4 This may be due to the fact that the rapid decline of birth rates in the last few decades attracted the attention of researchers more than the observed labor market outcomes during the 1990s. Among the analyses focusing on the relationship between business cycles and female employment status are Higuchi et al. (2004), Kitamura (2005), and Abe (2006). Higuchi et al. (2004) analyze the labor force status of JPSC respondents and find that the proportion of the women who change their employment status from regular to part-time employment was higher for the generation that graduated from school at the beginning of the 1990s than for older cohorts. In addition, they report that the proportion of women who continue to work as regular workers after having children was higher for those who completed their education in the period shortly after 1985, when the Equal Employment Opportunity Law was implemented, than for other cohorts. Also, among those who were classified as “freeters” (that is, young workers with tenuous labor market attachment), the proportion of part-time employees was higher, possibly due to the fact that these women were more likely to be single. Kitamura (2005) finds that the proportion of those leaving the workforce was higher for younger cohorts who had been in their twenties during the 1990s than for the older cohorts. He also finds that younger women had a lower annual income than older women since the younger workers were more likely to be part-time and other non-regular employees. Abe (2006) finds that among the married women in the JPSC, cohort C respondents accounted for a higher proportion of part-time employees than those in cohort B, which in turn had a higher proportion of part-time employees than those in cohort A. He also compares the annual income distributions of women and their spouses in cohorts A, B, and C and finds that couples in cohort C aged between 30 to 35 years had significantly lower minimum, median, and maximum wages compared to couples in the same age bracket in cohorts A and B. Such income differentials are hypothesized to be brought about by changes in a labor market in which

230

Globalization and economic integration

job openings for part-time employees increased and those for regular employees decreased. Finally, the labor supply behavior of married women has been analyzed by a number of researchers. For example, Matsuura and Shigeno (2001) use multinomial logit estimation and find that for married women the choice to work full-time, part-time or to not work at all is quite different. Hypotheses about Labor Market Transitions The above studies are certainly informative, but most of them do not look beyond the role of certain individual characteristics and the year of entering the labor market for understanding the impact on labor market transitions. In this chapter, we are particularly interested in the role of globalization and its impact on the growth of part-time employment. In part, this interest stems from our earlier work. A summary of the key findings in Gaston and Kishi (2007) is as follows. First, there is a relatively weak relationship between the hiring of part-time employees and economic activity. Second, the increase in the employment of part-time workers is unrelated to any broad capital–labor substitution. Third, the hiring of part-time workers has been strongest in those firms that convert part- into full-time jobs, that is, many Japanese firms use part-time work arrangements to screen workers for full-time positions. In other words, some workers do in fact make the transition to permanent, regular employment. Finally, and important for the present chapter, the growth of part-time employment is mainly accounted for by the wholesale trade, retail trade, and service industries. In turn, the negative effect for tradable goods industries is related to international competition and outsourcing. In fact, the growth of employment at overseas affiliates is strongly positively correlated with the growth in part-time employment in Japan. In our earlier paper, we speculated that Japanese multinationals respond to bleak economic conditions at home by increasing employment overseas, rather than by hiring workers for their domestic operations. In addition, domestic part-time workers, and lower-productivity workers more generally, have been replaced by unskilled workers at overseas affiliates. Our previous work leads us to the first hypothesis. That is, that firms in more internationally exposed sectors are the most likely to forgo traditional lifetime employment practices. As a corollary, while all firms are under financial pressure, firms in non-tradable goods sectors are better able to resist the need for more flexible workplace arrangements. H1 (International competition): Workers in internationally exposed sectors are more likely to make transitions to part-time employment or to

Labor market transitions for female workers in Japan

231

unemployment (or not in the labor force) than workers in non-tradable goods sectors. As a corollary, the transition probability to part-time employment in internationally exposed sectors is higher than in other sectors. A related hypothesis is that firms resort to hiring part-time workers as a direct response to the economic slowdown. For example, Rebick (2005) argues that it has been the slower growth that has pushed Japanese firms to increasingly rely on part-time workers and other secondary laborers. This hypothesis is related to H1, but it is not trade related. Hence, H2 (Demand and cost-cutting): In sectors with lower product market demand, the transition to part-time work and to unemployment (or not in the labor force) is higher. Related to the hypotheses about fluctuations in labor market demand is an argument that relates to structural change. In nearly every OECD country, part-time and casual forms of employment have risen in the last decade. This is due to structural change more generally, but may be due to globalization more specifically. On the other hand, the negative economic developments that so concern interested spectators of Japan probably have far more to do with the bursting of the bubble economy and the protracted economic downturn of the 1990s, rather than with globalization per se. In fact, one weakness of much current research is the way in which the effects of the long slump and the effects of globalization are often confounded. For example, a typical assertion is as follows “globalization has contributed to the country’s unemployment woes” (Nakamura et al., 2004, p. 107); likewise, “globalization of the world economy caused brutal competition across all areas of the Japanese economy and accelerated both domestic layoffs and the loss of many manufacturing and R&D jobs to FDI operations in Asia” (Nakamura and Horiuchi, 2004, p. 240). However, outsourcing has long been a characteristic of post-war Japan, and even an explicit part of the government’s industrial policy; it is not a recent phenomenon. Naturally, the rapid growth of trade with China has had, and will continue to have, profound effects on all its trading partners, Japan included. But little persuasive evidence or systematic analysis is available which sheds any light on the precise effects of globalization on labor market outcomes, such as the unemployment rate or economic growth as well as labor market transitions. Notwithstanding, our next hypothesis tests this common presumption: H3 (Globalization): The transition to part-time work has increased over time, as Japan has globalized.

232

Globalization and economic integration

Finally, we turn to the other hypotheses based on individual characteristics, which are largely motivated by the previous research summarized above. We simply list these with little further elaboration: H4 (Individual characteristics): The transition from full- to part-time work for female workers is higher when they: (i) are younger (that is, less specific human capital); (ii) have less human capital (that is, less general human capital or formal education); (iii) have more children (that is, higher opportunity cost of full-time work); and (iv) get married (that is, a cultural expectation). In the next section, we test these hypotheses by presenting first the empirical strategy and then the results.

METHODOLOGY AND RESULTS The Model In order to test the hypotheses, we perform two kinds of estimation. First, we estimate the relative probability that an individual employed in a full- or a part-time job in the (t 2 1) th period changes her labor market status in the tth period. Thus, the focus here is on the characteristics of the industry from which an individual moves. Second, in order to focus on the characteristics of the industry to which an individual moves, we estimate the probability of the transition from full- to part-time employment and vice versa. In both cases, the transition process is best approximated by a random-intercept proportional odds model. The proportional odds model with three outcomes The transitions from being in full-time employment to being in the three employment states – full-time, part-time, and housewives (or the unemployed) – are estimated using the proportional odds model as follows. The employment status of respondent i in the tth period is yi t and is defined as: 2, if a housewife (or unemployed) ; yit 5 • 1, if a part-time employee; 0, if a full-time employee.

(11.1)

We specify the latent variable (y*it ) underlying the outcome (yit) as a linear function of the explanatory variables, Xit21.5 The model is written as:

Labor market transitions for female workers in Japan

yit* 5 brXit21 1 ßt 1 eit,

233

(11.2)

where ßt is the random intercept which varies across individuals and eit is the error term. The continuous latent responses y*it in equation (11.2) are related to the ordinal outcomes yit via the threshold levels k1 and k2 as follows: 0, if y*it # k1 yit 5 • 1, if k1 , y*it # k2 2 , if k2 , y*it.

(11.3)

The relationship between the three outcomes and the explanatory variables is: ln e

Pr ( yit . s 0 Xit21, ßt) f 5 brXit21 1 ßt 2 ks, 1 2 Pr ( yit . s 0 Xit21, ßt)

where ßit 0 Xit21~N (0, y) ,

(11.4)

where s is either 1 or 2 and ks is the threshold level shown in equation (11.3). We produce estimates using two different specifications of Xit–1. The first simply uses the industry dummies. This enables us to determine whether the industry to which a worker belonged matters for making transitions from either full- or part-time work. The second replaces the industry dummies with industry openness characteristics. Unsurprisingly, we find that industry matters for labor market transitions. The hypotheses about openness are then tested using the second set of regressors. The explanatory variables are described in Table 11.1.6 The random-intercept logistic model with two outcomes In the previous subsection, we assume that individuals have a choice of three types of employment status. Here, we assume that the jth individual has just two choices in the tth period, that is, to be a part- or a full-time employee. That is, the outcome in the tth period, zjt, is the following binary variable: zjt 5 e

1, to be a part-time employee; 0, to be a full-time employee.

(11.5)

We specify the latent variables (z*it ) underlying the observed outcomes (zit) as linear functions of the explanatory variables. Denoting Wit as an explanatory variable matrix, then the latent-response model is written as follows:

234

Table 11.1

Globalization and economic integration

Explanatory variables

Label

Definition

Export

(Exports/total demand) for industry to which respondent’s employer belongs (%) (Imports/total demand) for industry to which respondent’s employer belongs (%) Growth rate of valued added for industry to which respondent’s employer belongs (%) KOF index of globalization (Foreign direct investment/gross value added) for industry to which the respondent’s employer belongs (%) Female unemployment rate (%) Dummy = 1 if the respondent married; = 0 otherwise Age in years Years of education (9 = junior high school, 12 = senior high school, 13 = technical school, 14 = junior college and technical college, 16 = university and 18 = higher education) Number of children less than 6 years

Import Growth KOF FDI UR Married Age Education

Kids

Dummy variables for industry (reference category: government, utilities and others) Primary Dummy = 1 if the respondent is in the primary sector; = 0 otherwise Const Dummy = 1 if the respondent is in the construction sector; = 0 otherwise Manuf Dummy = 1 if the respondent is in the manufacturing sector; = 0 otherwise Wsale Dummy = 1 if the respondent is in the wholesale and retail sector; = 0 otherwise Finance Dummy = 1 if the respondent is in the finance and insurance sector; = 0 otherwise Transport Dummy = 1 if the respondent is in the transportation and communication sector; =0 otherwise Service Dummy = 1 if the respondent is in the service sector; = 0 otherwise

z*it 5 grWjt 1 tt 1 ujt,

(11.6)

where tt and ujt are the random intercept and the error term, respectively. The continuous latent responses z*jt are related to the observed outcomes zjt via the threshold level l as follows: zjt 5 e

0, if z*jt # l; 1, if l , z*jt.

(11.7)

Labor market transitions for female workers in Japan

235

The relationship between the outcomes and the explanatory variables is: Pr (zjt 5 1 0 Wjt, tt) ln e f 5 grWjt 1 tt 2 l, 1 2 Pr (zjt 5 1 0 Wjt, tt) where tt 0 Wjt ~ N (0, f) .

(11.8)

Note that equation (11.8) differs from equation (11.4) since contemporaneous values of all the variables are used. That is, this model estimates the effects of the “destination” industry characteristics for job changers. As above, we use two different sets of explanatory variables and the STATA gllamm procedure is used. Econometric Procedures and Descriptive Statistics First, we estimate the model described by equation (11.4). The data used for the estimation are from Waves 2–12 of the JPSC. The descriptive statistics for the data are given in Table 11A.7.7 The most obvious feature is that more than half of the female workers in our sample are in the service and wholesale and retail trade industries. The estimates using the industry dummies are contained in Table 11A.9. The omitted industry categories are government and utilities; industries which are essentially closed and purely domestic. In addition to the personal characteristics and the measure of globalization, industry of affiliation is clearly important for labor market transitions from full-time work. Apart from the service and wholesale and retail trade industries, manufacturing is also important. Ignoring the mining industry, the last two industries are the most open and internationally exposed. The wholesale and retail trade industries, where a large percentage of women in our sample work, have very high import penetration, export orientation, and FDI. We turn next to an explicit consideration of these openness variables and the globalization index. First, consider panel I of Table 11.2. As the model makes the proportional odds assumption, the odds ratio for being in a part-time job or not working at all compared to being in a full-time job is the same as the odds ratio for not working compared to being in a full- or a part-time job.8 Hence, the result for imports indicates that a one percentage point increase in import penetration of the industry to which a woman belonged in the (t – 1)th period increases the odds ratio of leaving full-time work by 1.006 (= exp(0.006)) and the odds ratio of not working at all by 1.006. As for the other industry-level variables, a one percentage point increase in the female unemployment rate increases the odds of not working by 1.576. On the other hand, an increase in the KOF index

236

y^ 11 † y^ 11 1 p2/3

0.830

1.576 3.564 0.986

0.951

1.006

Odds ratio

−2.76*** −2.50**

2.13** 1.50 0.72 −1.30 −1.06 2.89*** 3.91*** −5.63*** −2.25** −4.77***

z-value

1.625 (0.353) −1658.720 4470 928 0.005

−4.806 −4.340

0.036 0.098 0.012 −0.449 −0.033 0.478 0.626 −0.106 −0.267 −0.250

Estimated coefficient

II. From FT jobs

1.613 1.870 0.899 0.766 0.799

1.037

Odds ratio

−0.27 2.88

−0.65 −0.59 0.91 −1.23 1.12 0.80 −5.56*** 5.91*** 3.70*** 1.13

z-value

1.601 (0.363) −1877.264 3060 812 0.005

−0.459 4.998

−0.009 −0.003 0.015 −0.043 0.028 0.133 −1.000 1.056 0.358 0.059

Estimated coefficient

III. From PT jobs

0.368 2.875 1.430

Odds ratio

Notes: Dependent variable for estimates in I and II: Odds ratios of yit, 2 = NLF, 1 = PT, 0 = FT; dependent variable for estimates in III: Odds ratios of yit, 2 = FT, 1 = PT, 0 = NLF. All explanatory variables are for the (t − 1)th period. † Estimated interclass correlation for the latent responses. *, ** and *** denote statistical significance at the 10%, 5% and 1% levels.

r5

Log likelihood Number of observations Number of groups

y

−3.04*** −0.48

0.16 1.68* 0.85 −2.27** −0.89 4.16*** 8.23*** −2.84*** −0.88 −3.66***

z-value

9.568 (0.707) −5054.079 7530 1450 0.744

−3.993 −0.619

k1 k2

Random part: variance

0.016 0.006 0.009 −0.050 −0.014 0.455 1.271 −0.014 −0.058 −0.186

Export Import Growth KOF FDI Female unemployment rate Married Age Kids Years of education Fixed part: thresholds

Estimated coefficient

I. From FT or PT jobs

Odds ratios of changing employment status (origin industry characteristics)

Explanatory variable

Table 11.2

Labor market transitions for female workers in Japan

237

increases the odds of working – contrary to the usual concerns about globalization. As expected, marriage increases the two odds ratios, that is, of not working versus being employed and of not working or working part-time compared to being in a full-time job. For married women, the odds of not working or working part-time as opposed to being in a full-time job are more than three times higher compared to unmarried women. The negative effect of the woman’s age on the odds ratio of leaving full-time work highlights the importance of job tenure and work experience. Education also has the expected significant negative effect on the transition from fulltime work. Now consider the next two panels of Table 11.2 where the estimates of the odds of leaving a full- or a part-time job are presented. There is a positive effect of age on the odds ratio of becoming a full- rather than a part-time employee. The negative effect of age on the odds ratios for losing jobs or obtaining part-time work observed in panel II is consistent with results from earlier research (for example, Kitamura, 2005 and Abe, 2006). Unsurprisingly, years of education have a negative effect on the odds both of being out of the labor force and of changing employment status from full to part time. Overall, the effects of openness characteristics of the ‘origin’ industries on the changes in employment status are important but less important than personal factors such as marriage and age. The latter finding, in particular, confirms the view that it was relatively uncommon for young, fulltime employees to be retrenched in Japan, even during the “lost decade”. Next, we estimate the model in which full-time employees face only two outcomes, that is, to work part or full time, that is, in the tth period. The estimates are in Table 11.3. The descriptive statistics for the data used to estimate equation (11.8) are given in Table 11A.8 and the results for the estimation with industry dummies are in Table 11A.10. The latter table once again reveals the importance of industry effects. Recall that one feature of this present procedure is that we now focus on the characteristics of a woman’s destination industry, rather than her origin industry. Panel I of Table 11.3 indicates that the import penetration ratio of the industry to which the woman’s employer belongs yields a positive effect on the odds ratio of becoming a part-time employee in the tth period. A one percentage point increase in the import ratio increases the odds ratio of having a part- rather than a full-time job by 1.015. A one percentage point increase in the female unemployment rate increases the odds ratio of having a part-time job by 1.74. FDI increases the odds of the transition to part-time jobs, but globalization lowers it. The results of the personal characteristics are similar to those for Table

238

f^ 11 † f^ 11 1 p2/3

0.897 1.126 1.740 12.962 1.120 1.246 0.458

1.015

−0.129

0.046 0.023 0.013 −0.016 0.022 0.813 1.085 −0.071 −0.167 −0.459

7.276 (1.852) −555.672 4150 856 0.022

−0.04

1.53 1.79* 0.39 −0.24 0.37 2.32** 3.39*** −1.90* −0.72 −4.44*** 0.632

2.255 2.959 0.931

1.023

Odds ratio

1.91*

−0.89 0.33 0.43 1.42 −2.61*** −1.31 −5.20*** 0.06 1.70* 0.14

z-value

3.133 (0.894) −645.387 2645 636 0.009

4.943

−0.022 0.004 0.129 0.077 −0.154 −0.370 −1.521 0.002 0.314 0.012

Estimated coefficient

1.012

1.369

0.218

0.857

Odds ratio

III. From PT (Dependent variable: Odds ratio of zjt, FT = 1, PT = 0.

Notes: Housewives and non-working persons are excluded; non-regular employees other than part-time employees are also excluded. † Estimated inter-class correlation for the latent responses. *, ** and *** denote statistical significance at the 10%, 5% and 1% levels.

r5

Log likelihood Number of level 1 units Number of level 2 units

f

−5.07***

1.56 2.33** 0.11 −3.43*** 4.22*** 3.17*** 13.33*** 4.50*** 1.77* −10.61***

14.146 (0.932) −2412.724 7509 1425 0.811

−8.112

l

Random part: variance

0.025 0.015 0.002 −0.109 0.119 0.554 2.562 0.113 0.220 −0.781

Export ratio Import ratio Industry growth rate KOF FDI Female unemployment rate Married Age Kids Years of education Fixed part: thresholds

z-value

Estimated coefficient

Odds ratio

Estimated coefficient

z-value

II. From FT (Dependent variable: Odds ratio of zjt, PT = 1, FT = 0

I. From FT, PT or NLF (Dependent variable: Odds ratio of zjt, PT = 1, FT = 0

Odds ratios of changing employment status (destination industry characteristics)

Explanatory variable

Table 11.3

Labor market transitions for female workers in Japan

239

11.2. In particular, the effects of being married are extremely strong. Comparing married and unmarried women, the odds of becoming a partrather than a full-time employee are about 13 times higher for married women. Panel III of Table 11.3 demonstrates that FDI in the industry to which a woman belongs in the tth period has a significant negative effect on the odds of changing employment status to full time. Summary First, starting with H4 (Individual characteristics) the results were generally highly supportive. Younger workers with less-specific human capital or fewer years of general human capital or formal education are more likely to make the transition to a part-time job. The effects of marriage are highly significant and, we conjecture, mirrors something idiosyncratically non-Western about Japanese society. Many Japanese women get married and leave full-time work. The only surprise is the somewhat ambiguous effects we found for the number of pre-school children. We are unwilling to provide any casual explanation here and this is a topic for further research. The second hypothesis was concerned with the increased need to cut costs during the lost decade; H2 (Demand and cost-cutting). Despite the presence of highly significant industry fixed effects on the probability of leaving full-time permanent work, these effects are not attributable to industry-specific growth rates. That is, industry growth has no significant effect on the probability of leaving full-time employment. On the other hand, the impact of a higher female unemployment rate has the predicted effect. The state of the macro economy matters, even if what happens at the industry level does not. As for our main hypothesis: H1 (International competition), we find quite strong evidence that women in internationally exposed sectors were more likely to make the transition to part-time employment or to exit the labor force compared to workers in non-tradable goods sectors. In addition, the transition probability of being in part-time employment in internationally exposed sectors is higher than in other sectors. These results are apparent using both estimation methods. Our results reinforce the perception that the most open segments of the Japanese economy are the more likely to make the greater adjustment to their employment practices. A caveat to our findings is that, while the measures of international exposure are often significant and have the hypothesized sign, they are not statistically significant in every model specification. In contrast, the effects of the industry dummies for internationally exposed sectors, particularly that for the wholesale and retail trade industries, are strong

240

Globalization and economic integration

and may reflect a time-invariant feature of employment practices in these industries.9 Finally, we tested H3 (Structural change and globalization). Is the specter of globalization really casting a dark shadow over the Japanese labor market? In Japan, as well as in most other countries, globalization is blamed for every imaginable economic malaise. However, it seems that globalization, at least as measured by the broadly defined KOF index, is not the ominous factor operating on labor markets as is commonly thought, at least as far as the labor market transitions for female workers are concerned. Arguably, the opposite may very well be the case. In our view, globalization may have actually served to offset some of the negative effects of the economic stagnation which affected all workers, part and full time, male and female.

DISCUSSION AND CONCLUDING COMMENTS Like nearly every other OECD country, Japan has experienced strong growth in part-time employment. In fact, the new jobs that Japan has added during its recovery from the lost decade have mainly been part-time jobs. Apart from some wistfulness about the capacity of the economy to once again generate high-quality and high-paying jobs, more practical concerns are normally raised about the low pay, poor training provisions, instability and general low quality of part-time jobs. The part-time jobs growth that Japan has experienced also raises questions about whether the new employees are, on average, less loyal and less productive than their predecessors, which might lower social efficiency and long-run growth. This chapter studied the labor market transitions of female workers for the period from 1993 to 2004; a period of unprecedented economic stagnation for Japan. It provided an excellent, albeit salutary, setting to investigate how workers are affected and how firms have adjusted their workforce during such trying economic times. To achieve this purpose we used data from the JPSC to investigate the determinants of labor market transitions. In particular, we focused on the transition from full- to part-time work. While a number of the customary demographic and sociological determinants such as marital status, children in care, and low levels of education are important, we also found that workers in trade-exposed industries were more likely to move to part-time jobs. The estimates in this chapter indicate that firms in the most internationally exposed sectors of the economy were the ones most under pressure to forgo traditional lifetime employment practices. In contrast, while all firms were under financial pressure during the economic slowdown, firms in non-tradable goods

Labor market transitions for female workers in Japan

241

sectors seemed better able to resist the need for more-flexible workplace arrangements. On the other hand, we found that the employment status in a previous period is important for what a person’s employment status will be in the next period. The women most likely to remain unemployed, or to not find full-time work, are those not in the labor force. While this result may not be particularly surprising, what it points to is the fact that transition probabilities to permanent full-time employment are significantly higher from part-time employment (compare, from unemployment). For reasons we are unable to completely fathom, this issue remains contentious in the literature, but the findings in this chapter support the research done for other countries (see, for example, Farber, 1999; Gaston and Timcke, 1999; Gaston and Kishi, 2005). In contrast to the specific findings for trade exposure, we find no evidence to support the commonly made claim that globalization is responsible for the changing nature of the Japanese employment relationship. In fact, globalization may have been a savior for female workers during the economic malaise of the 1990s in Japan. The growing integration of Japan into the world economy – as measured by an index capturing economic, social, and political globalization – has possibly softened the blow of a very severe recession. The rapid growth of ICT and high-tech service sector industries, inwards FDI, financial sector liberalization and political internationalization offer Japan’s workers full-time employment opportunities; they have not destroyed them.

NOTES *

The authors are grateful to the Institute for Research on Household Economics for agreeing to release confidential data. 1. In order to construct the indices of globalization, each variable (in Table 11A.6) is converted into an index on a zero to 10 scale. Higher values denote greater globalization. When higher values of the original variable indicate greater globalization, the formula ((Vi – Vmin)/(Vmax – Vmin)*10) is used for transformation. Conversely, when higher values indicate less globalization, the formula is ((Vmax – Vi)/(Vmax – Vmin)*10). The weights for the sub-indices are calculated using principal components analysis. The base year is 2000. For this year, the analysis partitions the variance of the variables used. The weights are then determined in a way that maximizes the variation of the resulting principal component, so that the index captures the variation as fully as possible. If possible, the weights determined for the base year are then used to calculate the indices for each single year back to 1970. Where no data are available, the weights are readjusted. See Dreher et al. (2008) for further details. 2. Some of the large movements are readily explicable, for example, outflows of FDI were enormous in the aftermath of the Plaza Accord, but then plummeted during the collapse of the bubble economy. 3. In the JPSC, apart from regular and part-time employees, dispatched employees (or

242

4.

5.

6. 7.

8.

9.

Globalization and economic integration temporary workers) and others are identified. However, the last and other types of employees are excluded from the following analysis. Some researchers have pointed out that it is difficult for married women to work as regular employees. In addition, female regular employees tend to have fewer children than part-time employees. Other researchers argue that both Japanese employment practices and an insufficient supply of childcare services hinder married women from simultaneously working as regular employees and rearing children. See Yashiro (1998), National Institute of Population and Social Security Research (2002) and Zhou (2003). Note that the lagged export and import ratios, FDI, and GDP growth rates are used. Obviously, the sample consists of respondents who were working (part time or full time) in period (t – 1); no industry-level data are available for non-working respondents for period (t – 1). Equation (11.4) is estimated using the STATA generalized linear latent and mixed model (gllamm) procedure (see Rabe-Hesketh and Skrondal, 2008). In Japan, transitions from full- to part-time employment “usually” proceed as follows: full-time employment to job quit to part-time employment. Changes in employment status within a given establishment are still uncommon. However, we do not distinguish job changers from those workers changing employment status within the same workplace. Whereas a binary logistic regression models a single logit, the proportional odds model models several cumulative logits. Therefore, since our ordinal outcome has three levels (0, 1, and 2), two logits are modeled, one for each of the following cut-off points: 0 versus 1, 2 and 0, 1 versus 2. The proportional odds assumption means that the two logits are equal. Based on 2004 data, it should be noted that, among the 10 industries covered by the JPSC, the wholesale and retail trade industries have the second highest import penetration (after mining); highest export orientation and second highest FDI (after mining).

REFERENCES Abe, M. (2006), “The effects of the deteriorated labor market on fertility”, in Y. Higuchi (ed.), Low Fertility and the Japanese Society and Economy, Tokyo: Nippon Hyoronsha (in Japanese). Dreher, A., N. Gaston and P. Martens (2008), Measuring Globalisation: Gauging Its Consequences, New York: Springer. Farber, H.S. (1999), “Alternative and part-time employment arrangements as a response to job loss”, Industrial Relations Section Working Paper 391, Princeton University, Princeton, NJ. Gaston, N. and T. Kishi (2005), “Labour market policy developments in Japan: following an Australian lead?”, Australian Economic Review, 38 (4), 389–404. Gaston, N. and T. Kishi (2007), “Part-time workers doing full-time work in Japan”, Journal of the Japanese and International Economies, 21 (4), 435–54. Gaston, N. and D. Timcke (1999), “Do casual workers find permanent full-time employment? Evidence from the Australian Youth Survey”, Economic Record, 75, 333–48. Higuchi, Y., K. Ohta and the Institute for Research on Household Economics (2004), The Recession in the Heisei Era and Japanese Women, Tokyo: Nippon Keizai Shinbunsha (in Japanese). Houseman, S. and M. Osawa (1995), “Part time and temporary employment in Japan”, Monthly Labor Review, October, 10–18.

Labor market transitions for female workers in Japan

243

Japan Institute for Labor Policy and Training (JILPT) (2006), The Labor Situation in Japan and Analysis: General Overview 2006/2007, Tokyo: JILPT. Kishi, T. (2003), Structural Changes in the Japanese Labor Market in the 1990s, Nagoya: Maruzen. Kitamura, Y. (2005), Panel Data Analysis, Tokyo: Iwanami-shoten (in Japanese). Koike, K. and T. Inoki (2003), College Graduates in Japanese Industries, Tokyo: Japan Institute of Labor. Matsuura, S. and Y. Shigeno (2001), Selection of Married Women and Household Savings, Tokyo: Nippon Hyoronsha (in Japanese). Ministry of Health, Labor and Welfare (MHLW) (2004), White Paper on the Labour Economy, 2004, Tokyo: MHLW. Ministry of Internal Affairs and Communications (MIAC) (2005), Labor Force Survey, Tokyo: MIAC. Nakamura, A., M. Nakamura and A. Seike (2004), “Aging, female and foreign workers, and Japanese labor markets: an international perspective”, in M. Nakamura (ed.), Changing Japanese Business, Economy and Society: Globalization of Post-Bubble Japan, Basingstoke, UK: Palgrave-Macmillan, pp. 107–43. Nakamura, M. and K. Horiuchi (2004), “The post-bubble Japanese business system and globalization: implications for Japanese society”, in M. Nakamura (ed.), Changing Japanese Business, Economy and Society: Globalization of PostBubble Japan, Basingstoke, UK: Palgrave-Macmillan, pp. 220–59. National Institute of Population and Social Security Research (2002), Childcare Services in an Aging Society with Low Fertility, Tokyo: Tokyo University Press (in Japanese). Organization for Economic Cooperation and Development (OECD) (1999), OECD Employment Outlook, Paris: OECD. Organization for Economic Cooperation and Development (OECD) (2007), OECD Employment Outlook, Paris: OECD. Rabe-Hesketh, S. and A. Skrondal (2008), Multilevel and Longitudinal Modeling using STATA, 2nd edn, College Station, TX: Stata Press. Rebick, M. (2005), The Japanese Employment System: Adapting to a New Economic Environment, Oxford: Oxford University Press. Shinozaki, T., M. Ishihara, T. Shiokawa and Y. Genda (2003), “A case for gaining consensus on wage differentials for part- and full-time workers”, Japanese Journal of Labor Studies, no. 512, 58–73 (in Japanese). Stratton, L.S. (1996), “Are ‘involuntary’ part-time workers indeed involuntary?”, Industrial and Labor Relations Review, 49 (3), 522–36. Toyoda, N. (2005), “The reason for an increasing share of non-regular workers”, Journal of Ohara Institute for Social Research, no. 556, 41–52 (in Japanese). Yashiro, N. (1998), “The economic factors for the declining birth rate”, Review of Population and Social Policy, 7, 129–44. Zhou, Y. (2003), “Childcare services and the female labor supply after childbirth”, in T. Tachibanaki and Y. Kaneko (eds), Structural Reforms for Enterprise-based Welfare Provisions, Tokyo: Toyo Keizai Shimposha (in Japanese).

244

Globalization and economic integration

APPENDIX 11A Table 11A.1 Wave

TABLES

The JPSC data

Year

Cohort A

1

1993

2 3 4 5

1994 1995 1996 1997

1500 entrants (aged 24–34 years) 1422 respondents 1342 respondents 1298 respondents 1255 respondents

6 7 8 9 10 11

1998 1999 2000 2001 2002 2003

1196 respondents 1137 respondents 1102 respondents 1059 respondents 1032 respondents 980 respondents

12

2004

944 respondents

Cohort B

Cohort C

500 new entrants (aged 24–27 years) 442 respondents 412 respondents 386 respondents 366 respondents 344 respondents 323 respondents 836 new entrants (aged 24–29 years) 311 respondents 724 respondents

245

Source:

Note:

9.5 51.5 0.0 4.8 20.6 0.6 5.6 0.0 1.8 0.0 3.8

0.0 1.7 0.0 3.7

0.0 2.0 0.0 4.2

9.0 53.9 0.0 5.4 24.2 0.5 5.5

1995

3

0.0 2.5 0.0 5.0

10.8 56.2 0.0 6.4 61.0 1.1 4.6

1996

4

0.0 2.6 0.0 4.8

11.0 57.6 0.0 6.6 60.5 1.1 4.3

1997

5

0.0 2.6 0.0 4.5

9.8 54.7 0.0 6.4 52.3 1.3 4.1

1998

6

0.0 2.1 0.0 4.4

8.7 55.2 0.0 6.5 42.3 0.5 3.9

1999

7

0.0 2.1 0.0 4.8

8.6 65.4 0.0 7.1 31.1 0.3 3.6

2000

8

0.0 2.3 0.0 5.0

9.0 66.0 0.0 7.6 32.4 0.4 3.5

2001

9

Cabinet Office (2007), National Accounts for 2006.

Imports are c.i.f.; figures for the finance and insurance industry include data from the real estate industry.

Agriculture and fisheries Mining Construction Manufacturing Wholesale & retail Finance & insurance Transportation & communication Utilities Services Government Average

1994

2

9.6 52.7 0.0 4.3 21.8 0.6 5.8

1993

Year

Industry

1

Data on import penetration

Wave

Table 11A.2

0.0 2.4 0.0 5.1

9.1 66.2 0.0 7.8 33.1 0.5 3.3

2002

10

0.0 2.1 0.0 5.3

9.5 68.4 0.0 8.1 30.1 0.6 2.9

2003

11

0.0 2.2 0.0 5.7

10.0 70.4 0.0 8.5 27.0 0.6 3.4

2004

12

246

Source:

Note:

0.2 0.2 0.0 9.3 6.7 0.8 8.9 0.1 1.3 0.0 5.2

0.3 9.4

0.1 1.2 0.0 5.2

0.1 1.0 0.0 5.1

0.3 8.0

0.2 0.2 0.0 9.0 7.3

1995

3

0.1 1.9 0.0 5.5

0.9 8.0

0.2 0.2 0.0 9.7 6.3

1996

4

0.1 1.9 0.0 5.6

0.7 8.0

0.2 0.2 0.0 10.6 9.1

1997

5

0.0 1.9 0.0 5.8

0.6 8.2

0.2 0.2 0.0 11.2 6.8

1998

6

0.0 1.8 0.0 5.5

0.3 6.9

0.2 0.2 0.0 10.9 5.3

1999

7

0.0 1.3 0.0 5.7

0.4 7.1

0.4 0.2 0.0 11.5 13.0

2000

8

0.0 1.9 0.0 5.6

0.3 7.3

0.8 0.2 0.0 11.1 12.8

2001

9

Cabinet Office (2007), National Accounts for 2006.

Exports are f.o.b.; figures for finance and insurance industries include data from the real estate industry.

Agriculture & fisheries Mining Construction Manufacturing Wholesale & retail trade Finance & insurance Transportation & communication Utilities Services Government Average

1994

2

0.2 0.2 0.0 9.1 7.1

1993

Year

Industry

1

Data on export orientation

Wave

Table 11A.3

0.0 2.0 0.0 6.0

0.3 7.4

0.4 0.2 0.0 12.2 15.4

2002

10

0.0 2.1 0.0 6.3

0.4 7.9

0.3 0.2 0.0 12.6 16.0

2003

11

0.0 2.3 0.0 6.9

0.6 8.9

0.3 0.2 0.0 13.5 19.2

2004

12

247

0.4 5.7 0.1 1.5 7.9 0.3 0.8 0.9 0.9

0.5 0.9

1.2 1.0

0.2 12.0 0.1 1.7 7.7 0.3 0.6

1995

3

0.7 1.2

0.3 20.4 0.1 2.2 9.1 0.2 0.6

1996

4

0.9 1.3

0.3 40.7 0.1 2.2 7.9 0.3 0.8

1997

5

0.6 1.2

0.1 15.2 0.1 1.7 8.9 0.3 0.8

1998

6

0.7 1.9

0.2 15.7 0.1 5.1 11.1 0.7 2.0

1999

7

0.4 1.6

0.2 11.3 0.0 1.9 9.7 3.3 9.7

2000

8

0.4 1.2

0.1 9.1 0.0 2.0 6.5 1.0 3.1

2001

9

0.5 1.3

0.1 7.4 0.0 2.6 10.5 0.3 1.1

2002

10

0.3 1.2

0.3 34.1 0.1 2.2 13.0 0.3 0.9

2003

11

0.4 1.5

0.1 45.9 0.1 1.5 4.6 1.0 2.7

2004

12

Sources:

Ministry of Finance (2008), Statistics on Foreign Direct Investment. Cabinet Office (2007), National Accounts for 2006.

Note: Figures are for the ratio of nominal FDI (both inward and outward) and nominal GDP. Data for the finance and insurance include that for real estate. For the utilities and government sectors, FDI is zero.

Agriculture & fisheries Mining Construction Manufacturing Wholesale & retail Finance & insurance Transportation & communication Services Total

1994

2

0.2 11.4 0.1 1.2 9.9 0.3 0.8

1993

Year

Industry

1

Data on FDI

Wave

Table 11A.4

248

7.5 −16.3 −18.8 −5.0 −1.5 4.6 7.7 2.5 2.7 0.1 2.5 0.8

−2.2 1.8 2.2 0.2

1.4 2.2 2.9 2.2

−4.7 −17.1 −4.7 −7.1 4.1 6.2 4.2 3.5

1995

3

7.8 5.8 2.4 3.2

1.5 5.2 7.1 −0.7 4.2 3.7 −1.3 1.0

1996

4

0.5 2.2 1.9 2.4

−4.2 −7.6 −4.3 −1.2 3.4 2.9 5.1 4.7

1997

5

3.1 3.9 2.2 −1.2

−1.6 −17.2 −5.1 −2.5 −5.9 −3.1 −1.6 2.4

1998

6

1.2 1.3 3.5 0.1

−4.0 0.9 −0.2 −1.5 1.7 −4.4 4.6 −1.7

1999

7

2.6 3.6 3.1 2.5

3.3 −5.1 12.3 −2.4 7.2 −3.0 2.4 2.2

2000

8

2.6 1.1 2.4 0.3

−4.4 −5.2 5.9 −2.6 −3.4 0.8 7.6 4.1

2001

9

0.8 0.4 1.9 −0.3

5.7 6.4 −3.0 −3.7 −2.6 −0.6 3.8 0.2

2002

10

0.4 1.7 1.2 2.7

−7.4 −11.5 2.5 −2.0 9.6 −1.5 3.4 1.6

2003

11

4.3 1.2 0.0 3.4

−13.7 4.8 −12.2 5.7 10.1 1.5 1.9 1.4

2004

12

Sources:

Cabinet Office (2005), National Accounts for 2004. Cabinet Office (2007), National Accounts for 2006.

Note: Figures for 2004 are obtained by multiplying the two series of the chain indices listed in National Accounts for 2004 and National Accounts for 2006; data for the finance and insurance industries include data for the real estate industry; figures for the fishery industry are separate from those for agriculture and forestry industries for real value added.

Agriculture Fisheries Mining Construction Manufacturing Wholesale & retail trade Finance & insurance Transportation & communication Utilities Services Government Average

1994

2

−11.5 −3.7 −9.0 −0.1 −3.7 2.0 4.1 1.8

1993

Year

Industry

1

Data on the real growth rate of industry value added

Wave

Table 11A.5

Labor market transitions for female workers in Japan

Table 11A.6

Components of the KOF index of globalization

Indices and variables

Weights

A. Economic globalization i) Actual flows Trade (percent of GDP) FDI, flows (percent of GDP) FDI, stocks (percent of GDP) Portfolio investment (percent of GDP) Income payments to foreign nationals (percent of GDP) ii) Restrictions Hidden import barriers Mean tariff rate Taxes on international trade (percent of current revenue) Capital account restrictions B. Social globalization i) Data on personal contact

Outgoing telephone traffic Transfers (percent of GDP) International tourism Foreign population (percent of total population) International letters (per capita)

ii) Data on information flows Internet hosts (per 1,000 people) Internet users (per 1,000 people) Cable television (per 1,000 people) Trade in newspapers (percent of GDP) Radios (per 1,000 people) iii) Data on cultural proximity Number of McDonald’s restaurants (per capita) Number of Ikea outlets (per capita) Trade in books (percent of GDP) C. Political globalization Embassies in country Membership in international organizations Participation in UN Security Council missions Source:

249

Dreher et al. (2008).

[36%] (50%) (16%) (21%) (23%) (19%) (22%) (50%) (24%) (28%) (28%) (20%) [38%] (29%) (14%) (8%) (27%) (25%) (27%) (35%) (20%) (24%) (20%) (14%) (23%) (37%) (40%) (40%) (20%) [26%] (35%) (36%) (29%)

250

Globalization and economic integration

Table 11A.7

Descriptive statistics for the estimation of equation (11.4) (origin industry characteristics) From FT or PT jobs Mean

Dependent variable Employment outcomes

Explanatory variables(lagged) Export ratio (%) Import ratio (%) GDP growth rate (%) KOF index FDI (%) Female unemployment rate (%) Married Age Children below 6 years Years of education Industry dummies (lagged) Primary Construction Manufacturing Wholesale & retail trade Finance & insurance Transportation machinery Services Note:

JPSC data, 7,530 observations.

Std dev.

NLF: 9.88% PT: 34.30% FT: 55.82% 4.861 10.935 1.419 58.870 2.937 3.950 0.586 32.567 0.317 13.645

5.025 16.609 3.282 3.595 3.913 0.807 0.493 4.986 0.622 1.388

0.014 0.045 0.163 0.230 0.084 0.026 0.316

0.116 0.208 0.370 0.421 0.277 0.160 0.116

Labor market transitions for female workers in Japan

Table 11A.8

251

Descriptive statistics for the estimation of equation (11.8) (destination industry characteristics) From FT, PT or NLF Mean

Dependent variable Employment outcomes Explanatory variables Export ratio (%) Import ratio (%) GDP growth rate (%) KOF index FDI ratio (%) Female unemployment rate (%) Married Age Children below 6 years Years of education Industry dummies Primary Construction Manufacturing Wholesale & retail trade Finance & insurance Transportation & communication Services Note:

JPSC data; 7,509 observations.

Std dev.

PT: 42.60% FT: 54.40% 5.186 11.080 1.532 59.816 2.891 4.143 0.617 33.736 0.298 13.633

5.389 16.441 3.388 3.704 3.804 0.687 0.486 4.950 0.604 1.384

0.012 0.044 0.161 0.234 0.082 0.027 0.318

0.110 0.205 0.368 0.424 0.274 0.162 0.466

252

Table 11A.9

Globalization and economic integration

Odds ratios of changing employment status (origin industry characteristics) From FT or PT Dependent variable: Odds ratios of yit, NLF = 2, PT = 1, FT = 0

Primary sector Construction Manufacturing Wholesale & retail Finance & insurance Transportation & communication Services KOF Female unemployment rate Married Age Kids Years of education Fixed part: thresholds k1 k2 Random part: variance y Log likelihood Number of observations Number of groups y^ 11 r5 y^ 11 1 p2 /3

Estimated coefficient

z-value

0.159 −0.768 0.475 0.563 0.323 0.256 0.561 −0.064 0.454 1.222 −0.032 −0.067 −0.174

0.52 −0.28 2.60*** 3.49*** 1.45 0.96 3.56*** −3.18*** 4.38*** 8.77*** −2.26** −1.01 −3.37***

−3.933 −0.553

−3.20*** −0.45

Odds ratio

1.608 1.756

1.752 0.938 1.575 3.394 0.969 0.840

9.394 (0.670) −5044.937 7530 1450 0.741

Note: Reference industries: government, utilities and others. ** and *** denote statistical significance at the 5% and 1% levels.

Labor market transitions for female workers in Japan

253

Table 11A.10 Odds ratios of changing employment status (destination industry characteristics) From FT (Dependent variable: Odds ratios of zjt, PT = 1, FT = 0

Primary sector Construction Manufacturing Wholesale & retail Finance & insurance Transportation & communication Services KOF Female unemployment rate Married Age Children Years of education Fixed part: thresholds l Random part: variance f Log likelihood Number of observations Number of groups f^ 11 r5 f^ 11 1 p2 /3

Estimated coefficient

z-value

0.470 −2.862 0.222 1.785 −2.167 −0.009 0.270 −0.133 0.594 2.596 0.131 0.167 −0.803

0.59 −6.53*** 0.62 5.65*** −5.46*** −0.02 0.89 −4.41*** 3.58*** 13.36*** 5.55*** 1.32 −11.15***

−9.178

Odds ratio

0.057 5.960 0.115

0.875 1.811 13.410 1.140 0.448

−5.56

14.905 (1.016) −2363.809 7509 1425 0.809

Note: Reference industries: government, utilities and others. *** denotes statistical significance at the 1% level.

12.

The effects of overseas operations on home employment of Japanese multinational enterprises Nobuaki Yamashita and Kyoji Fukao*

INTRODUCTION The controversy over the possible adverse effects of overseas production by multinational enterprises (MNEs) on home employment first arose in the United States in the late 1960s. It has increasingly gained attention in the policy circles of industrial countries in recent years with the growing importance of the international fragmentation of production (Lipsey, 1995; Harrison and McMillan, 2006). The possible substitution of home employment of MNEs with increased overseas production is known as “exporting jobs” (Kravis and Lipsey, 1988). It became the subject of heated policy debate in Japan under the label of “manufacturing hollowing-out” following a surge of Japanese foreign direct investment (FDI) outflow associated with the spread of production networks to low-cost countries in East Asia from the mid-1980s. In spite of the policy importance, only a few systematic empirical studies are available and they are based on FDI data at the industry level (Fukao, 1995; Fukao and Amano, 1998; Fukao and Yuan, 2001). There is virtually no evidence of how Japanese MNEs adjust home employment in response to changes in the production capacity of foreign affiliates at the firm level. This is certainly an area where studies on Japanese MNEs lag behind those of the US- and Swedish-based MNEs (Lipsey, 1995; Brainard and Riker, 1997a, 1997b; Braconier and Ekholm, 2000; Fors and Kokko, 2000; Desai et al., 2005; Harrison and McMillan, 2006). This chapter explores a panel dataset of operations of Japanese MNEs compiled from two unpublished firm-level surveys, the Basic Survey of Business Structure and Activity and the Basic Survey of Overseas Japanese Business Activity, collected by Japan’s Ministry of the Economy, Trade and Industry (METI) over the 1991–2002 period. The next section summarizes the existing empirical evidence. The third 254

Overseas operations and Japanese multinational enterprises

255

section describes the dataset used in the analysis, followed by the key patterns of trends of home and overseas operations of Japanese MNEs in the fourth section. The fifth section depicts the empirical model before explaining variable construction and the estimation methodology, and interpreting the results. The final section concludes.

THE EFFECT OF THE OVERSEAS OPERATIONS ON DOMESTIC OPERATIONS OF MNES In principle, there is little guidance from the theory of MNEs on the effects of overseas operations on home economic activity. One view argues that for a fixed level of overall production, including parent and affiliate production, any expansion in the overseas operations of MNEs simultaneously reduces domestic operations (the “substitution effect”). However, the substitution ignores the positive effects of overseas expansion on domestic activity. It is equally possible that increased overseas operations might enhance the scale of home economic activity due to better resource allocation and expanded overseas markets (the “scale effect”). Therefore, the net impact of increased overseas operations on home economic activity can be either positive or negative, depending on the magnitude of the scale and the substitution effects (Hanson et al., 2003). In early research, Kravis and Lipsey (1988) and Lipsey (1995) make initial attempts to examine the impact of foreign production on the home employment of US MNEs. A higher level of foreign affiliate production in developing countries is found to be associated with lower home employment for a given level of home production. Brainard and Riker (1997a) develop a more systematic analysis and estimate the foreign affiliate crosswage elasticities of parent firms’ labor demand for the 1983–92 period. They find evidence of a substitution relationship between foreign and domestic employment, although the degree of substitution is low. On the other hand, a strong substitution relationship is found among the various foreign affiliates of MNEs operating in developing countries. The evidence indicates that any employment substitution predominantly takes place between the foreign affiliates of MNEs operating in overseas locations rather than between parents and their foreign affiliates. If anything, parent firms adjusted employment very little in response to changes in foreign affiliate wages. Hanson et al. (2003) find that expansion in the sales of foreign affiliates of US MNEs raises the labor demand for their home operations, although the quantitative effect is small. This finding supports the hypothesis of a mild complementary relationship between increased overseas sales and

256

Globalization and economic integration

parent employment. Their second main finding is that the relationship between the parent and its foreign affiliates appears to depend on the skilled/unskilled labor costs of foreign affiliates. When the cost of skilled labor is lower in foreign affiliates, the demand for home labor appears to increase. This result suggests changes in the price of high-skilled employment in foreign affiliates tend to increase overall employment, both in foreign affiliates and in parent firms. On the other hand, where the cost of unskilled labor for foreign affiliates is lower, the US parent firms decrease the demand for home employment. Desai et al. (2005) find evidence of increased overseas operations of MNEs enhancing the scale of home operations. A 10 percent increase in the accumulation of foreign property plant and equipment is associated with a 2.2 percent increase in domestic net property plant and equipment. Similarly, a 10 percent rise in foreign employee compensation is associated with a 4 percent greater domestic employee compensation, and a 10 percent higher number of foreign employees with a 2.5 percent higher number of domestic employees. Overall, the results support the hypothesis that expanded operations of US MNEs’ foreign affiliates stimulate the domestic activity of US parent firms. Harrison and McMillan (2006) find strong evidence that the employment of foreign affiliates in developing countries substitutes for the home employment of parent firms in US manufacturing. However, the effect is quantitatively small. On the other hand, home employment in the US and employment in foreign affiliates in developed countries are found to be complementary. In other words, any decline in employment of foreign affiliates in developed countries leads to some contraction in employment of the parent firm in the US. Of the available studies on Japanese MNEs, a disproportionately large number have focused on the relationship between expanded overseas production and exports of home countries in Japan (for example, Fukao and Amano, 1998; Head and Ries, 2001; Kimura and Kiyota, 2006). Fukao (1995) makes an early attempt to examine the potential impacts of foreign affiliate production on domestic employment. Fukao and Yuan (2001) develop a 3–digit level of cross-industry data, concerning the impact of FDI on the employment growth rate over the period from 1989 to 1998. The unique feature of their study is the differentiation of FDI by investment motivation and region of the host country. They find that Japanese FDI in East Asia led to shedding around 600,000 jobs in the home country. They also find that market-oriented FDI in East Asia increased the amount of home-country employment.

Overseas operations and Japanese multinational enterprises

257

DATA The dataset is constructed by using the information on parent firms extracted from the Basic Survey of Business Structure and Activity (Kigyo Katsudou Kihou Chosa) and the information on their corresponding foreign affiliates from the Basic Survey of Overseas Japanese Business Activity (Kaigai Gigyou Katsudou Kihon Chosa).1 Both surveys are conducted by METI. For brevity, the former will henceforth be called the ‘METI Firm survey’ and the latter the ‘METI Foreign Affiliate survey’.2 The panel starts at 1991 when the first METI Firm survey was conducted. The second survey was undertaken in 1994 and it has been conducted yearly since then. The most recent data for both METI surveys available are for 2002 (note that 1992 and 1993 are missing since the METI Firm survey was not conducted in these two years). The panel data include parent firms that have both more than 50 employees and capital of more than ¥30 million. The industry classification is available at the 2–digit level of the Japan Standard Industrial Classification (JSIC). The panel set is unbalanced due to the ‘entry/exit’ of parent firms. Creating the matched panel data using the two METI surveys involves the following steps. First, information from both surveys is restricted to the manufacturing industry. This necessarily removed information on any foreign affiliates whose industry classification is not manufacturing. It is possible that this process underestimates the overseas operations of Japanese MNEs, since some parent manufacturing firms set up foreign affiliates in non-manufacturing industries. However, such downward bias is considered to be minimal. Next, the consistent 3–digit level of the manufacturing industry classification was assigned to each parent. Second, the two surveys are linked by using the permanent identifier assigned to each individual parent firm of the METI Firm survey to the same code reported by each individual foreign affiliate from the METI Foreign Affiliate survey. To ensure successful matching, cross-checking examines the name and the address of parent firms and the ownership structure. As a result, this procedure combines information on the overseas operations of Japanese MNEs and domestic economic activity. Third, following Hanson et al. (2003) and Harrison and McMillan (2006), sales weighted averages of foreign affiliate variables are constructed.3 This is essential because Japanese parent firms often own more than one foreign affiliate operating in multiple locations. Finally, about 1 percent of the data are excluded since some parent firms in the METI Firm survey report abnormally large or small values. Any parent firms are also dropped if at least one of the values of employment, sales, industry classification and identification code is missing.

258

Globalization and economic integration

The pooled data are disaggregated into four regions of host countries, East Asia, North America, the European Union (EU), and South America. A motivation for the regional disaggregation is to control for the level of the host country’s stage of development, the geographic proximity to Japan, and the possible characteristics of foreign affiliate production. Foreign affiliates of Japanese MNEs operating in developing countries (East Asia and South America) are most likely to be the vertical type of MNEs, whereas those in developed countries (North America and the EU) are more likely to be horizontal. Moreover, the postulated employment relationship between home and abroad critically depends on the location of foreign affiliates (Brainard and Riker, 1997a; Harrison and McMillan, 2006). One limitation of the panel set is that some parent firms disappear in one year in the data coverage and reappear in another, presumably because of varying sample restrictions imposed. This means that the entry and exit of firms in this survey do not necessarily correspond to the standard definitions of origin and termination of firms (Nishimura et al., 2005). The matched panel dataset also excludes small-scale Japanese firms, which do not meet the sample selection criteria of the METI Firm survey even if they do have foreign affiliates. However, their omissions do not affect the overall trends of MNEs’ operations.

PATTERNS AND TRENDS OF THE HOME AND OVERSEAS OPERATIONS OF MNES Home Operations Selected key indicators of home operations of MNEs, using the METI Firm survey data for the 1991–2002 period are summarized in Table 12.1. Total domestic sales by Japanese parent firms rose from ¥128 trillion in 1991 to ¥136 trillion in 2002 and the number of parent firms increased from 616 in 1991 to 1,114 in 2002. On the other hand, the employment figure contracted from about 2.2 million in 1991 to 2 million in 2002. This indicates that about 180,000 jobs were shed in the home employment of MNEs over the period. The share of parent firms of MNEs in total manufacturing accounted for an average of 6.6 percent over the 1991–2002 period. While this seems small, these parent firms contributed to the majority of economic activity to total manufacturing over 1991–2002. In 2002, parent firms of MNEs accounted for close to 55 percent of aggregate manufacturing output and over 40 percent of aggregate manufacturing employment as well as more

259

Source:

Note:

616 863 782 902 950 914 989 926 984 1144 907.0

Unit

Employment of:

128.8 124.2 128.5 143.2 142.7 131.4 138.6 144.8 139.3 136.8 135.8

¥ trillion 2245 2275 2267 2328 2292 2188 2261 2215 2121 2066 2225.8

’000

MNE parent firms

Number Sales of: of:

4.5 6.3 5.4 6.3 6.7 6.5 7.1 7.6 7.3 8.7 6.6

48.2 49.6 49.0 51.7 52.8 52.0 54.4 57.2 55.8 54.6 52.5

Number Output of of MNEs MNEs (%) (%)

Based on the METI database.

37.2 38.3 38.0 39.4 40.1 39.2 41.1 44.0 41.4 42.1 40.1

Employment of MNEs (%)

43.9 45.2 44.6 46.5 47.6 46.7 48.6 51.9 49.5 50.8 47.5

Worker earnings of MNEs (%)

47.1 49.3 48.7 50.5 51.9 50.4 52.4 55.7 53.5 52.9 51.2

Capital stock of MNEs (%)

– 86.2 80.8 82.5 82.5 82.8 84.8 86.5 83.2 86.1 83.9

Exports of MNEs (%)

Share in total Japanese manufacturing of:

Selected indicators of parent firms of Japanese manufacturing MNEs, 1991–2002

The survey data are not available for years 1992 and 1993.

1991 1994 1995 1996 1997 1998 1999 2000 2001 2002 Average

Year

Table 12.1

– 60.6 60.7 58.5 60.3 60.9 62.2 69.1 65.0 64.8 62.5

Imports of MNEs (%)

73.5 74.6 75.7 77.2 77.8 77.8 79.0 81.2 78.4 80.6 77.6

R&D of MNEs (%)

260

Globalization and economic integration

than half of the aggregate capital stock. Almost half of manufacturing workers’ compensation was also paid by MNEs. Not surprisingly, parent firms conducted the major proportion of international trade, accounting for over 80 and 60 percent of exports and imports, respectively, and contributed more than 75 percent of the research and development (R&D) expenditure in total manufacturing over the same period. These figures suggest that any effects on the operations of MNEs are likely to be deeply felt in Japan.4 Overseas Operations Table 12.2 summarizes the data on the key performance indicators of foreign affiliates of manufacturing MNEs. The number of foreign affiliates steadily increased from 2,656 in 1989 to over 10,000 in 2000, but the following two years (2001 and 2002) saw some declines in their number. Similarly, sales of foreign affiliates have achieved a fivefold increase between 1989 and 2000, although they dropped significantly between 2001 and 2002. Employment of foreign affiliates has expanded since 1989, and reached close to 3 million workers in 2002. This indicates that the number of workers employed in foreign affiliates is higher than the number of workers employed by the parent firms of MNEs (see Table 12.1). The data in Table 12.2 also indicate an increase in the size of foreign affiliates in terms of average employment and output over the period under study. On average, sales ratios both to Japan and to other countries have also been increasing since 1989, while the local sales ratio has remained stable at around 65–70 percent over the period. There is also some indication of upgrading in the technological capacity of the foreign affiliates of Japanese MNEs. Such foreign affiliates are heavily concentrated in general machinery, electronics, information and communication, and transport equipment industries (Table 12.3). There is an increasing share of sales in the transport equipment industry, growing from about 30 percent in 1989 to 37 percent in 2002. The similar expansion of sales can be seen for the information and communication industry. While the electronics machinery industry has been one of the most important participants in Japanese outward FDI, its sales share stagnated over the period. There was even a slight decline in the employment and sales share from 1989 to 2002. Foreign affiliates of Japanese MNEs were overwhelmingly concentrated in East Asian countries in the period under review (Table 12.4). About 60 percent of them were located in East Asia in 2002, up from 49 percent in 1989, with a corresponding employment share growth. Within East Asia, the rise of China as a destination for foreign affiliates is notable.

Overseas operations and Japanese multinational enterprises

261

Only 3.6 percent of Japanese foreign affiliates were operating in China in 1989, but that figure had jumped sharply to more than 19 percent by 2002. Accordingly, 1.8 percent of the employment share had grown to 22 percent in 2002. This geographical shift of overseas operation has facilitated the creation of international production networks by Japanese MNEs in East Asia (Kimura and Ando, 2005). Over the years, foreign affiliates of Japanese MNEs began to turn away from North America and the EU. For example, the employment share of foreign affiliates in the United States fell from 24 percent in 1989 to 16.5 percent in 2002; 22.5 percent were located in the US in 1989, but by 2002 this had fallen to 17 percent. Other developed country regions such as the EU experienced a slight decline or no change in foreign affiliate and employment share.

THE EMPIRICAL FRAMEWORK The baseline specification is based on a reduced form of labor demand function widely used in this literature (for example, Brainard and Riker, 1997a; Hanson et al., 2003; Harrison and McMillan, 2006): lnLiht 5 a0 1 b1lnwiht 1 b2lnQiht 1 b3lnrzht 1 b4lnR&Diht 1 b5lnIMPzht 1 b6lnLjft 1 b7lnQjft 1 b8lnPGDPft 1 fi 1 ␾ z 1 gt 1 ei,t

(12.1)

where subscripts i, h, z and t denote parent firm, home country, industry and time, respectively. Subscripts j and f represent foreign affiliate and host country; the symbol ln denotes the natural logarithm. The dependent variable (Liht) is the quantity of home employment and w and r refer to wage rate and user cost of capital for parent firms. Q stands for output either for parent firms or for foreign affiliates. R&D is the intensity of R&D expenditure in total outputs and IMP is the imports penetration ratio. PGDP is GDP per capita for host countries. f stands for the firm-specific effect and f and g represent industry and time-specific effects, respectively. Finally, e is a random error term representing other omitted influences. Estimation of equation (12.1) includes indicators of foreign affiliate employment and output. The estimated coefficient on these variables provides a test of the effect of overseas operations on home employment of MNEs. A complementary relationship between home and foreign affiliate employment is expected if the overseas operations of MNEs have scale effects. In this case, a positive sign is expected. On the other hand, a negative sign would indicate that home and foreign affiliate employment

262

Globalization and economic integration

Table 12.2

Year

Key indicators of foreign affiliates of Japanese manufacturing MNEs, 1989–2002 Number of

Employment of:

Sales of: Average Average sales employment

Average Average wage labor rates productivity

Foreign affiliate of manufacturing MNEs

1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 Level change % change

Unit

‘000

2,656 3,407 3,535 3,040 4,548 7,992 7,345 7,626 9,279 9,069 9,828 10,549 7,068 8,006 5,350

914 1,242 1,261 986 1,516 1,972 1,986 2,258 2,540 2,339 2,812 3,049 2,645 2,844 1,930

201.4

211.2

¥ trillion ¥ billion

Unit

¥ million

Output per worker

22.4 26.2 25.4 25.1 29.2 85.2 87.1 106.7 110.8 109.6 102.9 112.8 64.2 64.9 42.5

8.4 7.7 7.7 9.0 6.9 11.7 13.2 15.2 13.0 13.3 11.7 12.3 10.4 9.8 1.41

344.1 364.4 369.6 371.0 357.0 267.4 310.6 296.6 274.5 294.3 289.8 289.9 398.8 410.3 66.2

2.7 – – 3.0 – 3.1 4.2 3.7 4.2 4.1 3.6 4.1 3.7 3.4 0.7

24.5 21.1 20.8 24.3 19.4 43.6 42.6 51.2 47.4 45.2 40.2 42.4 26.1 24.0 −0.5

189.7

16.7

19.2

25.9

−2.0

Note: Average labor productivity is output per worker. R&D intensity is the ratio of R&D expenditures to output. Worker compensation and R&D expenditures for 1990, 1991, and 1993 are not contained in the original METI Foreign Affiliate survey. The METI Foreign Affiliate survey is available from 1989. Source:

Based on the METI dataset.

are substitutes for each other. We also examine the nature of the relationship between the scale of output by MNEs’ foreign affiliates and home employment. The wage rate of home employment is expected to be negatively related to home employment, given the downward-sloping labor demand curve. Product demand shocks both at home and in host countries are included (Brainard and Riker, 1997a; Braconier and Ekholm, 2000; Harrison and McMillan, 2006). These are measured by gross sales, time-specific dummies and the GDP per capita of host countries. The inclusion of the output scale of parent firms serves to hold the size of parent firms constant

Overseas operations and Japanese multinational enterprises

263

Average R&D intensity

Average local sales ratio

Average sales ratio to Japan

Average sales ratio to other countries

Average purchase ratio from local

Average purchase ratio from Japan

Average purchase ratio from other country

%

%

%

%

%

%

%

1.2 – – 3.7 – 0.6 1.0 3.0 1.1 1.1 1.2 4.2 3.3 1.7 0.5

74.8 69.5 67.0 75.1 65.8 68.7 74.5 70.8 68.7 71.6 72.2 70.7 68.6 66.4 −8.4

17.9 15.3 17.8 25.0 18.1 15.6 29.3 21.8 22.7 24.6 23.6 21.8 28.5 25.4 7.5

18.2 17.7 18.7 27.3 17.3 15.8 30.5 24.6 22.9 34.0 25.1 23.2 37.9 24.9 6.7

52.9 48.9 50.3 55.1 50.9 41.0 58.8 48.8 47.1 51.3 50.7 49.6 58.2 56.2 3.3

44.9 41.2 40.9 49.9 38.8 43.0 53.7 44.4 45.5 47.6 46.9 46.0 43.4 39.2 −5.7

13.4 11.1 12.5 22.9 11.4 16.3 21.1 25.9 23.7 32.1 25.8 23.3 32.8 20.0 6.6

41.7

−11.2

41.9

36.8

6.2

−12.7

49.3

(Kravis and Lipsey, 1988). Time-specific dummies are also included to capture shocks to the labor demand equation common to all firms. Similarly, foreign demand is proxied by the sales output of foreign affiliates as well as the GDP per capita of host countries. The positive impact of the product market in foreign countries should also translate positively to an increase in home employment (the market expansion effect). Labor demand also depends on the cost of capital services. The sign of the cross-factor price elasticity indicates the nature of relationship between labor and capital; a positive sign results if they are gross substitutes, and a negative sign is expected for a complementary relationship. The level of technology is proxied by the intensity of R&D and by unobserved firm- and industry-specific characteristics. The sign of R&D depends on the nature of technological progress. It can substitute for employment of parent firms since the introduction of technology may require fewer production workers. At the same time, technological

264

Source:

Unit

Based on the METI dataset.

5,834

8.5 13.6 6.4 3.0 10.6 6.3 16.9 12.1 3.4 19.4 100

1995

8,637

6.0 14.4 5.4 2.9 10.6 7.0 16.3 15.9 3.6 17.9 100

2002

8,014

Number of firms

6.9 13.6 6.5 3.2 11.0 7.2 15.6 12.1 3.8 20.1 100

1989

Firm

8.2 7.5 5.8 1.2 6.4 6.8 26.2 19.9 2.4 15.7 100

1995 5.5 6.3 3.9 1.2 5.5 7.3 27.1 23.7 3.1 16.3 100

2002

1,485

2,618

3,251

Employment in ’000

7.8 7.7 7.5 1.2 6.3 8.8 23.2 17.8 2.6 17.3 100

1989

Employment

3,2816

2.1 8.9 6.7 0.7 9.7 5.6 21.3 30.5 1.9 12.8 100

1989

50,591

Sales in ¥ bn

1.9 9.3 5.8 0.5 6.3 4.6 28.5 28.6 1.7 12.6 100

1995

Sales

Industry distributions of foreign affiliates of Japanese manufacturing MNEs (%), 1989–2002

Textiles Chemicals Primary metals Metal goods General machinery Electronics machinery Information & communication Transport equipment Scientific equipment Other manufacturing Total manufacturing

Table 12.3

72,295

1.3 10.2 3.9 0.5 5.5 4.6 24.2 36.7 2.2 10.9 100

2002

265

Source:

Unit

6,197

49.2 3.6 1.5 3.3 24.7 22.5 6.4 12.7 0.6 0.7 100

Based on the METI dataset.

East Asia China South Asia Oceania North America USA South America EU Eastern & Central Europe Africa World

1989

Units

9,042

56.8 16.4 1.2 2.9 20.2 18.6 4.7 12.0 0.8 0.4 100

1995

Number of:

8,265

59.0 19.3 1.7 2.4 18.4 17.2 3.7 11.6 1.1 0.4 100

2002

1,533

48.5 1.8 2.2 3.1 25.3 23.8 10.2 8.9 0.3 0.6 100

1989

’000

2,603

58.8 12.7 2.0 1.5 20.5 19.4 6.1 9.7 0.4 0.1 100

1995

3,285

64.6 22.2 1.9 1.0 17.3 16.5 3.7 8.5 0.8 0.5 100

2002

35,886

30.4 0.4 0.8 4.5 46.0 43.2 3.5 13.5 0.4 0.2 100

1989

foreign affiliates of manufacturing MNEs

Employment of:

Table 12.4 Geographical distribution of foreign affiliates of Japanese MNEs (%), 1989–2002

¥ bn

52,429

33.0 2.1 0.8 2.4 41.9 39.7 4.3 16.0 0.7 0.1 100

1995

Sales of:

73,886

32.6 6.1 1.1 2.4 41.9 38.9 3.4 16.5 0.9 0.4 100

2002

266

Globalization and economic integration

progress increases demand for skilled workers, engineers and IT-related personnel. Therefore, a priori the expected sign for R&D is ambiguous. Industry-specific effects are included to take into account industry-wide technological shocks. Another factor influencing labor demand is international competition. Tomiura (2004), Ito (2005) and Bernard et al. (2006) argue that manufacturing employment growth in developed countries is negatively related to rapid increases in imports from low-wage countries. To control for this effect, import penetration (IMP) at the industry level is included in the model. The expected sign of IMP is negative, because the entry of imports works against domestic manufacturing employment. Variable Construction Foreign affiliate variables Following the standard procedure in the literature, for each parent firm (i) in year t, a weighted average of employment of all foreign affiliates for a given parent firm is computed as: m

Li, j 5 a q j,iLj, f ,

(12.2)

j51

where subscript i refers to the individual parent firm and j denotes the corresponding foreign affiliate at multiple locations. The weight (q) is the share of foreign affiliate j in the aggregate foreign affiliate sales of parent firm i. The weighted averages of foreign sales and GDP per capita are computed in the same fashion. GDP per capita is taken from the World Bank Development Indicators. Parent firm variables Output (Q) is the reported total sales by parent firms. The nominal gross outputs are deflated by the industry wholesale price indices from the Bank of Japan.5 The home wage rate is computed by dividing the annual wages and salaries by the annual number of regular workers. Wages and salaries include bonus payments as well as non-wage compensation. The nominal wage series is deflated by the consumer price index and is also from the Bank of Japan. The user cost of capital (r) is proxied by the wholesale index of investment goods obtained from the Bank of Japan.6 The remaining variables for parent firms are obtained directly from the METI survey. R&D expenditure is the average value of R&D expenditure spent on knowledge creation and technological upgrading activity by firms, excluding R&D activities done by other firms. R&D intensity is computed by taking the share of R&D expenditure of the total sales of

Overseas operations and Japanese multinational enterprises

267

parent firms. The import penetration ratio (IMP) for each industry is computed taking the ratio of imports to apparent domestic absorption, that is, (Output + Imports) – Exports. Estimation Method Estimation of equation (12.1) takes into account the presence of the firmlevel and time-specific specific effects. Both the within-transformation and first-difference estimators of the fixed-effect model are employed to remove the firm-specific effects.7 The heteroscedasticity-robust standard errors clustering for each firm is used to compute the standard errors, which is also robust to serial correlation. The ordinary least squares (OLS) estimator is also performed to provide a benchmark comparison for results based on the other estimators. The most important estimation issue is the potential endogeneity problem for some of the explanatory variables in equation (12.1). MNEs make a decision on the overseas and domestic operations in terms of employment and output simultaneously rather than independently. Hence, a generalized method of moments (GMM) procedure is employed (Arellano and Bond, 1991). Instrument variables (IVs) for employment and output of foreign affiliates in a host country are the lagged employment output and wage rates of foreign affiliates, the percentage of the manufacturing labor force and the percentage of national income spent on education. The last two exogenous variables affect the supply side of labor in the host country and should only affect home labor market outcomes through their impact on the choice of employment in the host country.8 There is also concern about the possible correlation between the output variable (Q) of the parent firm and the error term in equation (12.1). The use of time dummies, industry- and firm-specific fixed effects to some extent alleviates the potential endogeneity problem (Roberts and Skoufias, 1997; Hasan et al., 2007). However, it is still possible that the output variable is correlated with some parts of the error term which are not covered by the fixed effects. In this case, the IV approach is employed to deal with this potential endogeneity problem on domestic output. Instruments include the lagged capital stock and the lagged intermediate inputs as well as lagged output. The first two variables are from the METI Firm survey. The capital stock is measured by the book value of the stock of tangible assets, such as capital, machinery and property. The nominal capital stock is converted into real terms using the wholesale price index of machinery and equipment. This deflator is obtained from the Bank of Japan. In the original METI Firm survey, there are no readily available data for

268

Globalization and economic integration

intermediate input expenditures. Hence, they are defined as the sum of the cost of goods sold and general administrative costs minus the wage bill, the rate of depreciation as well as the rental costs. Finally, in order to check the importance and nature of potential sensitivity bias due to the entry and exit of firms in the METI data, econometric estimations are performed for two sets of the data. One sample allows entry/exit (and reappearance) of parent firms at any time, hence, it is an unbalanced panel dataset. The other dataset is restricted to ‘surviving’ parent firms observed for the entire 1991–2002 period. Results Table 12.5 reports the summary statistics of variables in the regression analysis. Table 12.6a reports the estimation results of equation (12.1) based on the unbalanced panel dataset. Results based on the balanced panel set are reported in Table 12.6b for the purpose of comparison. Tables 12.7a to 12.7d present results for four regions – East Asia, North America, the EU, and South America. Time dummies and industry-specific dummies are included in all regressions performed except for OLS (Model 1), but the results are suppressed for brevity. Table 12.6a contains some evidence of a positive complementary relationship between foreign affiliate employment and output and home employment within Japanese manufacturing MNEs for the 1991–2002 period, but the magnitude of the estimated coefficient is very small. Model 2 (within transformation) suggests that a 10 percent increase in foreign affiliate employment leads to a 0.13 percent increase in home employment. Foreign affiliate output also indicates a statistically significant positive effect on home employment, although the magnitude of the estimated coefficient on this variable is also small. Additionally, foreign demand shocks, captured by GDP per capita, have no causal statistical relationship with the change in home employment (Model 2, Table 12.6a). The first-difference estimator (Model 3) in Table 12.6a also suggests a positive impact of expanded foreign affiliate sales on home, but not on foreign affiliate employment. The IV procedure in Model 4 in Table 12.6a improves the results for foreign affiliate employment, but the foreign affiliate sales lose statistical significance. The overidentifying test statistic for the instruments is 3.69, which does not reject the null hypothesis that all instruments are uncorrelated with the error term at a 5 percent significance level. In other words, the selected instruments are valid.9 Table 12.6b reports the same set of regression results, but the sample is restricted to the balanced panel data for 1991–2002. The estimation results

Overseas operations and Japanese multinational enterprises

269

Table 12.5

Summary statistics of selected variables used in regressions

Symbols of variables

Description

Obs.

Mean

Std dev.

Coeff. var.

L

Log parent firm employment Log wage rate Log output Log capital price Log R&D intensity Log import penetration Log foreign affiliates employment Log foreign affiliates sales Log GDP per capita of host countries

8432

6.81

1.36

0.20

8428 7837 8419 7179 7853

−2.84 5.36 4.57 −3.99 −3.56

0.33 1.71 0.06 1.31 1.03

8058

4.90

1.57

0.32

−4.91 10.53

8110

3.19

1.84

0.58

−9.48 10.41

7849

9.22

1.31

0.14

0.71 10.45

W Q K R&D IMP L (foreign)

Q (foreign) GDPP

Source:

Min

Max

3.91 11.32

−0.12 −5.65 −0.50 0.32 −1.13 11.21 0.01 4.35 4.65 −0.33 −10.81 −0.46 −0.29 −11.11 −0.66

Based on the METI dataset.

are somewhat different from the unbalanced panel data. The expanded overseas operations in terms of employment and output now have no impact on the home employment of parent firms. The only exception is that the estimated coefficient of foreign affiliate employment in Model 4 in Table 12.6b indicates a positive effect on home employment at a 5 percent statistical level. The contrasting findings between the balanced and unbalanced panel datasets suggest two interpretations. First, the estimation results obtained in the unbalanced panel data might be driven by the sample selection biases introduced by parent firms appearing and disappearing in the sample. Second, the surviving parent firms in the balanced panel dataset seem to adjust to changes in overseas operations without changing the scale of the home operations. Table 12.7 presents results for each region by dividing total foreign affiliates into four different regions: East Asia (Table 12.7a), North America (Table 12.7b), the EU (Table 12.7c), and South America (Table 12.7d). Employment and sales of foreign affiliates operating in East Asia are found to have no impact on home employment, despite the rapidly growing overseas operation in the region for the past 15 years. This suggests that international fragmentation of production in East Asia has no implications for Japanese employment adjustments. For North America,

270

Log foreign affiliates sales

Log foreign affiliates employment

Log import penetration

−0.257 (0.047)*** 1.013 (0.227)*** 0.689 (0.014)*** 0.154 (0.013)*** −0.033 (0.013)*** 0.088 (0.014)*** −0.051 (0.015)***

Model 1 OLS −0.117 (0.019)*** 0.367 (0.137)*** 0.135 (0.022)*** 0.022 (0.005)*** 0.001 (0.006) 0.013 (0.005)** 0.011 (0.005)**

Model 2 Within transformation −0.123 (0.015)*** 0.086 (0.103) 0.043 (0.014)*** 0.009 (0.003)*** 0.007 (0.003)* 0.004 (0.003) 0.006 (0.003)**

Model 3 First difference

−0.119 (0.016)*** 0.101 (0.136) 0.065 (0.064) 0.009 (0.005)* 0.009 (0.004)** 0.027 (0.014)** −0.009 (0.008)

Model 4 First difference, IVs

Dependent variable = log (home employment)

Labor demand by parent firms of MNEs, 1991–2002 (results from unbalanced panel data)

Log R&D intensity

Log output

Log capital prices

Log wage rate

Table 12.6(a)

271

0.081 (0.013)*** −2.657 (1.046)** 6170 1290 0.86 473.77***

0.002 (0.005) 4.161 (0.650)*** 6170 1290 0.30 170.11***

−0.004 (0.003) −0.048 (0.045) 4289 1023 0.07 9.06***

0.005 (0.005) −0.084 (0.052) 3691 952 − 7.43***

Note: Time- and industry-dummy variables are included for all estimations, but the results are suppressed here. Standard errors based on White’s heteroscedasticity correction clustered by individual firm are given in brackets, with statistical significance (two-tailed test) denoted as: *** 1 percent, ** 5 percent, and * 10 percent. The overidentifying test statistic for instruments used is 3.69, which does not reject the null hypothesis that all instruments are uncorrelated with the error term at the 5 percent significance level (c2q 54 = 9.49).

Observations Number of parent firms Adjusted R-squared F-statistics

Constant

Log GDP per capita of host country

272

Log foreign affiliates sales

Log foreign affiliates employment

Log import penetration

−0.539 (0.104)*** 0.769 (0.480) 0.687 (0.035)*** 0.195 (0.032)*** −0.027 (0.030) 0.088 (0.031)*** −0.040 (0.040)

Model 1 OLS −0.098 (0.035)*** −0.004 (0.255) 0.102 (0.040)** 0.018 (0.011) 0.010 (0.012) 0.012 (0.010) 0.014 (0.012)

Model 2 Within transformation −0.132 (0.027)*** −0.160 (0.164) 0.026 (0.016) −0.001 (0.004) 0.006 (0.006) 0.004 (0.006) 0.005 (0.008)

Model 3 First difference

Dependent variable = log (home employment)

−0.07 (0.028)** −0.064 (0.225) 0.162 (0.105) 0.005 (0.007) 0.017 (0.008)** 0.049 (0.019)** −0.005 (0.028)

Model 4 First difference, IVs

Labor demand by parent firms of MNEs, 1991–2002 (results from balanced panel data)

Log R&D intensity

Log output

Log capital prices

Log wage rate

Table 12.6(b)

273

0.107 (0.030)*** −2.461 −2.096 1459 178 0.90 165.72***

−0.011 (0.013) 7.076 (1.193)*** 1459 178 0.38 15.15***

−0.004 (0.004) 0.037 (0.034) 1254 177 0.08 4.63***

0.012 (0.010) 0.057 (0.043) 1070 168 – 4.25***

Note: Time- and industry-dummy variables are included for all estimations, but the results are suppressed here. Standard errors based on White’s heteroscedasticity correction clustered by individual firm are given in brackets, with statistical significance (two-tailed test) denoted as: *** 1 per cent and ** 5 per cent.

Observations Number of parent firms Adjusted R-squared F-statistics

Constant

Log GDP per capita of host country

274

(a) East Asia

Observations Number of parent firms Adjusted R-squared F-statistics

Constant

Log GDP per capita of host country

Log foreign affiliates sales

Log foreign affiliates employment

Log import penetration

Log R&D intensity

Log output

−0.258 (0.048)*** 1.043 (0.243)*** 0.737 (0.012)*** 0.173 (0.013)*** −0.026 (0.014)* 0.046 (0.014)*** −0.061 (0.015)*** 0.004 −0.011 −2.076 (1.145)* 4947 1058 0.88 458.61***

Model 1 OLS −0.123 (0.021)*** 0.303 (0.147)** 0.121 (0.025)*** 0.029 (0.006)*** 0.000 (0.006) 0.006 (0.005) 0.003 (0.004) −0.006 (0.005) 5.615 (0.732)*** 4947 1058 0.32 839.02***

Model 2 Within transformation −0.128 (0.017)*** 0.043 (0.100) 0.036 (0.015)** 0.011 (0.004)*** 0.002 (0.003) 0.001 (0.003) 0.002 (0.003) −0.002 (0.003) −0.004 (0.009) 3426 829 0.07 8.66***

Model 3 First difference

Dependent variable = log (home employment)

Labor demand by parent firms of MNEs by region, 1991–2002

Log capital prices

Log wage rate

Table 12.7

−0.125 (0.020)*** 0.125 (0.147) 0.111 (0.070) 0.015 (0.006)** 0.006 (0.004) 0.021 (0.011)** −0.009 (0.008) 0.000 (0.005) −0.011 (0.010) 2898 767 – 7.14***

Model 4 First difference, IVs

275

(b) North America

Observations Number of parent firms Adjusted R-squared F-statistics

Constant

Log GDP per capita of host country

Log foreign affiliates sales

Log foreign affiliates employment

Log import penetration

Log R&D intensity

Log output

Log capital prices

Log wage rate

Table 12.7

−0.276 (0.062)*** 0.425 (0.312) 0.688 (0.019)*** 0.16 (0.020)*** −0.014 (0.017) 0.097 (0.017)*** −0.052 (0.021)** −0.044 (0.031) 1.295 (1.416) 3996 812 0.84 252.06***

Model 1 OLS −0.121 (0.022)*** 0.200 (0.173) 0.100 (0.026)*** 0.015 (0.007)** 0.002 (0.007) 0.008 (0.007) 0.020 (0.008)** −0.048 (0.020)** 6.176 (0.852)*** 3996 812 0.25 386.67***

Model 2 Within transformation −0.121 (0.018)*** 0.082 (0.170) 0.031 (0.015)** 0.004 (0.003) 0.002 (0.004) 0.004 (0.005) 0.005 (0.005) −0.018 (0.010)* −0.004 (0.009) 2785 662 0.11 6.09***

Model 3 First difference

Dependent variable = log (home employment)

−0.102 (0.021)*** 0.058 (0.214) 0.063 (0.068) 0.004 (0.005) 0.003 (0.005) 0.001 (0.020) −0.002 (0.015) −0.002 (0.044) −0.011 (0.013) 2198 589 – 4.06***

Model 4 First difference, IVs

276

(c) EU

Observations Number of parent firms Adjusted R-squared F-statistics

Constant

Log GDP per capita of host country

Log foreign affiliates sales

Log foreign affiliates employment

Log import penetration

Log R&D intensity

Log output

Log capital prices

Log wage rate

Table 12.7

−0.201 (0.061)*** 1.155 (0.349)*** 0.695 (0.026)*** 0.21 (0.023)*** −0.013 (0.020) 0.074 (0.017)*** −0.039 (0.028) −0.04 (0.038) −1.606 (1.620) 2432 493 0.86 233.09***

Model 1 OLS −0.128 (0.027)*** 0.180 (0.216) 0.089 (0.032)*** 0.012 (0.008) 0.005 (0.008) 0.005 (0.006) 0.029 (0.015)** −0.063 (0.030)** 6.61 (1.107)*** 2432 493 0.29 2876.28***

Model 2 Within transformation −0.126 (0.022)*** 0.087 (0.189) 0.018 (0.020) 0.004 (0.005) 0.006 (0.005) 0.002 (0.004) 0.016 (0.008)** −0.04 (0.017)** 0.001 (0.011) 1715 399 0.07 3.85***

Model 3 First difference

Dependent variable = log (home employment)

−0.103 (0.024)*** 0.045 (0.272) 0.025 (0.085) 0.003 (0.007) 0.01 (0.007) 0.032 (0.018)* −0.011 (0.027) −0.02 (0.053) 0.002 (0.009) 1271 342 – 2.05**

Model 4 First difference, IVs

277

(d) South America

−0.482 (0.132)*** 0.677 (0.724) 0.731 (0.044)*** 0.156 (0.032)*** −0.021 (0.043) 0.023 (0.032) 0.045 (0.043) −0.122 (0.081) -0.344 (3.385) 764 154 0.88 108.16***

Model 1 OLS −0.242 (0.066)*** 0.842 (0.352)** 0.205 (0.077)*** 0.003 (0.015) 0.036 (0.013)*** 0.009 (0.012) 0.047 (0.019)** (0.069) (0.035)* 2.738 (1.819) 764 154 0.46 128.83***

Model 2 Within transformation −0.253 (0.072)*** 0.239 (0.261) 0.067 (0.075) 0.000 (0.007) 0.003 (0.010) 0.009 (0.008) 0.008 (0.008) −0.014 (0.019) 0.03 (0.012)*** 546 129 0.16 2.83***

Model 3 First difference

Dependent variable = log (home employment)

−0.176 (0.103)* −0.171 (0.505) 0.190 −0.209 0.001 (0.009) 0.010 (0.012) −0.002 (0.030) 0.016 (0.028) −0.017 (0.069) -0.038 (0.043) 320 96 – 2.14**

Model 4 First difference, IVs

Note: Time- and industry-dummy variables are included for all estimations, but the results are suppressed here. Standard errors based on White’s heteroscedasticity correction clustered by individual firm are given in brackets, with statistical significance (two-tailed test) denoted as: *** 1 per cent, ** 5 per cent, and * 10 per cent.

Observations Number of parent firms Adjusted R-squared F-statistics

Constant

Log GDP per capita of host country

Log foreign affiliates sales

Log foreign affiliates employment

Log import penetration

Log R&D intensity

Log output

Log capital prices

Log wage rate

Table 12.7

278

Globalization and economic integration

foreign affiliate operations do not have any impact on change in home employment, despite the country’s long history of foreign affiliate operations. By contrast, increased foreign affiliate output in the EU countries shows a small positive effect on home employment with a 5 percent significance (Models 2 and 3). In other words, the expansion of sales in the EU has scale effects on home employment in Japan. In summary, there is no clear-cut evidence of ‘exporting jobs’, despite the concerns expressed in public debates. Instead, some of the findings suggest that expanded overseas operations have actually helped to maintain the level of Japanese employment. The other determinants of labor demand by parent firms can be summarized as follows. The wage elasticity of labor demand consistently has the expected negative sign. This indicates that any increase in the wage rates results in a decrease in hiring more home workers. The own-wage elasticity is reported in the range of −0.1 to −0.5. The estimates are consistent with the findings reported in Hamermesh (1993) that the own-wage elasticity of labor demand usually varies from −0.3 to −0.6. The output elasticity is significant both in the within-transformation and the first-difference estimators. However, this result changes once we have corrected for the endogeneity problem in Model 4. The estimated coefficient of r (the user cost of capital) shows mixed results, making it impossible to infer whether capital and home employment are substitutes for, or complements to, each other. While the estimated coefficient for R&D intensity shows some significance in explaining changes in home employment in the unbalanced panel data (Table 12.6a), the statistical significance is lost when the analysis uses the balanced panel set (Table 12.6b). This could mean that the entry and exit of parent firms is highly correlated with their observable technological level. Import penetration is rarely significant. The signs of import penetration show an unexpected positive sign. This finding contrasts with US studies where there is a robust negative relationship to changes in parent employment with the degree of import penetration (Bernard et al., 2006; Harrison and McMillan 2006). Overall, we find that the expanded overseas operations of Japanese manufacturing MNEs have no adverse effects on home employment, in contrast to the “exporting job” concerns. By and large, there is no indication that overseas operations expand at the cost of workers in Japanese manufacturing over the 1991–2002 period.

Overseas operations and Japanese multinational enterprises

279

CONCLUSION This chapter has examined the hypothesis that an expansion of overseas operations of Japanese manufacturing MNEs reduces home employment within the MNEs’ operations. A labor demand equation is estimated by allowing for the effects of foreign affiliate employment and sales on home employment. The empirical findings are based on a newly constructed panel dataset, covering information for both home and foreign affiliates’ operations within matched manufacturing Japanese MNEs for the 1991–2002 period. Despite concerns expressed about the adverse effects of FDI, the evidence does not support the view that overseas operations expand at the cost of home employment in Japan. On the contrary, the findings suggest that overseas operations may have helped to maintain the level of home employment in Japanese manufacturing.

NOTES *

1. 2. 3.

4.

5. 6.

This study was conducted as part of the project on industry- and firm-level productivity in Japan undertaken by the Research Institute of the Economy, Trade and Industry (RIETI). The authors would like to thank RIETI and the Ministry of the Economy, Trade and Industry (METI) for making available the dataset to this study. We are grateful to Prema-chandra Athukorala and Sisira Jayasuriya for useful comments and suggestions. We also thank Noel Gaston for his careful editing of the earlier draft chapter and the seminar participants at the Asia-Pacific Trade Seminar (APTS) 2008, Yokohama National University, and Kobe University. Part of this research was undertaken when the first author was visiting the Institute of Economic Research at Hitotsubashi University, which provided a supportive and hospitable environment. During construction of this dataset, we referred extensively to Matsuura and Kiyota (2004) and the resources available from the RIETI website at http://www.rieti.go.jp/jp/ database/d02.html#01. For a detailed description of these two surveys, see Yamashita (2010). In principle, it would be possible to include variables for each host country where foreign affiliates potentially operate without aggregating foreign affiliate variables. However, this creates the problem of repeating the same information for the corresponding parent firms, making it difficult to interpret the estimated results (Brainard and Riker, 1997a). It would be particularly daunting to repeat the same home employment in the dependent variable. The MNE dominance in domestic manufacturing is not unique to Japan. For example, MNEs in the US manufacturing accounted for over 60 percent of total manufacturing sales, over 70 percent of total exports, almost 60 percent of manufacturing employment, and 82 percent of domestic R&D expenditure during 1982–99, although the number of US MNEs also looked small (Harrison and McMillan, 2006). Similar figures are reported for Swedish MNEs (Fors and Kokko, 2000). Compiled from the online database available at: http://www.boj.or.jp/type/stat/dlong/ price/cgpi/index.htm. They are available for the following industries: textile products, iron and steel, nonferrous metals, metal products, general machinery, electrical machinery, transport equipment, precision instruments, and other manufacturing industry products.

280

Globalization and economic integration

7. The first-difference estimator addresses the endogeneity problem more adequately, which is common to the firm-level data, as compared to the within-transformation estimator (Westbrook and Tybout, 1993). However, the first-difference estimator might suffer from the potential selectivity bias since it excludes firms not presenting in the period t and t − 1. It is also known that the first-difference estimator can exacerbate the bias due to measurement error by reducing the amount of systematic variation in the data (Griliches and Hausman, 1986). Therefore, the first-difference and within-transformation estimators are complementary estimation procedures. 8. These variables are taken from the online version of the World Bank Development Indicators for each host country, see http://devdata.worldbank.org/dataonline/. 9. The first-stage regression also finds a strong correlation between the selected instruments and the endogenous variables (the results are suppressed for brevity).

REFERENCES Arellano, M. and S. Bond (1991), “Some tests of specification for panel data: Monte Carlo evidence and an application to employment equations”, Review of Economic Studies, 58 (2): 277–97. Bernard, A.B., B. Jensen and P.K. Schott (2006), “Survival of the best fit: exposure to low wage countries and the (uneven) growth of US manufacturing plants”, Journal of International Economics, 68 (1): 219–37. Braconier, H. and K. Ekholm (2000), “Swedish multinationals and competition from high- and low-wage locations”, Review of International Economics, 8 (3): 448–61. Brainard, L. and D. Riker (1997a), “Are US multinationals exporting US jobs?”, NBER Working Paper 5958, National Bureau of Economic Research, Cambridge, MA. Brainard, L. and D. Riker (1997b), “US multinationals and competition from low wage countries”, NBER Working Paper 5959, National Bureau of Economic Research, Cambridge, MA. Desai, M., F. Foley and J. Hines (2005), “Foreign direct investment and domestic economic activity”, NBER Working Paper 11717, National Bureau of Economic Research, Cambridge, MA. Fors, G. and A. Kokko (2000), “Home country effects of FDI: foreign production and structural change in home country operations”, in M. Blomström and L. Goldberg (eds), Topics in Empirical International Economics: A Festschrift in Honor of Bob Lipsey, Chicago, IL: University of Chicago Press, pp. 137–62. Fukao, K. (1995), “Outward direct investment and jobs in Japan”, Monthly Journal of the Japan Institute of Labor, 37 (7): 2–12 (in Japanese). Fukao, K. and T. Amano (1998), “Outward foreign direct investment and manufacturing hollowing-out”, Keizai Kenkyu, 49 (3): 256–76 (in Japanese). Fukao, K. and T. Yuan (2001), “Japanese outward FDI and hollowing out”, RIETI Discussion Paper 003, Research Institute of the Economy, Trade and Industry, Tokyo. Griliches, Z. and J. Hausman (1986), “Errors in variables in panel data”, Journal of Econometrics, 31 (1): 93–118. Hamermesh, D.S. (1993), Labor Demand, Princeton, NJ: Princeton University Press.

Overseas operations and Japanese multinational enterprises

281

Hanson, G.H., R.J. Mataloni Jr. and M.J. Slaughter (2003), “Expansion abroad and the domestic operations of US multinational firms”, working paper, Tuck School of Business at Dartmouth, NH, available at: http://mba.tuck.dartmouth. edu/pages/faculty/matthew.slaughter/. Harrison, A.E. and M.S. McMillan (2006), “Outsourcing jobs? Multinationals and US employment”, NBER Working Paper 12372, National Bureau of Economic Research, Cambridge, MA. Hasan, R., D. Mitra and K.V. Ramaswamy (2007), “Trade reforms, labor regulations and labor demand elasticities: empirical evidence from India”, Review of Economics and Statistics, 89 (3): 466–81. Head, K. and J. Ries (2001), “Overseas investment and firm exports”, Review of International Economics, 9 (1): 108–22. Ito, K. (2005), “Import competition from low and middle income countries and firms’ growth: an empirical analysis using the ‘Basic Survey of Business Structure and Activity’”, RIETI Discussion Paper 028, Research Institute of the Economy, Trade and Industry, Tokyo. Kimura, F. and M. Ando (2005), “Two-dimensional fragmentation in East Asia: conceptual framework and empirics”, International Review of Economics and Finance, 14 (3): 317–48. Kimura, F. and K. Kiyota (2006), “Exports, FDI, and productivity: dynamic evidence from Japanese firms”, Weltwirtschaftliches Archiv (Review of World Economics), 142 (4): 695–719. Kravis, I.B. and R.E. Lipsey (1988), “The effect of multinational firms’ foreign operations on their domestic employment”, NBER Working Paper 2760, National Bureau of Economic Research, Cambridge, MA. Lipsey, R.E. (1995), “Outward direct investment and the US economy”, in M. Feldstein, J.R. Hines and G. Hubbard (eds), The Effects of Taxation on Multinational Corporations, Chicago, IL: University of Chicago Press: 7–33. Matsuura, T. and K. Kiyota (2004), “On the construction and utilization of panel data based on the ‘Basic Survey of Business Structure and Activity’: application to economic analysis and the data management issue”, RIETI Policy Discussion Paper 004, Research Institute of the Economy, Trade and Industry, Tokyo (in Japanese). Nishimura, K.G., T. Nakajima and K. Kiyota (2005), “Does the natural selection mechanism still work in severe recessions? Examination of the Japanese economy in the 1990s”, Journal of Economic Behavior and Organization, 58: 53–78. Roberts, M.J. and E. Skoufias (1997), “The long-run demand for skilled and unskilled labor in Colombian manufacturing industries”, Review of Economics and Statistics, 79 (2): 330–34. Tomiura, E. (2004), “Import competition and employment in Japan: plant start-up, shutdown and product changes”, Japanese Economic Review, 55 (2): 141–52. Westbrook, D. and J. Tybout (1993), “Estimating returns to scale with large, imperfect panels: an application to Chilean manufacturing industries”, World Bank Economic Review, 7 (1): 85–112. Yamashita, N. (2010) International Fragmentation of Production: The Impact of Outsourcing on the Japanese Economy, Cheltenham: Edward Elgar.

Index Abe, M. 229, 237 administered protection 82 Africa/African quotas 31 aging population (Japan) 16–17, 205–22 aggregation bias 174 Aghion, P. 192 agriculture 36, 50, 59, 64, 187, 190, 193, 220, 221 Ahmad, S. 195 Alter Chen, M. 193 Amano, T. 254, 256 Ando, M. 57, 261 anti-globalization 13, 69–84 Appelbaum, E. 178 arbitrage 4, 97 Arellano, M. 267 Ariff, M. 4 Arrow, K. (Arrow’s theorem) 70–71 ASEAN 98, 217 +1 FTAs 43, 61 +3 FTAs 46, 65 +6 FTAs 46, 65 Blueprint 58, 100 CER 43 -China FTA (ACFTA) 43, 49–51, 53, 55–6, 58, 63 Comprehensive Investment Agreement 58 Economic Community (AEC) 45, 57, 58–60 Economic Ministers (AEMs) 57 Framework Agreement on Services (AFAS) 57, 59 Free Trade Area (AFTA) 42–5, 48–61, 64–5 -India 43, 53 Investment Area (AIA) 58 -Japan FTA (AJCEP) 43, 50, 55–6, 58, 64, 68 -Korea FTA (AKFTA) 43, 49–51, 55–6, 58 Single Window 59

Asia-Pacific financial integration 13–14, 95–115 FTA networking 12–13, 42–65, 68 globalization (reversibility) 13, 69–84 globalization (vulnerability) 12, 25–39 globalization and integration 3–19 globalization and labor markets 16, 185–202 see also East Asia; Japan Asia-Pacific Economic Cooperation (APEC) 36–7, 46, 48, 62, 98 Asian Development Bank (ADB) 31–2, 38, 98, 101, 113, 186–90, 193, 195–6 Asian financial crisis 48, 95 contagion effect 4–6, 100 FTA policy and 60, 61 responses 13, 37–8, 39 Asian Monetary Fund (AMF) 38 asset holding 103–8 asset markets 97–8 asset prices 119–20, 122, 129, 137 Association of Southeast Asian Nations (ASEAN) see ASEAN Athukorala, P. 195, 196 Australia 42, 43, 186–9, 192 autoregression (long AR) 123–4, 127–9 Aw, B.Y. 174 Backus, D.K. 99 Balboa, J. 55–6 Baldwin, R. 12, 43, 60 Bank for International Settlements 142 Bank of Japan 266, 267 banking sector 4, 5–6, 37–8, 68 Barbera, F. 7 Barroso, José Manuel 36 Basel Accord (Basel II) 6

283

284

Globalization and economic integration

Basic Surveys (METI) 254, 257–9, 262–71 “Battle of Seattle” 32 Becerra, Xavier 36 Beladi, H. 154 benchmark bond 108 Bergsten, C.F. 35 Bernard, A.B. 266, 278 Betts, C. 154 Bhagwati, J. 17 bicycle theory 36 bilateral FTAs 42, 50, 55, 58, 63 bilateral holdings of assets 104–5 bilateral swap agreements 37–8 bilateral trade flows 108 Blonigen, B. 28 “Bloody Shirt” 81 Bloomberg 124 Bond, S. 267 bonds 104, 108, 160 Borjas, G. 74, 79 bound tariff lines 61, 62 Braconier, H. 254, 262 Brainard, L. 254, 255, 258, 261–2 Brecher, R. 209, 214, 221 Brecher–Diaz-Alejandro effect 214, 221 Bretton Woods institutions 31 Bubula, A. 6 Burgess, D.F. 175 business cycles 114, 229 call-money interest rate 142 Calvo-Pardo, H. 57 capital account openness 101–2 accumulation 99, 100, 102, 110, 111, 115, 154, 178 allocation 95, 99 arbitrage 4, 97 controls 6 cross-border flows 103–4, 111–12 markets 109–10, 112, 113 mobility 59, 209, 214 stock 267 capital asset pricing model (CAPM) 111 Carr, M. 193 Caselli, F. 112 casual employment 193, 223, 224, 231

CDAX 124, 142 Center for International Comparisons at the University of Pennsylvania 186 Central America FTA 36 certificate of origin 56 Chao, C-C. 154 cheaper foreign labor effect 17, 212–13, 216 Chen, J. 154 Chen, S. 189 Chiang Mai Initiative (CMI) 37–8, 48 China 3, 7, 82 Chinn, M.D. 102, 120 Choi, E.K. 154 Christiano, L.J. 154 citizen-agents 77, 78 civil rights 83 civil society 81 climate change 29 co-equal system 55 co-movement (returns on assets) 108 Cobb–Douglas production function 174, 210 Cobb–Douglas social utility function 209 Coe, D. 194 Cold War 27 Cole, H.L. 99 Common Effective Preferential Tariff (CEPT) 49–50, 51, 52 competition global 17, 223–41 performance 71, 153, 205, 208 policy 29, 59 Comprehensive Economic Partnership in East Asia (CEPEA) 45, 61, 63, 65 Consolidated International Portfolio Survey (CIPS) 103–4, 106–7 consumer price index 157, 159, 160, 169, 266 consumer protection 59 consumption 97, 100, 101 effects of tariffs 156–7, 162–3, 165–6 positive (effect) 17, 214 risk sharing 96, 99, 109, 114 smoothing 14, 98, 109–11, 114, 115 utility function of 157, 169

Index contagion effects 4–5, 100 Corley, M. 194 corruption 4, 82 costs -cutting, demand and 231, 239 transaction 25, 26, 27 covered interest parity 108 crime, migration and (Japan) 218–19 critical pluralism 80–81 Crook, C. 29 currency markets 6–8, 26–7, 95 exchange rate adjustment 14–15, 153–66 Davis, D.R. 174, 179 de facto economic integration 46–8 de jure economic integration 46–8, 101, 112 de minimis principle 56 de Rato, R. 5 debt 104, 108 deep integration 98 deindustrialization 4 democratic politics 69, 80–82, 84 deregulation 5 Desai, M. 254, 256 Devereux, M.B. 154 Diaz-Alejandro, C.F. 209, 214, 221 Diebold–Mariano test 137 Diewert, W.E. 176 diversification, incomplete 173–9 dividend/price ratio 124–5, 142 dividends (forecasting) 120–38, 142, 148–9 “dock-and-merge” strategy 46 Docquier, F. 191 Doha Development Agenda 12, 43 Doha Round 29, 30, 34, 35, 216 domestic operations (of MNEs) 254–79 domestic politics of globalization 70–72 domino effect (in regionalism) 61–3 Dornbusch–Frankel model 150–52 Dreher, A. 8, 227, 249 duty-drawback system 51, 54 Early Harvest Program 49, 51 earnings (forecasting stock prices) 120–24, 126–38, 142, 146–7

285

East Asia 12–13, 42–65, 68 FTA networking 43–60, 68 future trends 65 labor demand 269, 274 East Asian FTA (EAFTA) 45, 63, 65 Eatwell, J. 6 economic globalization 8 economic growth 3–4, 8–12 labor market and 16, 186–7, 189, 192 skills and 191–4 economic integration Asia-Pacific (prospects/risks) 3–19 de facto 46–8 de jure 46–8, 101, 112 extended East Asia 12–13, 42–65, 68 economies of scale 99 Edmonston, B. 79 education 27–8, 188, 193, 217, 267 employment status and 229, 232, 234, 236–40, 250–53 immigration and 71, 74, 78, 195 Edwards, J. 83 Eichengreen, B. 104, 105 Ekholm, K. 254, 262 Electronic Individual Action Plan 62 employees female see female workers full-time see full-time employment low-skilled 185, 190–96, 198 part-time see part-time employment skilled 15–16, 59, 79, 173–9, 256 unskilled 15–16, 173–9, 194, 195, 208, 222, 256 see also labor; labor market; unemployment; wages employment dividend 187 home (Japan) 254–79 lifetime 17, 240 shares (by sector) 187–8, 190 status (odds ratio) 235–8, 252–3 see also casual employment; fulltime employment; full employment; part-time employment; underemployment; unemployment Employment Status Surveys 228 Endriga, B. 198

286

Globalization and economic integration

engagement (Asian financial markets) 13, 96, 101, 102–8, 113, 114 Engel, C.M. 137 Equal Employment Opportunity Law 229 equities 104 markets 14, 119 prices 108, 138–9, 142 Estevadeordal, A. 55, 57 European Central Bank (ECB) 98 European Union 30–32, 36, 108–9, 114, 258, 269, 276, 278 exchange rate 6, 7, 100 adjustment (Japan) 14, 153–66 effects of tariffs 155, 160–66 forecasting 14, 119, 138–9 M–R model 14, 119–20, 122–4, 127–9, 134, 137–9, 150–52 random walk model 14, 119 exportables 209, 211 exports 4, 246 FTA utilization 51–4, 63 Japan 220, 221, 226 expropriation risk 95, 112 factor endowment 78, 194, 208–9 factor prices 208–9, 211 Falk, M. 175 Farber, H.S. 224, 241 Feenstra, R.C. 173, 178 female workers labor market transitions (Japan) 17, 223–41 labor shortages (Japan) 17, 205, 219, 220, 221, 222 fertility rate (Japan) 16, 17, 205 Feyrer, J. 112 final prediction error (FPE) 124 financial integration (Asia) 13–14, 95–115 benefits and costs 98–101 definitions 96–8 engagement 13, 96, 101–8, 113, 114 forecasting international prices 119–39 globalization and 3–19 measurement of 101, 108–9 openness 13, 96, 97, 100–102, 109, 113

policy implications 109–14 Financial Sector Assessment Program 6 financial services/sector 59, 98 Fiore, Q. 3 firm-specific effects 267 firm behavior, tariffs and 156–8 fiscal policy 100 Fisher, S. 6 Flood, R.P. 97, 108 forecasting international financial prices 14, 119–39 foreign affiliates (of MNEs) 260, 262–6, 268–9, 279 foreign assets/liabilities 103 foreign direct investment (FDI) 3, 4, 8, 99, 111 FTA networking 13, 48, 59–61, 63 Japan 18, 220–21, 226, 234, 236–8, 241, 247, 254, 256, 260 Fors, G. 254 forward rate (MR) 127, 150 Frankfurt Stock Exchange 142 Freeman, G. 78 Frenkel–Bilson model 150–51 FTAAP (APEC-wide FTA) 46 FTAs (free trade agreements) bilateral 42, 50, 55, 58, 63 extended East Asia 12–13, 42–65 impact of/attitudes to 32–6 utilization 50–54 FTSE All-Share Index 124, 142 Fukao, K. 254, 256 full-time employment 17, 223–4, 226, 228, 230, 232–3, 235–41, 250–53 full employment 161 Japan 153–4, 156, 160–62, 165, 170 fundamental-based models 14, 119–39, 142–52 G-7 countries 32 G-20 countries 32, 36 Gardner, B. 208 Gaston, N. 8, 223, 224, 230, 241 gateway communities 83 GDP per capita 262–3, 266 growth 186–7, 197–8 General Agreement on Tariffs and Trade (GATT) 27, 29, 61

Index General Agreement on Trade in Services (GATS) 57, 58 general exception list (GEL) 49 Ghose, A. 193 Ghosh, S. 104, 112 Gimpel, J. 83 global competition (labor market transitions) 17, 223–41 Global Financial Dataset 142 Globalisation and Development Centre (GDC) 10, 12, 18–19 globalization Asia-Pacific (prospects/risks) 3–19 challenges 12, 25–8 complexity 29–32 definitions 25 domestic politics of 69–84 economic growth and 3–4, 8–12 effects (snapshots) 8–12 Golden Age 25, 27, 38–9 KOF index 8–10, 17, 226–7, 234–6, 238, 240, 249–53 labor market transitions 17, 223–41 labor markets (Asia-Pacific) 16, 185–202 legitimacy 29–32 measurement of 226–7 pluralism 29–32 post-pax American 36–8 reaction against 12, 25–39 reversibility 13, 69–84 role of politics 69–84 gold standard 25 Golden Age of globalization 25, 27, 38–9 goods final 70, 173, 175 intermediate 15, 16, 173–5, 178, 179 non-traded 17, 208–9, 211–12, 214–15, 226, 230–31, 239 tradable 230 trade in (liberalization) 49–57 Gordon, M.J. 122 Gordon growth model (stock values) 120, 122, 127–9, 132–8, 146–9 Gorman–Diewert adaptation (Hotelling’s lemma) 176

287

Goto, J. 208, 221 Gourinchas, P-O. 99 governance problems 30–32 governments 158, 217–18 fiscal policy 100 monetary policy 14, 100 Granger–Newbold technique 134 gravity model/approach 105, 115 Great Depression 25, 153 Greenaway, D. 76 Grossman, G. 71 group politics model 69–71, 74–81, 83–4 group pressure models 72, 76, 77, 79 Hamermesh, D.S. 278 Hannan and Quinn information criterion (HQIC) 124 Hanson, G.H. 173, 178, 255, 257, 261 Harris–Todaro-type unemployment 154 Harris Interactive 36 Harrison, A.E. 254, 256–8, 261–2, 278 Hasan, R. 267 Hau, H. 154 Hawley–Smoot tariff 80, 82, 153 Hayakawa, K. 54 Head, K. 256 Heckscher–Ohlin (HO) mechanism 173, 178 Heckscher–Ohlin–Samuelson (HOS) framework 208 Helpman, E. 71 Higuchi, Y. 229 Hijzen, A. 175 home operations (of MNEs) 258–60 Hooper–Morton model 150–52 Horiuchi, K. 231 host country (welfare effect of immigration) 211–17 Hotelling’s lemma 176 household behavior, tariffs and 156–8 Houseman, S. 223–4 housewives (employment status) 232–3 Hurwicz bias 138 ideal bureaucrat 70 ideal state 72

288

Globalization and economic integration

Ikeda, S. 154 Imbs, J. 99, 100, 101, 112 immigration effects on welfare of host country (Japan) 16–17, 211–17 legislation (USA) 83 politics of 13, 69, 71, 73–4, 76–80, 82–4 import substitution 48 import tariffs see tariffs importables 209–10, 214 imports dutiable 54, 72 duty-drawback system 51, 54 Japan 15, 63–4, 153, 154, 220, 226, 237, 245, 261 most-favored-nation 51, 54, 55, 57, 61 penetration 226, 237, 245, 261, 266, 267, 278 trade barrier effect 17, 208–10, 212, 214, 215–17 wage inequality and 15–16, 173–9 inclusion list (IL) 49 income earnings (stocks) 120–24, 126–38, 146 KOF index by 9, 10 negative (effect) 17, 215 see also wage(s) India 3 individual characteristics (labor market transitions) 232, 239 industry data (female labor) 225–6, 245–8 distribution (MNEs) 260, 264 informal sector 193 information and communication technology (ICT) 3, 17, 241 information technology agreements (ITAs) 48, 51, 61 infrastructure 4, 59, 95, 112 Inoki, T. 228 insider trading 113 institutional infrastructure 4, 59 instrumental variables (IVs) 134–6, 267, 268, 270–71 integration deep 98

globalization and (prospects and risks) 3–19 see also economic integration; financial integration (Asia); regional integration intellectual property rights 29, 58–9 intensity index 104–5, 107 interest-group model 69–71, 74–81, 83–4 interest rates forecasting and 14, 120, 122, 124, 127, 130–31, 142 Japan 14, 154 intermediate goods 15, 16, 173–5, 178–9 intermediation functions 110 internal labor market 228 international financial institutions (IFIs) 31 International Financial Statistics (IFS) 103, 142 International Labour Organization (ILO) 187, 189, 192, 193, 197, 201–2 International Monetary Fund (IMF) 5, 6, 8, 101, 194 CIPS 103–4, 106–7 problems/challenges 12, 27, 28, 31, 32, 37, 38 internationalization (positive externality) 217 investment 98–9, 109 bilateral treaties 58 in infrastructure 59, 95 portfolio 4, 8, 59, 104, 105–8, 111 see also foreign direct investment Ito, H. 102 Ito, K. 266 Ito, T. 6 Japan-Indonesia EPA 58 Japan-Malaysia FTA 64 Japan-Mexico FTA 64 Japan-Singapore FTA 64 Japan 36, 104 aging and migration 16–17, 205–22 economic integration 61, 63, 64, 65 FTAs 43, 50, 55–6, 58, 61, 63–5, 68 labor market transitions 17, 223–41 MNEs 18, 254–79

Index protection 15, 50, 64, 153, 154 trade policy 14–15, 153–66 unemployment 15, 153 Japan External Trade Organization (JETRO) 44, 51–4 Japan Institute for Labor Policy and Training (JILPT) 228 Japan Standard Industrial Classification (JSIC) 257 Japanese Panel Data on Consumers 224–5, 228–30, 235, 240, 244, 250–51 Jeanne, O. 99 Johnson–Reed Act (1924) 83 Jones, G. 195 juggernaut effect 63–4 Kawai, M. 55 Kennedy Round 29 Khalid, A. 4 Kim, S. 99, 101, 108, 109, 111, 113 Kimura, F. 64, 256, 261 Kindleberger, C. 26 Kindleberger spiral 26 Kishi, T. 223, 224, 228, 230, 241 Kitamura, Y. 229, 237 Kiyoto, K. 256 Koebel, B.M. 175 KOF index of globalization 8–10, 17, 226–7, 234–6, 238, 240, 249–53 Kohli, U.R. 175, 178 Koike, K. 228 Koizumi, Junichiro 61, 63 Kokko, A. 254 Korea (FTAs) 35, 43, 49–51, 55–6, 58, 63, 65 Kravis, I.B. 254, 255, 263 Krugman, P. 14, 175 Kull, S. 35 Kuno, A. 64 labor 99 capital–labor ratios 112 cheap foreign 17, 212–13, 216 demand (Japan) 263, 266, 270–79 mobility 185, 195–6, 198, 209, 214 productivity 206, 219–20, 224 see also employees; female workers; skills; wages Labor Force Survey (MIAC) 224, 227

289

labor market 15, 77–8, 79 globalization and 16, 185–202 internal 228 mobility (Asia-Pacific) 195–6 transitions (females) 17, 223–41 Lane, P.R. 103, 113 Latin American crisis 5 least-absolute deviations (LADs) 134, 136 Lee, J.W. 99–101, 103–5, 108, 111, 112 legitimacy (globalization) 29–32 Lehman Brothers 3 less-developed countries 15, 18 liberalization 42 attitudes to 32–7 coverage 49–50 financial sector 3, 97 migration and 17, 220, 221 multilateral 43, 57 other policy modes 57–60 of trade in goods 49–57 unilateral 43, 48, 61 lifetime employment practices 17, 240 Lipsey, R.E. 254, 255, 263 liquidity preference 15 liquidity trap 14, 161 literacy rates 188, 193 Liu, H. 99 lobbying 78, 80, 82, 84 low-skilled workers 185, 190–96, 198 McDonald’s 9, 249 MacDougall, G.D.A. 99 McKinnon, R.I. 4 McLuhan, M. 3 McMillan, M.S. 254, 256–8, 261–2, 278 Mandelson, Peter 36 Mansoorian, A. 154 manufacturing sector 18, 51, 64, 173, 194, 223–4, 226, 231, 234–5 MNEs 256–7, 258, 260–61, 262–3 marginal productivity of labor 206 market adjustment (to tariffs) 158–9 market expansion effect 263 marriage/marital status 229–30, 232, 234, 236–9, 240 Marshall–Lerner condition 155, 163, 172 Matsuura, S. 230

290

Globalization and economic integration

Mattoo, A. 29 Mayda, A.M. 75–6, 77 Mayer, W. 71 Medalla, E.M. 55–6 median voter 75, 77 Meese, R. 14, 119 Meese–Rogoff model 14, 119–20, 122–4, 127, 128–9, 134, 137–9, 150–52 Meiji Restoration 220 Meissner, C. 7 migration aging and (Japan) 16–17, 205–22 alternative measures 219–21 economic impact 206–17 labor mobility 195–6 social effect 217–19 traditional view 206–8 Milesi-Ferretti, G.M. 103, 113 minimum wages 16, 196–202 Mishkin, F. 5 monetary policy 14, 100 monopoly power 73 Montiel, P.J. 109 Morgan, L. 208 most-favored-nation (MFN) tariffs 51, 54, 55, 57, 61 multilateral liberalization 43, 57 multilateralizing regionalism 12, 13, 42–3, 46, 57, 65 multinational enterprises 48, 60–61, 63 Japan 18, 230, 254–79 Mundell, R. 7 Munshi, F. 195 Mussa, M. 6 mutual recognition agreements 57, 59 mystery of missing protection 76, 78 Nakamura, M. 231 National Academy of Sciences 79 National Association of Securities Dealers Automated Quotations 97–8 National Income and Product Accounts of the US (NIPA) 177 networks FTA 12–13, 42–65 production 42–3, 47, 60, 65, 254, 261 Neumark, D. 201 Nikkei 225 124, 142

Nishimura, K.G. 258 non-stochastic discount rate 121 non-tariff barriers 59, 76–7, 209–10 non-traded goods 17, 208–9, 211–12, 214–15, 226, 230–31, 239 noodle bowl phenomenon 54–5 North American FTA (NAFTA) 32, 34, 36 Obama, Barack 34, 35, 36 Obstfeld, M. 99, 103, 109, 110, 154 OECD 4, 9–11, 15, 46, 99–100, 109, 191–3, 224, 227, 231 off-shoring 47, 194 Ono, Y. 14–15, 153, 154, 156, 161, 170 Oomes, N. 7 openness (financial markets) 13, 96, 97, 100, 101–2, 109, 113 Osawa, M. 223–4 Ostry, J.D. 113 Otker-Robe, I. 6 outcomes of financial flows 96 output shocks 110, 114 outsourcing 15, 47, 174, 178, 223, 230, 231 overnight bank rates 108 overseas operations (Japanese MNEs) 18, 254–79 Pacific Four (P4) 46 parent firm variables (MNEs) 266–77 Park, Y.C. 104, 105 Parreñas, J.C. 112 part-time employment 17, 223–4, 226–33, 235–41, 250–53 Pelosi, Nancy 36 Pew Center for the People and the Press 32–4 Pew Global Attitudes Project 28 pluralism 29–32, 80 plurilateral FTAs 42, 46, 55, 63 Political Action Committee (PAC) 77 political economy approach (regional integration) 113–14 models 69, 70, 71, 72–4 of regionalism 43, 60–64 politics of anti-globalization 69–84 democratic 69, 80–82, 84

Index domestic 70–72 of globalization 13, 70–74, 80–83 group (model) 69–71, 74–81, 83–4 Poonpatpibul, C. 103, 104, 109 portfolio investment 4, 8, 59, 104, 105, 106–8, 111 post-industrial economy 29 poverty 16, 189, 196, 201, 202 preferential trade agreements 113 price/earnings ratio 124, 126, 142 prices of assets 119, 120, 122, 129, 137 consumer price index 157, 159, 160, 169, 266 effect of tariffs (Japan) 153, 164–5 equity 108, 138–9, 142 of factors 208–9, 211 forecasting 119–39, 142–3, 164, 165 production networks 42–3, 47–8, 60, 65, 254, 261 specialization 173–4, 179 theory approach 175 productivity 206, 219–20, 224 protectionism 25–6, 55, 75–7 Japan 15, 50, 64, 153, 154 USA 13, 36–7, 71–2, 73, 82 pull factor (employment) 187–8 push factor (employment) 187–8 quotas 31 race to the bottom 60–61 Ramsey equation 158 random walk model 14, 119–21, 123–4, 127–37, 144–52 Ravallion, M. 189 Rebick, M. 223, 224, 231 Reciprocal Trade Agreements Act 71 referendum model 75 regional effects of globalization 9–11 regional integration 13–14, 61, 95–115 barriers to 112–13 political economy approach 113–14 welfare gains from 110–12 regional value content (RVC) test 55–6 regionalism FTAs 12–13, 42–65 multilateral 57, 65 research and development 259–63, 266

291

Ries, J. 256 Riker, D. 254, 255, 258, 261–2 risk-return profile 95, 114 risk sharing 96, 99–100, 109, 113–14 Roberts, M.J. 174, 267 Rodrik, D. 75–6 Rogoff, K. 14, 103, 109, 110, 119, 120, 154 Roldos, J.E. 154 Rongji, Zhu 63 root mean square forecast error 123, 127–30, 132–8, 144, 146, 148, 150–52 Rose, A.K. 97, 108 Rossi, B. 138 rules of origin 54–6, 59 Russian crisis 5 S&P 500 124, 142 Saget, C. 197, 198, 201 Samuelsonian social planner 70, 72 savings 95, 99, 109, 114, 115 Schattschneider, E.E. 80–81 Schiff, M. 208 Schott, P.K. 174, 179 Schwarz information criterion 124 seasonal dummies, OLS with 134, 137 self-employment 193 sensitive/highly sensitive list 49 services 59, 223, 224 GATS 57, 58 Shigeno, Y. 230 Shiller, R. 142 Shin, K. 99, 100, 111, 112 Shinozaki, T. 224 Short-Term Liquidity Facility 38 single currency 6 skills low-skilled workers 185, 190–96, 198 skilled labor 15–16, 59, 79, 173–9, 256 unskilled labor 15–16, 173–9, 194, 195, 208, 222, 256 wage inequality and 15–16, 173–9 Skoufias, E. 267 Slaughter, M.J. 173 Smith, J. 79 Smoot–Hawley Tariff Act 80, 82, 153 SNQ 175, 178, 179 social capital 208

292

Globalization and economic integration

social effect of migration 217–19 social globalization 8–9 social planning 70, 72, 99 social utility function 209 social values 70, 80–83 South America 269, 277 spaghetti bowl phenomenon 54–5 specialization 173–4, 179 specific-factors model 75 specified processes test 55 spillovers 99 stagnation dynamic (model) 154 effects of tariffs 161–5, 170–72 Japan 14, 17, 224, 229, 240 Standard International Trade Classification (SITC) 173–4 Stavrakeva, V. 120 stock market 14, 113 forecasting prices 119–39, 142, 144–9 Stolper–Samuelson theorem 75, 178 Stratton, L.S. 224 structural change 239–40 structural models 14, 119–20, 122, 127–34, 151–2 sub-prime loan crisis 5 Subramanian, A. 29 Sugiyarto, G. 198 Survey of Employment Diversification (MHLW) 228 tariff bindings overhang 61, 62 tariff heading criterion test 55–6 tariffs 59, 76–7, 209–10, 211, 214 CEPT scheme 49–50, 51, 52 GATT 27, 29, 61 Japan 15, 153–4, 155–6, 160–65 Smoot–Hawley Tariff Act 80, 82, 153 USA 13, 71–2, 73, 81–2 zero 49, 50, 54, 153–4 taxation 99, 113, 217 Taylor, L. 6 technological determinism 69 technological diffusion 99, 100, 111 technological innovation 26–7 technological progress (Japan) 15 technology transfer 110, 111, 115 temporary exclusion list (TEL) 49

temporary workers 211, 221, 224 terms of trade 15, 50, 154, 155–6 Tesar, L.L. 99 Thailand 51 “3–D” jobs 219 Timcke, D. 224, 241 Tokyo Round 29 Tombazos, C.G. 174, 175, 177, 178 Tomiura, E. 266 Tornqvist aggregations 177 Toronto Stock Exchange 98 Toyoda, N. 226 trade 7 barriers 17, 208–9, 212, 214–17, 221 bicycle theory 36 extended East Asia 12–13, 42–65, 68 FTAs 12–13, 32–6, 42–65 in goods (liberalization) 49–57 increased (global values) 26–7 inequality 15–16, 173–9 liberalization see liberalization negotiations 29–30 politics 13, 69–84 in services 57–8, 59, 223, 224 USA 32–8 Trade Expansion Act (1962) 82 trade policy (Japan) 14–15, 153–66 analysis (outline) 155–6 effects of tariffs 160–66 model 156–60 Trans-Pacific Partnership (TPP) 46, 65 transaction costs 25, 26, 27 Trefler, D. 173, 179 underemployment 192–3, 195, 202 UNDP 187 unemployment Asia-Pacific 16, 187–8, 190, 192, 193, 194–6 effects of tariffs 155–6, 160–65 Japan 15, 153–66, 223–4, 226, 231, 234–9, 241 unilateral trade liberalization 43, 48, 61 United Nations 9 Security Council 249 United States 30, 31 attitudes to trade 32–6 currency market 6–8, 26–7 financial crisis 5–6

Index globalization 36–8 group politics 80–83 immigration 71, 73, 78–9, 82–3 influence of 28 Japan and 14, 153 protection 13, 36–7, 71–2, 73, 82 tariffs 13, 71–2, 73 trade/wage inequality 15–16, 173–9 Washington consensus 4, 38 unskilled labor 208, 222, 256 wage inequality 15–16, 173–9, 194–5 Uruguay Round 29–30, 34, 64 US-Colombia FTA 35 value-added measure test 55, 56 Van Wijnbergen, S. 154 van Wincoop, E. 99, 110 Vietnam War 61 Viner, J. 43 wage councils 198, 199, 200 Wage Regulation Orders 199 wages economic impact of migration 206–7, 208, 210, 212–14, 278 effects of tariffs 153, 154, 159, 162, 163, 228, 229 inequality 15–16, 173–9 Japanese MNEs 255, 262–3, 278 labor market and 16, 188–9, 192, 194–202 low-skilled workers 188–9, 192, 194–5 minimum 16, 196–202 Wascher, W.L. 201 Washington Consensus 4, 38 Weberian model 69, 70, 71, 72–4, 79

293

Weinstein, D.E. 174, 179 welfare effect (financial integration) 99–100 effect (of immigration) 211–17 gains (regional integration) 110–12, 114, 115 welfare state 78–9 West, K.D. 137 Western Europe 30, 31, 32, 36, 108, 109, 114, 258, 269, 276, 278 Wignaraja, G. 55 Williamson, J. 4, 192 Wood, A. 178 working conditions (Japan) 219, 228 World Bank 6, 8, 27–8, 31–2, 103, 112, 189, 191 World Development Indicators 197, 266 World Commission on the Social Dimension of Globalization 202 World Trade Organization (WTO) Doha Development Agenda 12, 43 Doha Round 29, 30, 34, 35, 216 FTA networking 43, 46, 48, 49, 58, 60–61, 63, 65 problems/challenges 12, 27–32, 35, 37 Tokyo Round 29 Uruguay Round 29–30, 34, 64 WTO+ 58, 60, 65 Yip, C.K. 154 Yuan, T. 254, 256 zero tariffs 49, 50, 54, 153–4 Zheng, H. 99 Zhu, S.C. 173, 179