THE WZORLD BANK BCONOMVIC REVIEW ume 12 \I. Lbs _ l UM ON REGIONALISM AND DEVELOPMENT _ = c cs and Politics in Regional Integration _rrangements: An Introduction aurice Schiff and L. Alan Winters _egionalism: An Analysis of Nontraditional s from Regional Trade Agreements e uel Fernandez and Jonathan Portes _rrangements and Industrial Development lgo Puga and Anthony J. Venables - ial Integration and Economic Growth Athanasios Vamvakidis 'egional Integration as Diplomacy aurice Schiff and L. Alan Winters *.- * s Tariff Formation: The Case of Mercosur * rcelo Olarreaga and Isidro Soloaga A REVIEW ESSAY _ury of Development Economics: A Review _=e * he Handbook of Development Economics Jean Waelbroeck 15s\ /)_.)XA-() ()AWNE THE WORLD BANK ECONOMIC REVIEW EDITOR Moshe Syrquin CONSULTING EDITOR Sandra Gain EDITORIAL BOARD Kaushik Basu, Cornell University and Universitv of Delhi David Dollar Francois Bourguignon, DELTA, Paris Gregory K. Ingram Willem H. Buiter, University of Cambridge Lant H. Pritchett Gerald K. Helleiner, University of Toronro Jacques van der Gaag Mark R. Rosenzweig, University of Pennsylvania John Williamson L. Alan Winters The World Bank Economic Review is a professional journal for the dissemination of World Bank- sponsored research that informs policy analyses and choices. It is directed to an international readership among economists and social scientists in government, business, and international agencies, as well as in universities and development research institutions. The Review emphasizes policy relevance and opera- tional aspects of economics, rather than primarily theoretical and methodological issues. It is intended for readers familiar with economic theory and analysis but not necessarily proficient in advanced mathematical or econometric techniques. Articles will illustrate how professional research can shed light on policy choices. Inconsistency with Bank policy will not be grounds for rejection of an article. Articles will be drawn primarily from work conducted by World Bank staff and consultants. Before being accepted for publication by the Editorial Board, all articles are reviewed by two referees who are not mem- bers of the Bank's staff and one World Bank staff member; articles must also be recommended by at least one external member of the Editorial Board. The Review may on occasion publish articles on specified topics by non-Bank contributors. Any reader interested in preparing such an article is invited to submit a proposal of not more than two pages in length to the Editor. The views and interpretations expressed in this journal are those of the authors and do not necessar- ily represent the views and policies of the World Bank or of its Executive Directors or the countries they represent. The World Bank does not guarantee the accuracy of the data included in this publication and accepts no responsibility whatsoever for any consequences of their use. When maps are used, the bound- aries, denominations, and other information do not imply on the part of the World Bank Group any judgment on the legal status of any territory or the endorsement or acceptance of such boundaries. Comments or brief notes responding to Review articles are welcome and will be considered for publica- tion to the extent that space permits. Please direct all editorial correspondence to the Editor, The World Bank Economic Review, The World Bank, Washington, D.C. 20433, U.S.A. The W/orld Bank Economic Review is published three times a year (January, May, and September) by the World Bank. Single copies may be purchased at $12.95. Subscription rates are as follows: Individuals Institutions 1-year subscription USS30 US$50 2-year subscription US$55 US$95 3-year subscription US$70 US$130 Orders should be sent to: World Bank Publications, Box 7247-7956, Philadelphia, PA 19170-7956, U.S.A. Subscriptions are available without charge to readers with mailing addresses in developing coun- tries and in socialist economies in transition. Written request is required every two years to renew such subscriptions. (C 1998 The International Bank for Reconstruction and Development / THF WORLD BANK 1818 H Street, N.W., Washington, D.C. 20433, U.S.A. All rights reserved Manufactured in the United States of America ISBN 0-8213-4111-1; ISSN 0258-6770 Material in this journal is copyrighted. Requests to reproduce portions of it should be sent to the Office of the Publisher at the address in the copyright notice above. The World Bank encourages dis- semination of its work and will normally give permission promptly and, when the reproduction is for noncommercial purposes, without asking a fee. Permission to make photocopies is granted through the Copyright Clearance Center, Suite 910, 222 Rosewood Drive, Danvers, MA 01923, U.S.A. This journal is indexed regularly in Current Contents/Social & Behavioral Sciences, Index to Interna- tional Statistics, Journal of Economic Literature, Public Affairs Information Service, and Social Sci- ences Citation Index". It is available in microform through University Microfilms, Inc., 300 North Zeeb Road, Ann Arbor, MI 48106, U.S.A. THE WORLD BANK ECONOMIC REVIEW Volume 12 May 1998 Number 2 A SYMPOSIUM ON REGIONALISM AND DEVELOPMENT Dynamics and Politics in Regional Integration 177 Arrangements: An Introduction Maurice Schiff and L. Alan Winters Returns to Regionalism: An Analysis of Nontraditional 197 Gains from Regional Trade Agreements Raquel Fernindez and Jonathan Portes Trading Arrangements and Industrial Development 221 Diego Puga and Anthony J. Venables Regional Integration and Economic Growth 251 Athanasios Vamvakidis Regional Integration as Diplomacy 271 Maurice Schiff and L. Alan Winters Endogenous Tariff Formation: The Case of Mercosur 297 Marcelo Olarreaga and Isidro Soloaga A REVIEW ESSAY Half a Century of Development Economics: A Review 323 Based on the Handbook of Development Economics Jean Waelbroeck A SYMPOSIUM ON REGIONALISM AND DEVELOPMENT This symposium draws on papers originally prepared as part of the World Bank Development Economics Research Group's research program on regionalism and development. The articles in the symposium, which focus on dynamics and politics, were refereed in the usual way. The Editorial Board invited Maurice Schiff and L. Alan Winters, who led the research project, to write an introduc- tion to the symposium. THE WORLD BANK ECONOMIC REVIEW, VOL. 12, NO. 2: 177-95 Dynamics and Politics in Regional Integration Arrangements: An Introduction Maurice Schiff and L. Alan Winters Overwhelming evidence links openness and economic growth. In recent years many developing countries have attempted to liberalize their trade and investment regimes, mostly throuigh autonomous unilateral liberalization. At the same time, a growing number of governments have begun to explore and participate in regional trading agree- ments. The agreements grant reciprocal trade preferences to participating countries, resulting in discriminration against nonmembers. The causes and consequences of regional integration have given rise to an extensive and vigorous debate among both scholars and policymakers. However, the quality of this debate has been seriously hampered by the absence of clear analytical models and empirical evidence on many of the factors under discussion. Few of the recent argu- ments in favor of regional integration arrangements have been satisfactorily formal- ized or tested. To address some of these issues, a World Bank research program fo- cuses on new and developing country aspects of regionalism. The program explores iacunae in the traditional static analysis of regional integration arrangements; addresses the dynamic effects of integration, the economics of deep integration, and the politics and political economy of regional integration arrangements; and compares regional- ism with multilateralism. The articles in this symposium address the topics of dynami- ics, politics, and political economy in regional integration agreements. Overwhelming evidence links openness and economic growth. In recent years many developing countries have made efforts to liberalize their trade and invest- ment regimes. To a great extent these reform efforts have been consistent with the policy prescriptions that emerge from economic first principles: trade barri- ers should be low, more or less uniform across sectors, transparent, and nondiscretionary and should operate through the price mechanism. Most devel- oping countries have sought to apply these principles through a process of au- tonomous unilateral liberalization. At the same time, a growing number of governments have begun to explore and participate in regional trading agreements. The agreements grant reciprocal trade preferences to participating countries, resulting in discrimination against nonmembers. Indeed, nearly every country in the world is a member of-or in the process of discussing participation in-one or more regional integration ar- Maurice Schiff and L. Alan Winters are with the Development Economics Research Group at the World Bank. This symposium is made up of articles that were originally prepared as part of the group's research program on regionalism and development. Ši 1998 The International Bank for Reconstruction and Development/THE WORLD BANK 177 178 THE WORLD BANK ECONOMIC REVIEW, VOL. 12, NO. 2 rangements (RIAs), and some 55 to 60 percent of world trade now occurs within such trading blocs. Although most preferential trading arrangements are regional in the geographical sense, this is not necessary for most of the economic results we discuss. The term RIA loosely covers all reciprocal preferential arrangements. The causes and consequences of regional integration have given rise to an extensive and vigorous debate among both scholars and policymakers. How- ever, the quality of this debate has been seriously hampered by the absence of clear analytical models and empirical evidence on many of the factors under discussion. Few of the recent arguments in favor of RIAs have been satisfactorily formalized or tested. For example, analysts have not tested whether regionalism stimulates investment, whether it confers credibility on reform programs, or whether it leads automatically to multilateral liberalization. And no attempt has been made to weigh RIAs against one another in the circumstances of developing countries. Economists have not paid much attention to the noneconomic objec- tives that frequently underlie RIAs or to the role of trade preferences in achieving these objectives. Understanding the potential linkages between favoritism in trade and the pursuit of noneconomic political and social objectives can be crucial in a developing country's decision to participate in an RIA. To address some of these issues, we initiated a research program focusing on new and developing country aspects of regionalism. The program explores lacu- nae in the traditional static analysis of RIAS (see, for example, Schiff 1997). It also addresses the dynamic effects of integration, the economics of deep integra- tion, and the politics and political economy of RIAs. And it compares regional- ism with multilateralism. These six articles on the dynamics and political economy of regionalism and development constitute part of the output of the research program.' The articles do not so much develop new arguments for or against RIAs as analyze existing arguments that have figured in the popular and political debate. They deal with questions of industrial location, policy credibility, economic growth, political objectives, and pressure group politics. The authors make no claims to finality in these issues; rather they offer either the first formal analysis or the first rigor- ous empirical test of an argument. I. DYNAMICS Dynamics play an almost mystical role in many discussions of economic inte- gration. Having found small or even negative predicted static benefits, advo- cates of RIAS typically appeal to the dynamic benefits. However, what these con- stitute and how they come about are frequently rather vague, and evidence linking dynamic benefits with particular instances of integration are very difficult to pin down. The importance of economic growth for addressing poverty makes this a major area for investigation. 1. Details of the other research papers from the project, as well as many of the papers themselves, are available through the World Bank International Trade Team's Trade Web (http://wwwworldbank.org/html/iecit/iecit.html). Schiff and Winters 179 For concreteness we think of dynamics as anything that affects a country's rate of economic growth over the medium term. Thus we define dynamics to include both permanent increments to the rate of growth and temporary but long-lived increases of, say, more than five years as countries move from one growth path to another. In this section we consider briefly recent results on investment and credibility, industrial location, and the empirics of convergence and economic growth. Investment and Credibility Baldwin (1989, 1992) makes an early and striking application of neoclassical growth theory to regional integration. He models the effects of European inte- gration on capital accumulation and introduces the notion of a medium-term growth bonus. An RIA makes trade easier and hence tends to raise the returns to at least some factors of production, especially for deep integration that lowers real trading costs. If the RIA affects only tariffs, the benefits to factors affected by lower tariffs tend to be offset by the effects of replacement taxes on other fac- tors. If the cost of capital is unchanged, the economy responds with increased rates of return and thus increased capital stock. This increase leads to a tempo- rary increase in growth rates as the accumulation shifts the economy onto a higher trajectory. At the new steady-state level of capital stock, there are higher levels of output per head, but growth returns to its original level. Baldwin (1989) suggests that the medium-term bonus could double or even treble an RIA's static efficiency effects on output. Will an RIA raise or lower a developing country's rate of return to capital? A simple application of the Heckscher-Ohlin model suggests that in a North-South (industrial-developing country) RIA the rate of return to capital falls in the south- ern country because international trade tends to reduce the returns to the scarce factor. Mazumdar (1996) shows a similar problem if liberalization favors a labor- abundant commodity. However, the basic Heckscher-Ohlin model is probably too simple to ana- lyze the rate of return in this context. First, it applies only to a so-called square model with equal numbers of factors of production and goods. Second, this case of partial rather than complete liberalization could have rather different effects (see Falvey 1995). Third, the Heckscher-Ohlin model presumes homo- geneous products, whereas experience suggests that many markets are better represented by differentiated products and intraindustry trade. In the latter case, the degree of substitutability of domestic and foreign goods becomes very important. Building on these complications, Baldwin and a number of collaborators have suggested several reasons why economic integration might raise the rates of re- turn on capital in both partners regardless of capital abundance (see Baldwin, Forslid, and Haaland 1996 and Baldwin and Seghezza 1996a, 1996b). For ex- ample, an RIA typically reduces the transaction costs on tradable goods more than those on nontradable goods. 180 THE WORLD BANK ECONOMIC REVIEW, VOL. 12, NO 2 If, as is commonly believed, tradables are more capital-intensive than nontradables, trade liberalization increases the demand for capital relative to labor. In addition, an RIA may reduce tariffs and trading costs on imports of capital equipment or, by opening it up, improve efficiency in the financial sector and so reduce the cost of funds. Finally, an RIA may improve the atmosphere for investment by inducing greater credibility in the government's willingness or ability to pursue sound policies. Circumstantial evidence suggests that RIAS can generate investment booms, as occurred for example after the creation of the European Economic Community (EEC), the Iberian enlargement of the European Community, the European Com- munity 1992, the North American Free Trade Agreement (NAFTA), and Mercosur (a regional trade agreement among Argentina, Brazil, Paraguay, and Uruguay). Table A-1 provides a list of regional trade agreements and their member coun- tries. More formally, Brada and Mendez (1988) find significant effects when they estimate the effect on capital formation during 1960-77 in six RIAs-the European Free Trade Agreement (EFTA), the EEC, the Council for Mutual Eco- nomic Assistance, the Latin American Free Trade Area, the Central American Common Market (CACM), and the East African Common Market. Similarly, de Melo, Panagariya, and Rodrik (1992) find significant investment effects for the Central African Customs and Economic Union (UDEAC) and the Communaute economique ouest-africaine. However, neither study finds growth effects from the RIAS, possibly because of immiserizing investment in the presence of other distortions. An important share of investment in some developing countries is foreign direct investment (FDI). Many economists see inflows of FDI, first, as the harbin- ger of confidence in the economy and, second, although this is not uncontested, as the route through which an economy can modernize. For example, modern- ization occurs through access to modern technology, modern management, mar- keting networks, and sources of inputs (see Blomstrom and Kokko 1997a). A simple RIA may reduce FDI flows between member countries because it makes trade a more attractive option. Alternatively, FDI from outside the bloc may increase as foreigners seek to exploit new investment opportunities and to use one member as a platform for serving the whole bloc. Blomstrom and Kokko (1997b) suggest that, although the Canada-United States Free Trade Agreement had little investment effect, Mercosur and NAFTA both coincided with increased inflows of FDI. In more complete RIAS-for example, the Iberian accession to the European Community-FDI in nontraded sectors (and provisions for capital in- flows, repatriation of profits, and enhanced dispute settlement) may stimulate intrabloc investment flows. Overall, however, the principal requirement for at- tracting FDI is sound policies at home: the examples of China and Indonesia show that RIAs are not necessary for success; the example of Greece shows that they are not sufficient. Fernandez and Portes (this issue) combine elements of both dynamics and political economy to explore the argument that RIAs can improve the credibility Schiff and Winters 181 of members' policies through mechanisms that are not obtainable through uni- lateral or multilateral liberalization. The argument for increased credibility is now well known; see, for example, Whalley (1996) on NAFTA, Francois (1997) on the European Union-Mediterranean RIAs, and Baldwin, Francois, and Portes (1997) for empirical evidence. However, no other analyst has explained how RIAs enhance credibility, compared with other institutions and attitudes, or whether the arguments about credibility can be generalized beyond the classic cases of the European Union, the Europe Agreements, and NAFTA. Developing countries' reforms frequently lack credibility because of time in- consistency and asymmetric information problems. A government (or future government) that maintains policy discretion may be tempted to surprise the private sector, including foreign investors, through unexpected changes in fu- ture policy. An RIA can help to resolve these problems by "locking-in" trade reforms. Fernandez and Portes suggest that an RIA probably focuses the incen- tives to enforce liberalization commitments better than does the World Trade Organization (WTO) because the WTO has a larger constituency and thus retalia- tion has a larger element of public good. An RIA also offers more scope for punishment if it delivers benefits beyond the WTO in the form of, say, invest- ment. Similarly, developing countries may be able to achieve a measure of "lock- in" for their access to partner markets because, even if RIAs do not preclude the imposition of, say, antidumping duties or health restrictions, they do at least frequently offer special dispute settlement facilities. Whether RIAs discipline trade policy toward nonmembers is moot both theo- retically and empirically (see Winters 1997a). For example, Mexico responded to the peso crisis of 1994-95 by raising tariffs on 500 items against non-NAFTA suppliers. Bhagwati and Panagariya (1996) see this as diverting protectionist pressure onto third parties. Others argue that previous crises witnessed far worse protectionism and that NAFTA has induced restraint, to which Bhagwati and Panagariya respond that the intellectual atmosphere is far more liberal now than previously. It is even more difficult to find the source of credibility for policies that are not part of an RIA. Fernandez and Portes identify two possibilities. First, an RIA may raise the cost of macroeconomic laxity because it typically increases mar- ginal leakages to imports. However, Fernandez and Portes note that the RIA also increases the (temporary) returns to competitive devaluation that pushes the opposite way. Second, if entering an RIA entails (political) sunk costs, and if it requires liberal or sound policies to make sense, entry provides the government with a signaling device, for only a government with liberal intentions would sign. Thus in the presence of asymmetric information about the type of govern- ment, an RIA could improve credibility. Whether it is the best means of such signaling, however, is not obvious. Fernandez and Portes conclude by examining how these mechanisms apply in the cases of NAFTA and the Europe Agreements. They argue that the principal mechanism in NAFTA was Mexico's improved security of access to the U.S. mar- 182 THE WORLD BANK ECONOMIC REVIEW, VOL. 12, NO. 2 ket, while the time consistency argument appears much stronger in the case of the Europe Agreements. Although they do not pursue the point, Fernandez and Portes's analysis also leaves a strong impression that credibility effects are not likely to be very large in South-South RIAS. Industrial Location Many developing-country policymakers have contemplated the possibility that RIAS might have a dynamic effect on industrial location. Analytical interest in the topic originated mainly from efforts to predict the locational effects of the Euro- pean Union's Single Market Program. This interest led, in turn, to theoretical developments combining the insights from international trade and industrial organization that have reinvigorated the study of economic geography (Krugman 1991 and Krugman and Venables 1990, 1995). Puga and Venables (this issue) use techniques from the study of economic geography to extend their previous analysis (Puga and Venables 1997) to devel- oping countries. They assume one northern and two southern countries, each with two sectors. Agriculture is perfectly competitive and freely traded and uses both a specific factor-land-and a sectorally mobile factor-labor. Industry has increasing returns to scale and imperfect competition. Although their only primary input is labor, firms also buy inputs from one another. Because these transactions are costly if they cross national borders, agglomeration benefits accrue to firms located close to other firms. These benefits generate pecuniary externalities between firms that, in turn, induce cumulative causation such that as one firm relocates, it creates incentives for others to follow. As industry relo- cates, agriculture adjusts by releasing or absorbing labor and maintaining exter- nal balance. Given the fixed factor in agriculture, wages rise as industrialization causes agricultural employment to fall. The increase in wages ultimately pre- vents all industry from agglomerating in one location. Trade policy disturbs firms' locational decisions in three ways: through tariffs on inputs from abroad, through tariffs on sales (final and intermediate) abroad, and through the degree of competition in domestic markets. Puga and Venables (this issue) compare the effects of unilateral liberalization and various types of preferential liberalization. The unilateral liberalization of imports of manufac- tures by a southern country promotes the development of local industry by low- ering the cost of imported intermediates. However, the gains from an RIA with the North are likely to be greater because the South also benefits from improved access to the northern market. The gains from South-South RIAs are likely to be smaller than those from North-South arrangements. They depend essentially on whether the size of the combined southern market is large enough to attract industry: the smaller it is, the smaller the degree of industrialization and the later it takes place. If the southern economies are small, concerted nondiscriminatory liberalization offers larger gains than a South-South RIA because it frees up input supplies from the North even as it increases northern competition for local firms. All told, the best policy for a single southern country is to sign a North-South Schiff and Winters 183 RIA. However, this imposes the costs of even later industrialization on the south- ern country that is excluded. Thus developing countries have an incentive to be among the founders of an RIA, not only because the earlier they join the sooner they attract industry but also because being excluded is positively harmful, at least temporarily. Ethier's (1996) model also generates such competition be- tween developing countries as an incentive for joining RIAS. In the examples given by Puga and Venables, the North always loses and the South always gains from North-South RIAS. Some of the starkness of this result stems from their assumption that initially all industry is located in the North. Because, in this model, industry confers higher incomes, the South gains from any liberalization that permits industrialization-even southern countries left out of the RIA-while the North is likely to lose from most arrangements. These models are very stylized, but there is a little evidence of the effects they consider. Brulhart and Torstensson (1996) analyze the impact of European inte- gration and find that industries with increasing returns already tend to be highly localized and concentrated in the core countries in the European Union. They ar- gue that a further reduction in trade costs within the European Union will increase this concentration of scale-intensive activities, with the periphery specializing in constant-returns manufacturing and nonmanufacturing. They also argue, however, that the European Union has already experienced most of the scale-driven cluster- ing that it will see and that, in line with experience in the United States, future clustering is likely to occur in relatively small-scale industries in which peripheral regions have some advantage. This scenario should be particularly relevant in the event of an eastward enlargement of the European Union and also possibly the extension of RIAS to neighboring developing countries. Growth A large theoretical and empirical literature now exists on the relationship between a country's openness and growth-for example, Grossman and Helpman (1992); Edwards (1993); Sachs and Warner (199S); and Coe, Helpman, and Hoffmaister (1997). Although discussions of RIAS often casually appeal to this literature, doing so is rather risky (see Winters 1997a). Unfortunately, however, there are few RIA-specific analyses. Walz (1995, 1997) makes an important contribution by extending to regional integration Rivera-Batiz and Romer's (1991) model of growth with a research and development (R&D) sector. Walz relates growth to the static concepts of trade creation and diversion. If a country has a comparative advantage in R&D, an RIA that results in trade creation (that is, expansion of the sector with the comparative advantage) implies reallocation to the R&D sector and consequently faster growth. If the RIA results in trade diversion (expansion of the traditional sector), the R&D sector shrinks and growth falls. For a country whose compara- tive advantage lies in the traditional sector, trade diversion raises growth and trade creation lowers growth, although welfare need not change in the same direction as the growth rate. 184 THE WORLD BANK ECONOMIC REVIEW, VOL 12, NO. 2 The empirical evidence that RIAS stimulate growth is actually rather weak. Henrekson, Torstensson, and Torstensson (1997) use a cross-sectional regres- sion to suggest that European integration has enhanced members' growth rates. But others, including Brada and Mendez (1988) and de Melo, Panagariya, and Rodrik (1993), fail to find a positive association between RIAs and growth. Ben- David (1993) offers strong evidence that after signing RIAS, the EEC, the EFTA, and the Canada-United States Free Trade Agreement displayed marked increases in trade between member countries and dramatic increases in income conver- gence. He suggests that this convergence is upward, with the poorer members growing faster (Ben-David 1994), and that it owes more to convergence in rates of total factor productivity growth than in rates of investment (Ben-David 1996). Vamvakidis (this issue) attempts to resolve some of these issues. He addresses two aspects of regionalism. The first is geographical: the impact on a country's growth rate of the characteristics of its neighboring countries, such as their size, level of development, and degree of openness. The second is policy-based: the impact on growth of belonging to an RIA. Vamvakidis estimates cross-country and time-series growth regressions over 1970-90 and supplements the standard variables with variables on policies and neighbors' size and level of development. He finds several interesting results. For example, open economies grow faster; economies that have open and large neigh- bors grow faster, but the size of closed neighboring countries is of no account; economies that have open and developed neighbors also grow faster, but again the level of development of closed neighboring economies is not important; and the growth rate of neighboring economies has no significant impact on a country's growth rate. Vamvakidis also examines the impact on a country's growth of the economic size and level of development of non-neighboring economies in the same region and finds no significant impact. The contrast with the result for neighbors pre- sumably reflects the closer economic relations that countries have with neigh- bors through channels such as historical and cultural ties, similar languages and legal systems, and just plain familiarity. One of the implications of Vamvakidis's findings is that countries benefit from being located close to large, developed, open economies. This is regionalism in the geographic sense. Vamvakidis examines the impact of five RIAs-the Association of South East Asian Nations (ASEAN), the Andean Pact, the CACM, UDEAC, and the European Union-on the growth rate of its members. He finds no significant impact for any of the RIAS except the European Union, whose impact is marginally signifi- cant but vanishes once its members' openness is taken into account. He con- cludes that South-South agreements among small, closed developing countries are unlikely to have a positive impact on growth and that, although North- South agreements are more likely to have a positive growth effect on the south- ern partner, even this is far from guaranteed. An interesting question is whether countries can use RIAS to overcome the geographical misfortune of having closed neighbors-for example, by opening up a neighbor just to themselves or by forging Schiff and Winters 185 closer links with a distant partner as a substitute for the neighbors. However, there are so few examples to study that direct empirical tests are impossible. II. THE POLITICS OF REGIONAL INTEGRATION ARRANGEMENTS RIAs are far more than just economic policies. Winters (1997b), for example, argues that the commitment that stemmed from a political ideal was important in building an integrated Europe. Politics support many other RIAs, including NAFTA, Mercosur, the ASEAN free trade area, and the Southern African Develop- ment Community. The economics profession is not particularly well equipped to analyze the origins of such political motives and certainly is not qualified to comment on their legitimacy. It should, however, consider their economic impli- cations and examine whether the tools adopted for political purposes are effi- cient. Too often the declaration that an RIA (or any other policy) is political in intent is used to dismiss economic contributions to the debate, with the result being that economically more costly policy alternatives are frequently preferred to less costly ones. International Diplomacy Analysts argue that RIAS are an important tool of diplomacy in three ways. First, some RIAs help to stabilize neighboring countries and thus to reduce the probability that migrants or, indeed, bloodshed will spill across international borders. Second, RIAs respond to outside threats by cementing relations between the integrating partners. Third, RIAs between previously antagonistic states can potentially reduce tensions. The classic example of this is the precursor to the European Union, the European Coal and Steel Community of 1951. That com- munity was explicitly seen as a way to reduce Franco-German tensions, making war not only unthinkable but materially impossible. Similar objectives are said to be present, if not so centrally, in Mercosur and in ASEAN. These cases raise at least two questions. First, why would governments use trade as a diplomatic tool? Second, how do political objectives affect the RIA itself? On the first question, Mansfield (1993) argues that trade is a natural instru- ment because the higher income obtained from formation of the RIA enables allies to spend more on defense. A weakness with this argument is that RIAs need not raise income and may have the opposite effect. Also, trade is a civilizing influence that fosters understanding between partners; this venerable view is associated with Richard Cobden and Wilfredo Pareto in the nineteenth century and with Cordell Hull, among others, in the twentieth. As regionalism spreads, often pro- pelled by the rhetoric of international diplomacy, it is important for political scientists and economists to discover why (or whether) trade agreements domi- nate other, possibly economically better, approaches to rapprochement. In Schiff and Winters (this issue), we take a first indirect step toward identify- ing the political returns to RIAs and directly attack the question of how politics 186 THE WORLD BANK ECONOMIC REVIEW, VOL. 12, NO. 2 might affect the existence, shape, and evolution of RIAS. We accept at face value the premise that trade among neighboring countries provides security directly, for example, by raising the level of interaction and trust among the people of those countries, by increasing the stake that each country has in the welfare of its neighbor, or by increasing the security of access to the neighbor's strategic raw materials. These security effects are external to individual agents, so that private incentives do not induce their full exploitation. We show how, under these assumptions, a subsidy on intrabloc trade or, equivalently, an RIA accom- panied by appropriate domestic taxes, maximizes welfare by providing an opti- mal way to internalize the security externalities. In our article, we show that we can derive predictions about the development of RIAS. These predictions have implications for issues such as whether regional- ism encourages multilateral liberalization. Through these predictions, which re- fer to observable phenomena, we can, in principle, test whether security plays a dominant role in the formation of an RIA. Identifying the motivation for an RIA does not alter its economic and political effects, but it does allow a more ratio- nal discussion of the policy options. Our article presents the testable predictions that if security dominates the evolution of an RIA, its optimal external trade barriers decline over time and as member countries move toward deep integration. We also suggest that barriers tend to increase following enlargement of an RIA. Starting from a steady state, the institution of an RIA (accompanied by appropriate domestic taxes) increases the volume of trade among member countries and thus directly increases trust between their populations. As the level of trust rises, the marginal impact (exter- nality) of further trade on security and the marginal value of additional security are both expected to fall, and the optimal subsidy to intrabloc trade-that is, the optimal external trade barrier-falls as well. Deep integration, which lowers the costs of trade between member countries and raises the natural level of intrabloc trade, has a similar effect, although precisely how and when depends on the exact nature of the deep integration. The pattern of declining external tariffs is exactly that observed in the European Union. Our article analyzes the political dimension of RIAs by taking a popular argu- ment and subjecting it to formal analysis and testing. In welfare terms, however, it is very much a first step. Political cooperation is perfectly possible without tariff preferences, and free trade does not guarantee peace-witness the U.S. Civil War, which was partially caused by disagreements over trade policy. Thus to justify an RIA on political grounds requires showing that trade preferences contribute to political rapprochement, that such a rapprochement is valuable, and that it would not have happened if the RIA had not been formed. Internal Political Economy The term internal political economy describes how individual countries reach their decisions about what to seek in their relations with other countries. There are clear interactions between internal and external politics (see Putnam 1988). Schiff and Winters 187 In terms of economic analysis, internal politics can help to answer the question of what determines trade policy. This issue has been extensively studied both empirically and theoretically, although not in the context of RIAS. Economists have long understood that interest groups can affect international trade policy and, by extension, RIAS. Early empirical work was mostly rather intuitive and relied on regressing tariff rates or other indicators of trade policy on proxies for the political pressures exerted on different parts of government. Magee, Brock, and Young (1989) make a major advance in theoretical formality by carefully modeling electoral competition in the presence of lobbying groups. They find empirical support for their models in tariffs in the United States over both historical and more recent periods. Magee and Lee (1997) apply this broad approach in their study of tariff formation as the EEC deepened its integration. They consider French and Italian tariffs over 1968-83 and try to determine the amount of the change that was due to integration-induced changes in internal political economy forces. They attribute much of the tariff reduction to external forces (the General Agreement on Tariffs and Trade), but an average tariff of 7.5 percent remained in 1983. Within that remaining tariff, they find that tariff creation was responsible for increases of 1.7 percentage points; EEC industries were able to exploit high adjustment costs politically and to organize themselves more effectively as the number of firms fell. They find that tariff diversion was responsible for decreases of 1.1 percentage points; lobbying organizations be- came more complex in the larger political arena, and some industries benefited from larger markets. In a major advance in the theory of the political economy of trade policy, Grossman and Helpman (1994) model lobbying as the influence on governments in power regardless of their political hue. Their approach contrasts with that of Magee, Brock, and Young (1989) and Magee and Lee (1997) who model lobby- ing as the influence on political parties that have to fight elections. Helpman (1997) looks at the relationship between the two approaches (and others). The fact that lobbyists frequently contribute to both sides in an election as well as to incumbent politicians with large majorities lends credibility to Grossman and Helpman's view. In Grossman and Helpman's (1995) application of their theory to RIAS, they suggest, among other things, that free trade areas are likely to arise either if they provide overwhelming consumer benefits that allow governments to ignore the lobbies or if they tend toward increased protection. The latter result, which cor- relates with high levels of trade diversion, arises because export lobbies will support a free trade area that allows them to sell in the partner country at higher prices. Their support will help to offset the resistance of import-competing groups that will arise in both trade-creating and trade-diverting sectors. Hirschman (1981) notes the political attractions of trade diversion relative to trade creation. Grossman and Helpman (1995) also suggest that having exceptions to the free trade area-that is, sectors that retain their protection-will ease the politi- 188 THE WORLD BANK ECONOMIC REVIEW, VOL. 12, NO. 2 cal task of getting the free trade area accepted by helping to obviate resistance from the worst-hit sectors. The government will most likely exempt sectors that feature much trade creation. Any tariff reduction will encounter resistance by import-competing producers. With trade diversion, partner exporters will push hard for the inclusion of the sector. But only consumers and third-country pro- ducers-neither of which generally has much clout-will champion the liberal- ization of trade-creating sectors. Olarreaga and Soloaga (this issue) offer the first empirical operationalization and test of the Grossman-Helpman model. They implement the customs union version of Cadot, de Melo, and Olarreaga (1996, 1997) and apply it to Mercosur to explain the level of the common external tariff (CET), exceptions to it, and exceptions to internal free trade. Mercosur was created in 1991. By 1995 it had achieved internal free trade with relatively few exceptions. It had held an extended debate on the CET, which concluded only in 1994, and admitted exceptions on well over one-quarter of the tariff lines. Consistent with theoretical predictions, Olarreaga and Soloaga find variations in the CET over industries that are significantly related to those industries' labor/capital ratios (reflecting a tendency for protection to accrue where capital shares are higher), average wages (protecting unskilled labor), and industry concentration (reducing the costs of organized lobbying). Also in line with theory, they find that these variables perform best when used as production- weighted averages of the corresponding four national variables, although they cannot reject the hypothesis that the CET reflects only the political wishes of Brazil plus (any) one of the other partners. These results represent one of the few applications of endogenous tariff theory to developing countries and are unique in considering a CET rather than a na- tional tariff. The authors show that Mercosur's CET is well grounded in political realities and suggest that Mercosur will survive. (Unfortunately, the increase of one-quarter in the CET that occurred in late 1997 came too late to influence Olarreaga and Soloaga's analysis.) It remains to be seen whether Mercosur's CET will block further liberalization, as predicted by some in the debate comparing regionalism and multilateralism (see Winters 1998). Olarreaga and Soloaga find significant and plausible political economy ef- fects with respect to exceptions. Their most robust result is that strong labor unions lead to positive exceptions to both internal free trade and the CET. Unions are national bodies and may find it difficult at first to organize across national borders. Thus it seems plausible that although their positive effect on the CET is not robust, their effect on exceptions is. Olarreaga and Soloaga also find robust and significant that exceptions to internal free trade increase with the degree of (predicted) trade creation-the first direct confirmation of Grossman and Helpman (1995). And they find that, as with the CET itself, labor/capital ratios and industry concentration cause positive deviations from the CET. These results suggest that eliminating the current exceptions to the customs union is likely to be politically sensitive because those exceptions seem to reflect estab- Schiff and Winters 189 lished political forces. However, removal of exceptions is likely to be worth- while because it will reduce external protection somewhat. Also, convergence to internal free trade is likely to benefit sectors offering (welfare-enhancing) trade creation. III. CONCLUDING COMMENTS The articles in this symposium aim to advance the debate on RIAs by subject- ing existing arguments to rigorous theoretical analysis or empirical testing. The articles offer several important findings. The effect of an RIA on credibility de- pends on details such as its punishment mechanisms and the cost it implies for countries pursuing bad policy. An RIA may boost the industrialization efforts of a developing member but retard those of an excluded developing country. A country may derive growth benefits from being a neighbor to a large, developed, open economy, but South-South RIAs are unlikely to result in faster growth. When governments use RIAS to reduce conflict between neighboring countries, trade barriers are likely to fall over time and following deep integration. And the power of interest groups is quite evident in the patterns of developing-country RIAS' external tariffs and remaining internal trade taxes. In addition to their individual contributions, the articles touch on several common themes. For example, Fernandez and Portes analyze the economic mechanisms through which a developing-country member may gain policy credibility that is not obtainable unilaterally or multilaterally. Schiff and Winters's model contains a dimension about the credibility effects of solving the political problems caused by conflict between two neighboring countries. Credibility clearly also has a critical but implicit role in Puga and Venables's model of industrial location and in Vamvakidis's growth results. Moreover, Vamvakidis's conclusions on the benefits of desirable neighbors relate to cred- ibility changes considering the geographical spillover effects of recent cri- ses-the "tequila effect" in South America or the current "Asian flu." Com- bining these results suggests the possibility of virtuous and vicious circles in RIA formation. A strong, liberalizing RIA with the right partner may lead to a virtuous circle of increased credibility, increased investment and growth, more credibility and political stability, and so on. By contrast, a more-closed agree- ment or wrong choice of partner could lead in the opposite direction, with reduced credibility and lower investment and growth, resulting in less cred- ibility and more political instability over time. Another common theme concerns economic rents, the essential ingredient of Olarreaga and Soloaga's political economy and likely a significant outcome of the sort of processes analyzed by Puga and Venables. Relative size is an- other common theme. Ferndndez and Portes's and Puga and Venables's results clearly suggest the advantages of North-South RIAs over South-South ones, while Vamvakidis has a related finding about the benefits for developing coun- tries of having large, open neighbors. By contrast, while not considering the 190 THE WORLD BANK ECONOMIC REVIEW, VOL. 12, NO. 2 North-South issue explicitly, Schiff and Winters suggest a logic for South-South (or North-North) RIAs. That model is most relevant to countries with neigh- bors of around their own size. Olarreaga and Soloaga's results pertain to ques- tions of size, showing that if one partner is very large, it will predominate in the policymaking process. The articles in this issue are first steps and so lend themselves to several po- tentially fruitful extensions. Many more analytical and practical challenges re- main on the broad topic of regionalism and development. We hope that these articles and the other output of the World Bank research program encourage other scholars and policymakers to explore this issue further. Table A-1. Regional Trade Agreements Agreement Acronym Member economies Africa, Caribbean, and Pacifica ACP Angola, Antigua and Barbuda, Bahamas, (European Economic Community Barbados, Belize, Benin, Botswana, Fourth Lome Convention) Burkina Faso, Burundi, Cameroon, Cape Verde, Central African Republic, Chad, Comoros, Congo (Republic), Congo Democratic Republic, C6te d'lvoire, Djibouti, Dominica, Dominican Republic, Equatorial Guinea, Eritrea, Ethiopia, Fiji, Gabon, The Gambia, Ghana, Grenada, Guinea, Guinea Bissau, Guyana, Haiti, Jamaica, Kenya, Kiribati, Lesotho, Liberia, Madagascar, Malawi, Mali, Mauritania, Mauritius, Mozambique, Namibia, Niger, Nigeria, Papua New Guinea, Rwanda, St. Kitts and Nevis, St. Lucia, St. Vincent and the Grenadines, Sao Tome and Principe, Senegal, Seychelles, Sierra Leone, Solomon Islands, Somalia, South Africa, Sudan, Suriname, Swaziland, Tanzania, Togo, Tonga, Trinidad and Tobago, Tuvalu, Uganda, Vanuatu, Western Samoa, Zambia, Zimbabwe Andean Common Market (also ANCOM Bolivia, Colombia, Ecuador, Peru, Venezuela, called the Andean Pact or Andean Community) Asia-Pacific Economic Cooperation APEC Australia, Brunei, Canada, Chile, China, Hong Kong, Indonesia, Japan, Republic of Korea, Malaysia, Mexico, New Zealand, Papua New Guinea, Philippines, Singapore, Taiwan (China), Thailand, United States Association of South East Asian ASEAN Indonesia, Malaysia, Philippines, Singapore, Nations Thailand Schiff and Winters 191 Table A-1. (continued) Agreement Acronym Member economies Association of South East Asian ASEAN Free Brunei, Indonesia, Malaysia, Philippines, Nations Free Trade Area Trade Area Singapore, Thailand, Vietnam Caribbean Community and Common CARICOM Antigua and Barbuda, Bahamas, Barbados, Market Economic Community Belize, Dominica, Grenada, Guyana, Jamaica, Montserrat, St. Kitts and Nevis, St. Lucia, St. Vincent and the Grenadines, Suriname, Trinidad and Tobago Central African Customs and Eco- UDEAC Cameroon, the Central African Republic, nomic Union (Union douaniere et Chad, Congo, Gabon economique de l'afrique centrale) Central American Common Market CACM Costa Rica, El Salvador, Guatemala, Honduras, Nicaragua Central European Free Trade Area CEFTA Czech Republic, Hungary, Poland, Slovak Republic, Slovenia Communaut6 economique CEAO Benin, Burkina Faso, Cape Verde, Cote ouest-africaine d'Ivoire, The Gambia, Mali, Niger, Senegal, Togo Council for Mutual Economic Assis- CMEA Ghana, Guinea, Guinea Bissau, Liberia, tance (Common Market of the Mauritania, Nigeria, Sierra Leone Centrally Planned Economies, COMECON; dissolved in 1991) East African Community EAC Kenya, Tanzania, Uganda East Asian Economic Group (now EAEG Brunei, China, Hong Kong (China), called East Asian Economic Indonesia, Japan, Republic of Korea, Caucas) Malaysia, Philippines, Singapore, Taiwan (China), Thailand Economic Cooperation Organization Afghanistan, Azerbaijan, Kazakhstan, Kyrgyzia, Iran, Pakistan, Tajikistan, Turkey, Turkmenistan, Uzbekistan Europe Agreements Foreign trade agreement signed between the EU and EFTA with several Central and Eastern European countries European Community EC Austria, Belgium, Denmark, Finland, France, Germany, Greece, Ireland, Italy, Luxembourg, Portugal, Spain, Sweden, Netherlands, United Kingdom European Economic Area EEA EU and EFTA (Table continues on the following page.) 192 THE WORLD BANK ECONOMIC REVIEW, VOL. 12, NO. 2 Table A-1. (continued) Agreement Acronym Member economies European Free Trade Agreement EFTA Austria, Finland, Iceland, Liechtenstein, Norway, Sweden, Switzerland European Union EU Belgium, Denmark, France, Germany, Greece, Ireland, Italy, Luxembourg, Netherlands, Portugal, Spain, United Kingdom Gulf Cooperation Council GCC Bahrain, Kuwait, Oman, Quatar, Saudi Arabia, United Arab Emirates Latin American Free Trade Area LAFTA Argentina, Bolivia, Brazil, Chile, Colombia, Ecuador, Mexico, Paraguay, Peru, Uruguay, Venezuela Latin American Integration LAIA Argentina, Bolivia, Brazil, Chile, Colombia, Association Ecuador, Mexico, Paraguay, Peru, Uruguay, Venezuela Mercosur (Southern Common Mercosur Argentina, Brazil, Paraguay, Uruguay Market) North American Free Trade NAFTA Canada, Mexico, United States Agreement South Asian Preferential Trade SAPTA Bangladesh, Bhutan, India, Maldives, Nepal, Agreement Pakistan, Sri Lanka Southern African Development SADC Angola, Botswana, Lesotho, Malawi, Community Mauritius, Mozambique, Namibia, Swaziland, Tanzania, Zambia, Zimbabwe a. Nonreciprocal regional trade agreement. Note: The table lists only regional trade agreements among two or more countries. Schiff and Winters 193 REFERENCES The word "processed" describes informally reproduced works that may not be com- monly available through library systems. Baldwin, Richard A. 1989. "The Growth Effects of 1992." Economic Policy 9:247-81. .1992. "Measurable Dynamic Gains from Trade." Journal of Political Economy 100:162-74. Baldwin, Richard E., Rikard Forslid, and Jan Haaland. 1996. "Investment Creation and Investment Diversion: A Simulation Study of the EU'S Single Market Programme." CEPR Discussion Paper 1308. Centre for Economic Policy Research, London. Pro- cessed. Baldwin, Richard A., Joseph Francois, and Richard Portes. 1997. "The Costs and Ben- efits of Eastern Enlargement: The Impact on the EU and Central Europe." Economic Policy: A European Forum 24:127-76. Baldwin, Richard A., and Elena Seghezza. 1996a. "Growth and European Integration: Towards an Empirical Assessment." CEPR Discussion Paper 1393. Centre for Eco- nomic Policy Research, London. Processed. . 1996b. "Testing for Trade-Induced Investment-Led Growth." CEPR Working Paper 1331. Centre for Economic Policy Research, London. Processed. Ben-David, Dan. 1993. "Equalizing Exchange: Trade Liberalization and Economic Con- vergence." Quarterly Journal of Economics 108:653-79. - . 1994. "Income Disparity among Countries and the Effects of Freer Trade." In Luigi L. Pasinetti and Robert M. Solow, eds., Economic Growth and the Structure of Long-Run Development. London: Macmillan. -\ -. 1996. "Trade and Convergence among Countries." Journal of International Economics 40:279-98. Bhagwati, Jagdish, and Arvind Panagariya. 1996. "Preferential Trading Areas and Multilateralism: Strangers, Friends, or Foes?" In J. N. Bhagwati and Arvind Panagariya, eds., Free Trade Areas or Free Trade? The Econzomics of Preferential Trading Agree- ments. Washington, D.C.: American Enterprise Institute Press. Blomstrom, Magnus, and Ari Kokko. 1997a. "How Foreign Investment Affects Host Countries." Policy Research Working Paper 1745. International Economics Depart- ment, World Bank, Washington, D.C. Processed. =--. 1997b. "Regional Integration and Foreign Direct Investment: A Conceptual Framework and Three Cases." Policy Research Working Paper 1750. International Economics Department, World Bank, Washington, D.C. Processed. Brada, Josef C., and J. A. Mendez. 1988. "Exchange Rate Risk, Exchange Rate Re- gimes, and the Level of International Trade." Kyklos 41(2):263-80. Brulhart, Marius, and Johan Torstensson. 1996. "Regional Integration, Scale Econo- mies, and Industry Location in the European Union." CEPR Discussion Paper 1435. Centre for Economic Policy Research, London. Processed. Cadot, Olivier, Jaime de Melo, and Marcelo Olarreaga. 1996. "Regional Integration and Lobbying for Tariffs against Non-Members." CEPR Discussion Paper 1448. Cen- tre for Economic Policy Research, London. Processed. .1997 "Lobbying and the Structure of Protection." CEPR Discussion Paper 1574. Centre for Economic Policy Research, London. Processed. 194 THE WORLD BANK ECONOMIC REVIEW, VOL. 12, NO. 2 Coe, David T., Elhanan Helpman, and Alexander W. Hoffmaister. 1997. "North-South R&D Spillovers." The Economic Journal 107(January):134-49. de Melo, Jaime, Arvind Panagariya, and Dani Rodrik. 1992. "The New Regionalism: A Country Perspective." CEPR Discussion Paper 715. Centre for Economic Policy Re- search, London. Processed. . 1993. "The New Regionalism: A Country Perspective." In Jaime de Melo and Arvind Panagariya, eds., New Dimensions in Regional Integration. Cambridge, U.K.: Cambridge University Press. Edwards, Sebastian. 1993. "Openness, Trade Liberalization, and Growth in Developing Countries." Journal of Economic Literature 31(September):1358-93. Ethier, Wilfred J. 1996. "Regional Regionalism." Department of Economics, University of Pennsylvania, Philadelphia. Processed. Falvey, Rod. 1995. "Factor Price Convergence." Trade and Wage Dispersion Project Paper 2. International Economics Department, World Bank, Washington, D.C. Pro- cessed. Francois, Joseph. 1997. "External Bindings and the Credibility of Reform." In Ahmed Galal and Bernard Hoekman, eds., Regional Partners in Global Markets: Limits and Possibilities of the Euro-Med Agreements. London: Centre for Economic Policy Research. Grossman, Gene M., and Elhanan Helpman. 1992. Innovation and Growth: Techno- logical Competition in the Global Economy. Boston, Mass.: MIT Press. 1994. "Protection for Sale." American Economic Review 84(4):833-50. 1995. "The Politics of Free Trade Agreements." American Economic Review 85(4):667-90. Helpman, Elhanan. 1997. "Politics and Trade Policy." In David M. Kreps and K. F. Wallis, eds., Advances in Economics and Econometrics: Theory and Applications. Vol. 1. Cambridge, U.K.: Cambridge University Press. Henrekson, Magnus, Johan Torstensson, and Rasha Torstensson. 1997. "Growth Ef- fects of European Integration." European Economic Review 41(8):1537-57. Hirschman, Albert. 1981. Essays in Trespassing: Economics to Politics and Beyond. Cambridge, U.K.: Cambridge University Press. Krugman, Paul R. 1991. "Increasing Returns and Economic Geography." Journal of Political Economy 99:483-99. Krugman, Paul R., and Anthony J. Venables. 1990. "Integration and the Competitive- ness of Peripheral Industry." In Charles Bliss and Jose de Macedo, eds., Unity with Diversity in the European Community: The Community's Southern Frontier. Cam- bridge, U.K.: Cambridge University Press. .1995. "Globalization and the Inequality of Nations." Quarterly Journal of Eco- nomics 110:857-80. Magee, Stephen P., William A. Brock, and L. Young. 1989. Black Hole Tariffs and Endogenous Policy Formation. Cambridge, U.K.: Cambridge University Press. Magee, Stephen P., and Hak-Loh Lee. 1997. "Tariff Creation and Tariff Diversion in Customs Union: The Endogenous External Tariff of the EEC, 1968-1983." Nota di Lavoro, Fundazione Eni Enrico Mattei, 38(97), p. 1-39. Schiff and Winters 195 Mansfield, Edward D. 1993. "Effects of International Politics on Regionalism in Inter- national Trade." In Kym Anderson and Richard Blackhurst, eds., Regional Integra- tion and the Global Trading System. London: Harvester Wheatsheaf. Mazumdar, Joy. 1996. "Do Static Gains from Trade Lead to Medium-Run Growth?" Journal of Political Economy 104:1328-37. Puga, Diego, and Anthony J. Venables. 1997. "Preferential Trading Arrangements and Industrial Location." Journal of International Economics 43:347-68. Putnam, Robert D. 1988. "Diplomacy and Domestic Politics: The Logic of Two-Level Games." International Organization 4(2, Summer):427-60. Rivera-Batiz, Luis A., and Paul M. Romer. 1991. "Economic Integration and Endog- enous Growth." Quarterly Journal of Economics 106:531-55. Sachs, Jeffrey D., and Andrew Warner. 1995. Economic Reform and the Process of Global Integration. Brookings Papers on Economic Activity 1. Washington, D.C.: Brookings Institution. Schiff, Maurice. 1997. "Small Is Beautiful: Preferential Trade Agreements and the Im- pact of Country Size, Market Share, and Smuggling." Journal of Economic Integra- tion 12(3, September):359-87. Walz, Uwe. 1995. "Trade Liberalization, Factor Mobility, and Regional Growth." Journal of Institutional and Theoretical Economics 151 (September):5 05-25. . 1997. "Growth and Deeper Integration in a Three-Country Model." Review of International Economics 5(4, November): 492-507. Whalley, John. 1996. "Why Do Countries Seek Regional Trade Agreements?" NBER Working Paper 5552. National Bureau of Economic Research, Harvard University, Cambridge, Mass. Processed. Winters, L. Alan. 1997a. "Assessing Regional Integration Arrangements." Development Economics Research Group, World Bank, Washington, D.C. Processed. . 1997b. "What Can European Experience Teach Developing Countries about Integration?" World Economy 20:889-912. .1998. "Regionalism versus Multilateralism." In Richard A. Baldwin, David Cole, Andre Sapir, and Anthony Venables, eds., Regional Integration. Cambridge, U.K.: Cambridge University Press. THE WORLD BANK ECONOMIC REVIEW, VOL. 12, NO. 2: 197-220 Returns to Regionalism: An Analysis of Nontraditional Gains from Regional Trade Agreements Raquel Fernandez and Jonathan Portes The past decade has witnessed a renewed interest in regional trade agreements, with manzy policymakers and academics seeming to believe that these provide more than the traditional gains from trade. This article examines several possible benefits that re- gional trade agreements may confer on their partners, including credibility, signaling, bargaining power, insurance, and coordination. It assesses the necessary conditions for each possible channel to work, gives stylized examples of specific types of policy where the benefit might be applicable, examines cases where the explanation might be relevant, and discusses their overall plausibility. It concludes by examining the North American Free Trade Agreement and the Europe Agreements. The past decade has witnessed a renewed interest in regional trade agreements (RTAS), whose only precedent, though on a less dramatic scale, was the prolifera- tion of agreements in the 1970s. Table 1 lists the agreements notified to the General Agreement on Tariffs and Trade (GATT) in 1947-94 (see Schiff and Win- ters, this issue, for a table presenting the member countries in several RTAS). By the time the World Trade Organization (WTO) was established in January 1995, almost all GATT members (notable exceptions include Japan and Hong Kong) were signatories to at least one such agreement. These diverse agreements range from customs unions to free trade areas to nonreciprocal preferential agree- ments and differ substantially in their scope and treatment of issues such as labor and capital mobility, investment, and production sharing. Nonetheless, it is possible to characterize the growth in RTAs over the past decade as taking place in at least three separate, albeit related, dimensions. RTAS have deepened, as agreements that originally focused on hard trade re- strictions, like tariffs and quotas on manufactures and agriculture, have been extended to "soft" restrictions such as health and environmental standards or to other product areas such as services and intellectual property, where trade policy is typically far more complex. Issues such as investment and capital mobility Raquel Fernandez is with the Department of Economics at New York University, and Jonathan Portes is with National Economic Research Associates. The authors thank Kyle Bagwell for helpful conversations and Maurice Schiff, Alan Winters, and three referees for valuable suggestions. This article was produced as part of the World Bank Development Economics Research Group's research program on regionalism and development. Š 1998 The International Bank for Reconstruction and Development/THE WORLD BANK 197 198 THE WORLD BANK ECONOMIC REVIEW, VOL. 12, NO. 2 Table 1. Regional Integration Agreements Notified to GATT and WTO and in Force as of January 1995 Reciprocal regional integration agreements Europe Central European Free Trade Area Czech Republic and Slovak Republic Customs Union Czech Republic and Slovenia Free Trade Agreement EFTA Free Trade Agreements with Bulgaria, Czech Republic, Hungary, Israel, Poland, Romania, Slovak Republic, Turkey European Community European Community Association Agreements with Bulgaria, Cyprus, Czech Republic, Hungary, Malta, Poland, Romania, Slovak Republic, Turkey European Community Free Trade Agreements with Estonia, Iceland, Israel, Latvia, Liechtenstein, Lithuania, Norway, Switzerland European Free Trade Association (EFTA) Norway Free Trade Agreements with Estonia, Latvia, Lithuania Slovak Republic and Slovenia Free Trade Agreement Switzerland Free Trade Agreements with Estonia, Latvia, Lithuania North America Canada-United States Free Trade Agreement North American Free Trade Agreement Latin America and the Caribbean Andean Pact Caribbean Community and Common Market (CARICOM) Central American Common Market Latin American Integration Association Southern Common Market (Mercosur) Middle East Economic Cooperation Organization Gulf Cooperation Council Asia Australia-New Zealand Closer Economic Relations Trade Agreement Bangkok Agreement Common Effective Preferential Scheme for the Association for South East Asian Nations (ASEAN) Free Trade Area Lao People's Democratic Republic and Thailand Trade Agreement Other Israel-United States Free Trade Agreement Nonreciprocal regional integration agreements Europe Africa, Caribbean, and Pacific (AcP)-European Economic Community Fourth Lome Convention European Economic Community Association of Certain Non-European Countries and Territories (EEC-PTOM 1I) European Economic Community cooperative agreements with Algeria, Egypt, Jordan, Lebanon, Morocco, Syria, Tunisia Asia Australia-Papua New Guinea Agreement South Pacific Regional Trade Cooperation Agreement Source: WTO 1995. Fernc4ndez and Portes 199 that are not strictly within trade policy have also been added. The preeminent example of this change is the European Union, whose successive name changes (from the Common Market to the European Community to the European Union) describe its evolution from a customs union to a single market (with free move- ment of labor, capital, and services and substantial regulatory harmonization) and, eventually, to an economic union with a single currency. Widening has occurred as countries not previously a member of an RTA joined one or more such agreements. Examples include the North American Free Trade Agreement (NAFTA), which essentially involved Mexico's accession to the Canada- U.S. Free Trade Agreement, and the accession of some of the former European Free Trade Association (EFTA) countries to the European Union. Lastly, new RTAs have been launched and dormant ones revived. The most pub- licized of these RTAs has been Mercosur (Argentina, Brazil, Paraguay, and Uru- guay). Other examples include the Central American Common Market (founded in 1960, but effectively dormant until 1993) and the South Asian Preferential Trading Agreement (established in 1993 with the aim of forming a common mar- ket among Bangladesh, Bhutan, India, Maldives, Nepal, Pakistan, and Sri Lanka). The international trading system has also deepened and widened. GATT dis- cussions extended for the first time into trade in services and intellectual prop- erty rights and plans to embrace industries like telecommunications. Again, a name change signaled this deepening, as GATT became the wTo. The widening occurred largely as a consequence of the collapse of the formerly centrally planned economies. All but nine of the 33 agreements notified to GATT since 1990 have been concluded with Central and Eastern European countries (WTO 1995). The search for an economic explanation for this proliferation of RTAs begins with several negatives. Although the literature favoring free trade is as strong as ever, with the ability to rigorously model dynamic externalities, adding intellec- tual respectability to traditional arguments linking trade liberalization and growth (see Rivera-Batiz and Romer 1991a, 1991b and Grossman and Helpman 1991), it would be difficult to believe that the growth in RTAs represents the policy counterpart to the proliferation of academic papers on trade and growth. Nor has the debate as to whether RTAs are theoretically or empirically preferable to an admittedly imperfect global trading system been resolved in such a way as to encourage the spread of RTAS (see Winters 1996 for a survey of the literature debating this issue). In fact, the debate remains lively, although lately, with the Uruguay Round concluded, all parties seem to accept RTAs as a useful accom- paniment to the WTO. In any case, policymakers and bureaucrats seem to spend a disproportionate amount of time and effort concluding trade agreements com- pared with the gains or losses that, at least on the face of it, most RTAS can offer. NAFTA, for example, involved relatively small changes in Mexican trade policy and minuscule changes in U.S. policy yet managed to dominate U.S. political debate in late 1993.1 1. Prior to NAFTA, the average Mexican tariff on U.S. exports was about 10 percent, while the average U.S. tariff on imports from Mexico was less than 10 percent; see Hufbauer (1992). 200 THE WORLD BANK ECONOMIC REVIEW, VOL. 12, NO. 2 The question therefore arises: Is there more to an RTA than meets the eye? Why might a country create or join an RTA? Could the entry of a country into an RTA change the incentives, and hence the behavior, of that country, other coun- tries, or the private sector in ways beyond the actual provisions of the agree- ment? And insofar as the RTA does alter future incentives and behavior, how does it change the expectations of all parties involved? Many policymakers and academics seem to believe in a general and ill- defined way that RTAs provide more than the traditional gains from trade; hence the term New Regionalism. Summarizing the conclusions reached in a confer- ence on regional integration, de Melo and Panagariya (1993) report that re- gional integration is increasingly recognized as going beyond trade in goods, services, and factors. Nonetheless, remarkably little work examines rigorously either the theoretical or the quantitative plausibility of these claims. The few papers that, with varying degrees of rigor, take up this question include Ethier (1 996a, 1996b), Gould (1992), Panagariya (1996), Perroni and Whalley (1994), and Whalley (1996). In this article we attempt to classify and assess the relative plausibility and importance of nontraditional gains from RTAs. After a brief discussion of the traditional costs and benefits of regional agreements, we examine other possible routes whereby RTAS might benefit their members. These include helping with problems of time inconsistency, signaling, insurance, bargaining power, and coordination. In each case, we discuss the necessary conditions for such an ex- planation to make sense, give one or more stylized examples of specific types of policy where the benefit might be applicable, examine real-world cases where the explanation might be relevant, and discuss the overall plausibility of the argument. We conclude by comparing NAFTA with the Europe Agreements in light of the preceding analysis. I. THEORETICAL MECHANISMS This section discusses several possible theoretical mechanisms by which an RTA could have real effects beyond the actual provisions of the agreement. We start with a short review of the traditional gains and losses from RTAS. Traditional Gains In one of the classic demonstrations of the theory of the second best, Viner (1950) shows that a move to free trade between two countries that maintain their respective external tariffs toward the rest of the world could leave both countries worse off. If as a consequence of liberalizing trade with only a subset of its trading partners, one country switched from trade with a relatively lower-cost producer to trade with a higher-cost producer, this would potentially decrease welfare for all. This negative effect on world efficiency is given the term "trade diverting." Some observers have argued that enlargement of the European Community has had a trade-diverting effect on agricultural trade (see, for example, Krugman Fernindez and Portes 201 1991). Southern European countries now buy grain and other products from relatively high-cost European sources rather than from lower-cost providers on the other side of the Atlantic, in Australia, or in Asia. Northern European coun- tries buy Mediterranean products like olive oil or more labor-intensive products from Southern Europe rather than from cheaper sources in Northern Africa. This effect could be even more pronounced in the context of a customs union that requires the equalization of external tariffs. Indeed, in the United Kingdom one of the main arguments used against entry into the European Community was that it would result in price increases for dairy products previously pur- chased primarily from New Zealand (see Winters 1993 for an evaluation of European Community integration). Nonetheless, Krugman (1991) argues that most RTAs likely entail relatively low welfare losses resulting from trade diversion because the countries involved are often geographical neighbors and hence already engage in a sizable amount of trade. However, for that very reason, RTAs may result in substantial losses of tariff revenue for the country that is lowering its tariff barriers (as pointed out in Panagariya 1996; see also Bhagwati 1993). Most often the less-developed, smaller, or economically weaker countries reduce their protectionist structures the most and bear the redistributional cost. If smaller countries bear the distributional cost, why has the number of RTAS increased, especially those involving smaller countries with a larger, more powerful partner? Why do smaller countries make the bulk of the concessions in these agreements? TERMS OF TRADE. In addition to possible distributional effects caused by reducing tariff or quota revenue, RTAS may also cause distributional effects at the expense of third parties. To understand why, note that if an RTA leaves all prices unchanged, with tariffs eliminated on members but maintained on all other countries, then member countries will buy more from one another, each member country will substitute away from consumption of its own goods, and all member countries will substitute away from consumption of goods bought from nonmember countries. Although the net effect on the demand of each member country is ambiguous (because its own demand has fallen, but member demand has risen), if goods are sufficiently strong substitutes, the demand for third-party goods will decrease. In order to clear markets, the price of third- party goods will have to fall, which (as long as no member country's price decreases by too much) will create a positive terms-of-trade effect for the member countries. Mundell (1964) makes this point.2 The potential beggar-thy-neighbor effect can make RTAS an attractive proposition for potential members despite any negative trade-diversion effects for member countries. INCREASING RETURNS AND INCREASED COMPETITION. A largely unmitigated beneficial effect that may be expected from RTAS stems from the increased size of 2. See Winters and Chang (1996) for an assessment of the terms-of-trade effects from Spain's accession to the European Community. 202 THE WORLD BANK ECONOMIC REVIEW, VOL. 12, NO. 2 the market, leading to greater productive efficiency for any industry with economies of scale. Indeed, as Krugman (1991: 9) notes, When the European Common Market was formed in 1958, substantial trade diversion seemed a likely outcome. What turned the arrangement into a strong economic success was the huge intraindustry trade in manufactures, and the associated rationalization of production, that the Treaty of Rome made possible. Note that this factor would make RTAs relatively more attractive for small countries than for large ones, particularly if the country's firms produced solely for the domestic market prior to the RTA. Small countries would also see the traditional benefits associated with increased competition. Thus, one might ex- pect to see them attach a higher priority to joining an RTA. INVESTMENT. Many analysts cite the need to attract foreign investment and stimulate domestic investment as an impetus for RTAS (see, for example, Balasubramanyam and Greenaway 1993 and Blomstrom and Kokko 1997). Investment flows between countries have grown dramatically in recent years and are of particular relevance to developing countries, especially as the importance of public sector concessional lending has decreased. An RTA could stimulate investment flows both between its constituent mem- ber countries and from outside the RTA in several ways. By (potentially) reducing the distortions in production within two member countries, an RTA could in- crease the overall quantity of investment made by investors in those countries. By increasing the size of the potential market, an RTA could increase the quantity of investment made both by domestic and by outside investors. This effect is particularly important for lumpy investments, like a factory, that might only be economic above a certain size. In the case of a customs union, by creating a single market within a common external tariff wall, an RTA may increase the incentive for foreign investors to engage in tariff jumping, that is, invest in a member country in order to trade freely with all other members of the customs union. It is not the purpose of this article to examine these effects in detail. However, it should be noted that several explanations of nontraditional gains from RTAS are of particular importance in relation to investment. The incentive to invest, for both domestic and foreign investors, depends crucially not only on current trade policies but also on future trade policies, on the nature and level of uncer- tainty, and on the general macroeconomic and political environment. We next turn to an examination of nontraditional mechanisms by which a country may gain from an RTA. Time Inconsistency An RTA could provide a country with nontraditional benefits if the RTA allows the country to pursue policies that are welfare improving but time inconsistent Fernandez and Portes 203 in the absence of the RTA. A well-known idea in the economic literature, intro- duced in the seminal paper of Kydland and Prescott (1977), is that a government that maintains policy discretion will be tempted every so often to surprise the private sector and that this temptation will undermine the credibility of optimal government policies. Hence, adherence to binding rules can restore credibility and lead to superior outcomes. In the arena of international trade, the problem of time inconsistency occurs if the government faces the temptation to undertake surprise trade policy ac- tions when other first-best instruments are not available. This may lead to gov- ernments finding themselves in suboptimal equilibria if they cannot make a cred- ible promise not to intervene. Staiger and Tabellini (1987) examine the credibility issues that arise when, in the presence of a terms-of-trade shock, the government is faced with the tempta- tion to use tariffs to transfer income from workers with high income (and hence relatively lower marginal utility of income) to those with low income (and rela- tively higher marginal utility). In a model in which the relocation of workers involves a loss in productivity, a terms-of-trade shock will leave wages unequal across sectors. Consequently, the government may be tempted to intervene by using trade policy to reduce the wage differential between sectors. To the extent that workers anticipate the government's intervention, the wage differential that results in equilibrium will be left unchanged. The intervention serves only to decrease the number of workers that exit the injured sector. This occurs because the anticipation of a subsidy to those workers who remain in the injured sector decreases the number of workers who exit, leaving the equilibrium wage plus the subsidy equal to the net return that the marginal worker would receive from exiting the injured sector. For a large set of parameters, the optimal policy for the government to pursue in the situation described above is free trade. This policy, however, will be time inconsistent if the government cannot commit to it by, for example, setting up in advance credible rules for trade policy. The reason for this is that, while antici- pated trade policy does not change the income distribution, unanticipated pro- tection does. Therefore, government has an incentive to surprise the private sec- tor, which, knowing this, no longer finds the ex ante optimal policy of free trade credible. The time-consistent equilibrium is given by a tariff level that is fully anticipated and such that the government no longer has any further incentive (because of large distortions) to surprise. This time-consistent equilibrium is in- ferior to the free-trade equilibrium, which, in the absence of some binding mecha- nism, cannot be attained. This story shows one of many plausible time-inconsistency problems in inter- national trade policy (see Staiger 1995 for a review of this literature). A simpler example is the case of a country whose optimal policy is to open itself up to foreign investment. Once foreign investment has been made, however, the coun- try has the ability to confiscate it (in practice, confiscation might be through the imposition of a greater regulatory or fiscal burden). Suppose that it would be 204 THE WORLD BANK ECONOMIC REVIEW, VOL. 12, NO. 2 optimal for a sufficiently impatient government to engage in such confiscation. If the country is unable to commit to abstaining from this action, foreign inves- tors, perceiving this time-inconsistency problem, will not invest. The country will be worse off than if foreign investors had invested and the government had not confiscated the investment. What is the possible contribution of an RTA to solving any set of time- inconsistency problems? To be a valid mechanism for solving a time-inconsis- tency problem, the RTA must fulfill two conditions. First, policies in the absence of an RTA must be time inconsistent. That is, they must be optimal ex ante, but not ex post, at least with sufficient probability to be a material concern. Second, exit from the RTA must cost enough to outweigh the gains from simply abrogat- ing the agreement and returning to the time-consistent policy. Is there some set of policies for which an RTA could fulfill these conditions? Here we evaluate how an RTA might contribute to solving time-inconsistency problems within the con- text of two sets of policies: trade and other domestic reforms. Unilateral liberalization of a country's economy, even for a small economy, is unlikely to be a time-consistent policy. Perhaps more important, it is also un- likely to be a politically popular move with import-competing sectors (see Hillman and Moser 1996 for an analysis of limited trade liberalization based on political concerns and the consequent impetus for cooperative reciprocal liberalization). The government is likely to face large temptations to protect some sectors in the economy for income distributional reasons, for political economy concerns, or for terms-of-trade considerations. Although the government may extend protec- tion and subsidies in only a few sectors at a time, over longer periods of time the economy is likely to become severely distorted.3 An RTA, by making the cost of even a small deviation from an agreed trade liberalization large (either by forc- ing the country to exit from the agreement or by having members punish the deviating country), makes it easier to overcome small temptations that culmi- nate in a greatly distorted economy overall. The question that must be faced here, however, is whether some easily avail- able alternative policy-either through GATT or through some domestic alternative-could enforce a country's commitment to trade liberalization. The domestic alternative is relatively easy to argue against; it is difficult to think of ways that a government can easily and credibly bind itself to not protect a sec- tor. It could be argued, however, that the private sector can punish a govern- ment that undertakes gradual forays into protection. This punishment would be part of a trigger strategy, whereby a protectionist act by the government would trigger the expectation on the part of private agents that more protection would be swiftly forthcoming. As a consequence, the private sector would immediately distort its production activities, thereby obtaining the expected protection sooner rather than later. The government, understanding the expectations and actions 3. Status quo bias arises for many reasons and may prevent the removal of protection once it has been extended; see Fernandez and Rodrik (1991) for a model that provides one explanation for this. Ferncindez and Portes 205 that would be triggered by a small deviation into protection, would therefore refrain from protection altogether. The problem with the above argument is not its game-theoretic correctness, but rather the immense amount of expectational coordination required on the part of atomistic agents. Note that it is but one of many equilibria, and although explaining cooperative behavior by trigger strategies is enlightening when ap- plied to repeated interactions among several large agents, the argument is less persuasive in the context of an economy with many small agents.4 The reason why GATT in some cases will not serve as a commitment device whereas an RTA may is more subtle. The answer, it seems to us, must lie in the differing incentives for countries to punish a deviating member within the two organizations. Within GATT, the responsibility for singling out a culprit and, if the organization delivers a guilty verdict, delivering some retaliatory punish- ment lies with the country or countries that have been hurt by the action. In a large organization with a more diffuse trade structure, this incentive is likely to be much smaller for any single member, and the process likely to be slower and the outcome more uncertain, than within a regional agreement. In an RTA it is much clearer who has the responsibility to punish, and the reputational loss from not doing so should accordingly be greater. For example, suppose that, within an RTA such as Mercosur, Argentina does not (credibly) threaten to retaliate when Brazil attempts to protect some indus- try. The incentive for Brazil to engage in this behavior again, with some other industry in some other instance, will be correspondingly larger. Argentina, there- fore, faces a high future cost from not retaliating. Within GATT, the country that offends next time or the set of countries that are hurt could very well be differ- ent, and the country responsible for not retaliating or not retaliating at the cor- rect level may be less identifiable, making the cost of not retaliating correspond- ingly smaller. Thus the extent to which RTAS increase the importance of regional trade and make unambiguous the set of partners responsible for the retaliation contribute to making RTAs a superior enforcement mechanism than GATT. An alternative way in which an RTA may dominate GATT as a mechanism that can render policies time consistent relies not on an RTA having a superior punish- ment mechanism than GATT, but instead on an RTA changing incentives in a way that differs from GATT. A prime candidate for this explanation is increased do- mestic and foreign investment. Most likely, investment would increase for a partner in an RTA more than it would under GATT. This is because, if credibly in place, an RTA gives preferential access to a given set of markets, whereas GATT does not. Furthermore, precisely this preferential access and the additional in- vestment that accompanies it may turn trade liberalization from a unilaterally or multilaterally unattractive proposition for a given country (due to domestic po- litical concerns, for example) into a politically feasible proposition (see Ethier 4. In an economy characterized by few and large players, a trigger strategy is more plausible. Furthermore, for problems, such as inflation, that visibly affect the majority of agents in the economy, it is easier to imagine how atomistic agents can coordinate on extreme equilibria. 206 THE WORLD BANK ECONOMIC REVIEW, VOL. 12, NO. 2 1996a, 1996b and Maggi and Rodriguez-Clare 1996). Thus trade reform that is infeasible without an RTA becomes feasible with one. This conclusion is rein- forced if the RTA includes aspects of deeper integration, such as harmonization of the investment code, that are harder to negotiate multilaterally, but that can increase the attractiveness of freer trade. Some analysts believe that the benefits of an RTA go beyond extending cred- ibility solely in the area of trade reform and spill over into other micro- and macroeconomic reforms. Whalley (1996: 16), for example, asserts that a desire to increase the credibility of domestic reforms was a central preoccupation be- hind Mexico's negotiating position on NAFTA: As such, it led to the outcome that Mexican negotiators were less concerned to secure an exchange of concessions between them and their negotiating partners and were more concerned to make unilateral concessions to larger negotiating partners with whom they had little negotiating leverage as part of the bilateral negotiation. The idea was clearly to help lock in domestic policy reform through this process. An RTA that clearly stipulates domestic reforms and an appropriate punish- ment incentive will likely serve as a commitment mechanism. However, most RTAS, including NAFTA, do not (although the Europe Agreements do contain ele- ments of commitment to domestic policy reform). It seems unlikely that RTAS will help provide commitment to policies that are not explicitly part of the agreement unless the agreement itself has changed the underlying incentives. Hence an expectation of a reformed monetary or fiscal policy on the basis of an RTA alone seems mistaken. The extent to which an economy within an RTA is more open than before the agreement may help to discipline macroeconomic policies, although it may just as well provide greater temptations to manipulate exchange rates in the hope of increasing competitive- ness, if only momentarily. For example, the establishment of Mercosur coincided with radical domestic policy reforms, which led some analysts to suggest that Mercosur's member coun- tries are trying to entrench domestic policy discipline, particularly because these countries have a long history of failed attempts at domestic macroeconomic ad- justment. But it is not clear whether the Real Plan in Brazil gains or loses in credibility because of Mercosur. On the one hand, the greater openness of the economy may increase the incentive to maintain fiscal discipline because any increase in aggregate demand is likely to leak into imports rather than into the demand for domestic products. On the other hand, Mercosur may make the maintenance of a semifixed exchange rate more difficult because the loss in com- petitiveness resulting from the combination of a fixed exchange rate and domestic inflation will lead more quickly to a surge in imports and a possible balance of payments crisis. The experience of the French government in the early 1980s provides an ex- ample of an RTA increasing a government's credibility. The French government Fernaindez and Portes 207 discovered that it could no longer pursue macroeconomic policies that differed radically from those of the Federal Republic of Germany because domestic fiscal expansion simply leaked into an unsustainable current account deficit. The po- litical result was an impetus toward greater macroeconomic policy coordina- tion, now culminating in the European Monetary Union. Although it might seem relatively easy to argue that an RTA would have difficulty surviving without some degree of fiscal and monetary discipline among the member countries, the his- tory of earlier RTAs among developing countries gives lie to the belief that an RTA will provide sufficient incentive for countries to do so (see, for example, Genberg and de Simone 1993 for an analysis of how import substitution practices and fiscal and monetary policy led to the failure of the Latin American Free Trade Agreement). The difference between the two experiences-the success of the European Union and the failure of some RTAs-may lie with the differential in exit costs, which could very well increase with the longevity of the agreement. In the discussion so far, the time-inconsistency arguments assume implicitly that the objective function of the country does not change. This assumption does not hold in the more interesting case of political time inconsistency. This case occurs when a country's current government's policies are not time incon- sistent, but the current government fears that a subsequent government might not have the same objective function. In this case, the current government could use an RTA to tie the hands of an alternative government in the future. For example, consider a country with two political parties, liberal and protec- tionist. When the liberal party governs, it may wish to sign an RTA to prevent the protectionist party from reversing trade liberalization when it comes into power. Thus the policies guaranteed by the RTA could be consistent with the objective function of the government that signs it and inconsistent with that of some fu- ture government. Furthermore, the RTA may affect not only the incentives of the protectionist party if it takes power but also the probability of the protectionist party coming into power. If the voters perceive that a protectionist party could implement its policies only at great cost, RTA membership might reduce the prob- ability of electing a protectionist party. By contrast, if free trade is popular, but the government is generally unpopular, the voters might elect the protectionist opposition, knowing that it will not be able to implement protectionist policies in practice. The potential benefits of RTAS for helping with political time-inconsistency problems rely on similar, although less restrictive, basic conditions as those re- quired for helping with economic time-inconsistency problems. With political time inconsistency, government policy itself need not be time inconsistent for the given objective function. Rather, the policies contained in the RTA must be time inconsistent under some possible future course of events, the RTA must effectively constrain such policies, and the cost of exit must be large. The political time-inconsistency argument clearly has some relevance in the case of the European Union. Although the European Union is a special case because its coverage is far deeper than that of any other RTA, its recent and likely 208 THE WORLD BANK ECONOMIC REVIEW, VOL. 12, NO. 2 future widening has some broader relevance. Existing and prospective members have always seen European Union membership as a way of committing them- selves to "Europe." By this is meant not only, and not even primarily, trade liberalization within Europe but also membership in the European political sys- tem of liberal democracy. This political aspect applied to the initial formation of the European Union, which had the clear political objective of constraining Ger- many. It also applied to the accession of Greece, Portugal, and Spain (each for- merly governed by authoritarian regimes of the right), to the Europe Agree- ments (designed to lead to eventual accession) with the countries of Central and Eastern Europe (formerly governed by authoritarian regimes of the left), and to the conditionality attached to the European Union's recent customs union agree- ment with Turkey (covering human rights and free speech issues). A similar political motivation may help to explain the recent spate of Latin American RTAS. Newly democratic countries after many years of authoritarian rule have actively sought to cement a new relationship among governments by entering into trade agreements of various sorts. But besides being a celebratory event, does entering into an RTA increase a country's commitment to democ- racy? Some agreements, such as Mercosur, have followed the European Union's lead in making democracy a necessary condition for membership. That these clauses are not necessarily vacuous was demonstrated in April 1996 when Mercosur foreign ministers warned a Paraguayan general contemplating a coup that diplomatic and economic isolation would follow (see the Mercosur supple- ment in The Economist, vol. 341, no. 7987 [October 12, 1996]). In addition, RTAS may deepen ties and facilitate the exchange of information among demo- cratic neighbors, and they may help to avoid conflicts among traditionally more pugnacious members, such as Chile and Bolivia (see Schiff and Winters 1996 for an analysis of the formation of RTAs as an optimum response to security prob- lems among its members). Signaling An alternative potential benefit from RTAS is signaling. For an RTA to work as a signaling mechanism, an RTA does not have to help bind a country to policies that would otherwise be time inconsistent. In this case, the prime motivation for the RTA is not the provisions of the agreement itself, but rather the fact that entering the RTA is only worthwhile under certain circumstances, and the coun- try wants to persuade others that those circumstances do in fact prevail. For example, consider a country that could have either a liberal or a protec- tionist government. Suppose that it has a liberal government, but this is not immediately apparent to outside observers, and that potential investors are not particularly interested in the exact provisions of any RTA, but they do care about the type of government. By entering an RTA, the government signals to investors that it is in fact liberal. Alternatively, the government may wish to signal something about the under- lying condition of the economy, such as the prospective competitiveness of its Fernandez and Portes 209 industry or the sustainability of the exchange rate. Suppose, for example, that the country wishes to attract investment to the manufacturing sector, but inves- tors do not know if the sector is likely to be competitive. If entering the RTA iS very costly for a country with an uncompetitive manufacturing sector, then the RTA may signal to foreigners that the manufacturing sector is competitive and hence stimulate investment. Finally, the RTA may signal not the policies of individual governments but their future relationships. This would be relevant if private investors are uncer- tain about the relationship between two governments or if they fear that trade barriers will be reimposed between the two countries. Investors may only be prepared to invest in one or both countries if both signal their future good rela- tions by signing an RTA. This may be particularly relevant for RTAs between developing countries, which historically have not been very successful and are often motivated by the desire to attract foreign investment. Two conditions are needed for a signaling explanation to make sense. First, there has to be a significant information asymmetry. That is, the government has to have superior knowledge, either about its own preferred policies or about the economy, than other agents. The information asymmetry condition is most likely to be met in cases where there is a significant degree of doubt about the government's commitment to liberalization or reform. It is thus most likely to be relevant in countries that have previously had relatively protectionist policies or where liber- alization attempts have failed in the past, like the Mercosur countries. Second, unlike the time-inconsistency case, there need not be a large cost of exiting an RTA, but rather there has to be a significant cost to entering the agree- ment, at least for some hypothetical governments in at least some circumstances. This argument, at least in theory, has some degree of plausibility. In facing down domestic opponents to an RTA, governments may have to invest political capital that may represent a sunk cost. Moreover, joining an RTA typically requires an immense amount of negotiating time and effort, particularly for developing coun- tries that may have only a small amount of the necessary expertise, with compet- ing claims on this resource. For example, it appears that the binding constraint on European Union membership for several Central European countries is cur- rently neither political nor macroeconomic, but the institutional and legal changes needed to bring these countries up to European Union standards. Of course, some bureaucracies may give positive weight to using resources in this fashion, thus invalidating use of this mechanism as a signal in those cases. Insurance An RTA can improve the welfare of its members by providing at least one of them with insurance against possible future events. This may help explain why in some agreements, particularly those involving a large and a small country, the smaller country may enter on worse terms. For example, consider a country that faces macroeconomic uncertainty about the level of its imperfectly flexible real wages. If real wages increase more than 210 THE WORLD BANK ECONOMIC REVIEW, VOL. 12, NO. 2 anticipated, the country would like to devalue to improve competitiveness and avoid a fall in output and employment. However, if it does so, one or more of its trading partners may impose trade restrictions. The fear of such an eventuality can explain why this country could seek to enter an RTA with that trading part- ner on relatively unfavorable terms, but with the provision that devaluation would not provoke a protectionist response. The unfavorable terms of the RTA then constitute a type of insurance premium paid by one country to the other. The situation need not be asymmetric; whichever country experiences an ad- verse shock would be allowed to devalue. However, most RTAS, including NAFTA, allow the imposition of contingent pro- tection (such as antidumping duties) and make exemptions for national security (which, as the case of the U.S. Helms-Burton Act shows, countries may attempt to interpret broadly). Thus the insurance role of an RTA may be severely limited. Some of these actions, however, can be seen as the outcome of the multisided pro- vision of insurance in which all members, under special circumstances, are al- lowed to claim protection (which is what these contingent protection clauses al- low). Total arbitrariness in claiming the need for this protection would be restricted both by the greater accountability implicit in an agreement with a small number of parties (relative to the wro) and by the repeated game nature of the accord. Perroni and Whalley (1994) hypothesize that smaller countries might seek to join an RTA as insurance against a generalized (world) trade war in which they would lose the most. In return for this insurance, the small countries offer the larger countries largely nontrade benefits such as greater protection for intellec- tual property. Alternatively, another inducement to join an RTA could be to ob- tain insurance against a major trading partner arbitrarily imposing, say, health standards. A country could more easily appeal such standards in the context of an RTA than under GATT. Raff (1996) offers a related interpretation in which the Canada-U.S. Free Trade Agreement is insurance against a trade war between the two countries that is generated by multiple equilibria in a bargaining game. In this interpretation, an RTA is a means of binding the parties to a cooperative way of "playing the trade game." Whalley (1996) sees the Canada-U.S. Free Trade Agreement as a way for Canada to obtain arrangements that limit the application of U.S. safeguard mea- sures and, in turn, assure the United States that Canada will not return to poli- cies in energy and investment that were adverse to U.S. interests. Flam (1995) estimates that Austria, Finland, and Sweden, which recently joined the Euro- pean Union, will benefit very little from liberalized trade with their trading part- ners. Under the European Economic Area provisions, they already effectively had free trade with the European Union countries. Joining the European Union meant that they had to make significant net transfers to the European Union budget. Again, an insurance argument might explain the decision of these coun- tries to join the European Union. The insurance explanation may help to explain, in part, the flourishing of RTAs now and in the 1970s. If some group of countries, say the European Union, Fernandez and Portes 211 has independent reasons (for example, decreasing the likelihood of another Eu- ropean war) for forming an agreement, then the very fact of their doing so may cause a chain reaction. Other countries, fearing more protectionist responses from "fortress Europe" and seeing a reduction in the number of potential alter- native trade partners (and hence the gains from diversification), decide to form their own regional block, thus provoking yet more countries to do so. The re- cent deepening and widening of the European Union, at least in part as a re- sponse to political concerns, could have triggered the chain reaction.5 In a less benign view of the chain reaction theory, a period in which low-wage countries increasingly compete for markets in industrial countries presents an opportunity for industrial countries to obtain significant concessions from low- wage countries by playing off their fears of being left out in the cold. A small country might sign an RTA with a large country in an effort to become the large country's favored low-cost supplier with lower protectionist barriers. That small country will become significantly more attractive to investors than the remain- ing low-cost countries. Thus smaller countries scramble to enter RTAs with larger partners and make significant concessions in order to do so. Bhagwati (1991), for example, cites then-President of Mexico Salinas as stating that a factor push- ing Mexico toward an RTA was the fear that European investment would be diverted to Eastern Europe once it integrated with the European Community. As Salinas put it at an early stage of the NAFTA negotiating process (as cited in Perroni and Whalley 1994: 4, from The Globe and Mail, April 10, 1990, p. Bi), What we want is closer commercial ties with Canada and the United States, especially in a world in which big regional markets are being created. We don't want to be left out of any of those regional markets. The insurance explanation is less convincing in the context of agreements between developing countries. For example, in the event of a global trade war or of a resurgence of protectionism in the major export markets of the Caribbean Community and Common Market (CARICOM) or Mercosur, all of the member countries would suffer, as they would to a lesser extent from any major macro- economic shock. Thus small countries are not particularly logical candidates to insure one another, and the chain reaction theory does not explain why these countries would enter into agreements with one another. Bargaining Power Countries may wish to join an RTA to increase their bargaining power with respect to third parties. This makes sense if the countries would in fact have greater bargaining power combined than separately and if the RTA would reduce the transaction costs involved in reaching an optimal negotiating position. This explanation is more convincing for a customs union, which has a common ex- ternal tariff, than for a free trade area, which does not. For example, the coun- tries of Mercosur have a large incentive to coordinate because a large fraction of 5. Ethier ( 1996a, 1996b) attempts to endogenize regional responses. 212 THE WORLD BANK ECONOMIC REVIEW, VOL. 12, NO. 2 Mercosur's trade is still extra- rather than intraregional. The member countries could obtain significant gains from coordinating their trade policies with respect to third countries. A related but logically distinct position is that in order to spur the slow pace of negotiations in GATT, countries such as the United States played the regional card. This explanation implies that as multilateral negotiations pick up steam, the importance of regional agreements should abate, which, to date, they show no sign of doing. Furthermore, agreements entered into solely by relatively small countries are unlikely to pressure the pace or direction of multilateral negotia- tions. Multilateral negotiations in the WTO are driven almost entirely by the "Quad" group (Canada, Japan, the European Union, and the United States); developing countries only tend to have influence when they speak collectively on a global basis, led by larger countries such as India. Countries may also seek to join an existing RTA to enhance their bargaining power with its current members. The most obvious example of this is the pro- gressive movement of most of the former European Free Trade Association (EFTA) countries into the European Union. Although the European Economic Area Agree- ment effectively instituted free trade between EFTA and the European Union, it also obliged the EFTA countries to accept existing European Union standards and case law. Seeing that they were receiving most of the benefits and paying most of the costs of the European Union, but with no say in how those benefits and costs were determined or distributed, most of the EFTA countries quickly opted for full membership. These countries became net contributors to the European Union budget and were obliged to harmonize their agricultural policies with the Com- mon Agricultural Policy. In return, they gained voting rights in future European Union decisions. Of course, the strength of the relative bargaining positions meant that the EFTA countries had to join largely on the European Union's terms, but they evidently believe that the long-term gains will outweigh any costs. Coordination Device Finally, an additional political economy explanation for free trade agreements in general is that they serve as a coordination device for those who stand to gain from trade liberalization. It is often argued that the gains from liberalization are widely distributed, uncertain, difficult to quantify, and perhaps longer term, while losses are immediate and visible and fall on specific and identifiable sec- tors. These factors make coordination much more difficult for those favoring free trade than for those opposing it. A proposed RTA can therefore serve as a political focal point for the forces in favor of free trade. Why would an RTA serve as the coordination device rather than multilateral agreements such as GATT? Part of the problem in mobilizing forces in favor of free trade is the uncertainty as to who will gain from liberalizing. This uncer- tainty may be significantly reduced in a regional rather than a global context. The only problem with this argument is that the losers also would be easier to identify. Nonetheless, the reduction of uncertainty could make risk-averse indi- Fernindez and Portes 213 viduals more favorable to an RTA than a multilateral venture. Furthermore, RTAS have very clear internal reciprocity, making it easier for a country to ensure that a concession on its part will elicit a counterpart from another country, benefit- ing itself. A different way for an RTA to serve as a coordination device is by allowing tradeoffs between different policy areas. For example, the United States may fear that free trade with Mexico will place downward pressure on U.S. environ- mental and labor standards. This in turn could lead to political pressure in the United States to reintroduce protection against Mexican exports. Both sides might therefore benefit if the United States commits to continuing its relatively open trade policy with Mexico, while Mexico commits to improving labor and envi- ronmental standards. NAFTA can be seen as providing a mechanism for both sides to do so. Furthermore, it is far more difficult to pursue such initiatives in a global context, because it is necessary to deal with multiple issues as well as numerous countries, which may have widely varying standards. This is illus- trated by the current lack of success on the part of those countries-notably France and the United States-that are trying to put labor standards on the WTO agenda. II. COMPARISON OF NAFTA AND THE EUROPE AGREEMENTS Section I discussed the theoretical mechanisms that might affect a country's incentive to join an RTA and the plausibility of the conditions necessary for these mechanisms to apply. This section examines two recent RTAS: NAFTA and the Europe Agreements. NAFTA NAFTA is an example of particular interest in light of the preceding discussion. Several factors suggest that nontraditional gains played a significant role in the creation of NAFTA. NAFTA involved a relatively low degree of liberalization because trade flows between Canada, Mexico, and the United States were already relatively free. In particular, the extent of U.S. trade liberalization was very small; Mexico made the significant concessions (see, for example, Hufbauer 1992). As a con- sequence, NAFTA is estimated to have had tiny direct macroeconomic effects on the United States, on the order of a few tenths of 1 percent of gross domestic product over a decade and a few tens of thousands of jobs, less than the mar- gin of error in the quarterly national accounts statistics (see, for example, National Commission on Employment Policy 1992). Even so, NAFTA aroused strong political opposition in the United States, while it was relatively popular in Mexico. Mexican policymakers regarded NAFTA as by far the most impor- tant economic priority. In Canada, it was largely a nonissue because the prin- cipal political battle over trade liberalization had been fought over the Canada- U.S. Free Trade Agreement. 214 THE WORLD BANK ECONOMIC REVIEW, VOL. 12, NO. 2 Some analysts have argued that economic time-consistency and insurance mechanisms explain Mexico's desire to join NAFTA, despite the small direct gains found by most estimates (Whalley 1996). However, it is not clear what types of domestic policy reform were involved. NAFTA does not cover macroeconomic policy at all, so it is difficult to see how it could have operated as a commitment device. If it was intended as such, it does not appear to have been effective in this instance because NAFTA did not ultimately make Mexican fiscal and monetary policy more credible, as the 1994 devaluation demonstrated. Nor did NAFTA cover domestic microeconomic reforms such as privatization or deregulation. Indeed, Mexico has made relatively little progress in these areas since NAFTA, recently backtracking on its commitment to privatize the state pe- troleum monopoly. NAFTA does commit Mexico to some fairly modest provi- sions with respect to existing labor and environmental law. However, these pro- visions were inserted at the request of the United States and were the least attractive part of the package from the Mexican point of view. The political time-consistency argument has also been advanced as an impe- tus behind NAFTA. For example, Gould (1992) argues that NAFTA will "weaken domestic political pressures from special interest groups to reverse trade liberal- ization." But domestic political pressures to reverse reform in Mexico never focused primarily on trade in any case; policies such as privatization and labor law reform, not covered by NAFTA, were and are far more controversial. How- ever, after the recent peso crisis, Mexican tariffs on countries outside NAFTA increased sharply. It could be argued that, had the agreement not been in place, the tariff increase would have been across-the-board (although that would also have permitted the increase to be smaller). Overall, therefore, the time- consistency argument probably has some validity, although it may have been more of an unforeseen consequence than an intended effect. It is perhaps more plausible to argue that Mexican policymakers tried to signal, to U.S. and other foreign investors, that domestic policy reforms were likely to continue and that Mexico's private sector was in good health. The insurance explanation appears to fit Mexico's accession to NAFTA reason- ably well. Mexico joined NAFTA on rather unfavorable terms, securing very little in concrete tariff reductions or other concessions from the United States. Al- though this was not perceived as an insurance premium by U.S. policymakers, Mexican policymakers may have perceived it as such. Again, this is particularly relevant in the context of foreign investment. To persuade U.S. investors to take advantage of Mexico's low labor costs by investing in Mexico, it was necessary to reassure them not only that tariffs for Mexican exports to the United States were low but also that they would stay low and that contingent protection would be less likely to be imposed. Macroeconomic developments did indeed lead to a post-NAFTA devaluation of the Mexican peso, which in turn resulted in a dramatic expansion of Mexican exports to the United States, mitigating the recessionary effects of the downturn in Mexican domestic demand. So the insurance-if insurance it was-paid out. Of Ferndndez and Portes 215 course, no one knows how the United States would have responded to this expan- sion in the absence of NAFTA. But it must be at least plausible that there would have been greater pressure for a protectionist response. Indeed, it is arguable that without NAFTA, the United States would not have provided the financial support that enabled Mexico to avert default in early 1995 and perhaps a much worse economic downturn. In this case, NAFTA clearly yielded very large insurance ben- efits to Mexico (without any ultimate net cost to the United States, yet). However, it may be stretching the point to argue that this particular insur- ance mechanism was planned. There is no evidence that Mexican policymakers attached any significant probability to the devaluation of December 1994; in- stead, they were worried about potential protectionism in the United States or perhaps third countries (as argued in Whalley 1996). Such protection has not in fact materialized, despite the huge swing in the trade balance in Mexico's favor that has resulted from the Mexican recession and devaluation. Mexico may sim- ply have been "lucky" enough to find itself in the position of the homeowner who gets earthquake insurance thrown in with the normal fire and flood coverage. Nontraditional gains also played an important role from the point of view of the United States. NAFTA was designed to serve several purposes beyond direct trade benefits. It served as a political reward for, and an investment in the long- term economic stability of, a friendly regime that was perceived to be supportive of U.S. interests. NAFTA, it was hoped, would reduce the pressures of illegal im- migration as Mexico's relatively free access to U.S. markets would increase in- vestment and employment in Mexico and hence improve the Mexican labor market. And NAFTA might potentially improve the U.S. bargaining position with respect to the European Union and Japan in multilateral trade negotiations. Al- though Mexico's share of world trade is small, it represents a significant share of U.S. external trade, and NAFTA may at least have signaled that the United States could attempt to substitute the trading partners of the Organisation for Eco- nomic Co-operation and Development with countries elsewhere, especially in Latin America. However, U.S. opponents of NAFTA saw primarily the traditional losses from free trade; that is, the downward pressure on the wages of relatively low-skilled workers in the United States, and reduced investment in labor-intensive industries. The Europe Agreements The Europe Agreements between the Central and Eastern European countries (Bulgaria, the Czech Republic, Hungary, Poland, Romania, and Slovakia) and the European Union were concluded in 1993. They are similar to NAFTA in sev- eral respects: they are between much larger industrial economies and much smaller emerging economies, they contain significantly fewer concessions by the large countries to smaller countries than vice versa (Perroni and Whalley 1994), and the trade liberalization provisions have had relatively small direct effects. 216 THE WORLD BANK ECONOMIC REVIEW, VOL. 12, NO. 2 As with NAFTA, the pure trade liberalization provisions of the Europe Agree- ments were not likely to have a substantial effect on trade between Central and Eastern Europe and the European Union. Although Wang and Winters (1991) and Hamilton and Winters (1992) find much smaller exports from Central and Eastern Europe to the European Union-by a factor of 2 to 10- than expected on macroeconomic and geographic grounds, trade barriers were not the primary cause. Using 1989 data, Aghion and others (1992) estimated that removing trade barriers completely would potentially expand Central and Eastern European exports to the European Union by about 40 percent, more than half of which was in agriculture, while other estimates were even lower. However, between 1989 and 1992, without complete liberalization, trade ex- panded more than 50 percent, mostly in sectors like manufactures, where tar- iff barriers were not that high (Portes 1993). This suggests that European Com- munity trade policy was not the primary factor behind the observed expansion in trade. This does not in itself logically imply that European Community trade policy could not affect the volume of trade, but it does suggest that other fac- tors are probably more important. The Europe Agreements differ from NAFTA in some important ways, however. Unlike NAFTA and most other RTAs-and unlike other agreements between the European Union and less developed or emerging economies-the Europe Agree- ments contain major provisions that bind the domestic economic policies of the Central and Eastern European countries. The agreements effectively require the application of European Union competition policy law to trade between the European Union and the Central and Eastern European countries; questionable practices are to be judged by reference to European Union law. Given the shaky legal structures of the Central and Eastern European countries and the impor- tance of their trade with the European Union, this amounts to nearly wholesale extension of European Union competition policy to the Central and Eastern European countries. Because member states retain jurisdiction over competition policy if only the domestic market is affected, the Central and Eastern European countries will do so as well. One area where this extension is particularly important is that of state aid- government subsidies or tax breaks targeted to particular industries. The Euro- pean Commission has been taking an increasingly aggressive line against state aid within the European Union. Many likely recipients of state aid in the Central and Eastern European countries will be potential exporters to the European Union, and this policy could-after an initial transition period-be exported to the Central and Eastern European countries. The policy could help the Central and Eastern European countries avoid a downward auction of tax breaks and subsidies designed to compete for foreign investment. The Europe Agreements therefore could act as a binding mechanism for the Central and Eastern Euro- pean countries to commit not to engage in such an auction in a way that agree- ments among just the Central and Eastern European countries would not, be- cause the cost of exit would be much lower. Ferndndez and Portes 217 The Europe Agreements also had a much stronger political element than NAFTA. Both the European Union and the Central and Eastern European countries wanted to lock in a political commitment to democracy in the Central and Eastern Euro- pean countries. The promise of eventual European Union membership implied in the agreements (although it was not set out in them, it was endorsed by the 1993 Copenhagen Summit) was conditional on the continued democratization of the Central and Eastern European countries. Therefore, a reversion to authoritarianism would cost the countries not just the loss of the benefits, if any, of the agreements but also the loss of the prospect of eventual membership in the European Union. On both political and economic grounds, the time-consistency argument therefore appears much stronger for the Europe Agreements than for NAFTA. By contrast, the insurance argument appears less persuasive. Perroni and Whalley (1994) argue that the agreements represent a purchase of insurance on the part of the Central and Eastern European countries against a global trade war. However, in light of the text of the agreements, this does not seem persuasive. In the areas where the chances that an increase in European Union protectionism-whether for internal reasons or because of a global trade war-could genuinely damage the economies of the Central and Eastern Eu- ropean countries-most importantly agriculture and one or two other sensi- tive sectors like steel-the European Union has written in effective "get-out" clauses that enable it to reimpose protection under special circumstances (Portes 1993). III. CONCLUSIONS RTAs differ hugely in their scope, coverage, and motivation, and there may very well not be any one-size-fits-all explanation of their recent proliferation. However, some common themes do emerge. In particular, the increasing impor- tance of international investment flows of private sector capital, both for indus- trial and, especially, for developing countries, has heightened the need for coun- tries not just to put the right policies into place but also to provide certainty and credibility as to the direction of future policies and the economic environment more generally. It is no coincidence that this is the common thread running through the various theoretical mechanisms described in this article; commit- ment, signaling, and insurance mechanisms all have the practical effect of reduc- ing uncertainty or increasing credibility, whether about future national or inter- national economic policies and events or about political developments. RTAS can therefore serve a useful economic purpose above and beyond the direct gains from trade liberalization by reducing such uncertainties and enhanc- ing credibility. Such credibility would apply to a stable legal environment in Poland, continued access to U.S. markets for Mexican products, or a local mar- ket of sufficient size for a new plant in Uruguay. Increased credibility makes it easier for the private sector to plan and invest. Indeed, in some cases the reduc- 218 THE WORLD BANK ECONOMIC REVIEW, VOL. 12, NO. 2 tion in uncertainty resulting from an RTA may even be a necessary precondition to realizing gains from liberalization. Nevertheless, before drawing further conclusions about the qualitative and quantitative relevance of these nontraditional benefits from RTAS, analysts should examine closely the provisions of each RTA. What is its scope and coverage? What, if any, is the enforcement mechanism? How and in what circumstances can it be amended? How might it have changed the behavioral incentives of different agents in the economy? These questions are particularly important for assessing the effect of the RTA not directly on trade provisions, but indirectly on incentives, expectations, and future policies. This article concludes with a word of caution. Even if RTAs do provide non- traditional benefits, such benefits are not necessarily a good thing. Many per- suasive voices have argued the possible costs that a proliferation of RTAS may entail (for example, Bagwell and Staiger 1996 analyze the positive and negative effects of regional agreements). Nonetheless, independent of their benefits or detriments from a world welfare point of view, it is important to understand what benefits or costs they provide to their own members and the sources of their apparently increased attractiveness. REFERENCES The word "processed" describes informally reproduced works that may not be com- monly available through library systems. Aghion, Philippe, Ralph Burgess, Jean-Paul Fitoussi, and P. A. Messerlin. 1992. "To- wards the Establishment of a Continental European Customs Union." In John Flemming and J. M. C. Rollo, eds., Trade, Payments, and Adjustment in Central and Eastern Europe. London: Royal Institute for International Affairs. Bagwell, Kyle, and Robert Staiger. 1996. "Regionalism and Multilateral Tariff Coop- eration." Northwestern University and University of Wisconsin-Madison. Processed. Balasubramanyam, V. N., and David Greenaway. 1993. "Regional Integration Agree- ments and Foreign Direct Investment." In Kym Anderson and Richard Blackhurst, eds., Regional Integration and the Global Trading System. New York: St. Martin's Press. Bhagwati, Jagdish. 1993. "Regionalism and Multilateralism: An Overview." In Jaime de Melo and Arvind Panagariya, eds., New Dimensions in Regional Integration. Cam- bridge, U.K.: Cambridge University Press. . 1991. The World Trading System at Risk. Princeton, N.J.: Princeton University; London: Harvester Wheatsheaf. Blomstrom, Magnus, and Ari Kokko. 1997. "Regional Integration and Foreign Direct Investment." Policy Research Working Paper 1750. Policy Research Department, World Bank, Washington, D.C. Processed. de Melo, Jaime, and Arvind Panagariya. 1993. New Dimensions in Regional Integra- tion. Cambridge, U.K.: Cambridge University Press. Ethier, Wilfred. 1996a. "Multilateral Roads to Regionalism." University of Pennsylva- nia, Philadelphia. Processed. Ferndndez and Portes 219 1996b. "Regionalism in a Multilateral World." University of Pennsylvania, Phila- delphia. Processed. Fernandez, Raquel, and Dani Rodrik. 1991. "Resistance to Reform: Status Quo Bias in the Presence of Individual-Specific Uncertainty." American Economic Review 81(5):1146-55. Flam, Harry. 1995. "From EEA to European Union: Economic Consequences for the EFTA Countries." European Economic Review 39(3-4):457-66. Genberg, Haus, and F. N. de Simone. 1993. "Regional Integration Agreements and Macroeconomic Discipline." In Kym Anderson and Richard Blackhurst, eds., Regional Integration and the Global Trading System. New York: St. Martin's Press. Gould, David. 1992. "Free Trade Agreements and the Credibility of Trade Reforms." Economic Review, First Quarter, pp. 17-27. Dallas: Federal Reserve Bank of Dallas. Grossman, Gene, and Elhanan Helpman. 1991. Innovation and Growth in the World Economy. Cambridge, Mass.: MIT Press. Hamilton, C. B., and L. Alan Winters. 1992. "Opening up International Trade with Eastern Europe." Economic Policy (14):77-116. Hillman, Arye, and Peter Moser. 1996. "Trade Liberalization as Politically Optimal Exchange of Market Access," In Matthew Canzoneri, Wilfred Ethier, and Vittorio Grilli, eds., The New Transatlantic Economy. Cambridge, U.K.: Cambridge Univer- sity Press. Hufbauer, G. C. 1992. The North American Free Trade Agreement: Issues and Policies. Washington, D.C.: Brookings Institution. Krugman, Paul. 1991. "Move towards Free Trade Zones." Federal Reserve Bank of Kansas City Economic Review (November-December):5-25. Kydland, F. E., and E. C. Prescott. 1977. "Rules Rather than Discretion: The Inconsis- tency of Optimal Plans." Journal of Political Economy 85(3):473-91. Maggi, Giovanni, and Andres Rodriguez-Clare. 1996. "The Value of Trade Agreements in the Presence of Political Pressures." Princeton University, Princeton, N.J.; Univer- sity of Chicago. Processed. Mundell, Robert. 1964. "Tariff Preferences and the Terms of Trade." Manchester School of Economic and Social Studies 1(1):1-13. National Commission on Employment Policy. 1992. Special Report on NAFTA. Wash- ington, D.C. Panagariya, Arvind. 1996. "The Free Trade of the Americas: Good for Latin America?" University of Maryland, College Park. Processed. Perroni, Carlo, and John Whalley. 1994. "The New Regionalism: Trade Liberalization or Insurance?" NBER Working Paper 4626. National Bureau of Economic Research, Cambridge, Mass. Processed. Portes, Jonathan. 1993. "Trade Relations between the European Community and the Countries of Eastern and Central Europe." Central European Department, World Bank, Washington, D.C. Processed. Raff, Horst. 1996. "The 'New Regionalism' as Insurance." Indiana University, Bloomington. Processed. Rivera-Batiz, Luis, and Paul Romer. 1991a. "International Integration and Endogenous Growth." Quarterly Journal of Economics 106(2):531-55. 220 THE WORLD BANK ECONOMIC REVIEW, VOL. 12, NO. 2 . 1991b. "International Trade with Endogenous Technological Change." Euro- pean Economic Review 35(4):715-21. Schiff, Maurice, and L. Alan Winters. 1996. "Regional Integration as Trade Diplomacy." International Economics Department, World Bank, Washington, D.C. Processed. Staiger, Robert. 1995. "International Rules and Institutions for Cooperative Trade Policy." In Gene Grossman and Ken Rogoff, eds., Handbook of International Eco- nomics, vol. 3. Amsterdam: North-Holland. Staiger, Robert, and Guido Tabellini. 1987. "Discretionary Trade Policy and Excessive Protection." American Economic Review 77(5):823-37. Viner, Jacob. 1950. The Customs Union Issue. New York: Carnegie Endowment for International Peace. Wang, Z. K., and L. Alan Winters. 1991. The Trading Potential of Eastern Europe. CEPR Discussion Paper 610. London: Centre for Economic Policy Research. Whalley, John. 1996. "Why Do Countries Seek Regional Trade Agreements?" NBER Working Paper 5552. National Bureau of Economic Research, Cambridge, Mass. Processed. Winters, L. Alan. 1993. "The European Community: A Case of Successful Integration?" In Jaime de Melo and Arvind Panagariya, eds., New Dimensions in Regional Integra- tion. Cambridge, U.K.: Cambridge University Press. . 1996. Regionalism versus Multilateralism. CEPR Discussion Paper 1525. Lon- don: Centre for Economic Policy Research. Winters, L. Alan, and Won Chang. 1996. "Integration and Non-Member Welfare: Mea- suring the Price Effects." International Economics Department, World Bank, Wash- ington, D.C.; Columbia University, New York. Processed. wTo (World Trade Organization). 1995. Regionalism and the World Trading System. Geneva. THE WORLD BANK ECONOMIC REVIEW, VOL. 12, NO. 2: 221-49 Trading Arrangements and Industrial Development Diego Puga and Anthony J. Venables This article outlines a new approach for analyzing the role of trade in promoting in- dustrial development. It offers an explanation as to why firms are reluctant to move to countries with lower labor costs and shows how trade liberalization can change the incentives for firms to locate in developing countries. It models economic development as the spread of concentrations of firms from country to country. Different trading arrangements may have a major impact on this development process. By changing the attractiveness of countries as a base for manufacturing production, they can poten- tially trigger-or postpone-inzdustrial development. The analysis shows that unilater- ally liberalizing imports of manufactures can promote industrializationz but that mem- bership in a preferential trading arrangement is likely to create larger gains. South-South preferential trading arrangements will be sensitive to the market size of member states, while North-South arrangements seem to offer better prospects for participating south- ern countries, if not for excluded countries. How do different trading arrangements influence the industrialization process in developing countries? Can preferential trading arrangements (PTAS) be supe- rior to multilateral liberalization or, at least, be an alternative when multilateral liberalization proceeds slowly? If so, what form should the PTAS take? Should developing countries seek PTAs with industrial countries or among themselves? Traditional analysis answers these questions using the ideas of trade creation and trade diversion. For example, consider two developing countries. Based on their comparative advantage, they each produce agricultural products and a dif- ferent manufactured good, export only agricultural products, and import manu- factured goods from an industrial country. Can a PTA between these developing countries promote industrialization? The answer is yes: they can trade their manufactures instead of importing them from the industrial country. The two countries will increase their production of manufactures through trade diver- sion. The PTA and consequent trade diversion may reduce welfare if they create regional import substitution (see de Melo, Panagariya, and Rodrik 1993). The problem with this analysis is that it starts by assuming a pattern of com- parative advantage. In the initial situation, the analysis assumes that the devel- Diego Puga is with the Centre for Economic Performance, London School of Economics, and with the Centre for Economic Policy Research; Anthony J. Venables is with the London School of Economics and the Centre for Economic Policy Research. The authors thank Maurice Schiff, Alan Winters, and the anonymous referees for valuable comments on an earlier version. This article was produced as part of the World Bank Development Economics Research Group's research project on regionalism and development. C) 1998 The International Bank for Reconstruction and Development/THE WORLD BANK 221 222 THE WORLD BANK ECONOMIC REVIEW, VOL. 12, NO. 2 oping countries import manufactures only because the industrial country has a comparative advantage in manufactures. Given this assumption, the analysis concludes that PTAS promote industrialization in the developing countries by working against their comparative advantage. In sharp contrast, the apparently changing comparative advantage of newly industrialized countries suggests the need for an analysis of potentially flexible comparative advantage, in which developing countries can develop and converge to industrial economies in both income and economic structure. How can we extend the analysis to allow for the dramatic changes in relative income and in industrial structure that have occurred in some developing coun- tries? One way is to build a model of trade and growth and then see how trading arrangements change the incentives for factor accumulation and hence coun- tries' rates of growth and relative factor endowments. Although a small litera- ture analyzes the growth effects of PTAS, the writings in this area do not yet have sufficient microfoundations to discriminate convincingly between different types of trading arrangements (for a survey, see Baldwin and Venables 1995). An alternative approach is to suppose that countries have few fundamental differences that generate immutable patterns of comparative advantage. Instead, history determines most of the pattern of trade and development in the world economy. Cumulative causation has created concentrations of industrial activ- ity in particular locations (industrial countries) and left other areas more depen- dent on primary activities. According to this approach, economic development can be thought of as the spread of these concentrations from country to country, and different trading arrangements may have a major impact on this develop- ment process. By changing the attractiveness of countries as a base for manufac- turing production, they can potentially trigger-or postpone-industrial development. In this article we develop this alternative approach and illustrate how trading arrangements can shape economic development. We use building blocks from new trade theory and from somewhat older development economics. As in new trade theory, we focus on the location of firms using technologies with increas- ing returns and operating in imperfectly competitive environments. From devel- opment economics, we take the ideas of forward and backward linkages be- tween firms. Combining these linkages with imperfect competition creates pecuniary externalities between firms, thus providing the mechanism for cumu- lative causation. The pecuniary externalities support existing agglomerations of industrial activity and provide a mechanism for the "takeoff" of newly industri- alizing economies. Throughout the article, we concentrate exclusively on the trade flows gener- ated by these agglomeration forces and assume that countries have no underly- ing differences in technology or relative factor endowments. This is clearly an extreme position that abstracts from traditional comparative advantage. By ab- stracting from such differences, we do not mean to suggest that they are not important. Rather, we seek to focus on the way in which agglomeration forces Puga and Venables 223 can determine industrial location and on how trade policy may change this determination. Section I provides an overview of the analytical framework, and the appendix presents more detail. Section II runs through a series of experiments, simulating the effects of different trading arrangements on the industrialization process and showing how alternative arrangements can lead to quite different patterns of development. It shows how trade liberalization may have dissimilar impacts on similar member economies, creating internal tensions within a PTA. Section III draws out the policy implications of our findings and discusses evidence of the empirical relevance of the forces captured by this framework. Section IV sum- marizes the main conclusions. I. ANALYTICAL FRAMEWORK This section provides an informal overview of the key features of the model; the appendix gives further details. We assume that each country may have two sectors-agriculture and industry. The agricultural sector is a perfectly competi- tive commodity sector. It produces output using a sector-specific factor (land) and a sectorally mobile factor (labor). For simplicity we assume that the agricul- tural product is traded freely. A more realistic model of the agricultural sector including trade costs would not alter the main results of the article but would shift the focus away from our main concern here, the effects of trade policy on industrialization. Fujita, Krugman, and Venables (1997) explore the effects of agricultural trade costs in a model of this type. We also assume that agriculture does not use intermediates; Pande (1997) relaxes this assumption. We focus the analysis on the other sector, industry, although the two sectors interact in general equilibrium. As industry relocates, agriculture adjusts to re- lease or absorb labor and to maintain balance of payments; the wage in a coun- try is higher, the smaller is that country's agricultural employment. The industrial sector takes the form of a monopolistically competitive in- dustry in which firms produce differentiated products. We model this as Dixit- Stiglitz (1977) monopolistic competition, in the form applied to international economics by Helpman and Krugman (1985) and others. We generalize the model to include intermediate goods, along the lines of Krugman and Venables (1995) and Puga and Venables (1996, 1997). That is, each firm's output is used both as a final good and as an intermediate good, and each firm uses as inputs both labor and the output of other firms. The presence of intermedi- ate goods combined with imperfect competition generates the forward and backward linkages that are central to our approach. Rather than working with a full input-output structure (as in Puga and Venables 1996), we work with a single aggregate sector that uses its own output as input. Correspond- ing to our assumption that there is a single manufacturing sector, there is also a single trade policy instrument-tariffs on imports of the manufactured goods. 224 THE WORLD BANK ECONOMIC REVIEW, VOL. 12, NO. 2 Firms enter and exit in response to short-run profit opportunities, giving a long-run zero profit equilibrium that determines the level of industrial activity in each country. Four forces determine the short-run profitability of firms in a particular country. The first is factor market competition. A country that has a lot of industry will have higher wages, reducing the profitability of firms. The second is product market competition. Given some trade barriers, other things being equal, a country with more industry will have lower output prices, also reducing profitability. These standard neoclassical forces work to disperse activ- ity and to encourage firms to locate where labor is cheap and where there is little supply from other firms. The last two forces-cost (forward) and demand (backward) linkages-work in the other direction. Cost linkages occur because having more firms in a loca- tion means that more intermediate inputs are available locally, reducing costs and raising short-run profits. Demand linkages arise because having more firms in a location increases intermediate demand, raising the sales and profitability of other firms. Both these forces mean that firms want to set up in the same country as existing firms. They are centripetal forces working to concentrate industry in a single location. Tension among the four forces determines the equilibrium pattern of loca- tion. If the first two forces are more powerful than the last two, then it is gener- ally the case that industry will operate in all locations. This situation represents a standard "new trade theory" world. In this case, the model does not generate a distinction between industrial and developing countries. Assuming that all countries have the same relative endowments, technologies, and preferences, then they all will have similar industrial structures and patterns of trade. Differ- ences in market size may give rise to net trade (see, for example, Krugman 1980). If the last two forces are more powerful than the first two, then equilibrium will involve agglomeration of manufacturing in a subset of countries. Under these circumstances, the presence of an extra firm will raise the short-run profit- ability of existing firms in the country. Because firms enter in response to short- run profits, the increased profitability will attract more firms, generating a cu- mulative process of industrial agglomeration. In this case, equilibrium will not be unique. The model does not determine in which country or countries the industrial agglomeration will locate. However, equilibria have the property that, without assuming differences in underlying comparative advantage, some countries will have industry and other countries will not. Real income differences are associated with this uneven pattern of in- dustrialization. The countries with industry will be richer for two reasons: the demand for labor in industry raises wages, and the local supply of manufactures reduces the consumer price index. They will also have a larger market, arising both from consumer and intermediate demands. And they will have a better supply of intermediate goods, showing up as a lower price index for these goods. At equilibrium there may be quite large differences in wages and unit labor costs between the industrial and developing countries; however, despite these differ- Puga and Venables 225 ences, it is not profitable for a firm to relocate to a developing country. If a firm were to do so, it would benefit from lower wages and from being the only local supplier in this market (factor-market and product-market competition effects). But it would forgo the benefits of proximity to its suppliers and its industrial customers (the forward and backward linkages). How does trade liberalization affect this situation? In this model, whether agglomeration occurs depends critically on the level of trade barriers. With ex- tremely high barriers (autarky), agglomeration does not occur; each country must have industry to supply its consumers. And under perfectly free trade, ag- glomeration does not occur because proximity to industrial suppliers and con- sumers has no economic significance if trading across space is costless. Agglom- eration takes place only when trade barriers are between these extremes. Changing trade barriers may cause agglomerations to develop and to disappear (see Krugman and Venables 1995 and Puga 1998). In this article, we explore a variety of trade liberalization experiments. Three main mechanisms are at work. The first mechanism has to do with the barriers incurred in exporting from a developing country. Reducing these barriers re- duces one of the disadvantages of developing countries and makes it cheaper to export from the developing country to the large industrial-country market. There- fore, we expect to see reductions in import barriers in industrial countries, facili- tating the spread of industry to developing countries. The second and third mecha- nisms have to do with import barriers in developing countries. Opening markets to increased competition from foreign firms in product markets reduces the po- tential profitability of local firms. But lower import barriers make it cheaper to import intermediate goods and thus raise potential profitability. The combina- tion of these mechanisms often produces an effect that is, in some sense, greater than the sum of the parts. It can trigger cumulative causation, leading to quite large changes in levels of industrial activity. (The model does not allow trade liberalization to change the technology in use or to change the price markups through strategic interaction between firms. As a consequence equilibrium firm scale is constant.) Different PTAS offer a variety of combinations of reductions in trade barriers that affect differently countries with different amounts of established industry, different wage rates, and markets of different sizes. The next section looks at how the balance among market access, import competition, wage differentials, and linkages is affected by different PTAs. We study whether trade policy can make industry spread to developing countries and, if so, what trading arrange- ments are most conducive to this spread. II. TRADING ARRANGEMENTS In this section, we go fairly rapidly through a set of trade policy experiments and postpone a fuller discussion of policy issues until section III. In all our ex- periments, we work with four countries, assumed to be of equal size, that is, 226 THE WORLD BANK ECONOMIC REVIEW, VOL. 12, NO. 2 having the same factor endowments. We assume values for the parameters such that there is an initial equilibrium in which manufacturing is concentrated in just two of the countries. Within the formal structure of the model, which two countries is indeterminate. We simply label the two countries that have industry North, and the two that do not South. We base the analysis on four countries because for the questions addressed we require two southern economies and because having four countries allows for symmetry between regions. In the gen- eral case of a large number of different-size economies, although generally in- dustry will only operate in a subset of countries, there are many possible equilib- rium assignments of industry to countries. We set the following structure of trade barriers between economies. All trade flows in manufactures have an equal level of real trade costs per unit that can be thought of as the costs incurred when doing business at a distance. In the initial equilibrium, we assume that ad valorem tariffs are zero between the two north- ern economies and positive and equal on all other manufactured trade flows. The distinction between real trade costs and tariffs is that only tariffs generate revenue. Changes in either type of barrier have similar effects on the location of industry. In this section, we report the results of experiments that reduce some or all of the tariff barriers, corresponding to different trade liberalization pack- ages. In all the experiments, we assume that the two northern economies follow identical policies and keep identical economic structures (the reason for this is simply to focus on the South). We consequently refer to the North as a single policymaker. All the results presented are based on numerical simulations of the model. The appendix discusses the parameter values underlying the simulations. We do not report a systematic sensitivity analysis in the article, but we do comment at a number of points on how changing the model's parameters affects the results. Judging by the research we have undertaken, the qualitative conclusions we present are quite robust. A series of figures illustrates the outcomes of the experiments. The level of tariffs is denoted T, such that T = 1 is free trade and T- 1 is the ad valorem tariff rate. The initial value, T, is the same for all manufacturing trade flows involving a southern economy. Liberalization will reduce some (or, in the case of multilat- eral liberalization, all) of these tariffs, with those not affected by the liberaliza- tion held at T. The figures are constructed by reducing tariffs from their initial level in a series of small steps. At each step, there is a change in the short-run profitability of firms, and we let our entry and exit dynamic work until a new long-run equilibrium is established. The tariff is then reduced at the next step, and the procedure is repeated. The figures therefore trace values of the variables along a stable equilibrium path from the initial equilibrium, and we describe the evolution of the economy along this path. This procedure is necessary because, in general, the model exhibits path dependence. Each subsection presents a figure with two graphs. In each figure, the first graph plots the share of world industry in each of the two southern economies Puga and Venables 227 against the level of tariffs, and we use this graph to demonstrate the way in which liberalization causes industry to relocate. The second graph plots real income per worker against the level of tariffs. Real wages include tariff revenue (distributed to workers in a lump-sum manner) for the two southern countries and for the North. Real wages change because of changes in the demand for labor, the consumer price index in each country, and level of tariff revenue. The evolution of real wages excluding tariff revenue is similar to that presented in the figures. Multilateral Liberalization We take as the benchmark case multilateral trade liberalization among all countries. At the initial level of tariff barriers, T = 1.15 for all North-South and South-South trade, while there is free North-North trade. At this level, the whole of industry is agglomerated in the North; in figure 1, lines SI and S2 show that the share of each of the southern economies in world trade is zero. At this equi- librium, real wages in the South are approximately 65 percent of those in the North despite the fact that there are no differences in technology, labor skills, or relative endowments. As global tariffs T are reduced, at a point around 1.14 it becomes profitable for some firms to relocate in the South. This occurs when the more open market reduces the short-run profitability of southern firms (potential, if not yet actual firms), but the fall in the price of imported intermediate goods and easier access to the large northern market increase it. The last two forces-combined with the large initial wage difference-dominate, causing industry to move to the South. It is a general property that if trade barriers-real as well as tariff barriers-are low enough, then agglomeration in a subset of economies is unsustainable and industry will operate in all countries (see Puga 1998). The agglomeration of industry in the North will be sustainable over a wider range of tariffs if the northern economies are larger (relative to those in the South) or more integrated (that is, intranorthern trade barriers are lower). However, industry initially operates only in one of the southern countries. If the two southern countries are identical, the choice of which is entirely a matter of chance; we label it Sl. This uneven spread of industry occurs because the first firms to set up create cost and demand linkages to other firms in the same coun- try. They also raise wages, but the linkage effects are stronger than the wage effects, so the second industrial agglomeration forms in just one of the develop- ing countries. To understand further why industry spreads to just one country, note that there is a second equilibrium with industry operating in both SI and S2, but this equilibrium is unstable. If S, has slightly more firms than S2 then it will have higher short-run profits, attracting more firms and raising profits further. Other mechanisms may reinforce this result-most obviously a confidence factor cre- ated by the success of early entrants. The extent and form of divergence between southern economies depend on model specifications. In a full input-output 228 THE WORLD BANK ECONOMIC REVIEW, VOL. 12, NO. 2 Figure 1. Multilateral Liberalization Share of world industry in the southern countries Industry share 0.35- 0.30- 0,25- 0.20- SI 0.15- ~ 0.10-\ 0.05- 52 I I I~~~~~~~~~~~~~S 1 1.025 1.050 1.075 1.100 1.125 1.150 Tariffs, T Real income per worker in the southern countries and the North Real income per worker 1.0- 0.9- 0.8--\ S2\ 0.7- 1 1.025 1.050 1.075 1.100 1.125 1.150 Tariffs, T ANoe. S, and S, denote southern countries I and 2: Ndenotes the North. Source: Authors' calculations. Puga and Venables 229 structure in which not all sectors are tightly linked, each southern economy may develop agglomerations in different sectors. Even so, there is always a tendency for unequal development in the aggregate. In the range of tariffs from around 1.14 to 1.10 industry operates in one southern country and not the other, but as tariffs are reduced below this range, it becomes profitable for manufacturing firms to establish themselves in the other developing country, S2. This process is very abrupt and occurs partly at the ex- pense of Sl, which experiences a small fall in its share of world industry. At tariffs below this point, the two southern economies are identical, and further reductions in T bring a steady relocation of industry to the southern economies. At completely free trade, each of the now-developed southern countries has 25 percent of world industry (equal to its share of the world endowment). Figure 1 illustrates the corresponding real wage situation. Although welfare effects are not caused directly by the evolution in country shares of industry, they are closely related. Countries with more industry have higher labor demand and have to import fewer varieties of goods that are subject to trade barriers. Both the welfare and labor demand effects support the real income differences in figure 1. As wages increase in southern economies, real wages in the northern economies decline, although this result is not general. The combined effect of changing labor demand and price indexes on northern wages is ambiguous, with the balance of decline and increase depending quite sensitively on parameter values. Figure 1 illustrates two main messages. First, trade liberalization breaks down existing agglomerations of activity. Lower trade costs make firms more foot- loose and more sensitive to international differences in factor prices and drives the convergence in the figure. Second, the benefits do not necessarily flow equally to the two southern economies. It follows directly from the presence of agglom- eration forces in the model that as the northern agglomeration starts to break down, new agglomerations may develop. Unilateral Liberalization In the case of unilateral liberalization, a single southern economy (SI) engages in unilateral import tariff liberalization, with all other barriers held constant (at value T = 1.15). The solid lines in figure 2 (and all remaining figures) outline the process, and the dashed lines represent the reference case of multilateral liberalization. In figure 2, openness to imports of manufactures causes manufacturing pro- duction to start (see Venables 1996). Import competition obviously has a nega- tive effect through the product market, particularly because access to the north- ern market is not liberalized. But the cheaper supply of imported intermediate goods becomes the dominant force, enabling industry to become established. This result is not general-for example, if the southern economies are very small and face high export barriers, unilateral liberalization does not cause industry to develop. But providing this is not the case, we find that the combination of low 230 THE WORLD BANK ECONOMIC REVIEW, VOL. 12, NO. 2 Figure 2. Unilateral Liberalization Share of world industry in the southern countries Industry share 0.35 - 0.30 - 0.25 ......... ..... 0.20 - 0.15 - 0.10- 0.05 - ~ S2 j\ S S * o~~~~~~~~~ 1 1.025 1.050 1.075 1.100 1.125 1.150 Tariffs, T Real income per worker in the southern countries and the North Real income per worker 1.0- N ..- ............~ ~ ~ ~~~~~~.... 0.9. 0.89- 0.7 -= 1 1.025 1.050 1.075 1.100 1.125 1.150 Tariffs, T Note: S, and 52 denote southern countries 1 and 2; Ndenotes the North. Source: Authors' calculations. Puga and Venables 231 wages and low-cost intermediates (due to import liberalization) is sufficient to lead to industrialization. In the cases we have studied, industry will develop sooner and at larger scale (that is, the S, curve will be higher) the greater is the share of intermediates in production, and the larger is the market in the liberal- izing economy. Furthermore, the unilateral reduction in tariffs on imports unambiguously raises wages in the liberalizing country. Before industrialization takes off, there is a slight decrease in real wages in the liberalizing economy because of falling tariff revenue. But as soon as the country starts to attract some industrial pro- duction, real wages inclusive of tariff revenue rise unambiguously. The unilat- eral liberalization policy has no direct effect on the other southern economy because it has no industry to benefit from S1's liberalization. However, the other economy does experience a slight increase in real wages because the terms of trade improve due to the increased world supply of manufactures. Under unilateral liberalization the continuing barriers to developing coun- tries' exports mean both that it takes a lower tariff rate to start industrialization and that, once industrialization has started, S1 has a lower share of manufactur- ing than it would in the multilateral case. In addition, real wages are lower in SI than they would be in the case of multilateral liberalization. South-South PTA In the case of a South-South PTA, the two southern economies reduce trade barriers between each other, with import barriers to and from the North held constant. Figure 3 illustrates this case. Once again, the trade liberalization is sufficient to cause industry to become established in the developing countries, but the mechanism is completely differ- ent from the case of unilateral tariff reductions. In that case, industry started in response to cheaper intermediate inputs. In a South-South PTA, in the initial position tariff reductions do not affect intermediate goods. Instead, the effective market enlargement caused by reducing intra-South barriers drives the industri- alization. Similar to the multilateral case, industry spreads unevenly to develop- ing countries, initially developing in one of the countries and only spreading to the second at lower trade barriers. In the case of a South-South PTA, both the southern economies attract less industry than they would with multilateral liberalization because they do not benefit from better access to northern markets or intermediate goods produced in the North. Compared with unilateral liberalization, the South-South arrange- ment attracts industry later, although as tariffs become very small, the gain is larger. Similar to the unilateral case, with the South-South PTA, real wages are lower than they would be in the case of multilateral liberalization. Different mechanisms trigger industrialization in the case of a South-South PTA and in the case of unilateral liberalization. In the former, local demand trig- gers industrialization, which occurs earlier the higher is this demand. In the lat- ter, forward linkages from imports trigger industrialization, and the timing 232 THE WORLD BANK ECONOMIC REVIEW, VOL. 12, NO. 2 Figure 3. Soutb-South Preferential Trading Arrangement Share of world industry in the southern countries Industry share 0.35 - 0.30 - 0.25 . . ,... 0.20 - S2 0.15 0.10- S1 0.05 2 ~~~2 S 2 0- 1 1.025 1.050 1.075 1.100 1.125 1.150 Tariffs, T Real income per worker in the southern countries and the North Real income per worker 1.0 - ,--. - I S I , 1 1.025 1.050 1.075 1.100 1.125 1.150 Tariffs, T Note: S, and 52 denote southern countries 1 and 2; Ndenotes the North. Source: Authors' calculations. Puga and Venables 233 depends on the strength of these linkages. If the linkages are weak and southern demand is large, then a South-South PTA might attract industry at a higher value of T than would happen with a unilateral liberalization. Concerted Most-Favored-Nation Southern Liberalization In the case of concerted most-favored-nation (MFN) liberalization by both south- ern economies, all southern import tariffs are reduced, so only northern tariffs remain (see World Bank 1994). This case amounts to a South-South PTA plus liberalization of southern imports from the northern economy. Figure 4 illus- trates the results. The evolution of industry is similar to that in a South-South PTA. Industrial- ization starts first in one southern country, then in the other. The process of industrialization starts sooner (at higher levels of T) in one of the countries than it would in the case of unilateral liberalization. This difference occurs because the relationship between South and North is the same under both arrangements, but in the case of concerted MFN liberalization, there are the additional gains from southern liberalization. Compared with a South-South PTA, industrializa- tion occurs earlier in the example in figure 4. However, this result is not general. Liberalization with the North benefits the South through forward linkages, but it also raises the costs of import competition, whose net effect is ambiguous. At low tariff levels, concerted MFN southern liberalization gives a higher level of real income compared with unilateral liberalization, but real income is lower than it would be under a South-South PTA or multilateral liberalization. The difference in real income occurs because of the asymmetry in North-South trad- ing arrangements. With a concerted MFN southern liberalization, southern ex- ports to the North still face a tariff barrier, while the South's imports from the North are untaxed. North-South PTA In the case of a North-South PTA, one of the southern countries forms a PTA with the North. (Recall that the North is assumed to be a single policymaking agent, so the PTA is with both northern economies.) Figure 5 summarizes the results. A North-South PTA spreads a larger share of industry to the liberalizing southern economy and gives this economy higher real wages than any of the other ar- rangements we have considered. Figure 5 compares this case with the multilat- eral case given by the dashed lines. The spread of industry is larger because the southern economy benefits from both improved access to the large northern market and the low cost of northern intermediates. The liberalizing southern economy suffers from more competition from northern firms, but, because south- ern wages are lower, the balance of the improved reciprocal access to markets favors the South. This spread of industry is associated with a large fall in the North's share of industry (and also a fall in real wages in the North). Compared with the other arrangements, the other (not liberalizing) southern economy loses 234 THE WORLD BANK ECONOMIC REVIEW, VOL. 12, NO. 2 Figure 4. Concerted Most-Favored-Nation Southern Liberalization Share of world industry in the southern countries Industry share 0.35 - 0.30 - 0.25 --... 0.20- ....... ,,, l" 0.15- 0.10 SII 0.05- 1 1.025 1.050 1.075 1.100 1.125 1.150 Tariffs, T Real incotne per worker in the southern countries and the North Real income per worker 1.0- N ,_ - . ........ .. 0.9- 0.8- 1 1.025 1.050 1.075 1.100 1.125 1.150 Tariffs, T Note: S, and S, denote southern countries 1 and 2; N denotes the North. Source: Authors' calculations. Puga and Venzables 235 Figure 5. North-Soutth Preferential Trading Arrangement Share of world industry in the southern countries Industry share 0.35 - 0.30 - 0.25- ... -_ ......,, . 0.20- 0.15- 0.10- 0.05- SI SI 1 1.025 1.050 1.075 1.100 1.125 1.150 Tariffs, T Real income per worker in the southern countries and the North Real income per worker 1.0- N 0.9- 0.8- I I I~~~~~~~~~~~~~I S 1 1.025 1.050 1.075 1.100 1.125 1.150 Tariffs, T Note: S, and S, denote southern countries 1 and 2: Ndenotes the North. Soutrce: Authors' calculations. 236 THE WORLD BANK ECONOMIC REVIEW, VOL. 12, NO. 2 Figure 6. Hub-and-Spoke Arrangement Share of world industry in the southern countries Industry share 0.35 - 0.30 - 0.25 . ....... 0.05 S2 '2SSI 2 = 1 1.025 1.050 1.075 1.100 1.125 1.150 Tariffs, T Real income per worker in the southern countries and the North Real income per worker 1.0- 0.9 - 0.7 1 1.025 1.050 1.075 1.100 1.125 1.150 Tariffs, T Note: S, and 52 denote southern countries 1 and 2; N denotes the North. Source: Authors' calculations. Puga and Venables 237 because it does not attract any industry and experiences only a slight increase in real wages through the rise in world industrial production. Hub-and-Spoke Arrangement In the case of a North-South PTA, the North forms a PTA with just one south- ern economy. An interesting alternative is that in which the North forms a bilat- eral PTA with each of the southern economies, but the barriers between the south- ern economies remain unchanged. In this trade policy arrangement, the North (that is, both northern economies jointly) represents the "hub" in a hub-and- spoke arrangement. Figure 6 represents such a case. A hub-and-spoke arrangement brings relatively rapid and strong industrial- ization to the South for the same reasons that applied in the case of a bilateral North-South PTA. Both southern economies are affected, although initially they diverge. The spread of industry to the South is less pronounced than it would be under multilateral liberalization because location in each of the southern econo- mies is penalized by the barriers between the southern economies. The barriers in the South enable the North to maintain a higher real wage in this case than in either of the other two experiments involving northern liberalization. III. POLICY ISSUES Are the forces captured in this framework relevant in practice? Is there evi- dence to support the argument that PTAS cause such changes in the production structure of nations? We are aware of only one study that directly addresses these issues, Hanson's work on Mexico. Hanson (1998) uses data on Mexico and finds support for the hypothesis that agglomeration is associated with in- creasing returns. He also shows that integration with the United States has had strong effects on the location of industry in Mexico (Hanson 1997). Industry has shifted toward states with good access to the U.S. market (demand link- ages). At the same time, employment growth has been higher in regions that have larger agglomerations of industries with buyer/supplier relationships (cost linkages). We are not aware of any empirical work on the importance of demand and cost linkages under specific trading arrangements. Nevertheless, we believe that the experiments in section II shed light on some of the trade policy choices cur- rently facing developing countries. This section discusses the main implications. Unilateral and Concerted Liberalizations In recent years many developing countries (in particular East Asian econo- mies) have undertaken unilateral trade liberalization. However, others (includ- ing some members of the Asia-Pacific Economic Cooperation-APEc-process) have been reluctant to lower their tariffs without receiving reciprocal conces- sions. What are the benefits of unilateral liberalization? Can countries expect to do better by concerted action? 238 THE WORLD BANK ECONOMIC REVIEW, VOL. 12, NO. 2 Unilateral liberalization can attract industry and bring a real gain in income. Although more intense import competition has an adverse effect on profitability in the liberalizing economy, import supply creates beneficial forward linkages to domestic production and promotes industrialization. In the model, these link- ages arise just from the use of imported goods as inputs; in reality these linkages might come through several channels. A recent World Bank (1994: 7) study argues that By opening their economies, countries gain access to more affordable consumer goods and to technologies and intermediate goods that help reduce production costs. Thus, by improving the climate for investment, liber- alization also helps to attract foreign capital. Foreign investment, in turn, can provide the technology and financing required to establish a more efficient production structure. Tybout and Westbrook (1995) find that trade liberalization in Mexico has reduced average costs in most industries. In more export-oriented industries, these cost reductions have been due mainly to the type of forces captured by our model (falling prices of intermediates). In sectors with higher import penetra- tion, these cost effects apparently have been combined with relative productivity improvements. At the same time, increased import penetration has shifted down- ward the demand for domestic products. In our analysis, we find that the balance between import competition and cost linkages, combined with low initial wages, tends to work in favor of the liberal- izing economy, leading to industrialization, as in figure 2. However, the benefits of unilateral liberalization alone may be comparatively small; in our examples, full unilateral liberalization results in lower real income than any of the other experiments considered. What can developing countries do to promote industri- alization through trade policy? Our analysis suggests the strong likelihood of gains from concerted action, but two reservations have to be made. First, the gains from concerted action may not be divided equally between the members. Even in the case of concerted MFN liberalization, there may be some range of tariffs in which one country does worse than it would if it were the only country to liberalize. (We return to divi- sion of the gains in the subsection on southern disparities.) Second, even though all our simulations give greater gains from concerted action than from unilateral liberalization, there are no general theorems. All results are sensitive to specifi- cation of the model and of the experiment. In particular, North-South and South- South PTAs operate in quite different ways. South-South and North-South PTAs Should countries with highly developed industrial systems take part in con- certed liberalization or be excluded from it? South-South PTAS work essentially by enlarging market size, and their success depends on whether the combined market size is large enough to attract industry. Analysis indicates that smaller Puga and Venables 239 southern countries industrialize later and less (in the figures presented, the curves S1 and S2 move down and to the left). The mechanism is a form of trade diver- sion, but-unlike the traditional analysis outlined in the introduction-the di- version may successfully bring about a critical mass of activity that creates a viable, and welfare-improving, industrial base. Evidently, the market size of the group must reach a certain minimum size for this to work. The failure of many South-South PTAs can perhaps be attributed to inadequate scale. As Corden (1993: 457) puts it, It is far better for Argentina to go for the world market-i.e., to liberalize unilaterally and in a nondiscriminatory fashion, as she has been doing- than just go for the Brazilian market. Brazil has the largest economy in the Third World, and yet it is smaller than Canada's (as measured by the dollar value of GDP [gross domestic product]). And this applies even more to Brazil. North-South PTAs work quite differently on the basis of improved access to the large northern market and improved supply of intermediate goods, offset by increased import competition in domestic markets. In the cases we have exam- ined, North-South arrangements are better than South-South agreements from the point of view of the participating southern economies. The reasons for the success of these North-South agreements merit some thought. In many new trade models the argument is made that liberalization between economies of different sizes will draw industry into the country with the large market (the center) and away from smaller (peripheral) countries. However, the strength of these forces is greatest at intermediate levels of trade barriers, and at very low trade barriers, factor price differences can overturn these effects. How does this relate to our findings that liberalization will move industry out of the large economy to the small? Centripetal forces are present in the model we have developed. Indeed, they are amplified by forward and backward linkages. But these forces are pre- cisely those that make for the North-South divide in the initial equilibrium; they create the initial wage differentials. Given this initial position and these wage differences, further liberalization then moves industry from the large and devel- oped region to the less developed. The circumstances that are most conducive to the South benefiting from a North-South agreement of this type are, therefore, low remaining barriers to the northern market (secured, for example, by prox- imity, as in the North American Free Trade Agreement [NAFTA] or the European Union's southern regions and prospective eastern regions), combined with low unit labor costs. At the same time, the North may lose in this framework. The breakdown of the industrial agglomeration in the northern economy may cause a decline in the North's real income. Losses in the North are greatest in the case of South-South liberaliza- tion. In the arrangements where the North reduces barriers to southern imports, multilateral liberalization causes larger losses for the North than hub-and-spoke arrangements. Such losses in the North may explain why the European Union chose a bilateral rather than a multilateral approach to trade liberalization with its 240 THE WORLD BANK ECONOMIC REVIEW, VOL. 12, NO. 2 neighbors. The bilateral association agreements that the European Union has es- tablished both in Central and Eastern Europe and in the Mediterranean have in effect turned it into the hub of a large web of trading arrangements. The situation might not be better for the North if it did not reduce its barriers to southern imports. In fact, our analysis suggests three reasons why not liberal- izing may be a worse option. First, the northern losses in this context are not general. All our experiments start from an equilibrium in which the South has no industry, so the North has no manufacturing imports and there are large differences in unit labor costs. With higher initial levels of development in the South and smaller initial differences in unit labor costs, real wages in the North will tend to rise instead (see Krugman and Venables 1995). (Also, we have ex- cluded any gains from trade through comparative advantage that in practice would likely bring further benefits to northern consumers.) Second, even if the North were to lose from opening its market to southern imports, it would lose more from remaining closed while southern economies liberalize among themselves. A comparison of South-South and multilateral lib- eralization shows that in either case industry spreads to southern countries, but under South-South liberalization, northern firms and consumers have to pay higher prices on the growing number of goods produced in the South, so real wages are lower in the North. Third, falling real wage differences between the North and the South may reduce migration pressures. For example, proponents of NAFTA in the United States argue that it reduces illegal migration from Mexico. In fact, Hanson and Spilimbergo (1998) show that illegal migration from Mexico to the United States is very responsive to changes in relative wages. Nondiscriminatory Liberalization and Southern Trading Blocs The APEC process has raised hopes that integration in the Asia-Pacific region may develop in a less inward-looking way than in other geographical areas and may even catalyze deeper global trade liberalization. Calls for forming a re- gional trading bloc in Asia have received little support. The amount of trade covered by the Association of South East Asian Nations (ASEAN) remains small. The East Asian Economic Group (EAEG) has so far lacked the necessary backing to take off. Instead, inward-looking regional integration is giving way to APEC's vision of open regionalism. In the report presented at APEC's 1994 annual summit in Bogor, the APEC Eminent Persons Group explained APEC's vision of open regionalism as follows (APEC 1994). First, APEC members should liberalize intra-APEC trade flows on a nondiscriminatory basis. Second, APEC should, as a group, treat nonmembers as it does members, provided that nonmembers make reciprocal offers. Third, any individual APEC member should have the choice to waive unilaterally the reci- procity requirement and extend its APEC liberalization to all nonmembers. APEC'S members have been divided over the last point. Although East Asian countries have favored openness toward nonmembers, the U.S. president, Bill Puga and Venables 241 Clinton, made clear before APEC's Bogor summit that any trade concessions would be reciprocal and that there would be no free riders. One year later at the Osaka summit, Australia's trade minister, Bob McMullan, stressed that Australia would also give nothing for nothing. Outside Asia, other countries have also seen the need to open their markets reciprocally. The main argument was highlighted by The Economist ("Getting Together: South America. Mercosur Gets Bigger and Stronger," The Economist, 29 June 1996) after Chile signed its free trade agreement with Mercosur (a re- gional trade arrangement among Argentina, Brazil, Paraguay, and Uruguay): "Despite continued protectionist pressure from their weaker industries, Mercosur's leaders all know that, to attract investment they need to compete in the wider world, their firms want a bigger home market." However, Mercosur's agenda does not include unilateral liberalization. Instead, its member countries are advancing toward a regional free trade agreement that will liberalize trade flows between members, but not imports from outsiders. In the tradeoff between concerted MFN liberalization and a South-South PTA, nondiscriminatory liberalization brings beneficial cost linkages, but also more intense competition from outsiders. Comparison of figures 3 and 4 shows that the former effect dominates in the early stages of industrialization, and the latter dominates in determining real income once industry is established. Concerted MFN liberalization brings earlier industrialization, but at very low tariff levels a South-South PTA leads to more southern industry. These results are quite sensi- tive to parameter values. In particular, concerted MFN works better for small southern economies. With a small home market, most of their sales take place abroad, and protective tariffs provide little help. At the same time, extending liberalization unilaterally to nonmembers lowers the cost of intermediates and helps industrialization to take off. Thus it is not surprising that in Asia smaller countries push for unilateral liberalization, while larger ones insist on reciprocal concessions. The fact that more small than large countries have liberalized uni- laterally can also be explained by smaller countries having less bargaining power to extract reciprocal concessions. What is striking is that smaller Asian countries not only tend to have more open regimes but also have generally expressed their preference for a more open approach to trade liberalization even if larger coun- tries in APEC were not to do the same. Southern Disparities All the cases in which the two southern economies follow symmetric policies generate an asymmetric outcome for some interval of tariffs, with only one of the southern countries having industry. The theoretical model assumes that there are two identical southern countries, providing no basis for deciding which coun- try has industry. In practice, differences between the two countries will decide the issue (possibly quite small differences). The mix of factors obviously includes institutional, political, and geographical considerations. Here we highlight just a few factors. 242 THE WORLD BANK ECONOMIC REVIEW, VOL. 12, NO. 2 Geographical proximity to the existing industrial center benefits a country insofar as closeness is associated with lower natural trade barriers. (This pro- vides an interesting way of thinking about the spread of industry from Japan through the newly industrializing economies.) Low unit labor costs and a larger home market also benefit a country because, other things being equal, they in- crease the attraction of a country as a base for industry. The policy regime of the government may dominate the differences between the two southern countries. This situation creates several incentives for the gov- ernment to take policy action to obtain a "first mover advantage" and attract industry before it becomes established elsewhere. First, it creates an incentive for developing countries to establish trade links with industrial countries. North-South PTAs may effectively attract industry to the South. More generally, links with industrial countries may give a particular developing country the margin it needs to ensure that it becomes the first to industrialize in a South-South trading arrangement. Second, it creates an incentive for a country to be a founding member of a PTA. Late entrants to a PTA will not be the first to attract industry. Third, a reduction of barriers within a PTA can have a domino effect: it creates incentives for more countries to join the agreement. In figure 5, if SI forms a PTA with the North, tariffs on intra-PTA trade will gradually decrease. Once tariffs fall below T= 11, S2's real income will be higher if it also joined the agreement, because this will allow it to attract industry. Baldwin (1993) models a different, but related, mechanism driving the domino pattern of PTA membership. In his model, a fixed number of firms in each coun- try has a variable equilibrium scale of operation. Integration within the PTA in- creases the gains from membership for firms in excluded nations, leading them to exercise stronger pressure in favor of membership. Countries' idiosyncratic opposition to membership drives the sequence of entry. In our model, integra- tion within the PTA makes it profitable to locate industry in the next country to join the PTA. Entry eliminates profits, fully passing on to consumers the benefits of membership. Sequential entry arises with homogeneous countries: once a coun- try joins a PTA, it pays the next country to wait until barriers within the PTA are reduced further. In addition to creating incentives for countries to attract industry, the possi- bility that industry will agglomerate in a subset of member countries may also create real tensions within the PTA. In the history of southern PTAs, many schemes have failed, often because of internal disputes over the location of industry and the design of compensation schemes for perceived losers in the arrangement. A typical example, the Treaty of Brazzaville was intended to create a customs union and a common currency area with the former French Central African countries. Foroutan (1993) shows that the distortionary nature of compensations undid any benefits from this PTA. This article shows that differences between countries may be only transitional. In the figures, the differences disappear with sufficient reductions in tariffs. Puiga and Venables 243 However, there is no guarantee that the final liberalization will necessarily go far enough to iron out differences and secure the spread of industry to all par- ticipating southern economies, particularly if substantial differences underlie these economies. IV. CONCLUSIONS In this article, we have outlined a new approach for analyzing the role of trade in promoting industrial development. Interactions among imperfect com- petition, trade costs, and the input-output structure create incentives for firms to locate close to supplier and customer firms. Clustering of firms occurs so that only a few countries industrialize even if all countries have identical un- derlying structures. The industrializing countries have high wages, but the positive pecuniary externalities created by interfirm linkages compensate for the higher wage costs. Trade liberalization changes the attractiveness of coun- tries as a base for manufacturing production and can trigger-or postpone- industrial development. The process we describe abstracts from many important aspects of reality. The model disregards capital accumulation (physical or human) and interre- gional or international differences in technology and endowments. It has only a single policy instrument-a tariff that is unable to discriminate between differ- ent types of industrial imports. It does not explicitly model the political process that leads to a particular choice of policy or the policy games between govern- ments. Even within this framework, the model is simple. For example, firms are modeled as single-plant operations, so multinationality and foreign direct in- vestment are not considered. Also, firms are footloose and atomistic, which is helpful for focusing on long-run outcomes but abstracts from the costs of relo- cation and from strategic interaction. All these points create possible directions for future research, but, even at this level of abstraction, the model captures a number of important features of the world economy and provides new insights on the effects of trading arrangements on industrial development. The analysis offers an explanation as to why firms are reluctant to move to economies that have lower wages and labor costs and shows how trade liberal- ization can change the incentives for the establishment of firms in developing countries. The analysis provides a mechanism through which import liberaliza- tion can have a powerful effect in promoting industrialization. It suggests that import liberalization may create or amplify differences between liberalizing coun- tries, including possible political tensions. Although these features are consistent with the world economy, they of course fall far short of providing convincing empirical support for the approach. From this analysis, we derive several conclusions about the effects of trade liberalization. First, unilateral liberalization of imports of manufactures can pro- mote the development of local manufacturing industry. The development occurs through forward linkages from imported intermediates that may be part of a 244 THE WORLD BANK ECONOMIC REVIEW, VOL. 12, NO. 2 wider package of linkages coming from these imports. Second, the gains from liberalization through PTA membership are likely to exceed those from unilateral action. South-South PTAs will be sensitive to the market size of member states, and North-South PTAs seem to offer better prospects for participating southern economies, if not for the North and excluded countries. Third, the effects of particular schemes (on, for example, the division of benefits between southern economies) depend on the characteristics of the countries and cross-country dif- ferences in these characteristics. We have not yet conducted a systematic investi- gation of the sensitivity of our results to such differences. APPENDIX. DETAILS OF THE ANALYTICAL FRAMEWORK We consider a world with four regions, two northern (Nl and N2) and two southern (S1 and S2). Each region is endowed with L workers and K units of arable land and can produce agricultural and industrial output. Both primary factors are immobile between regions. Arable land is used only by the agricul- tural sector, while labor is used both by agriculture and by industry and is per- fectly mobile between sectors. Agriculture Agriculture is perfectly competitive and produces a homogenous output un- der constant returns to scale. We assume that the agricultural output is costlessly tradable and choose it as the numeraire. In each region, the agricultural produc- tion function is Cobb-Douglas in land and labor, with labor's share 0. If LA de- notes agricultural employment, agricultural output is (LA)0 K,1 -e), and the local wage is (A-1) wi = 0 (LA )(0-1) K(1-0). Industry The industrial sector has imperfectly competitive firms producing differenti- ated goods under increasing returns to scale. Production of a quantity xi(k) of any variety k in any country i requires the same fixed, ca, and variable, fixi(k), quantities of the production input. That production input is a Cobb-Douglas composite of labor and a constant elasticity of substitution (CES) aggregate of the differentiated industrial goods. The cost function of a firm producing variety k in country i is (A-2) C(k) = q- w("')[o + 3xi (k)] where qi is the price index of the aggregate, defined by (A-3) qi I fhp}(b)(Tj j)]L-`1db l =1 beN j Puga and Venables 245 The price index in each country depends on the local prices of individual varieties, which in turn are a function of the free on board (FOB) prices, real trade costs, and tariffs. The elasticity of substitution between varieties, a (> 1), is assumed to be the same in all countries. Ni is the set of varieties produced in location j, and p1(h) is the FOB price of variety h shipped from country j to country i. Real trade costs for the industrial goods take Samuelson's "iceberg" form: T units have to be shipped so that one unit arrives in another region. Industrial goods exported from j to i are also subject to an ad valorem tariff, TJ,- 1. Preferences Turning to the demand side, consumers have Cobb-Douglas preferences over the agricultural good and a CES aggregate of industrial goods. All industrial va- rieties produced enter consumers' utility function with the same constant elastic- ity of substitution with which they enter firms' technology. The indirect utility function of a worker in region i is then given by (A-4) V =q- F"-" wi. Landowners have the same preferences as workers but are assumed to be tied to their land. General Equilibrium Expenditure on manufactures in each region can be derived from equations A-2, A-3, and A-4 as (A-5) ei =y wiLi + (1 - 0)(LA)\ K J17t + f n(h)dh + Ri -+ F- |C(h)dh. heN, heN, The first term on the right-hand side is the value of consumer expenditure (in- cluding tariff revenue, denoted by Ri), and the second is the value of intermedi- ate demand, because consumers spend a fraction y of their income and firms a fraction 1i of their costs on manufactures. The division of consumers' and producers' expenditure on each industry be- tween individual varieties of industrial goods can be found by differentiating the price index with respect to the price of the variety. Total demand for a single variety produced in i, xi, is (A-6) xi(k) = T(1-0) [p (k)Ti ,lq I q ) e,. ji= Because the producer of an individual good faces an elasticity of demand 6, firms mark up prices over the marginal cost by the factor c / (6 - 1): 246 THE WORLD BANK ECONOMIC REVIEW, VOL. 12, NO. 2 (A-7) Pi= 6 i qwil The value of tariff revenue is 4 (A-8) Ri = , (T1,i - 1)n,p,-x j=l The profits of an individual manufacturing firm are, from expressions A-2 and A-7, P. (A-9) 7ci =(xi - x) where (A-10) x is the unique level of output giving firms zero profits. As usual in this type of model, equilibrium firm scale is a constant, depending only on demand and cost parameters. Turning to the labor market, we can write the labor-market-clearing condi- tion as (A-11) Li = (I - t)ni- wi where ni - #Ni denotes the mass (number) of firms in region i. The first term on the right-hand side of equation A-il is labor demand in manufacturing, ob- tained by applying Shephard's lemma to equation A-2, and the second term is labor demand in agriculture. This completes the description of short-run equilibrium. At any given mo- ment, we think of the economy as having a predetermined number of firms in each region. To this corresponds a short-run equilibrium defined as a set of wages and price indexes solving eight equations. The first four equations are the price indexes of manufactures in each of the four regions, obtained by substituting equation A-7 into equation A-3. The other four equations come from substituting A-1, A-2, and A-5 through A-O0 into A-11, which gives the labor-market-clearing condition in each region. We can then express profits at the short-run equilibrium in terms of the number of firms by substituting equations A-1 and A-2, A-5 through A-8, and A-10 and the short-run equi- librium values of wages and price indexes into equation A-9. Pupa and VL'PJenables 24- A long-run equilibrium obtains when the number of firms in each counrrv is such that there are zero profits in each country where there are a positive num- ber of firms and negative profits (for potential, if not for actual, firms) wherever the number of firms is zero: (A-12) 7cin, = 0, ni < 0, ni 2 0. We assume a myopic entry and exit process, according to which firms enter and exit in response to profit opportunities. This is described by differential equations, (A-13) hi = ozj, ni 2: Q where 8 (> 0) is the speed of adjustment. The system of equations A-12 may support multiple equilibriums, and under the dynamics of equation A-13, some are stable and others unstable. Complete analysis of the structure of equilibriums and of the dependence of bifurcation points of the system is undertaken for the case of symmetric changes in trade barriers in Puga (1998) and Fujita, Krugman, and Venables (1997) and for asym- metric changes in trade barriers in Puga and Venables (1997). Our approach in this article is to select the initial equilibrium in which two of the four economies have industry. For each experiment we take a small reduc- tion in tariffs and then let the model adjust according to the dynamics given by equation A- 13. Repeating this for successively lower tariffs traces out a path of stable equilibriums and gives the relationship between tariffs and the endog- enous variables illustrated in figures 1-6. Computation was undertaken using GAUSS code written bv the authors. and this is available on request. Parameters and Experinments Values of parameters are y = 0.5, 0 = 0.8, 1. = 0.55, and a = 4. At the starting point in all our experiments, there are real trade costs of T = 1. 1 between all four regions, an ad valorem tariff of 15 percent (T = 1.15) for all North-South and South-South trade, and free trade between the two northern economies (T = 1). These values are such that at the initial level of tariff barriers there is a stable equilibrium in which all industry is split between the two northern economies (qualitatively the same initial equilibrium can be supported by a range of values for either type of barrier). The experiments look at the evolution of this equilibrium as some (or, in the case of multilateral liberalization, all) of these tariffs are brought down to zero (T = 1), with those not affected by the liberalization held at T. Real barriers are kept unchanged (reductions in real barriers instead of reductions in tariffs have effects very similar to those presented, although no revenue is forgone). 248 THE WORLD BANK ECONOMIC REVIEW, VOL. 12, NO. 2 REFERENCES The word "processed" describes informally reproduced works that may not be com- monly available through library systems. APEC (Asia-Pacific Economic Cooperation), Eminent Persons' Group. 1994. Achieving the APEC Vision: Free and Open Trade in the Asia-Pacific. Singapore: Asia-Pacific Economic Cooperation Secretariat. Baldwin, Richard E. 1993. "A Domino Theory of Regionalism." Working Paper 4465. National Bureau of Economic Research, Cambridge, Mass. Processed. Baldwin, Richard E., and Anthony J. Venables. 1995. "Regional Economic Integration." In Gene M. Grossman and Kenneth Rogoff, eds., Handbook of International Eco- nomics. Vol. 3. Amsterdam: North-Holland. Corden, W. Max. 1993. "Round Table Discussion." In Jaime de Melo and Arvind Panagariya, eds., New Dimensions in Regional Integration. Cambridge, U.K.: Cam- bridge University Press. de Melo, Jaime, Arvind Panagariya, and Dani Rodrik. 1993. "The New Regionalism: A Country Perspective." In Jaime de Melo and Arvind Panagariya, eds., New Dimen- sions in Regional Integration. Cambridge, U.K.: Cambridge University Press. Dixit, Avinash K., and Joseph E. Stiglitz. 1977. "Monopolistic Competition and Opti- mum Product Diversity." American Economic Review 67(3):297-308. Foroutan, Faezeh. 1993. "Regional Integration in Sub-Saharan Africa: Past Experience and Future Prospects." In Jaime de Melo and Arvind Panagariya, eds., New Dimen- sions in Regional Integration. Cambridge, U.K.: Cambridge University Press. Fujita, Masahisa, Paul R. Krugman, and Anthony J. Venables. 1997. "The Spatial Economy: Cities, Regions, and International Trade." London School of Economics. Processed. Hanson, Gordon H. 1997. "Increasing Returns, Trade, and the Regional Structure of Wages." Economic Journal 107(440):113-33. . 1998. "Regional Adjustment to Trade Liberalisation." Regional Science and Urban Economics. In press. Hanson, Gordon H., and Antonio Spilimbergo. 1998. "Illegal Immigration, Border En- forcement, and Relative Wages: Evidence from Apprehensions at the U.S.-Mexico Border." American Economic Review. In press. Helpman, Elhanan, and Paul R. Krugman. 1985. Market Structure and Foreign Trade. Cambridge, Mass.: MIT Press. Krugman, Paul R. 1980. "Scale Economies, Product Differentiation, and the Pattern of Trade." American Economic Review 70(5):950-59. Krugman, Paul R., and Anthony J. Venables. 1995. "Globalization and the Inequality of Nations." Quarterly Jourinal of Economics 110(4):857-80. Pande, Rohini. 1997. "Transforming Economies: Agricultural Mechanisation and the Process of Industrialisation." London School of Economics. Processed. Puga, Diego. 1998. "The Rise and Fall of Regional Inequalities." European Economic Review. In press. Puga, Diego, and Anthony J. Venables. 1996. "The Spread of Industry: Spatial Agglom- eration in Economic Development." Journal of the Japanese and International Econo- mies 10(4):440-64. Puga and Venables 249 1997. "Preferential Trading Arrangements and Industrial Location." Journal of International Economics 43(3-4):347-68. Tybout, James R., and M. Daniel Westbrook. 1995. "Trade Liberalization and the Di- mensions of Efficiency Change in Mexican Manufacturing Industries." Journal of International Economics 39(1-2):53-78. Venables, Anthony J. 1996. "Trade Policy, Cumulative Causation, and Industrial De- velopment." Journal of Development Economics 49(1):179-98. World Bank. 1994. East Asia's Trade and Investment: Regional and Global Gains from Liberalization. Washington, D.C. THE WORLD BANK ECONOMIC REVIEW, VOL. 12, NO. 2 251-70 Regional Integration and Economic Growth Athanasios Vamvakidis The rapid economic growth of developing countries that opened their markets to free international trade during the past two decades has stimulated a large empirical and theoretical literature on the impact of trade on growth. This literature concludes that free trade and growth were positively correlated during the 1970s and 1980s. How- ever, most studies focus on nondiscriminatory openness. Does regional integration matter for economic growth? Do regional trade agreements have any impact on growth? This article presents empirical evidence that counztries with open, large, and more de- veloped neighboring economies grow faster than those with closed, smaller, and less de- veloped neighboring economies. The results are robust to different specifications of the empirical model and different definitions of openness, suggesting that small economies should grow faster when they form regional trade agreements with large and more devel- oped economies. However, testing for the impact of five regional trade agreements during the 1970s and 1980s finds that none led to faster growth. The main reason seems to be that most of these agreements were among small, closed, and developing economies. The rapid economic growth of developing countries that opened their markets to free international trade during the past two decades has stimulated a large empirical and theoretical literature on the impact of trade on growth. This lit- erature concludes that free trade and growth were positively correlated during the 1970s and 1980s (Dollar 1992, Edwards 1992, Levine and Renelt 1992, Ben-David 1993, Barro and Sala-I-Martin 1995, Sachs and Warner 1995, and Vamvakidis 1996, 1997). This result is robust to different proxies for openness and the specification of the empirical model. However, most studies focus on nondiscriminatory openness. Does regional integration matter for economic growth? Do regional trade agreements (RTAS) have any impact on growth? These questions remain unanswered. Their importance has become apparent recently, as many developing countries try to determine an appropriate liberalization strat- egy. Should they open their markets to countries from the same region before they open to the rest of the world? Or should they move directly toward nondis- criminatory liberalization? Athanasios Vamvakidis is with the Research Department at the International Monetary Fund. The author is grateful to Donald Davis, Ashish Garg, Robert E. Lucas, Maurice Schiff, David Weinstein, Jeffrey Williamson, and Alan Winters for their helpful comments. He would also like to thank the participants in the seminar series of the International Trade Division at the World Bank and three anonymous referees. This article was produced as part of the World Bank Development Economics Research Group's research program on regionalism and development. It is based on the author's research on regional integration for the World Bank and on his doctoral thesis at Harvard University. (C 1998 The International Bank for Reconstruction and Development/THE WORLD BANK 251 252 THE WORLD BANK ECONOMIC REVIEW, VOL. 12, NO. 2 The literature approximates openness by using trade figures or indexes of domestic trade policy. This methodology considers domestic trade policies but ignores trade policies in the rest of the world. However, openness depends on both domestic and foreign trade policies. As shown by Vamvakidis (1996, 1997), the benefits from domestic openness are a positive function of openness in the rest of the world. This result has important implications for regional integration and RTAs. This article considers three questions. First, how do the trade policies of coun- tries from the same region affect economic growth in the home country? Second, does a country benefit from positive spillover effects if it is in the same region as large, open, and developed economies? Third, have member countries of RTAs benefited from such spillovers, growing faster? I investigate the first and second questions without considering RTAs. If a country does benefit from being near open economies and if the size and development of these economies matter for growth, I examine whether RTAs, by increasing the openness of countries from the same region, increase these benefits. Even economies that are characterized as open have trade barriers; therefore, an RTA increases their openness toward member countries. All data in this article are from the Summers-Heston (1991) and Barro-Lee (1994) data sets. I include all countries with available data in each regression. As a result, the number of countries changes depending on the independent vari- ables included in each regression. However, the results remain the same if the regressions include only countries with data for all independent variables. Section I reviews the literature on international trade and growth. Section II discusses data and methodology issues. Section III estimates the growth benefits for a country whose neighbors have large and open economies. Section IV esti- mates spillover effects from the level of development and the growth of neigh- boring economies. Section V tests the impact of five RTAS on growth. Section VI concludes with some policy implications for developing countries. I. PREVIOUS LITERATURE ON TRADE AND GROWTH This section provides a brief review of the theoretical and empirical literature on trade and growth. Theoretical Studies The recent literature on trade and growth tries to explain why free trade fos- ters growth. This contrasts with earlier literature that used infant industry argu- ments to support protection. An important part of the recent theoretical litera- ture started with Lucas (1988), who considers a world in which human capital is the engine of growth and analyzes the effects of learning by doing. According to his analysis, the initial conditions determine the comparative advantage of each country and, thus, which products each country will produce under free trade. The model predicts that each country's comparative advantage increases through Vamvakidis 2S3 learning by doing. This implies that some countries are locked in sectors with relatively little learning by doing and diverge from the rest of the world. This model provides a plausible explanation for the absence of convergence in cross- country data. However, it predicts that only countries with initial comparative advantage in sectors with significant learning by doing will benefit from free trade. More recent theoretical literature focuses on the channels through which free trade leads to faster growth (Grossman and Helpman 1989, 1990, 1991; Rivera- Batiz and Romer 1991a, 1991b; Romer 1990; and Krugman 1990, ch. 11). Ac- cording to the new literature, trade increases innovation through economies of scale, technological spillovers, and elimination of the replication of research and development (R&D) in different countries. In the Grossman and Helpman stud- ies, innovation of new products is a positive function of past innovations, which represent the stock of knowledge. International trade provides access to a large international market, to advanced technology, and, therefore, to a larger stock of knowledge, leading to more innovations and faster growth. This implies that a country benefits from free trade with large economies and an advanced stock of knowledge, assuming that technological spillovers are absorbed to the same degree across countries. Indeed, Coe and Helpman (1997) provide empirical evidence for these arguments, showing that a country's total factor productivity depends not only on its own R&D capital stock but also on the R&D capital stocks of its trade partners. Also, Coe, Helpman, and Hoffmaister (1997) find that developing countries with limited R&D stock can boost productivity by trading with a more developed country that has a large stock of knowledge from its cumulative R&D activities. Finally, Vamvakidis (1996) presents evidence that the size of the domestic market is important for growth only for closed econo- mies, whereas a large international market fosters economic growth for open economies, as predicted by the theoretical models just mentioned. The new literature suggests that a country that is more open to free trade will have greater technological spillovers and, therefore, faster growth than a coun- try that is less open. However, the literature does not address the issue of re- gional integration. Should a country form or join an RTA or reduce trade barriers for all countries? Free trade is beneficial for growth, but what kind of free trade- regional or nondiscriminatory? Given that theory has not answered these im- portant questions, empirical evidence on the impact of regional integration on growth may provide stylized facts and show the direction that future theoretical work should follow. Empirical Studies The empirical literature has established that free trade fostered economic growth in the 1970s and 1980s. However, this was not true in earlier decades. Vamvakidis (1997) estimates the impact of international trade on growth from 1870 to 1990 and finds that free trade and growth are positively correlated only in the 1970s and 1980s. There is no correlation in any of the earlier decades in 254 THE WORLD BANK ECONOMIC REVIEW, VOL. 12, NO. 2 the sample, except for a negative correlation in the 1930s. Because the data used here are from the 1970s and 1980s, the other studies discussed focus on this period. Dollar (1992) estimates a cross-country index of distortion in the real ex- change rate. He uses the index to determine if a country is outward or inward oriented. Based on estimations for 95 developing countries, he concludes that more outward-oriented countries grow faster. The rigorous measure of exchange rate distortion and the focus on developing countries distinguish this study. Edwards (1992) estimates the impact of international trade on growth using nine indexes of openness proposed in the literature. All these openness measures show a positive correlation between openness and growth. Levine and Renelt (1992) test the robustness of the determinants of growth suggested by previous literature that performs sensitivity analysis. They find that free international trade indirectly affects growth through investment. Countries that have low trade barriers invest more and therefore grow faster. This result is robust to different specifications and to different indexes of openness. Ben-David (1993) shows that open economies converge and that the trade agreements of the European Union have resulted in the convergence of its mem- bers. It is a known puzzle in the empirical literature on growth that economies do not experience the convergence that the neoclassical growth model predicts. In general, growth of the world economy agrees more with divergence than with convergence. Ben-David's work shows that the only economies that converge are those that are integrated in the world economy through trade. Baldwin and Seghezza (1996) document the positive growth effects of the European Union for the medium term. Henrekson, Torstensson, and Torstensson (1997) show that a dummy for participation in the European Union has a posi- tive coefficient in cross-country growth regressions, but its significance is not always robust and depends on the specification of the empirical model. The results in this article confirm these findings. Sachs and Warner (1995) also confirm these results. They construct a dummy variable of openness based on five protection dimensions, including tariff and nontariff barriers, black market premia, and the role of the state in the economy. Using this index, they find that open economies grow on average 1.5 percent faster than closed economies and that unconditional convergence is true only for open economies. They also present evidence that economies liberalize only after a serious economic crisis. Bruno and Easterly (1996) present similar evidence. Alesina and Drazen (1991) provide the theoretical foundation of this stylized fact. Their model shows that an economic crisis can stop the war of attrition among economic groups that delay liberalization in an effort to avoid its costs. Barro and Sala-I-Martin (1995) find that protection has a negative impact on growth. Using tariffs on capital goods and intermediate inputs as a measure of protection, their results indicate that countries with high tariffs grow slower than those with low tariffs. Protection has no significant impact, however, when they use an index of nontariff barriers on capital goods and intermediate inputs. Vamvakidis 25S Most of the literature on trade and growth does not examine the impact of regional integration on growth. Ben-David (1993) is an exception, showing that trade agreements in Europe have caused convergence. Vamvakidis (1996) shows that the trade policies of countries ih the same region matter for growth, a result confirmed here, and compares the importance of a large international market for the growth of open economies with the importance of a large domestic mar- ket for the growth of closed economies. The present article focuses on whether the openness, size, and level of gross domestic product (GDP) per capita affect growth for countries in the same region. It also analyzes whether RTAs reinforce this impact. II. DATA AND METHODOLOGY Most RTAS in the 1970s and 1980s were among small developing economies and do not provide enough variation to test the theoretical predictions in section I. In this and the following section, I address this issue by looking at the impact of trade policy, market size, and level of development of countries in the same region on the growth of the home country, regardless of any existing RTA. Does it matter if a country is in the same region as open economies? Do the market size and the income level of neighboring economies foster economic growth in the home country? Most of the theoretical models reviewed in section I suggest that such spillovers should exist. For example, Grossman and Helpman (1991) suggest that a country with more advanced and bigger trading partners has greater spillover effects from them. The trading partners of a country are not always within the same region, but often countries in the same region trade a lot with one another if their economies are open. The model estimated here is similar to the models in Barro and Sala-I-Martin (1995), Sachs and Warner (1995), and Levine and Renelt (1992). The depen- dent variable is growth in average annual GDP per capita for 1970-90. The stan- dard independent variables are the log of GDP per capita in 1970, the average share of investment in physical capital over GDP in 1970-90, the secondary school enrollment in 1970, the average population growth in 1970-90, the average growth rate of the terms of trade in 1970-90, and the infant mortality rate in 1970. In this section, an economy is characterized as open if it has low trade barriers toward all countries. Section IV considers the case of RTAS in which a member country may not be open to nonmember countries. The analysis uses an openness dummy constructed by Sachs and Warner (1995), who define an economy as open if all of the following conditions are true: the average tariff rate is less than 40 percent, average nontariff barriers cover less than 40 percent of total trade, the black market premium is less than 20 percent of the official exchange rate, there is no communism, and there is no state mo- nopoly on major exports. I define a country as open based on the value of this dummy in 1970, the first year of the period considered, to avoid potential prob- lems of reversed causality. The results do not change when the dummy takes the Table 1. The Impact of the Size of Open Neighboring Economies on Growth, 1970-90 Regression Variable 1 2 3 4 5 6 7 8 Constant 0.030 0.155 0.149 0.033 0.153 0.146 0.153 0.164 (1.551) (5.809) (5.706) (1.664) (5.802) (5.670) (5.324) (6.320) Log of GDP per capita, 1970 -0.002 -0.019 -0.019 -0.003 -0.019 -0.019 -0.019 -0.020 (-1.023) (-6.028) (-6.033) (-1.061) (-6.063) (-6.128) (-5.535) (-5.317) Open neighbors' market, 1970' 0.155 0.106 0.070 0.132 0.130 0.102 0.07 0.07 (3.508) (2.238) (1.802) (2.287) (3.065) (2.537) (2.650) (2.574) Closed neighbors' market, 1970a -0.025 0.025 0.036 0.034 0.0002 (-0.670) (1.067) (1.509) (1.437) (0.005) Average annual share of investment in 0.113 0.102 0.113 0.099 0.080 0.118 physical capital over GDP, 1970-90 (3.659) (3.025) (3.677) (3.019) (2.287) (3.358) Secondary school enrollment, 1970 0.007 0.004 0.006 0.0013 -0.004 0.004 (0.528) (0.259) (0.426) (0.087) (-0.272) (0.232) Average annual growth in the terms 0.164 0.161 0.165 0.162 0.179 0.184 of trade, 1970-90 (3.051) (2.962) (3.084) (3.017) (3.249) (2.799) Infant mortality rate, 1970 -0.167 -0.172 -0.172 -0.180 -0.154 -0.164 (-3.440) (-3.350) (-3.625) (-3.664) (-2.742) (-3.024) Average annual population growth, 1970-90 -0.226 -0.078 -0.266 -0.120 -0.020 -0.222 (-0.998) (-0.308) (-1.092) (-0.456) (-0.076) (-0.762) Openness dummy, 1970 0.011 0.012 (1.842) (2.190) Share of years open, 1970-90 0.020 (3.407) Average annual trade share, 1970-90k -0.021 (-0.791) Africa dummy -0.008 -0.005 (-1.872) (-1.024) East Asia dummy -0.001 -0.0004 (-0.180) (-0.049) Number of observations 120 90 90 120 90 90 90 89 Adjusted R2 0.10 0.50 0.52 0.095 0.50 0.52 0.56 0.51 Note: The dependent variable is average annual growth in GDP per capita for 1970-90. Neighboring economies are listed in table A-1. Openness is based on Sachs and Warner's (1995) openness dummy. t-statistics based on heteroscedastic-consistent standard errors are in parentheses. a. The log of the sum of the GDP of all open or closed neighboring economies. b. (Exports + imports)/GDPl. Source: Author's calculations. Vamvakidis 257 value of 1 if a country is open for most years in the period 1970-90. In some regressions I use the share of the years a country has been open during 1970-90, and the results are also robust. The proxy for the openness of countries from the same region is the openness of neighboring countries. If the openness of neighboring countries matters for growth, then regional integration also matters if it leads to freer trade policies. For every country in the analysis, I consider the openness of countries with com- mon borders. The log of the sum of the GDP of all open neighboring economies in 1970 measures the size of the open neighbors' market. I construct a similar variable for neighboring economies with high protection, the closed neighbors' market. The estimates do not change if the GDP of the home country is also added to the open neighbors' market, as I do in some regressions, or if popula- tion is used as the definition of country size. Actually, GDP and population vary more within regions than the openness dummy. To test the robustness of the results, I use a continuous measure of openness instead of an openness dummy and estimate for each country the GDP-weighted average trade share [(exports + imports)/GDP] of the neighboring countries for 1970-90. This variable measures how much the neighbors of each country trade; therefore it shows whether proximity to countries that trade a lot has an impact on growth. The advantage of this variable is that it varies much more within regions than the openness dummy. For reasons of simplicity and comparison, I use the list of neighboring econo- mies that Chua (1993) and Barro and Sala-I-Martin (1995) use in their estimates of the effects of neighboring economies. Table A-1 provides the list of neighbor- ing economies for each country in their sample. III. THE IMPACT OF THE SIZE OF OPEN NEIGHBORING ECONOMIES ON GROWTH Table 1 shows that the openness of neighboring economies has a positive and statistically significant coefficient. Having neighbors with large open economies fosters growth. The coefficient in regressions 7 and 8 means that an increase in the log of GDP of open neighboring economies by one standard deviation (5.63) results in 0.39 percent faster annual growth.' Adding the GDP of all open coun- tries in the same continent to obtain the open market size for each country yields similar results. The estimated coefficient of the closed neighbors' market in table 1 is by far smaller than the coefficient for open neighboring economies and not statistically significant. This means that a country experiences significant posi- tive spillovers mainly from open neighboring economies. The results confirm those of Sachs and Warner (1995): the openness dummy remains significant in all regressions. Keeping everything else constant, open economies grew on aver- age 1.5 percent faster annually than closed economies during the 1970s and 1980s. 1. One standard deviation of the log of the market of open neighboring economies (5.63) times the estimated coefficient of this variable in regressions 7 and 8 (0.07). 258 THE WORLD BANK ECONOMIC REVIEW, VOL. 12, NO. 2 Table 2. The Impact of Regional Market Size on Growth, 1970-90 Regression Variable 1 2 3 4 5 Constant 0.133 0.139 0.131 0.144 0.176 (4.564) (4.732) (4.486) (5.040) (5.451) Log of GDP per capita, 1970 -0.020 -0.020 -0.020 -0.022 -0.019 (-6.801) (-5.333) (-6.686) (-5.386) (-4.628) Open neighbors' market plus GDP, 0.26 0.25 0.267 0.234 1970a (2.307) (2.407) (2.356) (2.267) Closed neighbors' market, 1970a 0.015 -0.023 (0.776) (-0.720) Average annual share of investment in 0.086 0.123 0.086 0.122 0.151 physical capital over GDP, 1970-90 (2.472) (3.819) (2.467) (3.667) (4.036) Secondary school enrollment, 1970 0.006 0.014 0.005 0.014 -0.011 (0.437) (1.060) (0.399) (1.008) (-0.635) Average annual growth in the terms 0.171 0.169 0.171 0.172 0.197 of trade, 1970-90 (2.907) (2.441) (2.902) (2.478) (2.872) Infant mortality rate, 1970 -0.166 -0.161 -0.168 -0.162 -0.291 (-2.868) (-2.960) (-2.902) (-2.955) (-4.291) Average annual population growth, 0.087 -0.160 0.064 -0.103 -0.290 1970-90 (0.305) (-0.600) (0.214) (-0.333) (-0.993) Share of years open, 1970 0.018 0.019 (2.744) (3.097) Average annual trade share, 1970-90b -0.013 -0.015 -0.043 (-0.581) (-0.633) (-1.584) Average annual trade share of 0.032 neighbors, 1970-90b (1.646) Number of observations 90 89 90 89 66 Adjusted R2 0.58 0.52 0.57 0.52 0.55 Note: The dependent variable is average annual growth in GDP per capita for 1970-90. Regional market size is measured by open neighbors' market plus GDP. Neighboring economies are listed in table A-1. Openness is based on Sachs and Warner's (1995) openness dummy. t-statistics based on heteroscedastic-consistent standard errors are in parentheses. a. The log of the sum of the GDP of all open neighboring economies plus the home country or of all closed neighboring economies. b. (Exports + imports)/GDP. For neighbors, the average is weighted by GDP. Source: Author's calculations. The results show that countries can benefit from being near large, open econo- mies. In contrast, the market size of closed neighboring economies matters much less for growth. This implies that keeping everything else constant, most coun- tries in Europe, North and Central America, and East Asia grew faster than those in Sub-Saharan Africa, which consists mainly of small and highly pro- tected economies. Sachs and Warner (1996) make a similar point about the coun- tries in Sub-Saharan Africa. Regression 6 in table 1 shows that, controlling for home-country openness, the coefficient of the open neighbors' market drops but is still significant. This indicates that the coefficient of open neighbors' market partly captures the effect Vamvakidis 259 of home-country openness. In other words, most open economies have open neighboring economies. Regression 7 tests the robustness of the results obtained by adding two dummy variables for Africa and East Asia and using the share of years a country was open during 1970-90 instead of the openness dummy. Regression 8 uses the trade share as the definition of openness instead. The results remain robust to these changes. The dummy variables for Africa and East Asia are not significant, which suggests that the other independent variables in the regression already explain the differences in these countries' growth compared with other areas of the world economy. Also, the trade share is not statistically significant, a result consistent with Levine and Renelt (1992). Table 2 presents some tests of robustness, including a variable that adds the home country's GDP to the open neighbors' market. This variable measures the available market for a country in its region, which is its own market plus the market of its open neighbors. If regional market size matters, domestic market size should also matter and should be considered. This variable has the advan- Table 3. The Impact of All Neighboring Economies on Growth, 1970-90 Regression Variable 1 2 3 4 5 6 Constant -0.054 0.113 -0.026 0.136 0.035 0.148 (-2.811) (3.425) (-1.741) (4.172) (1.924) (5.333) Log of GDP per capita, 1970 0.0006 -0.018 0.003 -0.017 -0.004 -0.018 (0.316) (-5.271) (2.039) (-5.061) (-1.487) (-5.661) Neighbors' GDP, 1970 0.350 0.157 (2.777) (1.696) Neighbors' population, 1970 0.145 0.0004 (1.776) (0.417) Open neighbors' population, 0.218 0.090 1970 (3.985) (1.761) Average annual share of 0.102 0.107 0.103 investment in physical (2.854) (2.894) (2.911) capital over GDP, 1970-90 Secondary school enrollment, 0.007 0.006 0.003 1970 (0.492) (0.374) (0.206) Average annual growth in the 0.161 0.158 0.154 terms of trade, 1970-90 (2.701) (2.622) (2.724) Infant mortality rate, 1970 -0.153 -0.163 -0.168 (-2.645) (-2.761) (-3.070) Average annual population -0.051 -0.105 -0.098 growth, 1970-90 (-0.191) (-0.379) (-0.369) Openness dummy, 1970 0.015 0.015 0.010 (2.067) (2.116) (1.556) Number of observations 118 90 120 90 120 90 Adjusted R2 0.06 0.52 0.02 0.51 0.15 0.52 Note: The dependent variable is average annual growth in GDP per capita for 1970-90. Neighboring economies are listed in table A-1. Openness is based on Sachs and Warner's (1995) openness dummy. t-statistics based on heteroscedastic-consistent standard errors are in parentheses. Souirce: Author's calculations. Table 4. Spillover Effects from the Economic Development of Neighboring Economies, 1970-90 Regression Variable 1 2 3 4 5 6 7 Constant -0.030 0.113 0.113 -0.012 0.142 0.137 0.110 (-2.507) (3.957) (4.217) (-0.903) (4.762) (4.954) (4.047) Log of GDP per capita, 1970 -0.002 -0.022 -0.022 0.003 -0.018 -0.017 -0.021 (-0.696) (-5.841) (-6.129) (2.113) (-5.045) (-5.470) (-6.054) Neighbors' GDP per capita, 1970a 0.787 0.757 0.683 0.643 (2.746) (3.730) (3.582) (3.453) Neighbors' average annual growth in -0.027 0.051 0.026 0.018 GDP per capita, 1970-90a (-0.672) (1.777) (1.007) (0.625) Average annual share of investment in 0.111 0.092 0.140 0.115 0.099 physical capital over GDP, 1970-90 (4.090) (2.954) (4.207) (3.150) (3.008) Secondary school enrollment, 1970 0.026 0.018 0.014 0.006 0.017 (1.877) (1.180) (1.020) (0.382) (1.069) Average annual growth in the terms 0.153 0.153 0.135 0.129 0.127 of trade, 1970-90 (3.055) (3.012) (2.484) (2.361) (2.555) Infant mortality rate, 1970 -0.143 -0.156 -0.161 -0.169 -0.159 (-2.984) (-3.090) (-3.048) (-3.014) (-3.055) Average annual population growth, -0.037 0.164 -0.330 -0.032 0.218 1970-90 (-0.178) (0.692) (-1.547) (-0.117) (0.857) Openness dummy, 1970 0.012 0.016 0.015 (1.963) (2.172) (1.964) Number of observations 117 90 90 115 87 87 87 Adjusted R2 0.08 0.52 0.55 0.008 0.49 0.52 0.56 Note: The dependent variable is average annual growth in GDP per capita for 1970-90. Neighboring economies are listed in table A-1. Openness is based on Sachs and Warner's (1995) openness dummy. t-statistics based on heteroscedastic-consistent standard errors are in parentheses. a. The sum of the GDP for all neighboring economies divided by the sum of their population. Source: Author's calculations. Vamvakidis 261 tage of capturing more variation within regions than the open neighbors' mar- ket variable. As regressions 1-4 show, the coefficient of the regional market measure is positive, statistically significant, and robust to different specifica- tions of the empirical model. These regressions do not include the Africa and East Asia dummies because they are not statistically significant and do not change the results. The regressions include the share of years a country was open during 1970-90; an openness dummy gives similar results. Regression 5 in table 2 uses a different definition for the openness of the neighboring countries-their GDP-weighted average trade share. The advantage of this variable is that it is continuous and varies a lot within and across regions. The estimated coefficient is positive and statistically significant at the 10 percent level. The estimate means that an increase in the average trade share of the neigh- boring countries by one standard deviation (9.6 percent) will result in 3.2 per- cent faster growth. Therefore, other things being equal, countries with neigh- bors that trade a lot grew faster during the period considered. Table 3 presents regressions that do not separate open from closed economies and that, for the most part, use population as the measure of the neighbors' market size. Neighbors' GDP and population vary more within and across re- gions than the open neighbors' market variable. The first two regressions show that countries with neighbors that have high GDP grow faster. However, when population is used as the measure of neighboring countries' size, without sepa- rating open from closed neighboring economies, the impact is not always signifi- cant. In contrast, when the regression includes only the population of open neigh- boring economies, the coefficient is positive and statistically significant. The results do not change in regressions using other definitions of openness from the previ- ous tables or dummies for Africa and East Asia. Based on regression 6, an in- crease in the log of the open neighbors' population by one standard deviation (4.9) will result in 0.41 percent faster growth. To summarize, countries with large and open neighbors grow faster, a result that is robust to different model specifications and measures of openness and is significant in both the statistical and economic meaning of the word. IV. SPILLOVER EFFECTS FROM THE ECONOMIC DEVELOPMENT OF NEIGHBORING ECONOMIES Section III focused on the market size of neighboring economies. This section estimates the impact that the stage of development and the growth of the neigh- boring economies have on domestic growth. It looks at the impact of all neigh- boring economies and also distinguishes between open and closed ones. Table 4 presents estimates of the impact of the level and growth of the popu- lation-weighted average per capita GDP in neighboring economies on growth of per capita GDP in the home country. (The results do not change if the estimation uses GDP-weighted averages instead.) The results show that countries benefit from having neighbors with more developed economies. The average per capita 262 THE WORLD BANK ECONOMIC REVIEW, VOL. 12, NO. 2 GDP of neighboring economies has a positive and statistically significant coeffi- cient. By contrast, the estimated coefficient of the growth rate of neighboring economies is not statistically significant. Therefore, being near developed coun- tries results in positive spillovers, but being near fast-growing countries does not. Barro and Sala-I-Martin (1995) also find positive spillovers from high GDP per capita in neighboring economies. The present article shows that these spillover effects essentially come from open neighboring economies. Table 5 presents estimates of the impact of the average GDP per capita of open and closed neighboring economies. The results show that the population-weighted GDP per capita of open neighboring economies has a positive and statistically significant impact on domestic growth. The effect of closed neighboring econo- mies is not significant except at the 10 percent level in regression 2. Regressions 3-6 include Africa and East Asia dummies, which do not have a robust impact. Regression 4 uses the share of trade as openness, and regression 5 uses the share of years a country was open during 1970-90. The results are robust for these alternative specifications. Based on regression 3, an increase in the log of aver- age GDP per capita in open neighboring economies by one standard deviation (3.86) results in 0.66 percent faster growth in the home country. Therefore, the spillover effects from the level of development of neighboring economies are important mainly if these economies are open. Regression 6 in table 5 includes the variable of open neighbors' market plus the home country's GDP. This coefficient is positive and significant as in specifications in table 2, but in this case the average GDP per capita of open neighboring econo- mies becomes insignificant. This result may imply that the market size of the neigh- boring economies is more important than their level of GDP per capita for home- country growth. However, given that the two variables are correlated, this interpretation should be approached with caution. If the regression includes the open neighbors' market variable without adding the home country's GDP, the two variables-open neighbors' market and average GDP per capita of neighboring economies-become insignificant, simply because they are highly correlated. V. REGIONAL TRADE AGREEMENTS AND GROWTH Most RTAs created during the 1960s and 1970s were South-South agreements and were part of the import substitution policies that member countries were following. As a result, they diverted trade from more efficient external sources of production. The member countries were typically small, highly protected, and similar in their economic endowments. To test the impact of RTAS on economic growth, I consider agreements that started during the 1970s or earlier and continued for most of the following two decades and for which data are available for most member countries (for more information, see United Nations Conference on Trade and Development 1994). Table A-2 defines these RTAS: Association of South East Asian Nations (ASEAN), Andean Common Market (ANCON), Central American Common Market (CACM), Vamu'akidis 263 Table 5. Spillover Effects from the Economic Development of Open and Closed Neighboring Economies, 1970-90 Regression Variable 1 2 3 4 5 6 Constant 0.038 0.145 0.151 0.163 0.150 0.146 (2.049) (5.357) (5.033) (6.341) (5.085) (5.001) Log of GDP per capita, -0.004 -0.019 -0.020 -0.020 -0.020 -0.023 1970 (-1.590) (-5.912) (-5.260) (-5.773) (-5.412) (-6.899) Open neighbors' GDP 0.279 0.18 0.17 0.21 0.15 0.090 per capita, 1970a (4.155) (2.411) (2.491) (3.054) (2.118) (1.073) Closed neighbors' GDP 0.12 0.09 0.069 0.091 0.099 per capita, 1970a (1.666) (1.343) (0.894) (1.387) (1.297) Open neighbors' market 0.25 plus GDP, 1970b (1.940) Average annual share of 0.103 0.100 0.125 0.087 0.087 investment in physical (3.051) (2.770) (3.776) (2.508) (2.585) capital over GDP, 1970-90 Secondary school 0.004 0.003 0.007 0.001 0.005 enrollment, 1970 (0.253) (0.174) (0.441) (0.086) (0.331) Average annual growth 0.147 0.151 0.173 0.165 0.173 in the terms of trade, 1970-90 (2.630) (2.610) (2.667) (2.867) (2.902) Infant mortality rate, 1970 -0.178 -0.1S3 -0.175 -0.152 -0.165 (-3.434) (-2.735) (-3.160) (-2.703) (-2.887) Average annual population -0.062 0.015 -0.135 0.059 0.147 growth, 1970-90 (-0.229) (0.054) (-0.537) (0.235) (0.527) Openness dummy, 1970 0.011 0.013 (1.893) (2.188) Share of years open, 1970-90 0.019 0.020 (3.055) (3.064) Average annual trade share, -0.025 1970-90c (-1.087) Africa dummy -0.007 -0.004 -0.008 -0.006 (-1.533) (-0.826) (-1.723) (-1.267) East Asia dummy -0.00 1 -0.002 -0.001 -0.006 (0.177) (-0.287) (-0.159) (-1.022) Number of observations 118 90 90 89 90 90 Adjusted R2 0.14 0.53 0.53 0.52 0.55 0.58 Note: The dependent variable is average annual growth in GDP per capita for 1970-90. Neighboring economies are listed in table A-1. Openness is based on Sachs and Warner's (1995) openness dummy. t-statistics based on heteroscedastic-consistent standard errors are in parentheses. a. The sum of the GDP for all neighboring economies divided by the sum of their population. b. The log of the sum of the GOP of all open neighboring economies plus the home country. c. (Exports + imports)/GDP. Source: Author's calculations. European Union, and Union Douaniere et Economique de l'Afrique Centrale (UDEAC). On the right side of the growth regression, dummy variables are in- cluded for countries that were members of any of the five RTAs. Table 6 presents the results. In the first regression, only the estimated coeffi- cient for the European Union is positive and statistically significant at the 10 percent level, while all other RTAs have no impact on growth. However, control- 264 THE WORLD BANK ECONOMIC REVIEW, VOL. 12, NO. 2 Table 6. The Impact of Regional Trade Agreements on Growth, 1970-90 Regression Variable 1 2 3 4 5 Constant 0.089 0.092 0.143 0.154 0.157 (4.758) (5.459) (4.99) (5.761) (6.024) Log of GDP per capita, 1970 -0.015 -0.015 -0.017 -0.019 -0.019 (-5.199) (-5.929) (-5.111) (-5.789) (-.5.970) Openness dummy, 1970 0.018 0.016 0.013 0.013 (3.039) (2.261) (2.031) (2.143) Dummy for membership in RTAa ASEAN 0.754E-2 0.203E-2 -0.261E-2 -0.442E-2 -0.598E-2 (0.614) (0.351) (-0.684) (-0.673) (-0.838) ANCON -0.264E-2 0.044E-2 -0.278E-2 -0.087E-2 0.023E-2 (-0.455) (0.075) (-0.520) (-0.162) (0.044) CACM -0.114E-2 -0.27E-4 -0.141E-2 -0.031E-2 0.345E-3 (-0.190) (-0.004) (-0.262) (-0.058) (0.064) European Union 0.598E-2 -0.048E-2 -0.183 -0.387E-2 -0.382E-2 (1.70) (-0.144) (-0.673) (-1.259) (-1.306) UDEAC 0.057E-2 -0.90E-4 -0.861E-2 -0.864E-2 -0.835E-2 (0.081) (-0.012) (-1.104) (-1.099) (-1.042) Average annual share of 0.155 0.143 0.100 0.092 0.091 investment in physical (5.394) (5.293) (2.840) (2.674) (2.660) capital over GDP, 1970-90 Secondary school 0.034 0.024 0.320 -0.55E-3 -0.004 enrollment, 1970 (2.605) (1.930) (0.219) (-0.037) (-0.255) Average annual growth in 0.184 0.192 0.191 the terms of trade, 1970-90 (2.860) (3.078) (3.099) Infant mortality rate, 1970 -0.177 -0.186 -0.185 (-3.036) (-3.429) (-3.485) Open neighbors' market, 0.078E-2 1970b (1.949) Open neighbors' GDP per 0.11SE-2 capita, 1970' (2.485) Average annual population -0.134 -0.013 -0.089 -0.082 -0.156 growth, 1970-90 (-0.S53) (-0.051) (-0.335) (-0.314) (-0.617) Number of observations 99 99 90 90 90 Adjusted R2 0.43 0.48 0.49 0.50 0.51 Note: The dependent variable is average annual growth in GDP per capita for 1970-90. Neighboring economies are listed in table A-1. Openness is based on Sachs and Warner's (1995) openness dummy. t-statistics based on heteroscedastic-consistent standard errors are in parentheses. a. Table A-2 provides the full name, country members, and date of formation for each of the five regional trade agreements (RTAS). b. The log of the sum of the GDP of all open neighboring economies. c. The sum of the GDP for all neighboring economies divided by the sum of their population. Source: Author's calculations. ling for openness and other independent variables, even the European Union dummy becomes insignificant. Although the estimated coefficients of all regional agreements' dummies are negative in most specifications, none is significant. These results indicate that these RTAs did not affect growth significantly. Although past RTAs did not result in faster growth, this does not necessarily imply that the RTAs formed during the 1990s will have no impact on growth. Vamvakidis 265 Some of the recent trade agreements are considered as part of a broader liberal- ization-especially in the case of North-South RTAs-and include open, large, and more developed economies. The results of section IV show that a country with such neighboring economies will experience positive spillovers. Thus, even though RTAs did not help member countries experience these benefits in previ- ous decades, some of the recent RTAS may have an important impact on growth. VI. CONCLUDING REMARKS AND POLICY IMPLICATIONS This article has examined whether the openness, market size, and level of development of countries in the same region foster growth in the home country. The results show that the economies of countries near large and open economies grow faster. Also, the level of development of neighboring economies, especially when they are open, has significant positive spillover effects. By contrast, the size and level of development of closed neighboring economies have little or no impact on domestic growth. These results suggest that trade agreements between developing countries and large and more developed countries may lead to faster growth. Given that no country has zero trade barriers, if an RTA increases the openness of the large and more developed economies toward less developed member countries, it will promote their growth. This article has found that RTAs had no impact on growth in the past. The most appealing theoretical reason is that these agreements were among small, developing, and very similar economies. Most of the time, the countries designed RTAS to divert trade, as a part of import substitution trade policies. However, this interpretation does not explain the empirical results for the European Union. It is possible that countries have designed recent RTAs as part of a broader liberalization and that these agreements, especially North-South RTAS, will af- fect growth. This article has shown that countries with open, large, and more developed neighboring economies do experience positive spillovers. (Table A-1 begins on the following page.) 266 THE WORLD BANK ECONOMIC REVIEW, VOL. 12, NO. 2 Table A-1. Neighboring Economies Country code and country Codes of neighboring economies 1 Algeria 29 26 43 31 27 2 Angola 45 46 12 3 Benin 32 42 5 31 4 Botswana 38 47 5 Burkino Faso 26 31 20 17 3 42 6 Burundi 41 33 45 7 Cameroon 32 10 9 12 15 8 Cape Verde 34 27 16 19 9 Central African Republic 45 10 39 7 12 10 Chad 39 9 31 7 32 11 Comoros 12 Congo 45 15 7 9 2 13 Egypt 39 89 14 Ethiopia 39 37 21 15 Gabon 12 7 16 The Gambia 34 17 Ghana 42 20 5 18 Guinea 26 36 20 23 34 19 19 Guinea Bissau 18 34 20 Cbte d'Ivoire 23 17 18 5 26 21 Kenya 44 14 41 37 39 22 Lesotho 38 23 Liberia 18 36 20 24 Madagascar 28 30 25 Malawi 30 46 41 26 Mali 27 1 5 18 31 20 34 27 Mauritania 26 34 1 28 Mauritius 24 29 Morocco 1 30 Mozambique 25 47 41 38 46 40 31 Niger 32 10 1 26 5 3 32 Nigeria 7 31 3 10 33 Rwanda 6 45 41 44 34 Senegal 27 16 26 18 19 35 Seychelles 21 37 41 36 Sierra Leone 18 23 37 Somalia 14 21 38 South Africa 4 22 30 40 47 39 Sudan 14 10 13 9 45 21 44 40 Swaziland 38 30 41 Tanzania 21 30 25 6 44 46 33 42 Togo 17 3 5 43 Tunisia 1 44 Uganda 21 45 39 41 33 45 Zaire 2 12 46 9 44 39 6 33 46 Zambia 45 2 25 47 30 41 47 Zimbabwe 30 4 46 38 Vamvakidis 267 Table A-1. (continued) Country code and country Codes of neighboring econiomies 48 Bahamas 66 49 Barbados 65 50 Canada 66 51 Costa Rica 62 52 Dominica 78 53 Dominican Republic 57 54 El Salvador 56 58 55 Grenada 78 56 Guatemala 60 58 54 57 Haiti 53 58 Honduras 61 54 56 59 Jamaica 57 60 Mexico 66 56 61 Nicaragua 58 51 62 Panama 51 71 63 St. Lucia 78 64 St. Vincent 78 65 Trinidad and Tobago 49 78 66 United States 50 60 67 Argentina 70 74 69 68 77 68 Bolivia 69 75 70 67 74 69 Brazil 68 78 71 75 74 67 77 76 70 Chile 67 68 75 71 Colombia 75 78 69 72 62 72 Ecuador 75 71 73 Guyana 69 78 76 74 Paraguay 67 69 68 75 Peru 71 69 72 68 70 76 Suriname 73 69 77 Uruguay 69 67 78 Venezuela 69 71 73 79 Afghanistan 97 87 80 Bahrain 99 0 81 Bangladesh 82 85 82 Myanmar 104 85 81 83 83 China 85 84 82 84 Hong Kong 83 103 85 India 81 83 97 95 82 86 Indonesia 94 134 87 Iran 88 97 128 79 88 Iraq 87 102 99 128 93 91 89 Israel 91 102 13 90 Japan 92 83 91 Jordan 99 102 89 88 92 Korea, Rep. of 90 93 Kuwait 88 99 (Table continues on the following page.) 268 THE WORLD BANK ECONOMIC REVIEW, VOL. 12, NO. 2 Table A-1. (continued) Country code and country Codes of neighboring economies 94 Malaysia 86 104 100 95 Nepal 85 83 96 Oman 99 105 106 97 Pakistan 85 87 83 79 98 Philippines 86 99 Saudi Arabia 106 91 96 105 88 93 100 Singapore 94 101 Sri Lanka 85 102 Syria 128 88 91 89 103 Taiwan 84 83 104 Thailand 94 82 105 United Arab Emirates 99 96 106 Yemen, Arab Republic 99 96 107Austria 113 118 127 115 108 Belgium 112 121 113 119 109 Cyprus 128 102 110 Denmark 113 111 Finland 122 126 112 France 125 108 127 118 113 119 113 Germany 107 121 112 127 108 119 110 114 Greece 128 130 115 Hungary 130 107 116 Iceland 122 129 117 Ireland 129 118 Italy 127 112 107 130 119 Luxembourg 108 113 112 120 Malta 118 114 121 Netherlands 113 108 122 Norway 126 Ill 123 Poland 124 Portugal 125 125 Spain 124 112 126 Sweden 122 111 127 Switzerland 118 112 113 107 128 Turkey 102 87 88 114 129 United Kingdom 117 130 Yugoslavia 114 118 131 Australia 133 86 134 132 Fiji 134 131 133 133 New Zealand 131 132 134 Papua New Guinea 86 135 Solomon Islands 134 136 Tonga 131 133 137 Vanuatu 131 138 Western Samoa 131 133 Vamvakidis 269 Table A-2. Regional Trade Agreements Agreement Year created Member countries Andean Common Market, ANCOMa 1969 Bolivia, Colombia, Ecuador, Peru, Venezuela Association of South East Asian 1967 Indonesia, Malaysia, Philippines, Singapore, Nations, ASEAN Thailand Central American Common Market 1960 Costa Rica, El Salvador, Guatemala, Honduras, CACM b Nicaragua European Union c 1958 Belgium, Denmark, France, Germany, Greece, Ireland, Italy, Luxembourg, Netherlands, Portugal, Spain, United Kingdom Union Douaniere et Economique 1966 Cameroon, Central African Republic, Chad, de l'Afrique Centrale, UDEAC Congo, Gabon a. Venezuela joined ANCOM in 1973. b. Costa Rica joined CACM in 1962. c. Denmark, Ireland, and United Kingdom joined the European Union in 1973. Source: UNCTAD (1994). REFERENCES The word "processed" describes informally reproduced works that may not be com- monly available through library systems. Alesina, Alberto, and Allan Drazen. 1991. "Why Are Stabilizations Delayed?" Ameri- can Economic Review 81(5, December):1170-88. Baldwin, Richard, and Elena Seghezza. 1996. "Growth and European Integration: To- wards an Empirical Assessment." CEPR Discussion Paper 1393. Centre for Economic Policy Research, London. Processed. Barro, Robert J., and Jong-Wha Lee. 1994. "Data Set for a Panel of 138 Countries." Department of Economics, Harvard University, Cambridge, Mass. Processed. Barro, Robert J., and Xavier Sala-I-Martin. 1995. Economic Growth. New York: McGraw-Hill, Inc. Ben-David, Dan. 1993. "Equalizing Exchange: Trade Liberalization and Income Con- vergence." Quarterly Journal of Economics 108(3, August):653-79. Bruno, Michael, and William Easterly. 1996. "Inflation's Children: Tales of Crises That Beget Reforms." American Economic Review 86(2, May):213-17. Chua, Hak B. 1993. "Regional Spillovers and Economic Growth." Ph.D. diss., Depart- ment of Economics, Harvard University, Cambridge, Mass. Processed. Coe, David T., and Elhanan Helpman. 1997. "North-South R&D Spillovers." The Eco- nomic Journal 107(440, January): 134-49. Coe, David T., Elhanan Helpman, and Alexander W. Hoffmaister. 1997. "North-South R&D Spillovers." The Economic Journal 107(440, January):134-49. Dollar, David. 1992. "Outward-Oriented Developing Economies Really Do Grow More Rapidly: Evidence from 95 LDCS, 1976-1985." Economic Development and Cultural Change 40(3, April):523-44. Edwards, Sebastian. 1992. "Trade Orientation, Distortions, and Growth in Developing Countries." Journal of Development Economics 39(1):31-57. Grossman, Gene, and Elhanan Helpman. 1989. "Product Development and Interna- tional Trade." Journal of Political Economy 97(6, December):1261-83. 270 THE WORLD BANK ECONOMIC REVIEW, VOL 12, NO. 2 . 1990. "Comparative Advantage and Long-Run Growth." American Economic Review 80(4, September):796-815. . 1991. Innovation and Growth in the Global Economy. Cambridge, MViass.: MIT Press. Henrekson, Magnus, Johan Torstensson, and Rasha Torstensson. 1997. "Growth Ef- fects of European Integration." European Economic Review 41(8):1537-57. Levine, Ross, and David Renelt. 1992. "A Sensitivity Analysis of Cross-Country Growth Regressions." American Economic Review 82(4, September):942-63. Lucas, Jr., Robert E. 1988. "On the Mechanics of Economic Development." Journal of Monetary Economics 22(1, July):3-42. Krugman, P. R. 1990. Rethinking International Trade. Cambridge, Mass.: MIT Press. Rivera-Batiz, Luis A., and Paul M. Romer. 1991a. "Economic Integration and Endog- enous Growth." Quarterly Journal of Economics 106(2, May): 531-55. . 1991b. "International Trade with Endogenous Technological Change." Euro- pean Economic Review 35(4):971-1001. Romer, Paul M. 1990. "Endogenous Technological Change." Journal of Political Economy 98(5, part 2, October):S71-S102. Sachs, Jeffrey, and Andrew Warner. 1995. "Economic Reform and the Process of Glo- bal Integration." Brooking Papers on Economic Activity 1:1-95. . 1996. "Sources of Slow Growth in the African Economies." Harvard Institute of International Development, Harvard University, Cambridge, Mass. Processed. Summers, Robert, and Alan Heston. 1991. "The Penn World Table (Mark 5): An Ex- panded Set of International Comparisons, 1950-1988." Quarterly Journal of Eco- nomics 106(2, May):327-68. United Nations Conference on Trade and Development. 1994. Handbook of Interna- tional Trade and Development Statistics. Geneva: United Nations. Vamvakidis, Athanasios. 1997. "GATT Was Right: Historical Evidence on the Growth- Openness Connection and Discriminatory versus Nondiscriminatory Liberalization." Economics Department, Harvard University, Cambridge, Mass. Processed. . 1996. "How Important Is a Large Market for Economic Growth?" Economics Department, Harvard University, Cambridge, Mass. Processed. THE WORLD BANK ECONOMIC REVIEW, VOL. 12, NO. 2 271-295 Regional Integration as Diplomacy Maurice Schiff and L. Alan Winters Regional integration agreements are examples of second-best policies and have an ambiguous impact on welfare. This article builds a model in which regional integra- tion agreements unambiguously raise welfare by correcting for externalities. It assumes that trade between neighboring countries raises trust between them and reduces the likelihood of conflict. The optimum intervention in that case is a subsidy on imports from the neighbor. The article shows that an equivalent solution is for the neighboring countries to tax imports from the rest of the world-that is, to form a regional integra- tion agreement-together with some domestic taxes. The article shows that (1) the optimum tariffs on imports from the rest of the world are likely to decline over time; (2) deep integration implies lower optimum external tariffs if it is exogenous; (3) optimum external tariffs are higher before deep integra- tion and lower thereafter if deep integration is endogenous; and (4) enlargement of bloc size (in terms of symmetric countries) has an ambiguous impact on external tariffs but raises welfare, and some form of domino effect exists. II n'y aura de paix en Europe si les Etats se reconstituent sur une base de souverainete nationale avec ce que cela entraine de politique de prestige et de protection &conomique ... la constitution de vastes armees sera a nouveau necessaire... L'Europe se recreera une fois de plus dans la crainte . . . moins que les Etats d'Europe ne se forment en une Federation ou une "entite europeenne" qui en fasse une unitW economique commune.' As has long been known, regional integration agreements (RIAs) are examples of second-best policies, and their impact on economic welfare is ambiguous. De- spite an enormous theoretical, empirical, and historical descriptive literature, no consensus on the desirability of RIAs has emerged. 1. "There will be no peace in Europe if the States reconstitute themselves on a basis of national sovereignty with its policies of prestige and economic protection.... The constitution of large armies will again be necessary . . . Europe will once again be recreated in fear . . . unless the States of Europe join in a Federation or a 'European entity' that results in a common economic unit." Jean Monnet, 5 August 1943, as quoted in Rieben, Monnet, and Chevallaz (1971) and translated by the authors. The idea that commercial ties will reduce the risk of European wars dates back at Least to the 1795 publication of Kant's Perpetual Peace (Kant 1992). Maurice Schiff and L. Alan Winters are with the Development Economics Research Group at the World Bank. The authors would like to thank the referees and participants in seminars at the World Bank and in the 1997 meetings of the American Economic Association for useful comments. This article was produced as part of the World Bank Development Economics Research Group's research program on regionalism and development. C) 1998 The International Bank for Reconstruction and Development/THE WORLD BANK 271 272 THE WORLD BANK ECONOMIC REVIEW, VOL. 12, NO. 2 One of the few unambiguous static results that has emerged is that, in the case of homogeneous products, an RIA between small countries whose terms of trade are given exogenously will result in a welfare loss for the RIA as a whole as long as its members continue to trade with the rest of the world after forming the RIA (Bhagwati and Panagariya 1996 and Schiff 1997). Even this result need not hold in the presence of smuggling (Schiff 1997). Given the ambiguity of the static welfare impact of RIAs and their generally small estimated size, many commentators have appealed to dynamic effects- such as those on foreign direct investment, economies of scale, and convergence- to justify them. Whether these dynamic effects actually take place as a result of RIAS, however, has not been demonstrated conclusively either analytically or empirically (see, for example, Vamvakidis, this issue). And even if they do occur, they need not result in welfare gains. For instance, increased foreign direct in- vestment may result in immiserization if it leads to the expansion of protected sectors. Related arguments have been made that regional trading arrangements might enhance developing countries' efforts to industrialize (see Puga and Venables 1997 and this issue for the first formalization of this idea). However, RIAs generate unambiguous welfare gains even in terms of static welfare theory when they correct externalities. For instance, it is frequently claimed that a developing country in the process of reforming its trade and other policies can benefit from an RIA with a large, developed country or region (such as the United States or European Union) because the RIA binds the country's reforms in an international treaty, weakens the groups that stand to lose from and oppose the reforms, and raises the reforms' credibility. If the beneficial reform- support effects outweigh the traditional deadweight loss effects, an RIA may in- crease economic welfare. (See Fernandez's article in this issue for an insightful discussion of these arguments.) Another kind of externality arises when an RIA improves security for its mem- ber countries. There are basically three types of situations in which RIAs may generate such positive externalities. First, there may be domestic security threats such as civil disruption or civil war. For instance, the Egyptian government has been concerned with the spread of fundamentalism, and the governments of Morocco and Tunisia have been concerned that fundamentalism in neighboring Algeria, with the associated risk of civil strife, may be contagious. These issues, which have also been of concern to the European Union, have provided one of the motivations for agreements between these Mediterranean countries and the European Union. Second, countries may respond to security threats from third countries by forming a regional arrangement. For instance, the Southern African Develop- ment Coordination Conference, which eventually developed trading arrange- ments under the Southern African Development Community, was formed to provide a united front against the Republic of South Africa.2 The Gulf Coopera- 2. See Table A-1 in Schiff and Winters' introduction to the symposium (this issue) for a list of regional trade agreements and their member countries. Schiff and Winters 273 tion Council was created partially in response to the potential threat of regional powers such as Iran and Iraq. And Central and Eastern European countries have applied for membership in the European Union partly as protection against the perceived threat from Russia. Third, security threats may move neighboring countries to form RIAS. For example, the European Coal and Steel Community (ECSC in 1951) and the Euro- pean Economic Community (in 1957) were created to reduce the threat of war in Europe (see the quote of Jean Monnet at the start of this article), the Associa- tion of South East Asian Nations (ASEAN) was created to reduce tensions be- tween Indonesia and Malaysia, and Mercosur (a regional trade agreement among Argentina, Brazil, Paraguay, and Uruguay) was created to reduce tensions be- tween Argentina and Brazil (see Bastian 1996). Page (1996) suggests that this motivation also prompted the formation of the Asia-Pacific Economic Coopera- tion Process and the Central American Common Market, which include poten- tial political/military opponents. Despite the frequency with which such motivation is claimed for RIAS, not only by politicians but by academic commentators as well (see, for example, Swann 1992), it has, to our knowledge, never been subjected to formal analysis and still less to empirical test. Our purpose is to start this process by analyzing a formal model of an RIA motivated by (intermember) security concerns and by deriving predictions about observable phenomena that could, in principle, be tested. Four major predictions are derived in this article. Assuming that trade can help to reduce frictions among hostile neighboring countries, we show that • An RIA is an optimum (first-best) arrangement in traditional, static welfare terms * The optimum external tariffs fall over time • The optimum external tariffs fall following deep integration (defined as integration measures that go beyond trade policy, including harmonization and mutual recognition of standards, investment codes, and other measures that increase market contestability and lower the cost of trade) * Enlargement raises member countries' welfare, and optimum external tar- iffs are likely to rise. Our purpose is not to advance the argument that trade is a route to rap- prochement between antagonistic states, but rather to take this argument at face value and explore whether an RIA can generate welfare gains under these conditions and to predict how such an RIA would evolve. First, we demonstrate that such welfare-enhancing RIAs exist. Second, we generate predictions about the development of regional integration. Finally, we identify theoretical pre- dictions about observable phenomena that might allow investigators to test whether particular RIAS were motivated by a concern for security. For example, the fact that during its 40-year history the European Union has regularly re- duced its tariffs on nonmembers is consistent with our predictions. Because other explanations exist, this obviously cannot prove our hypothesis, but it 274 THE WORLD BANK ECONOMIC REVIEW, VOL. 12, NO. 2 lends credence to the conclusion that intermember security was a factor be- hind European integration. Identifying the motivation for an RIA does not alter its economic and political effects but does allow a more rational discussion of policy options. Many poli- cies that are economically costly are justified on political or security grounds. If our test suggests that, in fact, behavior is not consistent with security concerns, it would be more difficult to dismiss economic considerations. See Winters (1989, 1991) for a similar approach to agricultural trade policies. We view this article as only a first step toward formalizing the link between intermember security and regional integration. We wish to show that formaliza- tion can generate hypotheses that can be tested empirically. We use the simplest framework that incorporates the links cited by advocates and explore their implications. Section I explores the argument that integration fosters security, and this is translated into a formal model in section II. (The model provided in this article abstracts from defense as an alternative means of providing security. This issue is on our research agenda, as is discussed in section VIII.) The solution is pro- vided in section III (under symmetry and under asymmetry). The relationship between both autonomous and endogenous deep integration and optimum ex- ternal trade policy is examined in section IV. Dynamic aspects are examined in section V. Bloc enlargement and domino effects are examined in section VI. Section VII concludes, and section VIII discusses possible extensions. Finally, technical derivations are included in appendixes A, B, and C. I. TRADE AND SECURITY The notion that trade is a civilizing influence is an old one (see Hirschman 1982). The notion that international trade can diffuse tension and bring nations together is also venerable. The nineteenth-century British politician Richard Cobden persistently advocated that Great Britain should trade freely with its neighbors as a means of convincing them of the advantages of free trade and also as a means of locking them more fully into the community of nations. In the twentieth century, Cordell Hull, U.S. secretary of state (1933-44) and an archi- tect of the postwar international trading order, advocated this view throughout his public life. His autobiography contains homely stories of trade reconciling warring neighbors along with discussions of international strategy (Hull 1948 ).3 However, both Cobden and Hull were strongly committed to nondiscrimina- tion. Thus, although the Cobden-Chevalier Treaty of 1860 between France and Great Britain was bilateral, Cobden's conception was multilateral, and the treaty 3. For a contrary position, see Keynes, who stated that "I sympathize ... with those who would minimize rather than those who would maximize economic entanglement between nations. Ideas, knowledge, art, hospitality, travel-these are the things which should of their nature be international" (see Skidelsky 1992: 476-80). Schiff and Winters 275 itself contained an unconditional most-favored-nation clause. Hull was explicit that trade discrimination bred rather than diffused tension (Hull 1948).4 For advocacy of regionalism on security grounds, we turn to continental Eu- rope. Wilfredo Pareto argued in 1889 that "Customs unions . . . [were] . .. a means to better political relations and eventual pacification" (Machlup 1977: 41), and Robert Schuman and Jean Monnet, the founding fathers of the Euro- pean Economic Community, were explicit that the ECSC was to make Franco- German war not only "unthinkable, but materially impossible" (see Swann 1992: 6). Monnet also argued that while Great Britain, the United States, and the So- viet Union could withdraw into their own spheres, France and Germany were inextricably linked and had no alternative than to solve the "European problem." Later echoes of Monnet's views are common. For example, Dr. Walter Hallstein, a former president of the European Community Commission, has put it clearly: "We are not in business at all; we are in politics" (see Swann 1992: ix). And Jones (1993: 83), referring to France and Germany, states that "Some trad- ing blocs may be advocated primarily to avoid military conflicts." Hirschman (1981: chap. 12) makes a similar point, although he argues that the European Community may have arrived a bit late. Jones (1993) also argues that, because open markets imply a loss of national control over the economy, countries may prefer to form RIAs with like-minded neighbors. In other words, the issue of cultural externalities (where people ben- efit from interacting with a larger number of similar people) may influence the choice of partner as well as the decision to form an RIA. For instance, one sugges- tion for why a number of North African countries have chosen to form RIAs with the European Union rather than liberalize unilaterally is that unilateral liberal- ization would subject them to open competition with Asia and possibly threaten their lifestyle. The analysis provided in this article applies to cultural externali- ties as well. Switzerland, by contrast, found an alternative way to generate secu- rity through neutrality and has so far been reluctant to join the European Union. Security and other noneconomic aspects of RIAs also seem to have played a role in the Southern Cone. Argentina and Brazil signed nuclear cooperation and economic agreements in the area of capital goods and automobiles in the mid- 1980s, with each quite possibly hoping that reducing external tensions would allow it to reduce the power of its military and strengthen its fragile democracy. The creation of Mercosur in 1991 continued this process and drew smaller neigh- boring countries into it. Recent rumors that a possible coup in Paraguay was laid to rest following reaffirmation by the presidents of the four member coun- tries that democracy was a necessary condition for membership in the bloc, based on a clause in the Treaty of Asunci6n establishing Mercosur, has been cited as evidence of the political effectiveness of Mercosur ("Survey on Mercosur," The Economist, October 12, 1996). 4. He advocated the principle of unconditional most-favored-nation status, which in U.S. usage implies offering most-favored-nation rariffs to any country offering most-favored-nation status to the United States. That is, it continues to permit discrimination against those outside the system. 276 THE WORLD BANK ECONOMIC REVIEW, VOL. 12, NO. 2 Srinivasan (1994: 7) argues that integration might help to reduce tensions in South Asia: "It is conceivable that promoting freer movement of goods, services, people, and capital in the region might also facilitate the resolution of political and territorial disputes." We focus on the movements of goods and services. A direct link between regionalism and security arises if a regional agreement makes access to the partners' supplies more secure by reducing the threat of a trade embargo. Arad and Hillman (1979) show how fear of being cut off from foreign sources of defense equipment can cause countries to overinvest (relative to comparative advantage) in their own defense industries. Similarly, Hillman and van Long (1983) consider the optimal exploitation of a mineral resource if the alternative foreign supply is uncertain. In both these cases, an RIA that as- sures partners' supplies would be advantageous through its effect on members' allocation of resources, which is a different effect from our pacification hypoth- esis. But although these authors do not consider this explicitly, an RIA that makes access to partners' supplies more secure could reduce the likelihood of a country using conquest to ensure access. As stated by Keynes (1919: 99), "In a regime of free trade and free economic intercourse, it would be of little consequence that iron lay on one side of a political frontier and labor, coal, and blast furnaces on the other." Political scientists have also discussed the use of trade diplomacy within a regional context, including whether and what type of RIAS might raise welfare of the member countries through intramural conflict avoidance and management. Bastian (1996) argues that trade negotiations between leaders of neighboring countries are likely to result in a higher degree of trust between them. He men- tions that RIAS may enable "political and/or economic elites to form coalitions for subsequent collaboration and consensual action" and that they may "carry their own language and discourse, thus being able to socialize the participants, e.g., when talking about a 'region."' Mansfield (1993) also recognizes the im- portance of security externalities in trade relations. However, he argues that countries will lower barriers with other countries that belong to the same alli- ance (for example, the North Atlantic Treaty Organization) because the increased trade will raise incomes, which can then be used to pay for defense expenditures. Contrary to Mansfield, we assume that the very action of trading generates se- curity benefits between adversaries or friends and note that Mansfield's analysis is weakened by the fact that trade preferences have an ambiguous impact on income. On this issue, see also Mansfield and Bronson (1997). Evidence is limited of the impact of trade on the likelihood of conflict be- tween countries. Numerous studies in the political science literature have con- firmed the results of Chan (1984) that conflict is less prevalent between two countries if both are democratic. Polachek (1992, 1996) explains this finding through the effect of democracy on trade. He finds that democracies trade more with each other than with other countries and, using data from the Conflict and Peace Data Bank, finds a significant and negative relationship between trade and conflict. He obtains an elasticity of his measure of conflict with respect to Schiff and Winters 277 trade of -0.15 to -0.19. Polachek tests the direction of causality by means of three-stage least squares and finds that the trade variable remains statistically significant and becomes more important (elasticity of -0.30) in the conflict equa- tion, while the conflict variable is not significant in the trade equation. These causality results are confirmed by Granger causality tests (Gasiorowski and Polachek 1982). In this article, we refer to RIAS in a geographic sense or to what Ethier (1996) has called "regional regionalism." This definition would seem to cover most RIAs because they are generally formed among neighboring countries. Examples include the European Union, Mercosur, the Andean Pact, the Central American Common Market, the North American Free Trade Agreement (NAFTA), ASEAN's Asian Free Trade Agreement, the South Asian Free Trade Agreement, the Gulf Cooperation Council, the Economic Community of West African States, and the Southern African Development Community. Our analysis is thus of potential relevance for the majority of RIAS. Although the rest of this article formalizes the role of trade in rapproche- ment, a similar role can be claimed for other forms of economic relations. For example, mutual investment or migration flows can stimulate trust, as can the meetings between national civil servants that are necessary for policy coopera- tion (see Winters 1997). RIAs frequently promote these forms of contact either directly or indirectly, and their analysis would be very similar to that which follows, that is, the acceptance of economic costs-investment from partners rather than from the rest of the world or constraints on policy discretion-in return for security benefits. Most of the literature and political discourse fo- cuses on trade, however, perhaps because capital mobility was not a strong feature of the 1940s or 1950s when the Europeans were formulating these ideas and perhaps also because resource ownership, migration, and the inter- mixing of political elites are more intrusive and harder to sell politically than trade. II. A TRADE MODEL WITH SECURITY BENEFITS The idea of the model is very straightforward: international trade with a po- tentially hostile partner generates a positive externality-security-that individual traders cannot appropriate. As a result, private decisions generate suboptimal levels of trade with such partners, calling for a subsidy on trade with partners. Taxing domestic commerce and trade with countries other than the partner can be equivalent to this subsidy, and so an RIA is born. We formalize this idea in this and the next section. Subsequent sections extend the idea to treat issues of deep integration, dynamics, and enlargement. Assume three countries (1, 2, and the rest of the world, ROW) and three normal goods (A, B, and R). Country 1 produces good A, country 2 produces good B, and ROW produces goods A, B, and R. We abstract from optimal taxa- tion issues related to economic power on the world market in order to focus 278 THE WORLD BANK ECONOMIC REVIEW, VOL. 12, NO. 2 exclusively on matters of security and trade diplomacy.5 Hence, we assume that countries 1 and 2 are small and that the world prices, PA*, PA', and PR, they face are determined in ROW. Without loss of generality, units of A, B, and R are selected such that PA* = PB' = PR = 1. Output in country 1 (2) is given exog- enously and equals national income Y, (Y2) measured at world prices. Denote consumer prices of goods A, B, and R in country i (i = 1, 2) by PA, PB, and PR, respectively. Assume a representative consumer in country i, whose con- sumption of goods A, B, and R is denoted by Ai, Bi, and Ri and who derives utility from the consumption of goods and from social capital in the form of (collective) security. Thus: (1) Ui = V(Ai,Bi,Ri) + Zi(SKi); i = 2, Vj,Z' > o,Vjj,Z< O,V(V=O ) =,vj =_ avl aj(j = Ai,Bi,Ri), and zo _ az1 I aSKi with Ui twice continuously differentiable. The assumption that Vi (0) = (j =i, Bi, Rj) ensures internal equilibrium in the sense that countries 1 and 2 trade both with each other and with ROW no matter what taxes or subsidies are applied. SKi denotes the level of security in country i that is associated with trust in the neighboring country or the level of security capital in country i. We assume that utility increases at a decreasing rate as the level of security increases (Z," < 0). Security SKi is a public good that our representative consumer takes as given when maximizing Ui. The welfare impact of incorporating a public good SKi (social capital) in the utility function is examined in the case of labor mobility in Schiff (1992) and in the case of international migration in Schiff (1996). Becker (1996) and Bliss (1994) also assume that social capital enters the utility function. It could equally well have been incorporated into the production functions for goods A and B rather than in the utility function, with a lower degree of security implying that more resources must be devoted to security matters and fewer are available for the production of goods A and B. This approach is taken in Schiff (1995, 1998), which examine the impact of policy reform under ethnic diversity. We refer below to countries 1 and 2 as partners whether or not they are members of an RIA. We assume that an increase in home-country imports from the partner reduces the level of insecurity in the home country. In other words, importing from the partner country increases interaction with individuals in the partner country, raises the level of information about them, and increases the level of trust and security in the home country.6 Alternatively, one might postu- 5. See Krugman (1991) for an analysis of customs unions' optimal common external tariffs associated with market power. 6. The level of home-country imports of the partner's good equals the level of home-country consumption of that good, which is not produced there. In fact, SK depends on total rather than per capita imports from the partner, so the term Li should be included in the equation. Expressing SK as a function of per capita imports does not affect the model's solution (although it matters in the case of population changes) and is done for notational simplicity. Schiff and Winters 279 late that larger home-country imports from the partner raise the importance of the home country as a market for the partner's exports and lowers the likeli- hood of a security threat by the partner country. Thus, (2) SK =SK (Bl); SK2 = SK (A2) with SK2 twice continuously differentiable and SK, (SK2) increasing in B1 (A2) and strictly concave. Keynes (1919: 30-32) describes how France continued to see Germany as a threat in 1919-even though Germany lost the war-because Germany's wealth and population had grown much faster than France's between 1870 and 1914. Thus a neighbor's relative power (such as wealth and population) also affects security. Such a term could easily be added in equation 2. It is not done for simplicity, although implications are discussed in section III. The discussion so far has been about the security benefits of trade with a neighboring country, while equation 2 expresses security as a function of im- ports rather than trade. Assuming that security is a function of imports rather than trade simplifies the analysis but has no effect on the main findings. This issue is discussed further at the end of section III and in section VIII. III. EQUILIBRIUM AND OPTIMAL TRADE SOLUTIONS Our representative consumer maximizes utility Ui with respect to Ai, Bi, and Ri, subject to the budget constraint Yi = PAAi + PhBi + PkRi, taking SKi as given exogenously (i = 1, 2). As usual, the individual's first-order conditions are (3) VA I VR = PA I PR; VB' / VR' = PB' / PR'; i =- 1,2. Given our choice of units, in the absence of domestic or trade taxes in countries 1 and 2, PJ= Pi<- 1 (i = 1, 2; j = Ai, Bi, R,), so that utility maximization implies (4) VAIV = VB /V=1. This maximizes national income LiVi, that is, the part of welfare W = LiUi, that depends directly on the consumption of goods and services (where Li is the population of country i). However, welfare also depends on benefits from secu- rity, LiZi, which is not a choice variable for the representative consumer. Thus while individual utility maximization requires that a VI/ aj = U / aj for Vj, social welfare maximization requires that av I / j < a U / aj for j = B1, A2, because the home country's social gain from imports from the partner exceeds the private gain by the positive security externality. It is well known that welfare maximization requires distortions to be attacked at the source (Bhagwati 1987) and that policy should internalize the externali- ties. In this case, because SK is a function of imports, welfare W, = LiUi in the home country is maximized with a subsidy Si on partner-country imports (as- sumed to be financed through lump-sum taxation) equal to the externality gen- 280 THE WORLD BANK ECONOMIC REVIEW, VOL. 12, NO. 2 erated by these imports. By standard manipulations, the optimal subsidy is equal to (5) Si = Li (aZi / aMi) / Xi = Li(Z'SK') / Xi = Li(Z'SK[') I R; i = 1,2, where the numerator equals the externality from an additional unit of imports in country i, Li (aZi / aMi) (M = imports from the partner), and ki is the marginal utility of income that equals VR /PR or VR1 (because PR = 1), where VR is the marginal utility of R. Because PA = P= 1, equation 5 also defines the optimum subsidy rate defined as a share of the world price. Because countries 1 and 2 continue to trade with ROW following the import subsidy Si (see equation 1), the import subsidy does not affect producer prices in countries 1 and 2. The only effect is to lower the consumer price of imports from the partner. This is shown for country 1 in figure 1 where all quantities are in per capita terms. DP(DS) measures the private (social) value of good B for country 1. The demand price on the social demand curve Ds (that is, the social value) is the sum of the private value and the value of the externality. Following the strict concav- ity assumptions made in equations 1 and 3, the vertical distance between the two demand curves falls with Bl. Private equilibrium in the absence of interven- tion is at point E, with imports equal to BP, whereas the social optimum is at point F, with imports equal to Bs. The latter is obtained with a subsidy on im- ports of good B from country 2, which lowers the consumer price by the amount of the externality. This is shown in figure 1 by the line PI, which is the world price, PB, minus the import subsidy (PB - S1 = 1 - SI), which intersects the private demand curve DP at point G, with subsidy S1 = FG and imports, Bs. The relative consumer prices following imposition of the optimal subsidy rate, Si, are: (6) PA IPR = 1, PB/ R 11 A R 2 B R As can easily be verified, the same relative prices can be obtained by trading B freely and imposing instead a tax rate Ti = Si I (1 - Si) on imports from ROW and on consumption of the home good (with the tax returned in lump-sum fashion). Thus welfare of country i (i = 1, 2) is maximized either by imposing a subsidy rate Si on imports from the partner or, alternatively, by countries 1 and 2 form- ing an RIA with a tariff on imports from ROW of Ti = Si ' (1 - Si) and an equal tax rate on consumption of the home good. Even if the tax is imposed exclusively on imports from ROW and not on consumption of the home good, there is a positive tax smaller than Ti at which welfare is higher than in the absence of interventions, although welfare is lower than with the optimum subsidy Si or with the optimum tax Ti applied to both imports from ROW and to consumption of home goods. The impact of the optimum intervention on per capita income and welfare can be seen in figure 1 (measured in currency units). Benefits from security Z1 Schiff and Winters 281 Figure 1. Equilibrium and Optimal Import Outcomes Price, P i\ > \ World supply, PA -Consumer price, P', Social demand, DA ________________r_____ _______ Consumer price, P' Social demand, DIB Private demand, DI 0 BP' Bs As Quantity imported Note. Subscripts A, B, and OPTdenote good A, good B, and the optimal value, respectively. Superscripts 1, 2, S, and Pdenote world values, country 1, countrv 2, social values, and private values, respectively. increase by area EGFK, income VI falls by area EFG, and welfare U1 rises by area EFK. Because the security externality is assumed to fall as imports increase, area EFK is larger than area EFG. The income loss for country 1 from the RIA can be approximated by S 2BB / 2P4, where cB is the elasticity of demand for good B in country 1. Thus, S2BIFB /2P4, the income loss measured as the loss from the RIA relative to a free trade situation, provides a lower bound of the welfare gain EFK that is generated by the optimum intervention. Symmetry If countries 1 and 2 are symmetric, and if the welfare function in goods, V, is symmetric in A and B, then B,1 = A2- M, SKA = SKB = SK, and UA = UB = V(M, M, R) + Z[SK(M)]. Then the optimum subsidy rate on partner imports (S) is the same for each country, and optimality can also be achieved through the forma- tion of a customs union between countries 1 and 2 with a common external tariff rate CET = S / (1 - S) and with a tax rate T = CET on consumption of the home good. In the symmetric case, trade between the partners is balanced and so is their trade with ROW. This is not the case under asymmetry. Asymmetry Now assume symmetry between countries 1 and 2, with the exception being that increasing security is more important to country 2 than to country 1. That 282 THE WORLD BANK ECONOMIC REVIEW, VOL. 12, NO. 2 is, for any SK1 = SK2, Z2 > Zl. This may be due to the fact that country 1 has a stronger tendency to resort to aggression to resolve disagreements or that coun- try 2 suffered more in previous wars than country 1. For instance, France may have been more concerned with security than Germany in the first half of the twentieth century, while the opposite may have been true in Napoleonic times. The equilibrium in the previous subsection is no longer sustainable because coun- try 2 now desires a larger subsidy: that is, S2 > S1 (see equation 5). Alternatively, countries 1 and 2 can maximize welfare by forming a free trade agreement with a tax rate Ti = Si / (1 - Si) on imports from ROW and on con- sumption of the home good and with T2 > T, (and rules of origin to prevent deflection of R from country 1 to country 2). Optimum tariffs on imports from ROW are higher for the country with the stronger security concerns. With S2 > Sl, B1 < A2. Thus, at the social optimum, country 1 runs a trade surplus with country 2 and a deficit with ROW (and vice versa for country 2). In other words, at the social optimum, the country with stronger security concerns imports more from its partner. This is shown in figure 1. Country 2's private demand curve, DA, is identical to DP. However, because security externalities are higher, the social demand curve for country 2 is D' > Ds, and the price function p2 (equal to the world price minus the optimum subsidy function S2) intersects the demand curve at point J. The subsidy S2= HJ > S, = FG, and B1 < A2. In both the symmetric and asymmetric cases, the welfare of both countries 1 and 2 would also be maximized if one of them applied the optimum subsidy on imports from the partner while the other applied the optimum tax on imports from ROW and on consumption of the home good. The formation of an RIA (customs union or free trade agreement) would require coordination to ensure that both countries 1 and 2 applied the tax. If we add to the model the assump- tion that policy coordination provides additional security benefits by improving trust and understanding among the leaders (and negotiators) of both countries (see the quote from Bastian 1996 in section I), then an RIA (accompanied by appropriate domestic taxes) would be superior to subsidies on imports from the partner, and welfare-maximizing governments in countries 1 and 2 would coor- dinate their policies. There are two additional reasons why coordinated formation of an RIA is likely to be superior. First, security may depend not only on the level of im- ports but also on the degree of certainty of access. The formation of an RIA requires policy coordination, including signing of an international treaty, and should therefore provide more security of access to the neighbor's goods and services than a situation in which each country determines policy entirely independently. Second, the model assumes that imports from-rather than trade with-the partner country generates security benefits in the home country. If one assumes that total trade (imports plus exports) with the partner generates security ben- efits, then-due to game-theoretic considerations- RIAs are likely to be superior Schiff and Winters 283 to intrabloc trade subsidies. Individual decisions in the home country generate externalities not only for other individuals of the home country but also for those of the partner country because the home country's imports are the partner's exports and vice versa. Thus intrabloc trade subsidies that maximize national welfare generate a Nash equilibrium, with welfare lower than in a cooperative equilibrium that could be obtained by forming an RIA. This section has shown that, in the presence of security externalities, the for- mation of an RIA is an optimal policy. We now examine the implications of the formation of such an optimum RIA for the evolution of external tariffs over time and the effects of deep integration and enlargement. The issue of whether coun- tries 1 and 2 are symmetric or asymmetric has no significant impact on results in the remainder of the article, so unless otherwise noted, symmetry is assumed. IV. TRADE PREFERENCES AND DEEP INTEGRATION It is often argued that the focus on border measures in regional integration may be misplaced if gains from deep integration, such as harmonization of tech- nical standards, investment codes, and legal principles and general facilitation of the movement of goods and factors, are substantial. Deep integration is ex- pected to lower trading costs and break down market segmentation between member countries. The most prominent recent example of deep integration is the European Community's single market program, which was predicted to have much larger economic effects than the mere removal of border policies-see, for example, Smith and Venables (1988), Gasoriek, Smith, and Venables (1992), and Harrison, Rutherford, and Tarr (1996). An unresolved question in the lit- erature is whether regional trade preferences are necessary for or supportive of deep integration on a regional basis. This section examines the relationship be- tween deep integration and trade preferences in the case of autonomous deep integration and in the case of endogenous deep integration. Autonomous Deep Integration Assume that trade is costly and that deep integration occurs autonomously and costlessly and is not related to present or past trade flows or policies. Ini- tially trade with ROW involves the same trading costs, C, as intrabloc trade, but following deep integration, intrabloc costs fall to C0 < C, while those with ROW remain unchanged at C. Because (by assumption, see equation 1) trade with ROW continues following deep integration, producer prices in countries 1 and 2 are not affected by deep integration, but consumer prices in the home country for goods imported from the partner (B1 in country 1 and A2 in country 2) fall by an amount C - CO. Consequently, consumption (imports) of B1 and A2 rises. Given that the marginal security value of imports falls as imports rise (equations 1 and 3), the optimum subsidy Si falls. Graphically, a reduction in trading costs shifts the PB line in figure 1 downward and thus moves both the intersection points E (of the PB line and the private demand curve, DI) and F (of the PB line 284 THE WORLD BANK ECONOMIC REVIEW, VOL. 12, NO. 2 and the social demand curve, Ds) to the right. Because the distance FG = S1 falls as B1 rises, the optimum subsidy falls. Equivalently, lower trading costs on intrabloc trade imply lower optimum taxes on trade with ROW and on consumption of home goods. Thus, in the case of exogenous deep integration within the region, trade preferences and deep integration are substitutes: lower intrabloc trading costs imply lower optimum trade preferences. Endogenous Deep Integration Alternatively, assume that lowering trade frictions on intrabloc trade is costly and requires a conscious policy decision. Assume also that the cost is negatively related to the level of trust or security capital SK. Because trust is positively related to imports from the partner, a larger volume of such imports reduces the cost of deep integration and thus is likely to result in more deep integration and in lower unit trading costs on trade with the partner. Haas (1958: 311) saw increased trade among members of the EC-6 (Belgium, France, Germany, Italy, Luxembourg, and Netherlands) positively affecting both international politics and deep integration: "It is inconceivable that the liberalization not only of trade but of the conditions governing trade can go on for long without 'harmoniza- tion of general economic policies' spilling over into the fields of currency and credit ... The spillover may make a political community of Europe." We examine two alternative scenarios. In the first, deep integration takes place every period. In the second, deep integration takes place once. Assume first that unit trading costs on imports from the partner are related to the current level of imports. In this case, two sources of positive externalities-rather than one-are associated with imports from the partner. Consumers who import from the part- ner generate both a security externality by raising utility directly (equation 1) and an indirect cost externality by lowering the unit cost of intrabloc trade. With a greater overall positive externality than in the previous cases without deep integration, the optimal subsidy or external tariff is higher. (See appendix A for a formal derivation of this result.) Alternatively, assume that deep integration takes place at a given point in time (for example, in 1992 for the EC-92) and is permanent. For instance, as- sume that trust and security, SKi, depend not only on the flow of current intrabloc trade but also on past flows (see section V) and that deep integration will take place when SKi is sufficiently large and therefore the cost of deep integration is sufficiently low. The optimum subsidy now changes through time. Because in- creased intrabloc trade initially generates externalities by lowering the cost of future deep integration, the optimum subsidy on intrabloc trade (or optimum CET) is higher before deep integration than it would be in its absence. However, once deep integration has taken place and intrabloc trading costs have fallen, imports are naturally higher, and the optimum subsidy (or optimum CET) is lower than in the absence of deep integration. (The results can be derived formally in a dynamic optimal-control-type framework such as in appendix B.) Schiff and Winters 285 To summarize, if the cost of deep integration is not related to the degree of trust between countries 1 and 2 (our autonomous case), not only are trade pref- erences not required in order to implement deep integration measures, but deep integration implies a reduction in optimum trade preferences. However, if deep integration needs to be replenished every period and its cost declines with the degree of security and trust, optimum trade preferences are larger. Finally, if deep integration takes place once and for all and is permanent, optimum trade preferences are higher before deep integration takes place and lower thereafter. V. DYNAMIC ADJUSTMENTS TO A STEADY STATE The European Community has reduced its external trade barriers on manu- facturing products with respect to ROW over time. The European Community's external trade barriers have fallen over time even as the number of member countries has grown from 6 to 15, that is, despite the increase in the optimum CET level based on market power considerations. Average tariffs on manufactur- ing products have fallen from about 13 percent in 1958 to about 3 percent after the Uruguay Round. Similarly, developing countries forming RIAs in the 1960s imposed high external trade barriers while recent RIAs and new incarnations of old RIAs have tended to impose lower external trade barriers. Is this gradual reduction of external trade barriers over time predicted by our model of regional integration? In our main model, security capital (SK) is assumed to depend on the current flow of imports from the partner and to depreciate fully at the end of each period. However, security could depend on past behavior as well, that is, it would not depreciate entirely at the end of every period. We can think of this as the level of SK, at time t depending positively on both the current flow of im- ports from the partner and on SK -1 at t - 1, with a rate of depreciation of SK equal to 8 < 1 per period. We provide here an intuitive solution to this problem; appendix B provides a formal derivation. This case is similar to that of endogenous deep integration except that instead of a single-step change in trading costs, and hence welfare, we postulate a con- tinuous relationship between accumulated trade and well-being. Two forces are at work, each leading the optimum subsidy (CET) to fall through time. The opti- mum subsidy (CET) would fall even if the government were only concerned with maximizing current (instantaneous) welfare. But it would fall even faster if the government were concerned with maximizing the present value of (instanta- neous) welfare. In the case of current welfare maximization, starting from a steady state un- der free trade, home governments would subsidize partner imports according to equation 6 above. This would raise intrabloc imports and thereby the level of security capital. In the next period, some of this security capital would persist and so, given the diminishing marginal products of intrabloc imports on security and of security on welfare, the optimum subsidy would fall. In other words, as 286 THE WORLD BANK ECONOMIC REVIEW, VOL. 12, NO. 2 the level of security rises over time, some of the benefits of security are taken in the form of an increase in utility V from a more efficient allocation of expendi- ture on consumption. This tendency to reduce the subsidy (or the CET) would continue until, in the new steady state, the new security capital generated each period just balances the depreciation of the old. In the case of maximization of the present value of welfare, the dynamic effect reinforces the effect just discussed. In any period prior to attaining steady state, a subsidy on intrabloc imports not only increases current welfare via increased security but also increases future welfare by creating persistent security capital. Thus a rational government will subsidize imports more than is called for by current welfare-that is, it will have a subsidy (CET) exceeding that based on current considerations alone. This dynamic effect does not affect the steady- state level of the CET but does cause it to be approached more rapidly-the opti- mal dynamic CET would start higher and fall more rapidly than a CET based only on current welfare considerations. Of course, new shocks could disturb these dynamic paths. For example, there may be times of crisis or reversal in which SK suddenly falls. Then, the optimum CET would suddenly increase before starting to fall gradually again. If the crisis is temporary, with no impact on behavior, the CET eventually returns to its pre- vious steady-state value. However, a deep crisis may affect security preferences or the Z-function in equation 1 as well as the security production function in equation 2, with a permanent change in optimum external barriers. VI. ENLARGEMENT AND DOMINO EFFECTS Here we examine the impact of bloc enlargement on welfare and on the opti- MUm CET. Assume that country 3 produces good C, exports it to countries 1, 2, and ROW, and imports goods A, B, and R. Symmetric Case First, assume that country 3 is symmetric with countries 1 and 2. Each coun- try is assumed to enjoy the same degree of trust (or to suffer the same degree of mistrust) with respect to the other two. Utility for the representative consumer in country i is now (1') Ui = V(Ai, Bi, Ci, Ri) + Zi (SKi); i =12,3 v;, Zi' > o, Vjj, Ztt< o, Vj (j = o) =°,Vj av I Di; j = Ai, Bi, Ci, Ri- The level of security in the home country is positively related to imports from both partner countries, with SK symmetric in both imports. That is, (2') SK1 = SK(B1,C1); SK2= SK(A2,C2); SK3 = SK(A3,B3). Schiff and Winters 287 Assume that the three countries impose optimum subsidies on trade with one another. Under symmetry, each country imposes the same subsidy rate S as the other two and imposes the same subsidy on its imports from both partners. The same result is obtained with a common external tariff rate CET = S / (1 - S) and a tax rate T = CET on the consumption of the home good. First, each member country's welfare rises when the RIA expands from two to three countries. In the two-country RIA case, the optimum subsidy is applied on the imports of one country only, while security benefits are obtained through imports from two countries. This case is equivalent to a constrained optimiza- tion in which one of the two subsidies is set (nonoptimally) equal to zero. This results in a level of welfare that is necessarily lower than under unconstrained optimization in the case of the three-country RIA, because in the latter case all externalities associated with the impact of imports on security are internalized. Consequently, starting from a two-member country RIA (say, countries 1 and 2), there is both a demand (by country 3) and a supply (by the RIA members) for enlargement. In this sense, there is a domino effect: if a two-country RIA does exist, the third country will join. However, if no two-country RIA exists, the three countries have an incentive to form a three-country RIA simultaneously. See Baldwin (1995) on domino effects. We now turn to the impact of enlargement on the optimum CET. Define the optimum subsidy S and optimum CET for a bloc with k members by Sk and CETk, respectively. What is the relationship between S2 and S3 (or CET2 and CEI5)? We show in appendix C that Sk need not increase with enlargement, although S2 < S3 iS likely to hold on average. This result generalizes to St - I < Sm for a bloc expansion from m - 1 to m member countries in a world of m symmet- ric countries (m 2 3), but not to Sk- 1 < Sk for a bloc of size k (k < m) in a world of m symmetric countries. In other words, the result generalizes to any number of symmetric countries as long as the enlargement results in a bloc that includes all symmetric countries. In a world of symmetric blocs with the optimum CET determined by market power, the CET increases with enlargement. In our model, the optimum CET need not increase with enlargement, although it is likely to do so on average. Asymmetric Case We now examine enlargement under asymmetry. Even though the main secu- rity concern in the 1950s in Western Europe was with France and Germany, four countries (Belgium, Netherlands, Luxembourg, and Italy) joined France and Germany to form the European Economic Community. These four coun- tries were smaller economically and weaker militarily. Similarly, Argentina and Brazil decided in the 1980s to integrate their economies, and Uruguay and Para- guay-two small neighbors-decided to join them and form Mercosur. And Chile and Bolivia recently signed free trade agreements with Mercosur. (On the im- pact on Chile of free trade agreements with Mercosur and NAFTA, see Harrison, Rutherford, and Tarr 1997 and Sapelli 1996.) 288 THE WORLD BANK ECONOMIC REVIEW, VOL. 12, NO. 2 Can our model predict anything about domino effects in such asymmetric situations? Imagine that countries 1 and 2 are symmetric and are large com- pared to country 3 (although they are still small economically compared to ROW). Assume that, given the small size of country 3, security in countries 1 and 2 depends on the level of trade between them and not on the level of trade with country 3. Countries 1 and 2 have an incentive to form an RIA. The RIA reduces the level of security in country 3 because countries 1 and 2 export less to it. Thus country 3 has a stronger incentive to join in a three-country RIA than it did to join with either country 1 or 2 before they formed a two-country RIA. However, countries 1 and 2 have no incentive to allow country 3 to join the RIA; consequently country 3 may have to offer some compensation to coun- tries 1 and 2 in order to be allowed to join. If the gains to country 3 from joining are larger than the losses (if any) to countries 1 and 2 (say, because they now trade more with country 3 and less among themselves), country 3 will compensate them and join the RIA.7 Whalley and Perroni (1994) note that compensation in the form of market access has become quite common recently, with large countries making significantly fewer market access concessions to smaller countries than vice versa. VII. CONCLUSIONS This article has examined the logical implications of positive security exter- nalities that may result from trade between suspicious neighbors. Adopting the common assumption that security with a neighbor increases as imports from that neighbor increase, and making no assumptions about the relative sizes of trade diversion and trade creation, we have shown that: e The formation of a customs union, accompanied by appropriate domestic taxes, provides an optimum economic arrangement under symmetry and that the same holds for a free trade agreement under asymmetry. i Deep integration (such as the EC-92) implies lower optimum external tariffs if it is exogenous, higher optimum external tariffs if it is endogenous and vanishes annually, and higher optimum external tariffs before deep integration but lower ones thereafter if deep integration is endogenous, takes place once and for all, and is permanent. * If the level of security depends on past as well as current trade flows and is in steady state in the absence of trade barriers, then external tariffs result in an increase in security over time and the optimum external tariffs decline over time. 7. In some cases, the gains for country 3 may be very large. For instance, Paraguay suffered a devastating defeat in the 1860s in its war (known in Spanish as "La Guerra de la Triple Alianza") with a coalition made up of Argentina, Brazil, and Uruguay. Most of its male population over 15 years of age perished in the conflict. Such a collective memory may provide a strong incentive to avoid potential conflicts in the future. Schiff and Winters 289 Enlargement of bloc size (in terms of the number of symmetric countries) implies higher welfare with an ambiguous impact on the optimum CET, although it is likely to be higher; and a form of domino effect exists. Although externalities associated with security imply that RIAs may maximize welfare, our model suggests that the RIA is a transitory arrangement in the sense that optimum trade preferences are highest when the RIA is formed (when secu- rity is low) and decline over time. In other words, the RIA'S optimal external trade policy becomes increasingly open over time. VIII. Two EXTENSIONS First, our model abstracts from defense expenditures as an alternative way to generate security. This was done for the sake of simplicity and to avoid game- theoretic issues at this stage. We recognize the importance of such issues, which are on our research agenda. In the case of defense expenditures, the cooperative solution is clearly optimal because the alternative may be a prisoner's dilemma situation with large defense expenditures, possibly without a higher level of se- curity. In this case, the "peace dividend" following formation of an RIA may be substantial. (See Srinivasan 1994 for a discussion of these issues in the case of India and Pakistan.) Second, our model assumes that imports from-rather than trade with-the partner country generates security benefits in the home country. Alternatively, one could assume that total trade (imports plus exports) with the partner gener- ates security benefits. In that case, RIAs are likely to be superior to optimum intrabloc trade subsidies. There are both game-theoretic and practical reasons for this. First, each country will use a subsidy to maximize national welfare. However, individual decisions in the home country generate externalities not only for other individuals in the home country but also for individuals in the partner country because the home country's imports are the partner's exports and vice versa. Thus intrabloc trade subsidies that maximize national welfare generate a Nash equilibrium, with welfare and Nash subsidies lower than in a cooperative equilibrium that could be obtained by forming an RIA. Second, the Nash equilibrium implies subsidies on both imports and exports. Export subsi- dies are illegal for members of the General Agreement on Tariffs and Trade so that even the inferior Nash solution cannot be implemented. The superior coop- erative solution implicit in the formation of an RIA (with appropriate domestic taxes) avoids this problem. APPENDIX A. ENDOGENOUS DEEP INTEGRATION Consider the case in which deep integration is related to current imports and has to be replenished every period. Assume that unit trading costs C on imports from the partner are 290 THE WORLD BANK ECONOMIC REVIEW, VOL. 12, NO. 2 (A-1) C = C[SK(M)],aC / aM = (aC / aSK)SK' < 0 where M is the volume of imports from the partner (and M = B1 = A2 under symmetry). There are two sources of positive externalities: a direct effect (equa- tion 1) and an indirect effect deriving from lower unit costs on intrabloc trade. On the latter, the total trading cost, TC, on imports, M, from the partner and the marginal social cost of trading, MSC, are (A-2) TC = C[SK(M)]M (A-3) MSC = C(1 + ) where E - (aC / aM)(M / C) < 0. The representative consumer takes the private cost, C, of importing from the partner into account but not the (negative) effect on unit costs, EC. This has to be added to the optimum subsidy, SE, such that (A4) SE = Z'(dSK / aM) /l x- EC > S (since Ł < 0) with S defined in equation 5. Equivalently, the optimum can be attained with a common external tariff, CETE = SE / (1 - SE) > CET = S / (1 - S) and a tax on home consumption TE = CETE > T = CET. APPENDIX B. DYNAMIC ADJUSTMENTS Assume that the level of SK, at time t depends on the current flow of imports from the partner (M) and on SK, - 1 at t - 1, with a rate of depreciation of SK equal to 5 < 1. Expressing variables in continuous time, equation 2 becomes, under symmetry, the law of motion (equation B-1): (B-i) SK = F(M) - 8SK; 8 < I where sk - aSK / at is the time derivative of SK, F(M) is the current gross addi- tion to the stock of security capital (with F'> 0), and the subscript t on SK and M has been deleted to simplify notation. Start at steady state under free trade, with Sk = 0. With a subsidy on intrabloc imports, SK accumulates, and its marginal utility Z' falls (equation 1). If govern- ments are myopic and choose S to maximize current welfare, W= LU, in each period, then from equation 5, S falls over time. However, optimizing govern- ments also take into account the impact of current subsidies on the level of SK and thus on future welfare. Assume that the governments of countries 1 and 2 select an optimum time path for the subsidy S on imports from the partner in order to maximize 4', the present value of W= LU. The Hamiltonian is (B-2) H = W(SK, S, t) + q SK(SK, S, t) Schiff and Winters 291 where q is the marginal value of SK, DOMAX / aSK, where DMAX is the maximum value of (D obtained with the optimum path for S. The solution (from the maxi- mum principle) is (B-3) aH I aq =SK = F(M) - iSK (BA4) aH / S= aW aS + q(aSKl DS) = 0 and (B-5) aHI aSK= aW aSK +q(aSklaSK) = -4. From equation B-3, aSK / aS = (aSK / DM) (WM / DS) > 0. Thus, from equation B-4, aW / as < 0-that is, along the optimal path (which maximizes the present value of welfare), the subsidy exceeds the value necessary to maximize current welfare. If a government myopically maximizes current welfare without regard for future welfare, it chooses a subsidy level Sw so that aW/ aSw = 0. However, an increase in the subsidy also generates future welfare gains because it raises M and thus raises SK. Thus the subsidy Sq, which maximizes (, is larger than the myopic SW, which maximizes W. Consequently, a W / aS < 0 at the optimum. Assume that SK is low (say, following a series of wars, such as in the 1950s for France and Germany) so that SK rises with imports M between the two partners (equation B-3) and the value of SK falls, (4< 0). From equation B-4, a reduction in q over time implies that aW/ aS falls over time in absolute value or becomes less negative because ask / as falls with SK as well in absolute value.8 In other words, along its optimum path, SD falls over time and approaches Sw. Sw also falls as SK rises because the marginal impact of SK on W falls with SK (Z" < 0). Equivalently, the optimum common external tariff (recall we are as- suming symmetry) falls over time. The decline in the CET will stop once a steady state is attained, that is, once SK = 0 (equation B-1). APPENDIX C. ENLARGEMENT We examine here the impact of enlargement on the optimum subsidy or CET. Equation 1' in section VI can be rewritten as (C-1) U = V(A, B, C, R) + Z[SK(B, C)] where the subscript i has been deleted for simplicity. Without loss of generality, we examine the issue from the viewpoint of country 1. Define Sk as the optimum subsidy in a bloc of k countries (k = 2, 3). From equation C-1, we have (C-2) dU/ dS = [(aV/ aA) (dA / dS) + (aVI /B) (dB / dS) + (DV / DC) (dC / dS) + (aV / DR) (dR I dS)] + Z' [(aSK / JB) (dB / dS) + (aSK / DC) (dC / dS)]. 8. asK / as = (asK / aSK) (aSK / DM) (aM / DS) = -5(aSK / DM) (aM / DS). The depreciation rate 6 is constant, aM / aS is independent of SK since U is additive, and aSK / aM _ SK' falls with M, so that aSK / as falls with SK in absolute value. 292 THE WORLD BANK ECONOMIC REVIEW, VOL 12, NO. 2 Because income is measured at world prices, Y1 = A = aL1 does not change with the subsidy, and because all world prices are unity, balanced trade implies that dA + dB + dC + dR = 0. Assume that we start with S = S3 and apply it to imports of both B and C. Because S3 is an optimum, dU / dS = 0. Given symme- try, B = C > A= R, aV I aA = aVI aR > aVI DB = DV/ DC, dB / dS = dC / dS = -dA / dS = -dR / dS > 0, and aSK I aB = aSK I ac. Substituting these in equation C-2, we have (C-3) dU / dS = 2(dB / dS)[(aV/ aB - avi DA) + Z'(aSK I aB)] = 0. Alternatively, assume that S = S3 applies only to imports of B and that the subsidy on imports of C is set equal to zero. Such a configuration of subsidies is not an optimum, so dU / dS  O. In this case, A = C R < B, dA / dS = dC /dS = dR / dS <0, dB / dS = -dA / dS - dC / dS - dR / dS =-3(dA / dS), and DV/ DA = aV / aC = aV / aR. Substituting in equation C-2, we have (C-4) dU/ dS =(dB dS) [(aVl aB - aV aDA) + Z' (aSK I aB)] -(Z'13) (aSKIaC). The objective is to evaluate dU I dS in equation C-4. If the sign were negative (positive), then S would have to be reduced (increased) to reach the optimum subsidy, S2, in the two-country RIA case, that is, S2 < S3 (S2 > S3). Note the similarities between equations C-3 and C-4. The term F(A, B, C, R) [(aV I aB - aV / DA) + Z' (aSK / aB)] = 0 in equation C-3. If both equations were evaluated at the same values for A, B, C, and R, then dU / dS in equation C-4 would be dU / dS = F(A, B, C, R) - (Z' /3) (aSK / DC) = - (Z' / 3) (aSK / ac) < 0, and thus S2 < S3. However, equations C-3 and C-4 are not evaluated at the same values for A, B, C, and R, so r(A, B, C, R) is not necessarily equal to zero in equation C-4. We now show that the value of F(A, B, C, R) _ (aVl aB - aV/ DA) + Z' (DSK I aB) can be either positive or negative when evaluated at the quantities prevailing in equation C-4. The difference between equations C-3 and C-4 is that the sub- sidy at the rate S3 in equation C-3 is set equal to zero on good C in equation C-4. Thus, B and A are both larger in equation C-4, aVI aB and av/ aA are both smaller, and the impact on (aV / aB - aV / DA) is ambiguous. The impact on Z' (DSK I aB) is ambiguous as well. At the larger value of B in equation C-4, aSK / dB is likely to be smaller, but because SK is also likely to be smaller, Z' is larger, and the impact on Z' (DSK I DB) is ambiguous.9 In other words, the sign 9. Comparing equilibria in equations C-3 and C-4, a second-degree Taylor expansion of the difference in SK is ASK = [AB(aSK / aB) + AC(dSK / aQC + [AB2(d2SK / dB2) + AC2(d2SK I aC2)] /2 + [ABAC(d2SK / DBDC)]. What is the sign of ASK? The first term in square brackets is negative. To show this, note that B = C in equation C-3 so that aSK / aB = aSK / aC, but AB < -AC. (The increase in B is smaller than the decline in C because removing the subsidy on C also leads to a rise in consumption of A and R, and from the budget constraint, -AC = AA + AB + AR > AB.) The second term in square brackets is negative Schiff and Winters 293 of F(A, B, C, R) in equation C-4 is ambiguous. For some functional forms of V, Z, and SK, [(A, B, C, R) will be negative, while for others, it will be positive or zero. One would need to know the exact functional forms of V, Z, and SK in order to determine the sign of [(A, B, C, R) in equation C-4. However, under symme- try of the positive and negative values for F(A, B, C, R) across all functional forms for V, Z, and SK, one may expect its average value to be zero. (Strictly speaking, whether the average value of [(A, B, C, D) in equation C-4 is zero or not depends on the definition of the domain of the functions V, Z, and SK.) And because -(Z' / 3) (dSK / AC) < 0, one may expect the average value of d U / dS to be negative. This implies that, on average, S2 < SI. It is easy to show that the result generalizes to Sm -1 < Sm for a bloc expansion from m - 1 to m member countries in a world of m symmetric countries (m > 3), but not, unfortunately, that it generalizes to Sk - 1 < SI for a customs union of size k (k < m) in a world of m symmetric countries. REFERENCES The word "processed" describes informally reproduced works that may not be com- monly available through library systems. Arad, Ruth, and Arye L. Hillman. 1979. "Embargo Threat, Learning, and Departure from Comparative Advantage." Journal of International Economics 9(May):265- 75. Baldwin, Richard E. 1995. "A Domino Theory of Regionalism." In R. E. Baldwin, Pertti Haaparanta, and Jaakko Kiander, eds., Expanding Membership in the European Union. Cambridge, U.K.: Cambridge University Press. Bastian, Jens. 1996. "Trade Diplomacy and Regional Integration." Report on a World Bank workshop, July, London School of Economics. Processed. Becker, Gary S. 1996. The Making of Preferences and Values. Cambridge, Mass.: Harvard University Press. Bhagwati, Jagdish. 1987. "The Generalized Theory of Distortions and Welfare." In Jagdish N. Bhagwati, ed., International Trade: Selected Readings, 2d ed. Cambridge, Mass.: MIT Press. Bhagwati, Jagdish, and Arvind Panagariya. 1996. "Preferential Trading Areas and Multilateralism: Strangers, Friends, or Foes." In Jagdish Bhagwati and Arvind Panagariya, eds., The Economics of Preferential Trade Agreements. Washington, D.C.: American Enterprise Institute Press. because the second derivatives are negative. (Intuitively, as C falls, aSK / aC rises so that the loss in SK from additional reductions in C increases; similarly, the gains in SK from increases in B decline with B.) However, a2SK / aBDC may be negative-if country 1 is less concerned about security with respect to country 2 (equation 3), if security with respect to country 3 (equation 2) is higher-and the third term in square brackets is then positive. Thus, although it may seem likely that removing the subsidy on imports of good C results in ASK < 0, this need not be the case. However, no matter what the sign of ASK, the fact remains that the difference in Z' (aSK I aB) is ambiguous. The reason is that with an ambiguous change in SK, the changes in both Z' and in aSK / aB are ambiguous. 294 THE WORLD BANK ECONOMIC REVIEW, VOL. 12, NO. 2 Bliss, Christopher. 1994. "Trade Blocks and Migration." Economic Theory and Policy for Trading Blocks. Manchester, U.K.: Manchester University Press. Chan, Steve. 1984. "Mirror, Mirror on the Wall . . . Are the Freer Countries More Pacific?" Journal of Conflict Resolution 28:617-48. Ethier, Wilfred. 1996. "Multilateral Roads to Regionalism." Department of Economics, University of Pennsylvania, Philadelphia. Processed. Gasiorek, Michael, Alasdair Smith, and Anthony J. Venables. 1992. "1992: Trade and Welfare: A General Equilibrium Model." In L. Alan Winters, ed., Trade Flows and Trade Policy after 1992. London: Cambridge University Press. Gasiorowski, Mark, and Solomon Polachek. 1982. "Conflict and Interdependence: East- West Trade Linkages in the Era of Detente." Journal of Conflict Resolution 26:709-29. Haas, Ernst. 1958. The Uniting of Europe. London: Stevens and Sons. Harrison, Glenn W., Thomas F. Rutherford, and David Tarr. 1996. "Increased Compe- tition and Completion of the Market in the European Union: Static and Steady-State Effects." Journal of Economic Integration 11(3):332-65. . 1997. "Trade Policy Options for Chile: A Quantitative Evaluation." Policy Re- search Working Paper 1783. Policy Research Department, World Bank, Washington, D.C. Processed. Hillman, Arye L., and Ngo van Long. 1983. "Pricing and Depletion of an Exhaustible Resource When There Is Anticipation of Trade Disruption." Quarterly Journal of Economics 98(May):215-33. Hirschman, Albert 0. 1981. Essays in Trespassing: Economics to Politics and Beyond. Cambridge, U.K.: Cambridge University Press. . 1982. "Rival Interpretations of Market Society: Civilizing, Destructive, or Feeble?" Journal of Economic Literature 20:1463-84. Hull, Cordell. 1948. The Memoirs of Cordell Hull. New York: Macmillan. Jones, Ronald. 1993. "Discussion." InJaime de Melo and Arvind Panagariya, eds., New Dimensions in Regional Integration. Cambridge, U.K.: Cambridge University Press. Kant, Immanuel. 1992. Perpetual Peace: A Philosophical Essay. Bristol: Thoemmes Press. Keynes, John Maynard. 1919. The Economic Consequences of the Peace. London: Macmillan. Krugman, Paul R. 1991. "Is Bilateralism Bad?" In Elhanan Helpman and Assaf Razin, eds., International Trade and Trade Policy. Cambridge, Mass.: MIT Press. Machlup, Fritz. 1977. A History of Thought on Economic Integration. London: Macmillan. Mansfield, Edward D. 1993. "Effects of International Politics on Regionalism in Inter- national Trade." In Kym Anderson and Richard Blackhurst, eds., Regional Integra- tion and the Global Trading System. New York: St. Martin's Press. Mansfield, Edward D., and Rachel Bronson. 1997. "Alliances, Preferential Trading Arrangements, and International Trade." American Political Science Review 91(1, March): 94-107. Page, Sheila. 1996. "Intensity Measures for Regional Groups." Paper prepared for EADI ninth general conference, September, Vienna. Processed. Polachek, Solomon W. 1992. "Conflict and Trade: An Economics Approach to Political Interactions." In Walter Isard and Charles H. Anderton, eds., Economics of Arms Reduction and the Peace Process. Amsterdam: North-Holland. Schiff and Winters 295 1996. "Why Democracies Cooperate More and Fight Less: The Relationship between International Trade and Cooperation." Review of International Economics. Puga, Diego, and Anthony J. Venables. 1997. "Preferential Trading Arrangements and Industrial Location." Journal of International Economics 43:347-68. Rieben, Henri, Jean Monnet, and G. A. Chevallaz. 1971. Jean Monnet. Lausanne: Cen- tre de Recherches Europeennes. Sapelli, Claudio, ed., 1996. Chile en el NAFTA: Acuerdos de libre comercio versus liberalisaci6n unilateral. Santiago: International Center for Economic Growth. Schiff, Maurice. 1992. "Social Capital, Labor Mobility, and Welfare: The Impact of Uniting States." Rationality and Society 4(2, April):157-75. . 1995. "Capital socio-culturel, diversite ethnique et reforme en Afrique." Revue d'Economie du Developpement 3:41-53. . 1996. "Social Capital, Trade, and Optimal Migration Policy." International Economics Department, World Bank, Washington, D.C. Processed. . 1997. "Small Is Beautiful: Preferential Trade Agreements and the Impact of Country Size, Market Share, and Smuggling." Journal of Economic Integration 12(3):359-87. .1998. "Ethnic Diversity and Economic Reform in Sub-Saharan Africa." Journal of African Economics 17(3, May). Forthcoming. Skidelsky, Robert. 1992. John Maynard Keynes: The Economist as Saviour 1920-1937. London: Macmillan. Smith, Alasdair, and Anthony Venables. 1988. "Completing the Internal Market in the European Community: Some Industry Simulations." Discussion Paper Series 233. Centre for Economic Policy Research, London. Processed. Srinivasan, T. N. 1994. "Regional Trading Arrangements and Beyond. Exploring Some Options for South Asia: Theory, Empirics, and Policy." Report IDP-142. South Asia Office of the Chief Economist, World Bank, Washington, D.C. Processed. Swann, Dennis. 1992. The Economics of the Common Market, 7th ed. Harmondsworth: Penguin. Whalley, John, and Carlo Perroni. 1994. "The New Regionalism: Trade Liberalization or Insurance?" NBER Working Paper 4626. National Bureau of Economic Research, Harvard University, Cambridge, Mass. Processed. Winters, L. Alan. 1989. "The 'So-Called Non-Economic Objectives' of Agricultural Policy." OECD Economic Studies 13:237-66. -. 1991. "Digging for Victory: Agricultural Policy and National Security." The World Economy 13:170-90. . 1997. "What Can European Experience Teach Developing Countries about In- tegration?" The World Economy 20:889-912. THE WORLD BANK ECONOMIC REVIEW, VOL. 12, NO. 2: 297-320 Endogenous Tariff Formation: The Case of Mercosur Marcelo Olarreaga and Isidro Soloaga Mercosur appears as an interesting case study for analyzing the determinants of excep- tions in regional trade agreements. Its member countries-Argentina, Brazil, Para- guay, and Uruguay-intended to make Mercosur a full customs union by January 1995. This goal turned out to be too ambitious, and the Protocol of Ouro Preto and other agreements signed in December 1994 led to a hybrid solution. Overall, out of a total of 9,119 tariff lines, around 30 percent are subject, in at least one member coun- try, to either external deviations from the common external tariff or internal devia- tions from free trade. Thus an important set of holes remains under the existing agree- ment, leading some authors to consider Mercosur an incomplete customs union. This article compares the results of the theoretical literature on endogenous tariff formation with evidence from Mercosur. The results show that Mercosur's common external tariff and member countries' deviations from it and from internal free trade can be explained by sector or industry lobbying as predicted by the endogenous tariff literature. If a viable political economy is a key to success, then Mercosur is here to stay. Several analysts argue that regional trade agreements (RTAS) allow for more far-reaching liberalization, in a regional context, than multilateral trade agree- ments. Some suggest that the North American Free Trade Agreement or the European Community would not have achieved their degree of intrabloc liber- alization if member countries had relied on multilateral negotiations because smaller groups find it easier to agree. Although it seems theoretically valid, Hoekman and Leidy (1992) challenge this view on empirical grounds. They argue that the same political forces that block far-reaching liberalization in the multilateral context are also present in regional negotiations and that the out- comes are relatively similar. Mercosur appears as an interesting case study for analyzing the determinants of exceptions ("holes" in Hoekman and Leidy's terminology) in RTAS. Mercosur's Marcelo Olarreaga is with the Economic Research Division at the World Trade Organization, Geneva, and the Centre for Economic Policy Research in London. Isidro Soloaga is with the Agricultural and Economic Resources Department at the University of Maryland and the Development Economics Research Group at the World Bank. The authors thank Olivier Cadot, Jaime de Melo, Isidoro Hodara, Maurice Schiff, Alan Winters, and three anonymous referees for thorough discussions and suggestions. They are also grateful to Sanoussi Bilal, Gary Pursell, and David Tarr for helpful comments on an earlier version. This article was produced as part of the World Bank Development Economics Research Group's research program on regionalism and development. C 1998 The International Bank for Reconstruction and Development/ THE WORLD BANK 297 298 THE WORLD BANK ECONOMIC REVIEW, VOL. 12, NO. 2 member countries-Argentina, Brazil, Paraguay, and Uruguay-intended to make Mercosur a full customs union (a common market as specified in the Treaty of Asunci6n) by January 1995. This goal turned out to be too ambitious, and the Ouro Preto protocol and other agreements signed in December 1994 led to a hybrid solution. A common external tariff (CET) was established, but countries were allowed to deviate (upward and downward) in some tariff lines until the beginning of the next century. Intra-Mercosur free trade has not been achieved. Overall, out of a total of 9,119 tariff lines, around 30 percent are subject, in at least one member country, to either external deviations from the CET or internal deviations from free trade. Thus an important set of holes remains under the existing agreement, leading some authors to consider Mercosur an incomplete customs union. This article compares the results of the theoretical literature on endogenous tariff formation with evidence from Mercosur. Does Mercosur's CET reflect sec- tor or industry lobbying? Are member-country deviations from the CET and from internal free trade consistent with the literature's predictions? The answers to these questions may provide some insight about the potential durability of the Mercosur agreements. Tariffs and trade agreements are subject to change, but if today's Mercosur structure of protection and its expected evolution to the CET (and internal free trade) reflect the interests of the private sector and average voters, then Mercosur will face fewer political challenges and may be politically viable in the long run. To our knowledge, this article is the first attempt to explain empirically de- viations from the CET and from free trade within an RTA. It also analyzes the structure of the CET in light of recent findings in the theoretical literature on the endogenous formation of trade policy. It emphasizes these new theoretical re- sults in the empirical sections. Section I describes Mercosur's tariff structure and deviations from the final objec- tives of internal free trade and a CET. Section II discusses the theoretical literature's predictions. Section III sets out the empirical model to be tested, and section TV presents the empirical results. Section V offers some concluding remarks. 1. SOME FACTS ABOUT MERCOSUR The Treaty of Asunci6n, signed in March 1991, established a framework for achieving a common market among Argentina, Brazil, Paraguay, and Uruguay, known as Mercosur. The region represents 75 percent of the gross domestic product (GDP) of South America (excluding Guyana), 60 percent of its popula- tion, and 65 percent of its geographic area. This makes Mercosur geographically the largest customs union in the world. Mercosur's geographic area is almost four times larger than that of the European Union. Its GDP per capita in 1994 was around $3,800. The relatively small size of member countries' national markets (Paraguay's GDP is smaller than that of the canton of Geneva) was obviously an important Olarreaga and Soloaga 299 factor calling for regional integration. Mercosur's internal market remains rela- tively small (around half of France's GDP). Its member governments recognize the necessity of employing Mercosur's market as a tool to integrate their econo- mies into the world economy. The governments have made an official commit- ment to further trade liberalization. Mercosur signed free trade agreements with Bolivia in December 1996 and Chile in October 1996. Bolivia cannot enter Mercosur as a full member because it has been a member of the Andean Pact (recently renamed the Andean Com- munity) since its beginnings. Chile, which has a uniform tariff structure, was not interested in adapting to Mercosur's nonuniform CET structure. In December 1995 Mercosur signed a Framework Cooperation Agreement with the Euro- pean Community, which, among other things, is intended to lead toward recip- rocal liberalization of all mutual trade. Apparently, even China and Russia have approached Mercosur authorities to engage in bilateral trade agreements. Since the agreement was signed, intraregional trade has increased at an aver- age rate of 28.5 percent a year. The increase represents three times the rate of growth of the region's total trade and five times the rate of growth of world trade (Mercosur Secretariat 1996). Table 1 gives the World Bank's indicator for Mercosur's rate of integration with the world, Mercosur, and the rest of the world. This indicator is calculated as the difference between the rate of growth of total trade and the rate of growth of GDP. Mercosur's GDP grew at an average rate of 1 percent in the 1980s and 3.8 percent in 1991-95. For Mercosur's member countries, the rates of integration with the world and with Mercosur were 10 times larger in 1991-95 than in the 1980s. Thus Mercosur has been a success in terms of volume of trade. The Treaty of Asunci6n established nontrade-related objectives including co- ordination of macroeconomic and investment policies and free factor mobility. Member governments also see as important objectives cooperation in education and transport policies and consolidation of the democratic process in the region. Moreover, as suggested by Schiff and Winters (1997), trade within Mercosur may reduce regional tensions in the framework of the nonaggression pact signed by Argentina and Brazil. From the last column of table 1, in absolute terms, the rate of regional integration has increased twice as much as the rate of integra- tion with the world. This may be explained by the fact that Mercosur members are geographically natural trading partners, that is, close neighbors. Yeats (1997) Table 1. Mercosur's Rate of Integration, 1980-90 and 1991-95 Region 1980-90 1991-95 Change World 1.1 13.1 12.0 Mercosur 2.4 24.7 22.3 Rest of the world 0.9 11.0 10.1 Note: The Mercosur countries are Argentina, Brazil, Paraguay, and Uruguay. The rate of integration is estimated as the difference between total trade and GDP growth. Source: Mercosur Secretariat (1996) and Camara de Industrias del Uruguay (1996). 300 THE WORLD BANK ECONOMIC REVIEW, VOL. 12, NO. 2 provides an alternative explanation based on internal and external tariff discrimination. The Protocol of Ouro Preto and other agreements signed in December 1994 quantified objectives and completed procedures for establishing the customs union aspects of the proposed common market. The governments negotiated a CET for all products but implemented it for only 75 percent of the total tariff lines in 1995. The remaining 25 percent includes mainly capital goods, computer prod- ucts and telecommunications equipment, automobiles, and sugar, for which spe- cial regimes have been negotiated. Convergence to the CET should be achieved at the latest by the year 2006 for telecommunications equipment. For sugar and automobiles, there has been no agreement on dates for final convergence. A list of other goods-a maximum of 300 tariff lines from a universe of 9,119 (3.3 percent) for Argentina, Brazil, and Uruguay and a maximum of 399 (4.3 per- cent) for Paraguay-should also converge to the negotiated CET by the year 2000. Total deviations from the CET represent 16.9 percent of all tariff lines for Argen- tina, 17.6 percent for Brazil, 23.0 percent for Paraguay, and 21.0 percent for Uruguay. Because the CET has been implemented for 75 percent of the tariff lines, most member-country deviations from the CET occurred for the same items (essentially under the special regimes). Deviations from the CET go beyond the ones specified at Ouro Preto; they include dozens of special regimes allowing for tariff-free or tariff-reduced im- ports as part of several promotional schemes existing in the member countries. This is a result of preexisting preferences granted to other members of the Latin American Integration Association (LAIA), except Chile and Bolivia, where recent agreements have regionalized concessions formerly granted by individual Mercosur countries, or the peculiar status of Manaos and Tierra del Fuego (tax- free areas) that has been preserved until 2013. To make things harder, Mercosur has made little progress in applying a truly common external trade policy. More- over, a certain number of import quotas are still in place at the national level, in spite of the commitment taken in the Treaty of Asunci6n to eliminate quantita- tive barriers to trade or to harmonize them when necessary, as for example within the Multifibre Arrangement of the World Trade Organization (WTo). However, due to the lack of reliable data, this article only analyzes the Ouro Preto tariff agreements. The textile sector illustrates the differences from the CET. Argentina as- sesses tariffs for more than 600 textile items, employing "minimum specific duties" that result in specific duties for the lower price range and ad valorem duties for the top of the price scale. These duties are under WTO examination, at the request of the United States. Uruguay assesses tariffs for more than 100 textile items using "precios minimos de exportaci6n" that result in a combination of variable levies, specific duties, and ad valorem tariffs (see Changanaqui and Messerlin 1994). Brazil originally included a number of textile tariff items on a list of exceptions to the CET with duties up to 70 percent. It has recently begun to use the special safeguard provision of the Olarreaga and Soloaga 301 textile agreement that results in quotas for specific products from specific origins. Article 5 of the Treaty of Asunci6n calls for the elimination of nontariff re- strictions or equivalent measures. The elimination of quantitative barriers to trade is probably explained by the commitments made during the Uruguay Round and not by endogenous political choice. For a discussion of the political choice of the means of protection, see chapter 7 in Hillman (1989). Internal tariffs (tariffs applied to imports from other member countries) pre- exist the Ouro Preto protocol under the list of exceptions of the Treaty of Asunci6n. Almost all tariff lines outside the list of exceptions reached the Treaty of Asunci6n's target of free trade on schedule at the beginning of 1995 (see Laird 1997). The Ouro Preto protocol decided, under what has been renamed the "adecuaci6n" regime, that items in the list of exceptions should converge to internal free trade by 1999 for Argentina and Brazil and by 2000 for Paraguay and Uruguay. In 1996 deviations from internal free trade only corresponded to 0.2 percent of total tariff lines for Brazil, 2.5 percent for Argentina, 3.3 percent for Paraguay, and 4.4 percent for Uruguay. Table 2 reports average external and internal tariffs and the average CET for 1996 (simple averages of the whole and exempted universe and import-weighted averages). The tariffs indicate that Argentina and Brazil are on average converg- ing downward to the CET, whereas Paraguay and Uruguay are converging up- ward to it. The simple average external tariff rates over their exempted universe are 6.83 percent in Paraguay and 5.92 percent in Uruguay; they reach 21.39 percent in Brazil. Mercosur's members have experienced a significant liberalization effort in the past decade. In 1986 the average tariff was close to 80 percent in Brazil, 41 Table 2. Average Tariffs in the Mercosur Countries, 1996 Average import-weighted Average unweighted tariff Average tariff tariff on exempt items Country External Internal External Internal Externala internalb Argentina 11.78 0.36 13.37 0.86 14.33 11.69 Brazil 13.14 0.02 15.44 0.02 21.39 10.20 Paraguay 8.79 0.80 5.18 0.37 6.83 24.91 Uruguay 10.78 0.88 11.01 1.77 5.92 19.73 Mercosur 11.15 0.00 11.09 0.00 n.a. n.a. n.a. Not applicable. There are no exemptions for Mercosur. Note: Import-weighted tariffs tend to be downward biased because sectors with high tariffs tend to import less. For Mercosur, external tariff data are the common external tariff (CET) to be achieved once the Ouro Preto objectives are reached; internal tariff data reflect the objective established in Ouro Preto, that is, free trade. a. For Argentina, 1,540 tariff lines (16.9 percent of the total universe); for Brazil, 1,605 (17.6 percent); for Paraguay, 2,101 (23.0 percent); and for Uruguay, 1,961 (21.5 percent). b. For Argentina, 231 tariff lines (2.5 percent of the total universe); for Brazil, 17 (0.2 percent); for Paraguay, 293 (3.2 percent); and for Uruguay, 407 (4.4 percent). Source: See appendix B. 302 THE WORLD BANK ECONOMIC REVIEW, VOL. 12, NO. 2 percent in Argentina, 20 percent in Paraguay, and 36 percent in Uruguay (Mendoza 1996). Table 2 indicates that tariff levels averaged between 9 and 13 percent for 1996. Internal tariffs were particularly low for Argentina and Brazil, calculating averages over the entire universe of more than 9,000 tariff items. However, considering the averages over the exempted universe only, they were close to 10 percent and reached a high of 25 percent in Paraguay. These internal tariffs may be subject to some discussion, as suggested by an anonymous referee. Smuggling accounts for a large share of Paraguay's trade. Connolly, Devereux, and Cortes (1995) estimate that the unreported trade in Paraguay accounts for 58 percent of its total exports. However, the convergence to internal free trade will reduce the incentives in Paraguay to smuggle goods to its larger neighbors. Figure 1 illustrates deviations from the negotiated CET in percentage points in Argentina, Brazil, Paraguay, and Uruguay for the 27 sectors of the International Standard Industrial Classification (IsIc) three-digit codes. (See appendix A for the sectors and appendix B for data sources.) The negotiated CET differs from the external tariff due to national exceptions. Figure 1 shows that Argentina tends to deviate upward for wearing apparel, footwear, furniture, paper and paper products, printing and publishing, plastic products, and iron and steel. Argentina is relatively close to the CET for the rest of the isic three-digit classifi- cation. Brazil deviates upward for footwear, plastic products, machinery, elec- tric machinery, and transport equipment and remains close to the CET for the rest of the isic three-digit classification. Paraguay deviates upward for wearing Figure 1. Deviationsfrom MVercosur's Common External Tariff, 1996 Percentage points 15 E} Argentina r Brazil 10 * Paraguay El Uruguay 5 |- -1 flA ffi I PInsal, i ; ;r -5 -10 Industrial sector, by istc product code Note: See Appendix A for the industries corresponding to the ISIC product codes. Source: Authors' calculations. Olarreaga and Soloaga 303 apparel, footwear, and wood products. Uruguay deviates upward for textiles, wearing apparel, and nonmetallic mineral products. Paraguay and Uruguay de- viate downward for machinery, electric machinery, transport equipment, and professional and scientific equipment. Figure 1 also shows the asymmetry between the deviations in Brazil and the deviations in Paraguay and Uruguay for the last four products of the isic three- digit classification. Indeed, Brazil is a relatively large producer of these products and deviates upward from the CET, whereas Paraguay and Uruguay are rela- tively small producers and deviate downward from the CET. Not surprisingly, these products have been relatively sensitive in the Mercosur negotiations on the CET. These industries mainly correspond to the sectors where special regimes have been negotiated, for example, automobiles, computer products, and capi- tal goods. Uruguayan negotiators apparently put great effort in trying to lower the CET on capital goods because Uruguay's external tariff was set at zero for most of these products before the Ouro Preto negotiations. Internal deviations from free trade for the four member countries are illus- trated in figure 2. Brazil's deviations are almost nonexistent at this level of ag- gregation. Argentina, Paraguay, and Uruguay have relatively high internal tar- iffs. Internal deviations seem to be more important for the seven categories of items from textiles to paper and paper products (see appendix A), with a high of 9 percent for footwear in Paraguay. Internal tariffs are also relatively high for iron and steel and transport equipment in Argentina. In general, tariff levels may seem relatively low, but these are simple average tariffs at a high level of aggregation (27 sectors). For example, using the two- digit harmonized system that includes 97 sectors, Brazil's external tariff is 70 percent for articles of apparel and clothing and 50 percent for vehicles, foot- wear, and prepared vegetables. Internal deviations from free trade reach a high Figure 2. Deviationsfrom Internal Free Trade, Mercosur, 1996 Percentage points 9 8 [- Argentina 7E Brazil U Paraguay 6 El Uruguay 4 Industrial sector, by isic product code Note: See Appendix A for the industries corresponding to the ISIc product codes. Source: Authors' calculations. 304 THE WORLD BANK ECONOMIC REVIEW, VOL. 12, NO. 2 of 28 percent for vehicles in Argentina and for articles of apparel and clothing in Argentina, Paraguay, and Uruguay. II. PREDICTIONS OF THE ENDOGENOUS TARIFF LITERATURE The theory of endogenous protection describes how a combination of agents' preferences on trade policy and the weight given to different groups' preferences on policy determination may translate into deviations from first-best trade poli- cies. This literature is extremely rich and has followed different approaches. Rodrik (1995) reviews the empirical and theoretical literature, and Rama (1994) provides an empirical application to the region. For alternative approaches to the theory of endogenous protection based on social insurance, see, for example, Hillman (1989). As recently suggested by Helpman (1995), however, the different approaches tend to generate the same predictions. This section summarizes the main results of the theoretical literature, starting with the general results, then turning to the new predictions of the theoretical literature on the endogenous formation of RTAs. General Results Here we focus on seven predictions of the correlates of expected cross-sectoral variations in tariff protection. These results are well documented in the empiri- cal literature on endogenous tariff formation (see Rodrik 1995). However, both the theoretical and empirical results are somewhat partial equilibrium results because they do not necessarily account for simultaneity bias. For an empirical study that accounts for the simultaneity bias between imports and tariffs, see Trefler (1993). Other things being equal, the level of protection received by an industry is higher if any of the seven conditions hold. The first condition is that the level of industry concentration is high. Rodrik (1987) provides a theoretical justification, and Trefler (1993) and Marvel and Ray (1983) present empirical examples. This condition captures the type of free- riding incentives in Olson (1965). However, both empirical and theoretical evi- dence indicates that this condition need not hold. On the one hand, industry concentration might solve the free-riding problem. On the other hand, an in- crease in group size may result in higher contributions of the group (see Cornes and Sandler 1996). Moreover, the theory is not well founded in empirical mea- sures of industry concentration, as shown by Hillman (1991) and van Long and Soubeyran (1996). Baldwin (1984) presents ambiguous evidence on the relation between protection and industry concentration. Bilal (1995) reviews the litera- ture on seller concentration and protection. In general the literature presumes that industry concentration leads to higher levels of protection, and this is con- firmed in section IV. The second condition is that the import penetration ratio is low. The lower the import penetration ratio, the lower is the relative weight of consumers com- Olarreaga and Soloaga 305 pared with producers in the government's objective function.' Grossman and Helpman (1994) provide a theoretical justification for this condition. This result has been generally challenged on empirical grounds, as discussed by Rodrik (1995). For empirical examples, see Anderson (1980) or Finger and Harrison (1994). The third condition is that the proxy for input sales-the share of sector pro- duction that is purchased by other sectors as intermediates-is low. Cadot, de Melo, and Olarreaga (1997) provide a theoretical justification for this condi- tion, and Ray (1991) and Marvel and Ray (1983) present empirical examples. This condition captures lobbying rivalry. If sector j purchases goods from sector i, then sector j will counter-lobby any increase in sector i's level of protection. Thus the higher the share of sector i production that is purchased by other sec- tors, the smaller is the endogenous tariff. As long as consumers are not orga- nized, other things being equal, consumer goods receive higher levels of protec- tion than intermediate goods. The fourth condition is that the labor/capital ratio is high. For empirical evi- dence, see, for example, Finger and Harrison (1994) and Rodrik (1995). Cadot, de Melo, and Olarreaga (1997) explain this condition. They show that tariffs are higher in sectors where the share of capital remuneration in value added is large, after introducing lobbying rivalry in the labor market. A higher labor/ capital ratio, other things being equal, has two opposing effects on the share of capital remuneration in value added. On the one hand, the direct effect tends to reduce it, because a higher labor/capital ratio obviously implies a smaller capi- tal/labor ratio. On the other hand, a higher labor/capital ratio implies a higher marginal productivity of capital relative to labor, which in turn raises the share of capital remuneration in value added. Under suitably general conditions, the latter effect dominates the former if the elasticity of substitution between labor and capital is smaller than 1 (which is a generally accepted value in the empirical literature).2 The fifth condition is that the share of intraindustry trade is small. Cadot, de Melo, and Olarreaga (1997), Levy (1997), and Marvel and Ray (1987) provide theoretical explanations; Marvel and Ray (1987) provide an empirical example. Cadot, de Melo, and Olarreaga (1997) argue that the larger the share of intraindustry trade in total trade, the larger is the elasticity of import demand for goods produced in the domestic economy. Following the Ramsey pricing rule, this larger elasticity will make the tariff lower because the efficiency cost of a tariff is relatively large compared with the producers' gain. Marvel and Ray I. To see this, m/y = (c - y) I y = (c I y) - 1, where m denotes imports (or net imports), c is consumption, and y is the level of production. 2. In a two-factor sector, the share of capital remuneration in value added is given by: b = rk I (rk + wi) + 1 / [(wl I rk) + 1), where r is capital wage, k is the amount of capital, w is labor wage, and I is the amount of labor. Then ab I a(l/k) = [I / [(wl /rk) + 1112 (wI[r(1 + s)]), where s is the elasticity of substitution between labor and capital and the right-hand side is larger than zero if Isl < 1. The empirical estimation of the elasticities of substitution between labor and capital generally yield values below 1. 306 THE WORLD BANK ECONOMIC REVIEW, VOL. 12, NO. 2 (1987) suggest an alternative explanation based on intermediate inputs and counter-lobbying. They argue that intraindustry trade essentially arises among producers (who purchase intermediate goods), and because producers are more concentrated than consumers, they tend to be more efficient in combating pro- tectionist pressures. Levy (1997) argues that an increase in intraindustry trade benefits all agents, whereas an increase in interindustry trade has the usual Stolper- Samuelson redistributive effects and therefore is subject to more conflict and higher lobbying pressures. Assuming that labor markets are segmented in the sense that labor is better conceived as being mobile across a particular group of industries than across the economy as a whole, two additional conditions are relevant. The level of protec- tion received by an industry is higher if either the sixth or the seventh condition holds. The sixth condition is that the equilibrium wage in the sector is low. Cadot, de Melo, and Olarreaga (1997) provide a theoretical justification, and Anderson and Baldwin (1987) and Ray (1991) provide empirical examples. Cadot, de Melo, and Olarreaga (1997) show that the optimal endogenous tariff for each sector is positively related to the share of specific capital in total sales. The larger the wage in sector i (controlling for output and labor/capital ratios), the smaller is the share of capital in total sales and therefore the smaller are the incentives to lobby in the political game. The seventh condition is that the share of labor in this sector is large relative to total employment in the economy. Cadot, de Melo, and Olarreaga (1997) provide a theoretical justification, and de Melo and Tarr (1994) present an em- pirical example. If labor unions are organized, then the share of employment in the sector will be larger, the larger is the political weight of this sector's labor union. Alternatively, it may be the case that votes matter and that a high labor/ capital ratio indicates the presence of voters (see, for example, Potters, Sloof, and van Winden 1997). New Results on the Endogenous Formation of RTAS The endogenous trade policy literature has recently shed some light on the endogenous formation of RTAS. (There are other results in that literature that are not subjected to testing here; see, for example, Cadot, de Melo, and Olarreaga 1996, Ethier 1996, Hillman, Long, and Moser 1995, and Winters 1996 for a review.) The two results that are to be examined in the empirical section are listed here. First, Grossman and Helpman (1995) suggest that deviations from internal free trade in RTAs are more likely to occur in sectors where trade creation is more likely. This is explained by two interrelated forces. First, and perhaps most important, from the perspective of the importing country, the political cost of trade diversion is higher than the political cost of trade creation. This is due to the fact that trade creation entails larger domestic price reductions in the importing country than trade diversion, others things being equal. (Alter- Olarreaga and Soloaga 307 natively, one can argue that trade-creating sectors create the import competi- tion that harms the domestic producers' interests in a country and that are accordingly resisted by protectionist pressures if the endogenous deviations from internal free trade are determined by the producers' interests.) Second, and for the same reason, from the perspective of the exporting country, the political gains (producer gains) are higher in the case of trade diversion than in the case of trade creation. Thus, when countries negotiate over which sectors to exclude from internal free trade, they prefer to create exceptions for trade- creating sectors that would result in higher costs for the importing country and lower gains for the exporting country. To illustrate the factors that influence the political costs of trade creation and trade diversion, assume that the production of an exporting country in an RTA is sufficiently large to satisfy the whole demand at world prices for an importing country in the RTA. After the RTA is formed, domestic prices in the importing country drop to world levels, which implies that producer surplus in the import-competing industry is significantly reduced if this industry was protected before the RTA is formed. By contrast, exporters in the RTA partner receive the same price as before. If industry directly links its political contribu- tions to the price that producers receive, such a trade-creating RTA may have high political costs. The political cost would be smaller if the exporting coun- try could not satisfy its partner market at world prices, in which case producer prices in the RTA may be above world prices and some trade diversion may exist. The idea is that trade creation, unlike trade diversion, offers no extra benefits to the exporting producer. Ray (1987) indirectly tests this result for the U.S. preferential agreements. He shows that these agreements have failed to offset the protectionist bias in the United States against competitive imports from developing countries. Thus the first new result to be verified is that trade-creating sectors tend to be exempted from internal free trade. Cadot, de Melo, and Olarreaga (1996) suggest that the negotiated CET in a customs union is a weighted sum of member countries' politically optimal tariff vectors. The endogenous weight given to country j in the determination of the CET in sector i equals the size of sector i in country j relative to its size in other member countries. A country with large output in one particular industry would lobby more aggressively to protect its large industry in the negotiations for a CET. Technically, this result occurs because lobbies determine their contribu- tions such that the marginal gain from a tariff increase equals its marginal cost. The marginal gain is given by the derivative of the profit function with respect to prices, which, in a competitive environment, equals the level of production (by Hotelling's lemma). Thus the second new result to be tested is that the CET is determined by the production-weighted sum of the political economy variables in member coun- tries. The CET in sector i mainly reflects the preferences of the member country that has the largest level of production in sector i. 308 THE WORLD BANK ECONOMIC REVIEW, VOL. 12, NO. 2 For other new results in the literature that are not tested here, see, for ex- ample, Cadot, de Melo, and Olarreaga (1996); Ethier (1996); Hillman, van Long, and Moser (1995); and Winters (1996). III. THE EMPIRICAL MODEL The first step of the empirical study is to verify that a set of Mercosur's politi- cal variables can explain the structure of Mercosur's CET. The CET equation is: n U (1) log CETi = a:o + Ycak log ,OiPYik + HLi k=1 j=A where subscript i refers to the industry (from the 27-industry aggregation of the isic three-digit classification, see appendix A); CETi is the CET in sector i; as are parameters; Of, is the share of country j = A (Argentina), B (Brazil), P (Paraguay), or U (Uruguay) in total production of good i in Mercosur; PV'Ik is the political economy variable k in sector i in country j; and p,u is the error term. The political economy variables are listed in appendix C, along with an explanation of how the variables are constructed and their expected signs. The weights given to each country's political variables should not be imposed but should be estimated empirically (Cadot, de Melo, and Olarreaga 1996). However, given the lack of industrial data at a high level of aggregation, this is empirically infeasible. In section IV we discuss some alternative determinations of the CET. The second step is to verify that member countries' deviations from the CET can be explained by deviations in the political structure of each country from Mercosur's political structure (weighted averages). The equation to be estimated for each country is: (2) log CET = o+ fklog k + FI CETi OI4 where ET', is the external tariff in country j for sector i. There are four CET- deviation equations to be estimated, one for each member country. The last step is to verify that member countries' deviations from internal free trade can also be explained by the political structure of each member country. The equation to be estimated for each country is: n (3) log ITi' = Yo + E Yk log PVi7k +y y +TC' + 1)i k=1 where ITI is country j's internal tariff in sector i, and TC, is the expected level of trade creation in the importing country j in sector i. Ideally, there are four inter- nal tariff equations to be estimated, one for each country. Olarreaga and Soloaga 309 IV. RESULTS There are nine equations to be estimated. Equation 1 determines the CET, equa- tion 2 (for the four countries) determines member countries' deviations from the CET, and equation 3 (for the four countries) determines member countries' devia- tions from internal free trade. Because the error terms in these different equations turn out to be correlated, we use a seemingly unrelated regressions (SUR) tech- nique. The correlation between the error terms is probably due to omitted vari- ables, like the influence that the World Bank, the International Monetary Fund, or the wTro may have over trade policy and other environmental variables in these countries. However, the industrial data lack disaggregation: each member country has only 27 sectors (observations). Therefore, a panel estimate for external devia- tions from the CET and internal deviations from free trade seems more appropri- ate. We estimate these panel regressions using a SUR technique. Panel estimation is not possible for the CET because, by definition, the CET is common to all member countries. We estimate equation 1 using a SUR technique including the external and internal deviation equations for each country. All equations are estimated over the whole tariff universe (that is, the full range of tariff lines aggregated to the 27 sectors of the isic three-digit classification). The endogenous variables in all equations reported here are import-weighted average tariffs. The explanatory power of the whole set of regressors improves in all equations when import-weighted tariffs are used rather than simple averages to translate the eight-digit harmonized tariff system data into the three-digit classifi- cation system of the United Nations Industrial Development Organization (UNIDO; see appendix A for sectors and appendix B for details on data). The results of the analysis exclude primary products and only correspond to industrial products. In the four member countries, industrial imports account for no less than 80 percent of total imports. All regressions are run in a double-log form except for the net import penetration term (which can take negative values). Determining Mercosur's CET This section explores alternative determinations of the CET to test indirectly the prediction that the relevant variables in the determination of the CET are the production-weighted political economy variables (Cadot, de Melo, and Olarreaga 1996). It then reports the SUR estimation of the CET equation. ALTERNATIVE DETERMINATIONS OF THE CET. To examine alternative deter- minations of Mercosur's CET, we test whether the Mercosur political variables have a better explanatory power than any other combination of member countries' political variables. To this end, we use the Davidson and MacKinnon (1981) J non-nested test. (An unfortunate feature of this test is that, in testing whether one set of regressors is more appropriate than another, it allows rejection or acceptance of both sets of regressors.) 310 THE WORLD BANK ECONOMIC REVIEW, VOL. 12, NO. 2 Table 3 gives results of the Davidson and MacKinnon test. The first row, for example, tests the hypothesis that Argentina's political economy variables taken alone are as appropriate a set of regressors as Mercosur's weighted average of political economy variables. In the first column, we report the t-statistic coeffi- cient obtained for the first hypothesis, that is, assuming that Argentina's politi- cal economy variables are a better set of regressors than Mercosur's. In the sec- ond column, we report the t-statistic coefficient obtained for the second hypothesis, that is, assuming that Mercosur's political economy variables are a better set of regressors than Argentina's. In the case of Argentina, the first hy- pothesis can be rejected, whereas the second hypothesis cannot be rejected with 99 percent confidence. This implies that Mercosur's set of regressors has better explanatory power. When both t-statistics are statistically significant or insig- nificant, the test is inconclusive. When only one of these tests is significant, a conclusion can be drawn. The last column in table 3 specifies the better- performing sets of regressors. Two main conclusions follow from the results of the Davidson and MacKinnon non-nested tests for the determination of the CET. First, Mercosur's political variables (weighted averages) have a higher explanatory power than those for Argentina, Paraguay, or Uruguay taken one by one. Mercosur's political vari- ables also have greater explanatory power than any combination of these three countries. Second, the analysis does not reject the hypothesis that the set of regressors for Brazil by itself, or any combination that includes Brazil, is as ap- propriate as Mercosur's political economy variables (weighted averages). This result from the non-nested tests indicates that Brazil had an important role in determining Mercosur's CET. Because Brazil represents, in all sectors, at least 70 percent of Mercosur's production, it is not surprising that the CET closely reflects Brazil's interests. These results may be seen as an indirect test of the proposition that the CET should reflect a production-weighted average of member countries' political varn- Table 3. The Explanatory Power of Different Sets of Political Economy Variables in Determining Mercosur's Common External Tariff The political economy variables that are assumed to be the better set of regressors The better-performing Country Country set Mercosur set set of regressors Argentina 0.69 3.47-*' Mercosur Brazil 0.91 1.06 Inconclusive Paraguay 1.44 3.85$** Mercosur Uruguay 0.37 3.46*** Mercosur Argentina and Brazil 1.13 1.47 Inconclusive Argentina, Paraguay, and Uruguay 0.33 2.49" Mercosur Significant at 95 percent. * Significant at 99 percent. Note: Results are based on Davidson and MacKinnon (1981) J non-nested tests. Source: Authors' calculations. Olarreaga and Soloaga 311 Table 4. Estimation of the Common External Tariff for Mercosur Regression including Regression including only Variable' all variables significant and robust variables Constant 2.47T 2.49- (6.26) (6.57) Laborlcapital ratio 0.67- 0.58- (3.42) (3.74) Wage -0.40" -0.41 (-2.02) (-1.77) Industry concentration index 0.21- 0.14** (2.37) (2.02) Net import penetration ratio 0.43b (1.23) Input sales proxy -0.33b (-0.84) Labor union proxy 0.06' (0.57) Intraindustry trade proxy 0.03b (0.42) R2 0.58 0.51 F-test of SUR 222- 224"** Number of observations' 27 27 Significant at 90 percent. Significant at 95 percent. * Significant at 99 percent. Note: The dependent variable is log (cET). All independent variables are in logs except for the net import penetration ratio. The regression is estimated using a SUR technique. Standard errors are White- robust. t-statistics are in parentheses. The R2 is estimated using the covariance matrix of the residuals. a. For variable construction, see appendix C. b. The coefficient changes sign when performing an outlier analysis over the 27 observations. c. The 27 observations are for the 27 industries listed in appendix A. Source: Authors' calculations. ables (Cadot, de Melo, and Olarreaga 1996). Therefore, in the remainder of this article, we consider the CET for Mercosur as being the production-weighted av- erage of member countries' political economy variables. ESTIMATION OF THE CET. Table 4 reports the results of the SUR estimation of the CET.3 The first column in table 4 shows the results when the regression includes all the explanatory variables discussed in section II, whereas the second column gives the results when only the significant and robust variables are included. It appears that the most statistically significant variables in the determination of the CET are the labor/capital ratio, the wage level, and the index of industry concentration. The proxy for input sales is statistically insignificant, as are the import penetration ratio, the labor union proxy, and the intraindustry trade variable. All coefficients have the expected sign except for the import penetration 3. The SUR estimates for internal and external deviations are relatively less efficient than the ones obtained with a panel technique and are not reported here. 312 THE WORLD BANK ECONOMIC REVIEW, VOL. 12, NO. 2 ratio and the intraindustry trade variable. In addition, all the insignificant variables change signs in an outlier analysis over the 27 observations. The explanatory power of the political variables is relatively high and ac- counts for 58 percent of the CET total variation in the first column of table 4 and 51 percent in the regression on only the significant and robust variables. Deviations from the CET and from Internal Free Trade We estimate the deviation equations using a fixed-effects model and a SUR technique. Table 5 reports the results. The overall explanatory power of the political variables oscillates between 21 and 51 percent for internal deviations, depending on whether industry and country dummies are included. For external deviations, the explanatory power is 49 percent. The relatively low explanatory power of the internal deviations regression is probably due to the fact that Brazil's Table 5. Deviations from Mercosur Objectives, 1996 Deviations from internal free trade Deviations from CET, Industry and No industry and country Variable4 country dummies dummies dummies Constant 0.14 3.45 -0.27" (0.03) (1.37) (-1.76) Labor/capital ratio -0.87 -0.41 0.15-' (-1.07) (-0.54) (2.85) Labor union proxy 2.35-- 2.04- 0.09- (3.25) (3.29) (2.42) Net import penetration ratio -0.01 -0.02" 0.0001"lb (-0.56) (-2.10) (2.09) Industry concentration index 1.43' 0.08k 0.11" (1.71) (0.12) (2.13) Trade creation term 0.85' 1.18-" (1.72) (4.75) R2 0.51 0.21 0.49 F-test of SUR 216".. 216 - 216- Number of observationsc 108 108 108 F on country dummies 1.86 2.99- F on industry dummies 1.12 1.69- F on all dummies 1.49 1.78" Significant at 90 percent. Significant at 95 percent. * Significant at 99 percent. Note: CET is the common external tariff. The dependent variable is the log of the internal tariff or the log of the external tariff/cET ratio. All independent variables are in logs except the net import penetration ratio. The regression is estimated using a fixed-effects model and a SUR technique. Standard errors are White-robust. t-statistics are in parentheses. The R2 is estimated using the covariance matrix of the residuals. a. For variable construction, see appendix C. b. The coefficient changes sign when performing an outlier analysis over 12 potential outliers identified following Belsley, Kuh, and Welsch (1980). c. For each of the four countries, there are 27 industries (see appendix A). Source: Authors' calculations. Olarreaga and Soloaga 313 (and to some extent Argentina's) internal deviations from free trade are rela- tively rare (Brazil deviates in only 3 of the 27 sectors). For the same reason, we performed an outlier analysis following Belsley, Kuh, and Welsch (1980) to check for the robustness of results reported in table 5. INTERNAL DEVIATIONS FROM FREE TRADE. The first column in table 5 reports the results of the internal deviations regression including country and industry dummies. The second column reports results without dummies, because neither country-specific nor industry-specific effects are significant at a 95 percent confidence level, according to F-tests. Industry-specific effects may be insignificant because the adecuacion regime for internal deviations from free trade has been inherited from the Asunci6n Treaty, where no special regimes per sector existed and different member coun- tries could deviate from internal free trade in different products. Country- specific effects may be insignificant because this term is correlated with the trade creation variable, given the difference in size of the domestic markets of Mercosur's members. The trade creation term, the labor union proxy, and the index of industry concentration are the significant political economy variables in the explanation of internal deviations from free trade (all above the 95 percent level, not includ- ing dummies). The statistically significant coefficient obtained for the trade cre- ation variable confirms Grossman and Helpman's (1995) view that sectors with high levels of trade creation tend to be exempted from internal free trade. The precision of the trade creation term is reduced (although the coefficient is still significant at the 90 percent level) when dummies are introduced. This result is probably due to the fact that the potential trade creation term is correlated with country dummies. Given the size of its internal market, Brazil faces much lower trade creation than the other member countries. The only significant dummy variable is the one for Brazil, which has a negative coefficient. This result seems reasonable because the effect of internal trade barriers should depend on the size of the domestic market in each member country. For instance, Brazilian produc- ers can easily inundate Uruguay's domestic markets, whereas the reverse is not possible. Both the import penetration ratio and the labor/capital ratio are insignifi- cant in the first column in table 5. The labor/capital ratio is probably insignifi- cant because Mercosur's smaller members (Paraguay and Uruguay) set the barriers to internal free trade, and these countries tend to be relatively less capital abundant than the larger members (Brazil and Argentina). Smaller members set barriers in capital-intensive sectors to protect their markets from larger members. Thus the labor/capital ratio may capture something different than what was expected in section II. Results are reported for regressions ex- cluding the three variables-the wage variable, the proxy for intraindustry trade, and the proxy for input sales-that are insignificant at the 90 percent level in all the regressions. 314 THE WORLD BANK ECONOMIC REVIEW, VOL. 12, NO. 2 DEVIATIONS FROM THE CET. The third column in table 5 reports the results of the external deviations regression with both industry and country dummies. The presence of both industry and country-specific effects cannot be rejected at the 95 percent confidence level. The statistical significance of industry dummies may suggest that Mercosur members mainly deviate from the CET under the special regimes in the industries discussed in section I. (It is not possible to introduce dummies for these regimes given the level of aggregation of industrial data.) The statistical significance of the country dummies suggests that different members tend to deviate differently, as discussed in section 1. The labor/capital ratio, the index of concentration, the labor union proxy, and the import penetration ratio are the important political economy variables in the explanation of deviations from the CET. The import penetration ratio does not have the expected sign, but this is a rather common empirical result, as discussed in section II. We also introduce a variable calculated as the share of extra-regional imports in total imports (not reported here). It corresponds to the trade creation term in table 5, although it is defined differently than in the inter- nal deviation equations. The idea is that high levels of extra-regional protection may occur in sectors with high shares of imports from non-Mercosur countries. This term is statistically insignificant; therefore we report results without in- cluding this variable. The labor union proxy has high explanatory power in determining the devia- tions from internal free trade and from the CET. As the deviations were negoti- ated within the countries, labor unions may have had more weight in determin- ing them than in determining the CET, which was negotiated at the regional level. Alternatively, the difference in explanatory power may indicate that the relative size of the industry helps to explain not the formation of the CET but only devia- tions from internal free trade and from the CET within countries. All industry dummies have a positive coefficient (not reported here). The dummy for the scientific equipment sector of the isic classification is dropped from the regression, implying that all other sectors have a relatively higher de- viation from the CET than that sector. The coefficient is significantly different from zero at the 99 percent level in wood products, printing and publishing, petroleum refineries, and transport equipment. Because these sectors do not nec- essarily reflect the comparative advantage of Mercosur countries, the (down- ward) convergence to the CET tariff probably tends to reduce trade diversion in the region, as preference margins are reduced. The downward convergence to the CET is not necessarily true for Paraguay and Uruguay. However, these countries are relatively small trading partners com- pared with Argentina and Brazil, and, therefore, their influence on the overall picture can be minimized. The upward convergence of Mercosur's small mem- bers to the CET is confirmed by the signs of the country dummies in the external deviation equation. Indeed, the country dummies for Paraguay and Uruguay have negative coefficients, whereas the country dummies for Argentina and Bra- zil have positive coefficients. This confirms the fact that generally Brazil and Otarreaga and Soloaga 315 Argentina are downward-converging to the CET and that the opposite is true for the small members of Mercosur. A striking feature of table 5 is that estimated elasticities tend to be much higher in the internal deviations regressions than in the external deviations re- gression. The coefficient of the labor union proxy, for example, is 20 times larger in the internal deviations regressions. It might be that smaller members of Mercosur tend to deviate more often than larger members (internally and exter- nally), and lobbies in smaller members are certainly more concerned about in- ternal barriers than external ones. In other words, it might be useless for Uru- guayan producers, for example, to obtain a high external tariff if Brazilian producers can enter their market without facing any barrier. However, this result may be better explained by the statistical techniques employed and the data structure. Internal deviations are calculated in absolute terms because the optimal situation implies internal tariffs that are nil. Exter- nal deviations are calculated from an optimal situation in which the CET is always positive, and thus the endogenous variable in the external deviation regression is taken in relative (percentage) terms. The standard deviation of the endogenous variable in the internal deviation equation is 20 times larger than the standard deviation of the endogenous variable in the external devia- tion regression, which in turn may partly explain differences in the size of the estimated coefficients. Because some results may be driven by a small number of large deviations, we perform an outlier analysis following the procedures of Beisley, Kuh, and Welsch (1980). The main results in all regressions are robust to the exclusion of potential outliers (results that are not robust are noted in tables 4 and 5). To summarize, political variables seem to explain the deviations from the CET and from internal free trade in Mercosur countries. Further, achieving the Ouro Preto convergence (that is, internal free trade and convergence to the CET) would probably make the region more trade creating and more open. V. CONCLUSIONS In its first five years, Mercosur achieved an important degree of integration not only within the region but also with the rest of the world. It also succeeded from a volume-of-trade perspective. Intraregional trade increased at an average annual rate of 29 percent and total trade increased at 9 percent, compared with a 6 percent increase in world trade since 1991 when the Treaty of Asunci6n was signed. Mercosur seems to enjoy a relatively high level of protection against the rest of the world. Its negotiated CET is around two times higher than the tariff levels in the Organisation for Economic Co-operation and Development (OECD). However, Mercosur's average CET as established in Ouro Preto is only 0.1 percentage point higher than Chile's average tariff (this decade's Latin Ameri- can example) and eight times lower than Brazil's average tariff in 1986. The 316 TIHE WORLD BANK ECONOMIC REVIEW, VOL. 12, NO. 2 larger members of Mercosur are adjusting downward to the CET. In 1996 aver- age external tariffs were 0.8 and 2.0 percentage points higher in Argentina and Brazil, respectively, than in Chile. This adjustment signals a liberalizing effort. It also indicates that the region will be more open once it achieves the Ouro Preto objectives. The new results of the literature on the endogenous formation of RTAS seem to be reflected in Mercosur's structure of protection. First, as predicted by Cadot, de Melo, and Olarreaga (1996), Brazil has an important weight in Mercosur's tariff structure, given Brazil's significant share of Mercosur's production (and the intergovernmental aspect of the decisionmaking process). Second, as pre- dicted by Grossman and Helpman (1995), internal barriers to free trade tend to be higher in the potentially trade-creating sectors. Convergence to the Ouro Preto objective of internal free trade will therefore lead to more trade creation. This may turn out to be a difficult process as deviations from internal free trade reflect lobbying pressures in member countries. The Mercosur countries need to make a serious political commitment to the Ouro Preto objectives. However, insofar as Mercosur's tariff structure and espe- cially the CET reflect the political economy of the member countries, they may achieve convergence more easily. Tariffs are subject to changes, but ensuring that they respect political preferences in member countries will facilitate con- solidation of the integrating effort. In sum, whether Mercosur is a step in the right direction may remain an open question, but this article shows that Mercosur is apparently here to stay. APPENDIX A. ISIC CODES FOR THE 27 INDUSTRIES IN THE ANALYSIS 311 Food products 313 Beverages 314 Tobacco 321 Textiles 322 Wearing apparel, except footwear 323 Leather products 324 Footwear, except rubber or plastic 331 Wood products, except furniture 332 Furniture, except metal 341 Paper and paper products 342 Printing and publishing 351 Industrial chemicals 352 Other chemicals 353 Petroleum refineries 354 Miscellaneous petroleum and coal products 355 Rubber products 356 Plastic products 361 Pottery, china, earthenware Olarreaga and Soloaga 317 362 Glass and glass products 369 Other nonmetallic mineral products 371 Iron and steel 372 Nonferrous metals 381 Fabricated metal products 382 Machinery, except electrical 383 Machinery, electrical 384 Transport equipment 385 Professional and scientific equipment APPENDIX B. DATA The Mercosur Secretariat provided data on common external tariffs (official tariffs for 1996, announced in December 1995). We obtained member coun- tries' external and internal tariffs from the country governments: for Argentina, decree no. 998/95 (December 29, 1995) and resolutions no. 649/96, 370/96, 111/96, and 735/96; for Brazil, decree no. 1767 (December 29, 1995); for Para- guay, decree no. 12056 (December 29, 1995); and for Uruguay, decree no. 466/ 95 (December 29, 1995) and decrees no. 242/996, 282/996, and 316/996. Tariff data are aggregated at the eight-digit level of the harmonized system (9,119 items). Trade data sources are national accounts (COMTRADE), also aggregated at the eight-digit level of the harmonized system (1994 data). We obtained the best (and more aggregated) industrial data for Mercosur countries from UNIDO'S three- digit database (average from 1987 to 1993). They are disaggregated into 27 sectors (see appendix A). To convert data from the harmonized systems into UNIDO three-digit data, we used a filter provided by Jerzy Rozanski at the World Bank. APPENDIX C. VARIABLE CONSTRUCTION AND NOTATION Endogeneity problems can be important, as suggested in a study by Trefler (1993), because most of the exogenous variables may also be functions of tar- iffs. Due to data restrictions, the empirical section does not deal with endogeneity problems. All of Mercosur's political economy variables are constructed as the sector-production-weighted sum of member countries' political variables as dis- cussed in section II. Alternative specifications for Mercosur have been tested and are discussed in section IV. The following variables are used in the analysis. All variables are in logs ex- cept for the net import penetration ratio. * Tariffs, endogenous variables, are constructed from import-weighted averages in the eight-digit harmonized system that correspond with the UNIDO three-digit classification. 318 THE WORLD BANK ECONOMIC REVIEW, VOL. 12, NO. 2 * The industry concentration index is calculated as (number of firms in the whole economy)/(number of firms in sector i). The expected sign is positive. * The net import penetration ratio is calculated as (imports - exports)/(gross output). The expected sign is negative. In estimating the equations for external tariffs, we use extra-Mercosur trade to calculate this variable, whereas in dealing with internal deviations from free trade, we employ intra-Mercosur trade. * The proxy for input sales (the share of production sold to other sectors as intermediate goods) is calculated as (value added)/(total output) in sector i divided by the economywide average (value added)/(total output). The expected sign is negative. The idea is that a relatively high share of value added in total output in sector i (with respect to the economywide average) indicates that sector i purchases a relatively small amount of intermediate goods, which suggests that, on average, the rest of the economy purchases a relatively high amount of good i as an intermediate. Although this is not a good proxy, data restrictions make it impossible to construct a better one. This may explain why the variable does not perform well in the estimation. * The labor/capital ratio is calculated as (number of employees)/(value added - labor costs). The expected sign is positive. e The intraindustry trade proxy is calculated as [(imports - exports)2/(imports + exports)210 . As with the net import penetration ratio, intra- and extra- Mercosur data are used when necessary. The expected sign is positive. * Wages per sector are calculated as (labor cost)/(number of employees). The expected sign is negative. [ The labor union proxy is calculated as (number of employees in sector i)l (total number of employees). The expected sign is positive. • The trade creation term is calculated as (intra-Mercosur imports)/(total output). The expected sign is positive. REFERENCES The word "processed" describes informally reproduced works that mayT not be com- monly available through library systems. Anderson, Kym. 1980. "The Political Market for Government Assistance to Australian Manufacturing Industries." The Economic Record 56(153):132-44. Anderson, Kym, and Robert E. Baldwin. 1987. "The Political Market for Protection in Industrial Countries." In Ali M. El-Agra, ed., Protection, Cooperation, Integration, and Development. New York: Macmillan. Baldwin, Richard. 1984. "Trade Policies in Developed Countries." In R. W. Jones and P. B. Kenen, eds., Handbook of International Economics. Amsterdam: North- Holland. Beisley, David A., Edwin Kuh, and Roy E. Welsch. 1980. Regression Diagnostics: Iden- tifying Influential Data and Sources of Collinearity. New York: Wiley. Oiarreaga and Soloaga 319 Bilal, Sanoussi. 1995. "The Political Economy of Regional Integration: The Demand for Trade Protection." Working Paper 95-26. University of Birmingham, England. Pro- cessed. Cadot, Olivier, Jaime de Melo, and Marcelo Olarreaga. 1996. "Regional Integration and Lobbying for Tariffs against Non-Members." CEPR Discussion Paper 1448. Cen- tre for Economic Policy Research, London. Processed. .1997. "Lobbying and the Structure of Protection." CEPR Discussion Paper 1574. Centre for Economic Policy Research, London. Processed. Camara de Industrias del Uruguay. 1996. Boletzn Estadistico Trimestral 3. Montevideo. Changanaquf, Federico, and Patrick Messerlin. 1994. "Administered Protection: Refer- ence Prices and Minimum Export Prices." In Michael B. Connolly and Jaime de Melo, eds., The Effects of Protectionism on a Small Coutntry: The Case of Uruiguay. Wash- ington, D.C.: World Bank. Connolly, Michael B., John Devereux, and Mariluz Cortes. 1995. "The Transshipment Problem: Smuggling and Welfare in Paraguay." World Development 23(6):975-85. Cornes, Richard, and Todd Sandler. 1996. The Theory of Externalities, Puiblic Goods, and Club Goods. Cambridge, U.K.: Cambridge University Press. Davidson, Russell, andJames G. MacKinnon. 1981. "Several Tests for Model Specifica- tion in the Presence of Alternative Hypotheses." Econometrica 49(3):781-93. de Melo, Jaime, and David Tarr. 1994. "Industrial Policies in the Presence of Wage Distortions: The Case of the U.S. Auto and Steel Industries." International Economic Review 34(4):833-51. Ethier, Wilfred J. 1996. "Regional Regionalism." Department of Economics, University of Pennsylvania, Philadelphia. Processed. Finger, Michael, and Ann Harrison. 1994. "The MIFA Paradox: More Protection and More Trade?" International Economics Department, World Bank, Washington, D.C. Processed. Grossman, Gene, and Elhanan Helpman. 1994. "Protection for Sale." Amierican Eco- nomic Review 84(4):833-50. . 1995. "The Politics of Free Trade Agreements." American Economic Review 85(4):667-90. Helpman, Elhanan. 1995. "Politics and Trade Policy." CEPR Discussion Paper 1269. Centre for Economic Policy Research, London. Processed. Hillman, Arye L. 1989. The Political Economy of Protection. London: Harwood Aca- demic Publishers. . 1991. "Market Structure, Politics, and Protection." In Elhanan Helpman and Assaf Razin, eds., itzternational Trade and Trade Policy. Cambridge, Mass.: MIT Press. Hillman, Arye L., Ngo van Long, and Peter Moser. 1995. "Modelling Reciprocal Trade Liberalization: The Political Economy and National Welfare Perspectives." Swiss Jour- nal of Economics and Statistics 131(3):503-15. Hoekman, Bernard, and Michael Leidy. 1992. "Holes and Loopholes in Regional Trade Arrangements and the Multilateral Trading System." Aussenwirtschaft 47(3):325- 60. Laird, Sam. 1997. "Mercosur: Objectives and Achievements." Staff Working Paper TPRD- 97-002. World Trade Organization, Geneva. Processed. 320 THE WORLD BANK ECONOMIC REVIEW. VOL. 12, NO. 2 Levy, Paul I. 1997. "A Political Economy Analysis of Free Trade Agreements." Ameri- can Economic Review 87(4):506-19. Marvel, Howard P., and Edward J. Ray. 1983. "The Kennedy Round: Evidence on the Regulation of Trade in the U.S." American Economic Review 73(1):190-97. . 1987. "Intraindustry Trade: Sources and Effects on Protection." Journal of Po- litical Economy 95(6):1278-90. Mendoza, Mario. 1996. "Which Mercosur Anyway?" Bolet[n de Integracdo Latino- americana, edi,cao especial (December):32-36. Mercosur Secretariat. 1996. "Informe de actividades, 1995-1996." Montevideo. Olson, Mancur. 1965. The Logic of Collective Action: Public Goods and the Theory of Groups. Cambridge, Mass.: Harvard University Press. Potters, Jan, Randolph Sloof, and Frans van Winden. 1997. "Campaign Expenditures, Contributions, and Direct Endorsements: The Strategic Use of Information and Money to Influence Voter Behaviour." European Journal of Political Economy 13 (1):19-32. Rama, Martin. 1994. "Endogenous Trade Policy: A Time-Series Approach." Economics and Politics 6(3):215-32. Ray, Edward J. 1987. "The Impact of Special Interests on Preferential Trade Conces- sions by the U.S." Review of Economics and Statistics 69(2):187-95. .1991. "Protection of Manufactures in the United States." In David Greenaway, ed., Global Protectionism: Is the U.S. Playing on a Level Field? London: Macmillan. Rodrik, Dani. 1987. "Policy Targeting with Endogenous Distortion: Theory of Opti- mum Subsidy Revisited." Quarterly Journal of Economics 102(4):903-10. - . 1995. "Political Economy of Trade Policy." In Gene Grossman and Kenneth Rogoff, eds., Handbook of International Economics. Vol. 3. Amsterdam: North- Holland. Schiff, Maurice, and L. Alan Winters. 1997. "Regional Integration as Diplomacy." CEPR Discussion Paper 1690. Centre for Economic Policy Research, London. Processed. Trefler, Daniel. 1993. "Trade Liberalisation and the Theory of Endogenous Protection." Journal of Political Economy 101(1):138-60. van Long, Ngo, and Antoine Soubeyran. 1996. "Lobbying for Protection by Heteroge- neous Firms." European Journal of Political Economy 12(1):19-32. Winters, L. Alan. 1996. "Regionalism versus Multilateralism." Paper presented at the Centre for Economic Policy Research (London) conference on regional integration in La Corunia, Spain, April 1996. Processed. Yeats, Alexander. 1997. "Does Mercosur's Trade Performance Raise Concerns about the Effects of Regional Trade Arrangements?" Policy Research Working Paper 1729. Policy Research Department, World Bank, Washington, D.C. Processed. A REVIEW ESSAY THE WORLD BANK ECONOMIC REVIEW, VOL. 12, NO. 2: 323-52 Half a Century of Development Economics: A Review Based on the Handbook of Development Economics Jean Waelbroeck The Handbook of Development Economics provides an unmatched perspective on the past half century of research in the field. This article reviews some of the significant findings of the Handbook and identifies areas of development economics not covered there. It describes the evolution of the field, beginning with the postwar Indian Con- gress consensus, which emphasized state planning. That paradigm was complemented in the 1970s by the dual economy paradigm and the work of the Latin American structuralists. As the costs of protectionism came to be recognized, many economies adopted an outward orientation with respect to both trade and foreign investment. This outward orientation later formed part of a new consensus on macroeconomic management, sometimes referred to as the Washington consensus. This consensus- more accurately dubbed the One World consensus-is now widely accepted through- out the world, as developing and industrial countries alike have embraced its emphasis on fiscal discipline, tax reform, outward orientation, privatization of state enterprises, deregulation, and the safeguarding of property rights. Despite 50 years of research, large gaps in the understanding of development remain. Adopting Sen's framework of beings and doings and focusing on collective action and the roles played by individuals may help to fill some of these gaps. The three volumes of the Handbook of Development Economics summarize a vast economic literature and provide an unmatched perspective on half a cen- tury of development research (Chenery and Srinivasan 1988, 1989 and Behrman and Srinivasan 1995). The contributions are thorough, and the spectrum of views is fairly presented, with contributions from both policymakers and academic economists. Development economics is a very applied science, which policymakers and academics have enriched. Politicians emphasize their differences but tend to think alike to a surprising extent. Some of the reason must be frequent contacts, the influence of the media, and each individual's keeping close watch of policy inno- vations elsewhere. These "consensuses," as I shall refer to them, do not exist at all times, but when they do, it is usually possible to characterize by one name the concept that inspired them. For example, the creation of a unified European Jean Waelbroeck is with the Universite Libre de Bruxelles and the European Center for Applied Research in Economics. Š 1998 The International Bank for Reconstruction and Development/ THE WORLD BANK 323 324 THE WORLD BANK ECONOMIC REVIEW, VOL. 12, NO. 2 market has been the focus of much that European Union governments have done in recent years. Williamson (1997) has drawn attention to the importance of consensuses. Academics disagree about everything, but there is an anchor in this swirling, and it is the agreement, so strong that it borders on intolerance, of an over- whelming majority of academics about what Kuhn (1962) has called a "para- digm," a shared view about assumptions, the discipline's frontiers, and the top- ics that merit attention. This article reviews some of the significant findings of the Handbook and identifies areas of development economics not covered there. The appendix pro- vides a list of the authors and chapter titles for the three volumes of the Hand- book. For a longer version of this article, see Waelbroeck (1998). Section I summarizes the ideas of Jawaharlal Nehru on economic planning and development. Section II provides an overview of the first quarter century of development economics. It describes the main ideas in the field during that pe- riod, including the dual economy development paradigm, the views on popula- tion, and other views. Section III looks at some of the questions raised about the early literature. Section IV discusses neoradical ideas, including dependency theory and Latin American structuralism. Section V describes the progress that has been made in economic theory, data, and empirical methods. Section VI examines the move toward outward orientation and its effect on North-South capital flows. Section VII describes structural adjustment and the One World consensus. Section VIII addresses some issues not addressed in the HIandbook, including Sen's concept of beings and doings, the theory of collective action, and the role of the entrepreneur. The last section presents some final thoughts on development economics. L. THE INDIAN CONGRESS CONSENSUS In their introduction to volume 3B of the Handbook, Behrman and Srinivasan call attention to the marvelous pages on economic planning that Nehru wrote while in jail in Ahmadnagar in 1944 and 1945 (see Nehru 1989). They set out the conclusions of a large group of businessmen, trade unionists, and officials of provincial governments who worked on a plan that could serve as a model for Indian planning after independence. What is striking is that, except for big busi- ness, all of the groups involved reached a consensus on economic policies. This Indian Congress consensus defined the program that Nehru implemented. The ideas it embraced seemed so obviously right that the leaders of other newly independent developing countries, many of whom had not read Nehru's book, adopted them as cornerstones of their policies. Most of the ideas of early development economics are there, laid out in 14 pages with the clarity and brevity of an extraordinary mind. Truly, as Srinivasan observes, those pages marked the beginning of development economics. Nehru's ideas can be summarized as follows: Waelbroeck 325 * Although Asia long represented the advancing spirit of man, standards of living in the industrial world (the North) shot ahead of those in the developing world (the South) over the past 150 years. * The industrial revolution transformed Europe. To experience a similar transformation the South must acquire the technology that made the European miracle possible. * Southern countries need to modernize their societies. Traditional modes of thought and institutions should be replaced, and the masses should be educated. * Industrialization could take place rapidly in the South, and the South could now acquire far better technology than that which propelled the industrial revolution. * Agriculture requires attention because it is so large and important and because the most abject poverty is found in that sector. * Because the marginal product of agricultural labor is zero, moving workers to the cities would produce a large boost in output. * Overcoming poverty is a fundamental task that cannot be achieved without rapid growth. But much can be accomplished by eliminating exploitive institutions. * Developing societies should be based on collective action rather than on the acquisitiveness of capitalist society. The influence of big business should be curbed. State enterprises and cooperatives could realize important scale economies. i The state should seize control of the economy and induce all citizens to participate in the effort. Clearly, Nehru was unaware of the success of the very different Japanese development model, which inspired the Four Dragons (Hong Kong, the Repub- lic of Korea, Singapore, and Taiwan [China]) consensus, which modified so pro- foundly our understanding of development. Nehru provided a fascinating account of his debate with Gandhi, who feared that the emphasis on high technology and heavy industries would lead to wide- spread unemployment. Gandhi started the debate on appropriate technologies; had he lived, his thinking might have provided a useful counterweight to Nehru's. II. THE FIRST QUARTER CENTURY OF DEVELOPMENT ECONOMICS The discipline of economics has changed radically since the end of World War II. Just after the war, doubts about the effectiveness of market incentives were widespread, and the new welfare economics emphasized the costs of mar- ket failures. Keynes himself announced that business people are driven by ani- mal spirits (a conclusion that Oxford Keynesians claimed was borne out by sur- veys suggesting that interest rates had no effect on the investment decisions of businesses). Wages were believed to be rigid downward; prices were believed to 326 THE WORLD BANK ECONOMIC REVIEW, VOL. 12, NO. 2 be set by adding standard markups to costs. Econometric research had demon- strated, many economists believed, that foreign trade elasticities were very low so that exchange rates had little effect on the quantities exchanged, and price elasticities for many commodities were so low that prices had little effect on the quantities exchanged. This skepticism over market incentives accounted, for the popularity of analytical approaches that took no account of prices. Economists felt comfortable basing policy analysis on such devices as input-output tables, Harrod-Domar models, and incremental capital-output ratios. Meanwhile, claims of extraordinary growth in the Soviet Union were taken more or less at face value. Only later did Gershenkron (1951) and researchers directed by Bergson identify how the data had been manipulated to distort ac- tual rates of growth. Most economists saw no reason why large state firms should be less efficient than private ones, since both were run by bureaucracies. That politicians govern in the general interest was taken for granted. The Dual Economy Development Paradigm The whole economy replacement concept-which suggests that successful development replaces traditional economies with economies based on northern technology-is the cornerstone of development economics. Its most visible mani- festation is the massive population shift from rural areas to cities. Ranis's (1:4) reformulation of the Fei and Ranis (1961) dual economy model analyzes the economic flows between industry and agriculture (where agriculture stands for the subsistence sector, in which wages are set according to sharing principles rooted in custom rather than market principles).' Agricultural wages exceed the marginal product of labor (which is greater than zero in this version of the model). Such an economy has considerable slack, whose elimination yields a large gain. Trade is discussed only in a brief section in Ranis (1:4) on the logic of import substitution policies, which notes that the few countries lacking natural resources may resort to export substitution by selling simple consumer goods abroad. Other chapters cite the suggestion of Soviet economist Preobrazhensky (1924) that because food output does not respond to prices, taxing agriculture provides a ready way of generating primitive socialist accumulation. Another view (still widespread among policymakers in the developing world) is that rural-urban migration is supply driven and that, in the absence of control, masses of people move to the cities, where they are driven to crime and political unrest by the lack of employment opportunities. Hoselitz (1953) and later Bairoch (197.5) contrib- ute to this thinking by comparing the experience of developing countries with that of industrial countries in the nineteenth century, suggesting that urban popu- lations in the developing world were outrunning industrial jobs. Williamson (1:11) shows that the idea is incorrect. Nurkse (1953) and Johnston and Mellor (1961) claim that the rural sector should be able to feed the cities and to absorb the cities' manufactured goods in 1. Citations from the Handbook are identified by their volume and chapter numbers. For example, Ranis (1:4) refers to volume 1, chapter 4. Waelbroeck 327 return. In their view, improved nutrition raises the productivity of workers, trig- gering a virtuous circle of growth. Empirical support for such a belief, it was later found, is quite weak. Population Growth as Market Failure Rapid population growth, driven by medical advances, was recognized early on as a policy problem. It would prove difficult to accumulate enough capital to catch up with the industrial countries even if population were stable. With rising populations, the task might prove impossible. Analysis showed the problem to be more subtle than had been thought. Boserup (1981) suggested that population growth expanded the food supply by causing endogenous technical change. Economists subjected the problem to the logic of welfare economics and showed that market failure is involved: population growth, determined by the actions of parents, is optimal only if parents take account of the interests of their progeny as well as their own, if information about contra- ception is available, and if property rights are well defined (see Sen 1:1, Birdsall 1:12, and Dasgupta and Maler 3A:39). Tragedy of the commons problems arise when property rights are poorly defined. Pecuniary externalities are a possible source of market failure (as, for example, when population growth reduces wages). The decision to have children is often based on self-interest, as Dasgupta and Maler (3A:39) note in their discussion of the demographic transition in rural settings, where having children (at least male children) may be the only way to provide for old age. Development loosens the bond between parents and chil- dren by offering children an Hirschmanian exit to the cities. In addition, unless runaway inflation destroys quasi-monetary assets, a wider range of low-risk sav- ings assets becomes available, which also makes child-bearing less attractive from this point of view. Many births reflect ignorance about birth control, as shown by the impact that the provision of information on birth control has on the birth rate. Evi- dence of market failure is strong. The measures taken to discourage births in China, Singapore, and India under Indira Gandhi may nevertheless be repug- nant to economists. Other Early Views The two-gap model of Chenery and Bruno (1962) also implies market failure, in the sense that market prices may differ from the scarcity prices that matter to society. A linear programming formulation reveals that the gap may be large. Foreign aid is apt to be extremely valuable; policies that seem costly may be socially very profitable. Lack of scale economies was believed to be a possible source of market fail- ure. (The licensing policy introduced in India after independence reflected this belief.) Countries with large domestic markets were assumed to enjoy an advan- tage over countries with smaller domestic markets. Perkins and Syrquin (2:32) find that large countries do indeed have a larger share of heavy industries (typi- 328 THE WORLD BANK ECONOMIC REVIEW, VOL. 12, NO. 2 cally characterized by scale economies) and grow faster than smaller ones. The latter finding could reflect the fact that many small countries are located in Sub- Saharan Africa, where growth rates are low. III. QUESTIONING THE EARLY LITERATURE It was 1979, and Ian Little and I were chatting. The conversation drifted to that favorite topic of academics, the award of the Nobel Prize, granted that year to Theodore Schulz and Arthur Lewis. "An obvious one I should think," said Ian. "Clearly, Lewis got it for stating that the marginal product of agricultural labor is zero, and Schulz for showing that it is not." A quip but a deep one. The initial model of development was flawed, as was Nehru's and other leaders' understanding of their countries. The developing world would have grown much faster had early errors been avoided. Schulz (1964) first questioned the belief that subsistence agriculture does not respond to market forces. In fact, a true subsistence sector exists almost no- where. As Rosenzweig (1:15) has shown, about 30 percent of the income of farm families is of nonagricultural origin. Schulz's findings were accepted with some hesitation by the academic community when he published them in 1964. Even today many policymakers in developing countries do not believe that suc- cessions of bad harvests may be the result of inadequate prices rather than of persistent drought. Little, Scitovsky, and Scott (1970) and Balassa (1971) identify a second error in the early work, the belief that developing countries have limited access to world markets. With the benefit of hindsight, we see what a strange idea it was. A fundamental assumption of development economics is that people everywhere have similar abilities. If the technology used in many industries is easy to ac- quire, why are developing countries unable to produce goods that satisfy con- sumers in the industrial world? It seems that a Nobel prize should be awarded to the policymakers who took a chance on trade. Why did the governments of the Dragons set out oln a course that, 40 years later, so many countries are trying to emulate? Japanese ideas had a decisive impact on Korean decisionmakers and on Taiwanese business leaders who spoke Japanese and understood why Japan had been so successful. A num- ber of Taiwanese policymakers were influenced by American ideas; the ideas of C. S. Tsiang were also influential. The Republic of Korea was subjected at an early stage to U.S. pressure to liberalize. After listening to the advice of northern development economists, Lee Kwan-yu of Singapore decided that their advice was wrong and set his country on its successful course. IV. NEORADICAL IDEAS I call the economists whose work is discussed in this section neoradicals be- cause they share a conviction that the aims of mainstream economics are mis- Waelbroeck 329 guided. The authors of the Handbook chapters that present this work have sym- pathy for that school but do not all belong to it. Dependency Theory and Exploitation Bardhan (1:3) and Evans (2:24) examine dependency models that describe how anonymous economic forces stifle national efforts to escape from the trap of underdevelopment. The best-known model is the Prebisch-Singer one, ac- cording to which differences in income elasticities of demand for primary prod- ucts and industrial goods, combined with the struggle for higher wages by orga- nized workers in the industrial countries, allow the industrial world to capture the benefits of productivity growth in the developing world. Both chapters note the weak empirical foundations of those models (and the logical inadequacies of the Prebisch-Singer model). Lewis's (1969) elegant neoclassical model of trade between England and tropical countries rigorously analyzes a situation in which such a transfer may occur, but its relevance is limited to the rare situations in which the South produces commodities (such as cocoa) that are not produced in the North. Taylor and Bacha's (1976) model describes a situation in which rich groups have a high marginal propensity to consume luxuries they produce, en- gendering feedbacks that widen inequalities (Taylor and Arida 1:6). The authors note, however, that the assumption about propensities to consume is contra- dicted by the empirical findings of Clark (1975). Exploitation often entails manipulation of the legal framework by ruling groups. Evans (2:24) discusses Roemer's use of the concept of the core of a game to characterize how the prohibition of specific contracts may induce exploita- tion. Nehru complained repeatedly about the British practice of preventing In- dian industrial projects from going ahead, a practice that narrowed the set of subcoalitions that Indians could form and changed the core of the game in favor of British businessmen. Using a different device, both Taylor and Arida (1:6) and Evans (2:24) use the concept of closure to show how political power enables a group to modify mecha- nisms in its favor. Closure is best known as a technical device that keeps general equilibrium models manageable (see section V). It can, however, also express the ability of a group to disable a mechanism to achieve its goals. "Maintained hypotheses about the direction of causality," Taylor and Arida conclude, "are the key to the results of most models described herein.... One causal structure may be more appropriate than another. However, the decision is almost always external to the models" (p. 189). Latin American Structuralism Latin American structuralism has drawn some of its strength from the hostil- ity of noneconomists to the realities of economic management, in particular to the conditionalities imposed by the International Monetary Fund (IMF) and the World Bank. Developing countries tend to balk at implementing structural ad- justment programs, arguing that investment cannot be cut back because it is 330 THE WORLD BANK ECONOMIC REVIEW, VOL. 12, NO. 2 needed to eliminate bottlenecks, that high real wages sustain demand, that de- valuation reduces purchasing power and is deflationary, and that high interest rates raise costs and cause inflation. Such arguments are incomplete rather than incorrect, because they take account of only the first-round effects of policy decisions that confirm what the common man wishes to believe. Such reasoning has led to shocking policy errors: it was not so long ago that the U.S.-educated prime minister of a large developing country reduced interest rates to combat inflation. In Latin America a school developed-at the universities and in the Eco- nomic Commission for Latin America-that sought to transform this thinking into a coherent theory. That literature is discussed in several chapters of the Handbook. Arida and Taylor (2:17, p. 863) provide the best exposition. The authors acknowledge that "summarizing these contributions ... is not pos- sible because so many 'effects' are involved." Together with a number of bi- zarre ideas, the structuralists identified important issues that remain unresolved even today. Before mainstream economists recognized the problem, for ex- ample, the structuralists noted that stopping inflation would inflict enormous losses on banking systems; the term structuralists is meant to highlight the very painful adjustment that solving Latin America's problems by orthodox policies would impose. What accounts for structural rigidities was never spelled out carefully, but it has indeed turned out that freeing economies from govern- ment interference has proved slow and much more painful in economic and even more in political terms than many had expected. I return to this topic in the last section of the article. Taylor's explanation of inflation inertia in terms of contracts seems a bit elementary today, but this phenomenon, too, is poorly understood. Seeking to avoid the pain of adjustment, Latin America experimented with a variety of quick-fix schemes, involving the manipulation of exchange rates, in- dexation schemes, or freezing of key prices, wage rates, or interest rates, meant to solve problems without requiring macroeconomic cutbacks and structural reform. All failed, almost all ending in catastrophe. Ignoring the structuralists' advice, Latin American governments have adopted the simple policies advocated by the IMF. These have produced results, albeit slowly. Krueger (3B:40) and Corbo and Fischer (3B:44) discuss the impact of those policies, finding that se- lection biases make it difficult to measure their costs or benefits but that there is no evidence that those policies have either been very costly or led invariably to a swift resumption of growth. Latin American structuralism is widely considered to have been a failure. It is possible, however, that economists of the next generation will look more kindly on the contribution of that school of thought. What the structuralists focused on, the economics of disequilibrium, is the weakest area of modern economics. It seems possible that by setting up a sound theory of disequilibrium, Latin American economists, using better analytical tools than their elders, could make a fundamental contribution to our discipline. Waelbroeck 331 V. PROGRESS IN ECONOMIC THEORY, DATA, AND EMPIRICAL METHODS Data have driven the progress of development economics-data and the imagi- nativeness of economists who realized that the theory was inconsistent with the empirical results. Development economics is indebted to the economists who built up the macroeconomic record of development. Syrquin's chapter (1:7) re- minds us of the insights owed to these pioneers. Deaton's chapter (3A:33) pro- vides a thorough presentation of the theoretical progress in the econometrics of microeconomic data and reviews what has been learned through experience about the control of selection and other biases. Deaton (3A:33) and Strauss and Thomas (3A:34) show that the quality of much of the macroeconomic data is very poor.2 In their introduction to volume 3B of the Handbook, Behrman and Srinivasan note that Summers and Heston assign a score of D+ to D to the quality of the data in 66 of 138 countries. Gross national product (GNP) growth rates are dangerously dependent on guesses about informal sector output, and it is difficult to overstate the weakness of data for Sub-Saharan Africa, which accounts for so many data points in regressions. Even apart from data problems, cross-country regressions must be viewed critically. As Syrquin (1:7) notes, cross-country regressions are reduced-form estimates of more complicated models and do not provide reliable evidence of causality. Unless complemented by deeper study at the country level, such re- gressions reveal little about development. Sample surveys are making an increasingly important contribution to devel- opment research, thanks partly to the initiatives of the Indian statistical school, which has also made major contributions to statistical theory. The World Bank's living standards surveys, which cover social as well as budget variables, have also become a major source of information on the development process. Surveys in developing countries appear to be as accurate as in industrial countries; with a few exceptions, this is hardly true of national accounts data. The number of such surveys has been rising, but too many have been terminated after a few years. The most influential data by far have come from what econometricians in the Cowles Commission have called experiments by nature. Principally in response to these natural experiments, a revolution has taken place in the paradigm that guides economic research. The excellent chapters by Krueger (3B:40) on post- war experience and by Corbo and Fischer (3B:44) on structural adjustment are particularly valuable. Gone is the confidence instilled by Keynes that macroeconomic equilibrium can be restored by simple devices; economists now understand that rational ex- pectations make economies difficult to control and that such phenomena as lack of credibility rob policies of their effectiveness. Policymakers are coming to un- derstand that although many price elasticities are low in the short run, relative 2. These chapters draw on Srinivasan (1994) and other sources. On the much discussed China/India growth race, see Timberg (1997), which comments on Srinivasan (1993). 332 THE WORLD BANK ECONOMIC REVIEW, VOL. 12, NO. 2 prices cannot be manipulated harmlessly. Long-term elasticities are often far higher than short-term ones. Agents have ways of eroding and eventually de- stroying controls through black markets, unexpected technical change, retalia- tory measures by foreigners, and massive speculative swings, so that policies that seemed safe become the source of escalating complications. The Dragons' natural experiment has shown that long-run trade elasticities are virtually infi- nite for dynamic countries that diversify exports as they expand them. Quantifying Theoretical Results by General Equilibrium Modeling Samuelson (1947) assigns theoreticians the goal of developing meaningful theorems that predict the sign of impacts and can be disproved by observations. Applied general equilibrium modeling allows the magnitude of impacts to be predicted as well. Informative chapters by Robinson (2:18) and by Gunning and Keyzer (3A:35) present the methodology. Gunning and Keyzer pay special at- tention to computational issues. Their chapter reveals that quantitative analysis can be applied to a wide range of theoretical problems. It is well to keep the limitations of the approach in mind. In addition to data problems, theoretical problems remain, chief among them the problem of clo- sure, first noted by Sen (1963), and the difficulties in modeling disequilibrium. Both problems reflect weaknesses of economic theory. Closure reflects the diffi- culty of conceptualizing situations in which markets are incomplete. The device bypasses the difficulty by replacing whole market segments with a few equa- tions. The widely used classical closure equation, for example, assumes that investment absorbs the savings available at full employment, replacing all capi- tal and the money market system. As for modeling disequilibrium, no theory of the business cycle exists that has both rigorous theoretical roots and is compat- ible with observed behavior. In a pivotal chapter, Stiglitz (1:5) illustrates how today's economic organiza- tion theory has the potential to deepen our understanding of markets. Econo- mists have long been puzzled over the so-called facts that improved nutrition increases productivity and that workers accept jobs in the informal sector when higher-paying jobs are available in the formal sector. Stiglitz uses efficiency wage theory in a general equilibrium context to examine whether these phenomena reflect market failure. In both cases, he finds, market failure is not at work. Since Stiglitz's study, empirical work has shown that neither phenomenon ex- ists: improved nutrition has only a small effect on the productivity of labor, while the Harris and Todaro (1970) assumption that urban migrants can secure formal sector jobs is belied by evidence that they are able to secure jobs only in the informal sector. The value of the chapter thus now lies in its methodology, which can be used to analyze other questions. Stiglitz also elucidates the puzzle of why sharecropping arrangements are so prevalent in poor countries, even though the incentives they offer do not moti- vate farmers to extract maximum output from their land. Does the practice re- flect exploitation or the lack of rationality of rural agents? Economists today Waelbroeck 333 believe that sharecropping is prized because it provides insurance in incomplete markets in which rents cannot be conditioned on realized harvest and selling prices. Stiglitz also accounts for the interlinking of contracts-the fact that farmers carry out a variety of transactions with the same individual. Because of informa- tion asymmetries, both parties prefer to know with whom they are dealing. Such knowledge requires frequent interaction. Both devices reflect optimal adapta- tion to market imperfections. Econometric Results The emphasis on individual rationality in the modern paradigm heightens the importance of the very careful ongoing work of analyzing microeconomic data. Progress in methods has been swift. Lack of space makes it impossible to do more here than convey the flavor of some of the results, whose scope is broaden- ing rapidly. Observations that stem from aggregate data are presented first. EDUCATION AND HEALTH. The finding that the rate of return to investment in education is high (Schulz 1961) is confirmed by sophisticated econometric methods. Strauss and Thomas (3A:34) stress the important point made by Schulz (1975) that the main contribution of education is to teach individuals how to process information in a rapidly changing world. Econometric results strongly confirm the multifaceted benefits that flow from the education of women, including improving the health and nutrition of families and the academic success of their children, as well as moderating population growth. Indicators of school quality turn out to be less significant than expected. With respect to health, the results are so far more limited. The impact of nutrition on productivity is far smaller than once thought; the elegant work on nutrition efficiency models would seem to be empirically irrelevant. Selection biases make it impossible to secure robust results with respect to benefits from investment in health facilities. PRIVATE SAVINGS AND ITS INTERMEDIATION. Early planners created market failures by nationalizing or limiting the freedom of banks and insurance companies and allowing runaway inflation to destroy savings. As Besley (3A:36) notes, what economists designate as trading over time and over states of nature is crucial to welfare. Who will take care of me in old age, who will pay my hospital bill, and how can I finance a tubewell are enormously important questions for individuals. Politicians who view banks and insurance companies only as sources of funds have caused great harm. Besley presents a theoretical overview, a discussion of the (scant) available empirical literature, and a study of how institutions such as the Raiffeisenkassen banks created in Germany in the nineteenth century and the Grameen Bank in Bangladesh have met specific needs. Besley cites Townsend's (1994) fascinating study of the informal system of insurance that sharing traditions have established in parts of India. He also discusses the countervailing role of the Susu men in parts of Africa, who enable 334 THE WORLD BANK ECONOMIC REVIEW, VOL. 12, NO. 2 market women to shelter their capital from the greed of relatives who, invoking sharing traditions, might ruin them. Why do savings rates differ so widely across countries? Evaluation of the Chilean pensions scheme, which is run by private institutions along accumula- tion lines and is being copied in several Latin American countries, could shed light on the issue. Most economists agree that the scheme has helped to increase the savings rate in Chile from 6 to 26 percent of gross domestic product. The success of the system is a striking demonstration of the importance of good policies for the trading over time that Besley emphasizes. Financial repression has diverted the natural course of development. Besley discusses Gershenkron's (1962) conjecture about the relationship between eco- nomic development and the growth of financial intermediaries-a relationship borne out by the data collected by Goldsmith (1969), which show that the ratio of the value of financial instruments to wealth rises as development proceeds. Now that financial repression has been reduced almost everywhere, a rapid catch- ing up process is under way, revolutionizing the financial sectors in the develop- ing world. TECHNOLOGICAL PROGRESS. Ranis (1:4, p. 89) describes the developing economy as "picking and choosing from the formidable shelf of technology." Rostow's stages of economic growth assume implicitly that technology is somehow embedded in capital goods, so that achieving an adequate investment rate is the key to development; the postwar fashion for buying turnkey plants reflected the belief that they would prove as productive as plants in the North. The view that development can be reduced to a capital accumulation process is questioned in Stiglitz's chapter (1:5), in which he examines the belief of such authors as Rostow (1965) that achieving a takeoff investment ratio brings about self-sustained growth. If lack of capital were the main cause of low productivity, the return on capital in the South should be much higher than empirical studies indicate. Using Arrow's suggestion that education both selects ability and com- municates knowledge, Stiglitz notes that estimates of the productivity of human capital are biased upward, so that education cannot generate the higher output required for the South to catch up with the North.3 Underdevelopment itself could be seen as a gigantic market failure in which agents in the South fail to use technology that is routinely used in the North; Stiglitz suggests that Schumpeter's entrepreneur is needed to correct the situation. 3. Stiglitz assumes a low (0.2) capital exponent and may overemphasize the bias implied by Arrow's (1973) selection effect. An article by Krugman (1994) has given visibility to results by Young (1994) that suggest that all of the Dragons' growth is accounted for by a colossal capital accumulation. The result is generated by a simple regression based on cross-country regressions of the type criticized above. The 0.45 elasticity of output with respect to capital is too high; technical progress is negative. Egypt, Pakistan, and Botswana rank at the top of the league, while Singapore ranks fourth from last-anomalous results that cast doubt on the calculations used to produce them. The implicit production function estimated by Young implies vastly higher returns to capital in the South than in the North, which are not borne out by the data. Waelbroeck 335 In a key chapter, Evenson and Westphal (3A:37) try to clarify the mystery of why some countries adopt technology and others do not. They describe technol- ogy as circumstantial and tacit. Firms must spend money to acquire the tacit knowledge necessary to make the invention work in its new environment. Pack (1:9) observes that foreign trade contacts are important. Strauss and Thomas (3A:34) note that education provides individuals with the ability to solve the many practical problems that arise when a new technology is introduced. The authors say little about the entrepreneur, a topic addressed here in section VIII. Poverty and Equity The problem of poverty was little addressed by early development econo- mists, although its importance was recognized. Nehru believed that rapid growth would pull up the poor with the rest of society but feared that this process would not be swift enough. Accordingly, the first Indian development plan placed a higher priority on poverty than on growth. The stand taken by McNamara in his 1973 Nairobi speech reflected the evolution that came to be. Lipton and Ravallion's pivotal chapter (3B:41) surveys the history of think- ing on poverty and policy. The trickle-down idea is confirmed, although there is much interest in the case of the Indian state of Kerala, which has demon- strated that much can be achieved in slow-growing states. Some early ideas have turned out to be incorrect or not useful. One is Kuznets's (1955) cel- ebrated inverted U, according to which development first increases differences in income distribution and later reduces them. The disastrous effect on the rural poor of financing development through primitive capital accumulation extracted from agriculture is now understood. The current belief that the poor suffer disproportionately from macroeconomic adjustment policies is not con- firmed (but is useful in stimulating efforts to reduce the effect that these poli- cies do have on the poor). Because of difficulties in targeting, poverty is difficult to combat, even though the poverty gap is not large in absolute terms: according to Lipton and Ravallion (3B:41), it amounts to about 1 percent of total consumption for the poorest fifth of the developing world's population. Excellent governance might help, if it could be achieved by the wave of a magic wand. Combating poverty by spending state funds is impractical, given the need to pay for development. Ahmad and Stern's chapter (2:20) points out, however, that budget tightness is in part governments' own mak- ing. Shifting part of the tax burden to land would redistribute income and mini- mize the distortionary effect of taxation. The power of landowners prevents this change from occurring, however. Lipton and Ravallion (3B:41) and Jimenez (3B:43) emphasize that much more could be done to impose user charges to offset the cost of roads and other utilities, which benefit mainly well-off individuals. Binswanger, Deininger, and Feder (3B:42) analyze the historical evolution of property rights in agriculture, which are often the legacy of brutal conquest and the manipulation of property rights legislation by landowners. Sweeping land reforms have often led to disappointing results. Work like theirs on the history 336 THE WORLD BANK ECONOMIC REVIEW, VOL. 12, NO. 2 of land relationships could prevent policy errors from being made when such rights are redesigned. Dasgupta and Maler (3A:39) examine rural property rights and poverty. They show that much can be done by controlling the gradual privatization of local commons, which contribute significantly to the livelihood of the poor. VI. OPENING THE ECONOMY TO FOREIGN TRADE AND INVESTMENT Several chapters in the Handbook examine the role of outward-oriented poli- cies, including liberal trade policy and openness to foreign investment. Outward Orientation According to trade theory, free trade is optimal for countries that are small in international trade so that their import and export prices are given. Balassa (2:3 1) points out that the spectacular increase in trade of the Dragons has not harmed their terms of trade; hence the assumption holds. The export pessimism of the early days is discredited and not questioned anymore. Yet the debate goes on, the dispute now being about the impact of outward-oriented trade policies on growth. What trade theory teaches us can be set out in terms of a so-called welfare triangle; the triangle's height equals the distortion induced by the tariff, while its base equals the impact on the quantities traded. As a convenient approximation, demand and supply curves are assumed to be linear so that, to use the language of geometry, the triangles corresponding to different protection levels are simi- lar. (Experiments of mine have taught me that back-of-the-envelope calculations using welfare triangles provide a surprisingly good approximation of those pro- duced by large nonlinear general equilibrium models.) A first conclusion is that the welfare cost of protection is proportional to the square of the tariff-induced distortion (hereafter, the distortion). The cost of a small departure from free trade is second-order small; for example, almost all of the welfare gains secured by industrial countries as a result of the Uruguay Round stemmed from cuts in the high protection of agriculture and a few service sec- tors. Proponents of protection often say that in the nineteenth century, France and Germany did not suffer visibly from their high tariffs. In the 1970s, protec- tion levels in many developing countries reached levels that amounted to qua- druple those of nineteenth-century France and Germany; these would imply a welfare cost that was 16 times larger. Closer examination shows that this exor- bitant protection was riddled with holes as particular importers were granted exemptions. This generated, however, a climate of corruption that worsened income distribution and a climate of cronyism that harmed governance. General equilibrium models have been used to assess the welfare cost of pro- tection. They suggest that the welfare cost of protection is significant but not exorbitant. A volume edited by Srinivasan and Whalley (1986) presents results under standard assumptions. As Rodrik notes (3B:45), the estimated welfare Waelbroeck 337 losses do not exceed a couple of percentage points. This is far from negligible but the stakes of development are on a different scale. India's GNP, for example, grew on average by 4 percent less than Korea's from 1980 to 1995. Accepting the 2 percent estimate, India's protectionism would account for a bare thirtieth of the growth gap between the two countries. The problem for the models-and for orthodox trade theory-is that the facts that they describe seem to be belied by the facts. As mentioned by Krueger (2:31), Korea's exports to GNP ratio, for example, grew from 3.8 to 31.8 percent from 1960 to 1990. None of the models in the Srinivasan-Whalley volume is capable of generating such a result without resorting to arbitrary changes in coefficients or other tricks. The small welfare gains may reflect these models' lack of real- ism. This has encouraged model builders to embed in general equilibrium mod- els components that take into account "dynamic" gains stemming from the stimu- lus of keener competition, increasing returns, and reductions in rent-seeking costs. Rodrick (3B:45) discusses surveys of such work by Balassa that imply welfare losses of more than 5 percent of GNP in highly protectionist countries. Yet even on this basis, differences in trade policy would account for less than a tenth of the growth gap between India and Korea. The conclusions change radically when, setting aside theory and the models that embody its logic, the analyst uses econometrics to study growth rates. There exists by now a very large number of studies of this type, which both Balassa (2:31) and Rodrik (3B:45) discuss. The first concludes that the impact of protec- tion on growth rates is quite large; the second finds this work "fragile." Having looked at many such studies, I find him too skeptical. Forget econometrics and international trade theory, some might say, look at the numbers. Even those are hard to interpret, however. No one calculates wel- fare; the only data are GNP figures at constant, tariff-distorted domestic prices. Helleiner (2:27) cites studies by Lall and Streeten and by Encarnaci6n and Wells showing that between 25 and 45 percent of investments by multinationals low- ered welfare in the host country. That investment that increases GNP at constant domestic market prices may be welfare-reducing is no theoretical curiosity. The contradiction may account for the fact that trade liberalization in highly pro- tected countries in Latin America and in the former Soviet bloc, which often resulted in a protracted period of slow GNP growth, was so popular that the policy change was irreversible. North-South Capital Flows Since volume 3 of the Handbook was published, a remarkable shift in thinking has taken place in the South regarding foreign control of enter- prises. In a number of developing countries, foreign banks account for a sizable share of loans and deposits. Governments have learned to value the economic role of small stockholders, and they have fostered the growth of stock exchanges, opening them up to foreign investors. Privatization has pro- vided these exchanges with a large supply of sound and well-known stocks, 338 THE WORLD BANK ECONOMIC REVIEW, VOL. 12, NO. 2 helping them to attain critical mass. Foreigners have more easily obtained permission to develop natural resources. Multinationals are viewed with less suspicion than in the past. These changes may accelerate development. It is surely desirable that more of the pool of northern savings in industrial coun- tries should flow to the South, as it did in the nineteenth century, when West- ern European capital contributed to growth in the Americas, Eastern Eu- rope, and Oceania. Cardoso and Dornbusch (2:26) present valuable data on the history of capital flows. The accumulated stock of postwar flows from North to South amounted to 30 percent of U.S. GNP in 1987. Of this, only 15 percent took the form of direct investment; 85 percent was loans, of which official credits accounted for 39.4 percent. There is room for some growth in foreign investment, but a large increase is unlikely. Investing abroad has always been risky, and Cardoso and Dornbusch remind us that even in the pre-World War I era, debt crises were not unusual. There is reason, however, to believe that much of the postwar North-South capital flow was used ineffectively. The overwhelming share that went to public borrowers was largely wasted. Large amounts funded budget deficits, directly or indirectly, by financing investments that governments meant to carry out any- way, thus freeing up funds that could be offered to politically influential indi- viduals or groups. Even money that was invested in projects that would not otherwise have been financed was often squandered, because it was directed to state enterprises, which, as Krueger (3B:40) notes, have with few exceptions been extraordinarily inefficient. The use of capital flows has probably improved since 1987. Recent years have witnessed the birth of a significant but unstable flow of capital into emer- gent stock exchanges, which has facilitated the privatization of public enter- prises. Investment by multinationals has increased, particularly in mining, bringing in advanced technology and organizational know-how. Governments have used the borrowed funds better, cutting their deficits and improving budgetary control. Still, the high risk associated with investment in the South remains. The 1994 Mexican crisis, which struck a country that Corbo and Fischer (3B:44) viewed as a successful adjuster, was a spectacular setback. In 1997 four of the East Asian miracle countries-Indonesia, the Republic of Korea, Malaysia, and Thai- land-were hit by a financial crisis that few had foreseen. North-South capital flows are problematic, and domestic savings will remain the prime motor of southern growth. Nevertheless, North-South capital flows should rise and be used more wisely. Good national and international gover- nance is the key. Does the Handbook, written before the Mexican and East Asian crises, present the analytical tools required to analyze those crashes? Are those crises a signal that outward-oriented policies cause instability, as their opponents have always claimed? Waelbroeck 339 The first tool, examined in the chapters by Eaton (2:25) and Corbo and Fischer (3B:44), is the rule that borrowing is sustainable if the country's rate of growth exceeds the real interest rate (see Domar 1950). As Eaton notes, the reasoning assumes that today's rate of growth and interest rate will last forever. This analysis could have provided a warning to the IMF when, before the 1982 debt crisis, real interest rates rose steeply while growth in parts of the developing world flagged. The second tool is Krugman's (1979) theory of foreign exchange crises, dis- cussed by Corbo and Fischer, which asserts that these crises begin as soon as it becomes clear that the continuation of current policies will exhaust foreign ex- change reserves at some point in the future. The message is that governments must take into account that market forces anticipate the future and promptlv modify policies that are unsustainable (an adjective that must be interpreted more broadly than in the Domar theory). Corbo and Fischer use the third tool, the Salter (1959) model, to study the policies that make it possible to deal with crises caused by external shocks or policy errors. Edwards and van Wijnbergen (2:28) also use the model, embed- ding it in an elegant framework that uses disequilibrium theory. The model im- plies that foreign borrowing, often undertaken in response to a nascent disequi- librium, causes a temporary improvement in the terms of trade and so engenders complacency. This delays the government's response and makes the crisis more painful than if action had been taken in time. The model also identifies the mecha- nisms through which expenditure cuts and expenditure and production switch- ing through relative price changes achieve adjustment. The Handbook does not set out what contract theory teaches us about the causes of bank crises of the type that heightened the 1982 crisis in Chile, con- tinue to hold back Mexico's recovery, and are a major cause of today's difficul- ties in Indonesia, the Republic of Korea, and Thailand. Such crises have oc- curred in industrial countries, where they have entailed large budget costs but not caused macroeconomic imbalances; they seem to be caused by the changing economic role of banks. Were outward-oriented policies the cause of those mishaps? Certainly out- ward orientation does not insulate countries from such blows. In 1982 both inward- and outward-oriented countries were badly hit. The difference came later, as the outward-oriented East Asian countries, which had good macroeco- nomic management, were able to rebound much faster than the inward-oriented countries (mostly in Latin America), which did not reform their macroeconomic policies and shift to a more open trade stance. Behind it all is presumably able governance, which produced both the good macroeconomic management and the outward-oriented policies that made the trade balance responsive to cur- rency devaluation. The cause of these financial crises is the rising power of international capital markets in today's global economy. Serious policy errors were made by the gov- ernments hit by the current East Asian crisis and by the government of Mexico in 1994, but it is doubtful that these errors justified the "visceral, engulfing 340 THE WORLD BANK ECONOMIC REVIEW, VOL. 12, NO. 2 fear" that, in Alan Greenspan's characterization, spread among speculators. Faith in the rationality of agents is a component of the paradigm to which most econo- mists adhere today, but a list of episodes is beginning to accumulate in which markets lose contact with reality, as operators let themselves become mesmer- ized by fashionable ideas (examples include the Japanese assets bubble, the dol- lar boom in the early days of the Reagan administration, and the speculation against the pound and the French franc in 1992). Such episodes will recur. The fears they will engender will reduce the flow of capital from North to South and limit the extent of financial liberalization that developing countries would be wise to adopt. VII. STRUCTURAL ADJUSTMENT AND THE ONE WORLD CONSENSUS The Indian Congress consensus had serious flaws, although it made sense for historical reasons. It is now obsolete. A new consensus has been formed that reflects a worldwide change in views. In every developing country, a vigorous policy debate is carried on that draws on the memory of past problems and the example of successful countries. That debate has shifted thinking in the South from the Indian Congress consensus to the contemporary one. In volume 3, several authors comment on this convergence of policies, which crystallized af- ter the 1982 debt shock. In a well-known paper, Williamson (1990, updated in 1997) labels this set of beliefs the Washington consensus. The tenets of the Washington consensus can be summarized as follows: * Governments should exercise fiscal discipline, obviating the need for an inflation tax. * Public expenditure priorities should be shifted from politically sensitive areas to neglected fields with high economic returns and the potential to improve income distribution. * Tax reform should broaden the tax base and reduce marginal rates. Ways should be found to tax flight capital. * Financial markets should be partially liberalized. * Exchange rates should be kept competitive. * Import quotas should be replaced by tariffs. * Barriers impeding the entry of foreign firms should be removed; foreign and domestic firms should compete on equal terms. * State enterprises should be privatized. * Governments should remove regulations that are not justified by such criteria as safety, environmental protection, or prudential supervision of financial institutions. * Property rights should be safeguarded. Williamson's list of policy reforms was presented at a conference in Latin America. The reaction by Latin American economists revealed that there had Waelbroeck 341 indeed been a sea change in Latin American attitudes; Williamson's list of poli- cies "that Washington could agree with" in effect described common ground between Washington and Latin American economists. Williamson is said to have some regret over having called this set of beliefs the Washington consensus-and well he might, as will be shown. The Wash- ington designation is not groundless. Three years before the debt shock, the World Bank transformed itself into a major provider of adjustment credits, while the IMF began to condition its lending to a greater extent than before on market-oriented reforms. In 1979 the World Bank, in close cooperation with the IMF, decided to devote a substantial portion of its resources to struc- tural adjustment loans in support of policy reform in the developing world. Pressure from the World Bank and IMF must have been effective, because the Bank and the IMF were the only sizable providers of funds to victims of the debt crisis.4 Corbo and Fischer (3B:44) survey empirical appraisals of adjust- ment lending, all of which, as they note, suffer from econometric selection biases. They also report case studies describing the adjustment experience of Chile, Ghana, and New Zealand. Krueger (3B:40) and Rodrik (3B:45) also evaluate the Bank's adjustment lending. The term Washington consensus is nevertheless regrettable, because it rein- forces the common man's delusion that the world is run by a few thousand bureaucrats and politicians around Washington's Mall. Words matter in poli- tics. Moreover, the term is incorrect. The Bank and IMF are intellectual sponges rather than leaders. Economists from both institutions absorb the latest aca- demic ideas, including those from the South. Both institutions' policy of recruit- ing mainly staff with experience brings in practical men and women, some of whom have held important posts abroad. During Bank and IMF missions, staff exchange ideas with civil servants and politicians in the South-a fruitful ex- change for both sides. The best illustration of the sponge theory is China. No conditionality inspired Deng's reforms, and it is the World Bank and IMF econo- mists who are trying to learn from China's experience to enrich their dialogue with other countries. It is not obedience to Washington institutions that kept Mexican and Argen- tine reforms on track in the face of severe difficulties in the recent past. India's 1991 balance of payments crisis is history, but reform has continued through several changes of government, although by now large foreign exchange reserves insulate the country from outside pressures. The domestic roots of the new con- sensus-referred to here as the One World consensus-are illustrated by the conversions of such radical politicians as the late Michael Manley of Jamaica, Jerry Rawlings of Ghana, Yoweri Museveni of Uganda, and Laurent Kabila of the Democratic Republic of the Congo, who fought with Che Guevara to up- hold Patrice Lumumba's ideas. 4. The power of conditionality should not be exaggerated. As Edwards (1989) and Corbo and Fischer (3B:44) note, compliance has been low; once devaluation and budget cuts have fended off a crisis, who can force borrowers to put into effect commitments they dislike? 342 THE WORLD BANK ECONOMIC REVIEW, VOL. 12, NO. 2 The world has been shifting to a consensual system in which countries ac- cept international obligations more willingly and pay greater heed to interna- tional public opinion. The Uruguay Round, whose scope India and Brazil ini- tially wished to limit, has stiffened the trade discipline that members of the World Trade Organization must respect. Supported by public opinion, the Bank, the IMF, and other donors are becoming more forthright in pressing coun- tries to fight corruption and promote democracy, while a variety of nongov- ernmental organizations are fostering an awareness that government and pri- vate decisions have worldwide implications. The world government described by Streeten (2:22) is not in sight, but the world is moving toward it. It is cer- tainly remarkable that when the euro is in place, the European Union will perform all of the functions that Streeten thought would be deemed "unrealis- tic and utopian."5 In the long run the European Union could expand as far as Vladivostok, NAFTA could include Central and South American countries, and the Association of South East Asian Nations (ASEAN) could at last acquire a meaningful economic dimension. Mercosur (a regional trade arrangement among Argentina, Brazil, Paraguay, and Uruguay) appears to be establishing solid roots and is open to new members. Streeten identified very natural lines along which trade integration tends to induce a transfer of functions to a com- mon body. His world state might come into being through the cooperation and later the merger of a variety of regional organizations. VIII. BROADENING THE FRAMEWORK OF DEVELOPMENT ECONOMICS Clearly by design, the Handbook opens on an agnostic note with respect to conventional wisdom. The first three chapters-by Sen, Lewis, and Bardhan- openly or implicitly criticize traditional thinking. Sen sets out an alternative vi- sion of the development process, Lewis criticizes economists' long disregard of development, while Bardhan warns of the limited usefulness of traditional eco- nomics. Taylor and Arida's two chapters (1:6 and 2:17) are no more orthodox. Other chapters also challenge standard approaches. This section examines a few of the important challenges to conventional thinking in development economics. Sen's Concept of Beings and Doings Sen (1:1) proposes a new way of evaluating development (see also Dr&ze and Sen 1989). The key innovation is that economics should be concerned not with commodities but with actions, represented by the twin concepts of capabilities and functionings. Sen argues that individuals are interested in beings and do- ings. Neither a starving man nor a fasting one eats; both function in the same way. But the two men have completely different capabilities: one has the free- dom to eat, while the other does not. 5. A central bank; redistribution of incomes among members; control of investment in long-lasting plant and equipment (the coal and steel treaty and the control of subsidies); foreign aid; and control of primary commodity prices (agricultural products, coal, and steel). Waelbroeck 343 Sen's formulation captures the essence of human experience better than tradi- tional economics: what one has been and done is more important than the com- modities one has consumed. The formulation also opens our discipline to con- cerns that it does not normally recognize, providing a richer picture of development than the orthodox view. Through its link with freedom, the concept of capabilities encompasses the concerns of the Handbook's neoradicals. Attributing value to freedom recog- nizes the concern of dependency theory as it examines how countries in the periphery may feel helpless as the fruits of their efforts accrue to countries in the center. Other examples of the usefulness of Sen's concept include Roemer's (1982) use of the concept of the core to describe economies in which prohibitions limit the capabilities of some agents and Taylor and Arida's (1:6) use of the concept of closure to express restrictions on economic freedom. Structuralists examine countries that are trapped by rigidities of their structure. The concept of capabilities also sets the discussion of interlinked contracts in a novel perspective. Such contracts are an optimal adaptation to asymmet- ric information. The capabilities concept highlights the fact that agents with many partners negotiate with agents whose choices are far more restricted. Replacing interlinked contracts with standard contracts has a liberating effect. Remittances from migrants strengthen the bargaining situation of their fami- lies and eliminate the fear that crop failure might make them dependent on powerful neighbors. The perspective is also useful for analyzing the costs of bureaucratic overregulation, which exposes villagers to possibly corrupt local bureaucrats, who can exploit them like the zamindars who served the Mogul Empire. Sen's concept also casts the Handbook's findings on health and education in a new light. Stating, for example, that education enhances the capabilities of women and that health services protect individuals from infirmities that would prevent them from playing respected social roles is an enlightening way of view- ing their benefits. Those notions could have been expressed within the tradi- tional confines of economics. The added value of Sen's contribution is that it allows problems to be viewed from a new and interesting perspective. The Theory of Collective Action "I am for myself, for if not, who is for me? And being for myself, who am I?" wrote Rabbi Hillel in the Talmud. Hillel's question underscores a gap in the current paradigm of economics and reveals the usefulness of Sen's terminology. The standard approach in microeconomic theory that takes account of individu- als' concern for others by injecting altruistic terms into utility functions is artifi- cial. In contrast, generosity and the respect earned through giving are natural components of Sen's beings and doings. Man is a social animal, as Edward Wilson (1975), the father of sociobiology, points out. The colossal ant hill that mankind has erected is not sustained by prices alone; concern for our fellows is necessary to its functioning. 344 THE WORLD BANK ECONOMIC REVIEW, VOL. 12, NO. 2 Both Olson's (1965) work on collective action, which Yifu and Nugent (3A:38) discuss, and the ideas of Hirschman (1970) deal with social behavior. Olson examines how the temptation of group members to free ride can be eliminated by punishing free riders according to agreed rules. Invoking Sen's beings and doings extends the range of punishments and rewards that can induce group cohesion; the desire for honor and the fear of shame, so readily expressed in terms of Sen's functionings, fosters the cohesion that, for example, wins strikes. Olson's concept has come to be associated too closely with his work on inter- est groups. Collective action is an element of all political and market institu- tions, of which the state is as much an example as trade unions. It plays a role in firms (witness the emphasis on firm culture). The Catholic Church, a powerful economic institution, represents collective action as does the Grameen Bank's weekly meetings, which feature rituals intended to enhance cohesiveness among borrowers. Extended families, which play such a crucial role in the developing world, also represent collective action. The effectiveness of collective action is tacit and circumstantial, to borrow Evenson and Westphal's term-tacit because even the members of a group can- not predict how colleagues will react to a challenge; circumstantial because it coalesces most strongly in the face of an exceptional challenge. It is both poten- tially powerful and fragile. Collective action has three characteristics. The perception of a threat to the group or to its honor is important (the threat rule); cornered groups are hard to defeat. A sense of fairness is essential; participants in collective action want clearly defined rules that detect free riders and ensure that their own actions are judged fairly (the fairness rule). Finally, circumstances are important. Trust in a leader matters, as does the memory of past success or failure (the circumstantial rule). The emphasis on fairness in the public debate on trade policy has puzzled trade economists, who note that theory does not account for the strength of public support for the most-favored-nations clause. Public support for that clause and for antidumping duties is accounted for by a preference for transparent rules and assurance that they are respected. The World Trade Organization needs and has a judicial system, which is becoming more elaborate. Also worthy of note is the connection between the threat rule and Corden's (1974) conservative welfare function, a concept that reflects his observation that, to be successful, trade reform must be designed so that no group suffers severe losses. The same concern accounts for the success of Chenery's (1974) concept of redistribution with growth, praised by many Handbook contributors who accept instinctively that poverty policies should reserve for the poor a share in the fruits of growth rather than a share in wealth itself. As Vogel (1991) and others have noted, the Four Dragons performed their economic miracle under foreign threat; now that the foreign threat has disap- peared, some of these economies are suffering. Collective action produced the cooperative spirit that made their populations submit to state guidance; imple- menting the methods of the Republic of Korea's Park Chung Hee would almost Waeljbroeck 345 surely have failed elsewhere. Examples abound of countries that have achieved remarkable success when faced with great national danger. The prosperity of Athens was established when a third Persian invasion seemed possible; Venice prospered as it won a 16-year war against Turkey; and the golden age of the Netherlands coincided with an 80-year war against Spain, the most powerful nation in Europe. Governance in the Republic of Korea and Japan has not been as prescient as is often said: only in 1997, for example, did Japan close coal mines that pro- duced fuel at three times the world price. Krueger (3B:40) reminds us that the success of Korea's heavy investment in steel, heavy chemicals, and automobiles was not remarkable (the current spate of bankruptcies there makes her judg- ment prophetic). Hard work, induced by a sense of national danger or destiny, not governance, produced remarkable results. Collective action accounts for the effectiveness of Kornai's (1990) hard bud- get constraint. "When a man knows he is to be hanged in a fortnight, it concen- trates his mind wonderfully," Samuel Johnson said. It is the willingness of all to set aside personal concerns that accounts for the often spectacular recovery of enterprises on the verge of bankruptcy. This mechanism also accounts for the virtues of outward orientation. In the import-substituting days, Chilean compa- nies faced no true hard budget constraints, because their few domestic competi- tors were partners in a cartel, while the government could be relied on to hike protection to keep out foreign competition. The Republic of Korea's govern- ment was very interventionist, but it granted subsidies reluctantly; support was provided mainly through loans from banks that were closely controlled by plan- ners. Exports were large, and, knowing that subsidies might trigger antidump- ing duties in key markets, firms knew they must be competitive. Corbo and Fischer draw attention to the excruciating slowness of structural adjustment. It was deep insight that led the Latin American structuralists to predict that running the economies of their countries on market lines would be quite difficult, inducing them to devise other ways of solving problems. Their recommendations did not work, and now everywhere in Latin America govern- ments are reducing their interference in markets. This is producing results only slowly, and progress has been marred by unexpected setbacks. Adjustment to the market has proved even more difficult in the former Soviet bloc, especially in the core countries of the former Soviet Union. Nehru too knew that modernizing society would involve profound adjust- ments in behavior, as he predicted that "In India especially, where we have been wedded far too long to past forms and modes of thought and action, new expe- riences, new processes, leading to new ideas and new horizons, are necessary. Thus we will change the static character of our living and make it more dynamic and vital, and our minds will become active and adventurous" (Nehru 1989: 408-09). Here also, Sen's perspective is useful in providing a framework for consider- ing how roles, a key component of the beings and doings that he emphasizes, 346 THE WORLD BANK ECONOMIC REVIEW, VOL. 12, NO. 2 make the social structure more rigid and hamper transition to a new organiza- tion of the economy. Human capital theory is also important. In adolescence, all individuals compare different roles that they will play as an adult, which they prepare for mentally and by investing in training. That role can be changed later, but only at a cost. There is the sunk cost of the training required. Success also involves fashioning appropriate attitudes and building up the credibility that makes this personal discipline effective; here Sen's doings and beings frame- work is relevant. An efficient market economy can function only if served by businessmen, labor union leaders, civil servants including judges, the police system, and tax officials, who understand its mechanisms and respect its ethics. Rodrik's "gov- ernance" will not work otherwise. Students of the transition from socialism are wont to remark that those who were more than 30 when the Soviet bloc broke up cannot adapt fully to the new system. Kornai (1990: 52, 54) expresses this by stating that "embourgeoisement is a lengthy historical process." As implied by the citation of Nehru, flexibility in roles is very beneficial. It is good to start from a situation where individuals are torn from their roots: Ja- pan, with its large number of demobilized soldiers and repatriates just after the war; Korea, where a quarter of the population were refugees in 1953 (Vogel 1991: 43); Hong Kong, with its large inflow of immigrants including ruined Shanghai industrialists; the arrival of troops and other migrants into Taiwan (China). Last but not least, Mao created a gigantic blank page through the cul- tural revolution, which forced tens of millions of people to migrate to the coun- tryside and caused vast internal migration flows. It is interesting that the Ger- man postwar miracle also coincided with large movements of uprooted individuals. And of course many examples abound in history of the dynamism of immigrant societies, from Carthage and the Greek colonies to the United States. The Role of the Entrepreneur Schumpeter's (1934) entrepreneur fits neatly into Sen's scheme of beings and doings and illustrates the importance of role selection. Being an entrepre- neur requires abilities and character traits, which Schumpeter describes in de- tail, concerned that readers might not appreciate what sort of person drives development. The market for entrepreneurs should receive much more attention from de- velopment economists and policymakers. In China the key element of Deng's reform program was the creation of an environment in which individuals could become rich by competing in the market and earn esteem to boot. In India the post-1991 reforms reduced the bureaucratic obstacles that made it difficult to set up new ventures. In both countries, supply responded strongly to improved incentives. Schumpeter's theory is well known. The engine of development is the entre- preneur, who conceives a new combination of factors of production that is more Waelbroeck 347 effective than current ones and may bring in large profits. Innovations may be small; the farmer who discovers that growing tomatoes is profitable in his vil- lage is an entrepreneur. The funds required are usually borrowed. The profit is eventually eliminated by imitation, returning the system to its state as a static circular flow. Neoclassical theory has struggled to model Schumpeter's vision of develop- ment. Grossman and Helpman (1991) describe North-South competition in in- novation, but their model is so hamstrung by marginal conditions and by the constraint of balanced growth that it does not capture the richness of Schumpeter's concept. Evenson and Westphal (3A:37), who do not let themselves be tied down to the procrustean bed of formal theory, come much closer. Other Schumpeterian models include Caves's (1983) model of the product cycle. The work that won Douglas North the Nobel prize is close to Schumpeter's thinking, although North does not spell out the connections (North and Thomas 1973). Schumpeter's model needs updating. It is not clear whether he was con- scious of the transactional character of innovations, which both North and Evenson and Westphal emphasize. Schumpeter's model relates to North's em- phasis on the large returns to scale that characterize the transaction sector of the economy. According to North, these returns account for the cumulative character of development as an initial reduction in transaction costs induces a cascade of innovations. Schumpeter, who was thinking of industrial countries in the .1920s, did not consider the possibility that the entrepreneur's profit could be plucked from his hands. This happens routinely in the South, where gatekeepers (the officials who stand between willing buyers and sellers) abuse their authority to appropriate a slice of entrepreneurial profits. If it is successful, the fight against corruption, which the World Bank and IMF have taken up, will boost growth as well as promote justice. Competition also matters. As Basu (1989) notes, the village lender, if able to operate as a discriminating monopolist, can also skim off entrepreneurial profit (see Besley 3A:36). So can agents who control such resources as water or a right- of-way. IX. FINAL THOUGHTS Development economics has made remarkable progress in the past 50 years. There is greater dominance than formerly by one school of thought, but the range of that school's research has become much broader. Mainstream eco- nomics has hardened, as its proponents have learned to use data more care- fully and to reason more rigorously. Much early postwar work seems sloppy in retrospect. The policy message has been turned upside down. Gone is the idea that devel- opment means industrialization and that the main policy problem is to manage the interface between country and city. Today development is viewed as an inte- 348 THE WORLD BANK ECONOMIC REVIEW, VOL. 12, NO. 2 grated transformation, of which urbanization and industrialization are but two components. The expansion of foreign trade has become a central issue. Eco- nomic organization theory has taught us to view traditional institutions with far more understanding than in the past; blind imitation of northern institutions may be counterproductive. More than ever, development is seen as a whole replacement process, whose key is to master northern technology. Such technology is now understood to be both simpler and more complex than once thought-simpler because much tech- nology is uncomplicated, more complex because even simple technology requires ingenuity and a costly investment in adaptations. The most important change has been the radical shift in economists' view of market agents and policymakers. Gone are the days when policy advice was directed primarily at planners. Policymakers are utility maximizers, too; the employees of state enterprises coalesce into powerful interest groups that block efforts to raise productivity. The new thinking is sometimes challenged by criticisms that highlight the somewhat vague concept of governance, ac- cording to which the task of economists is to help design a system of inter- acting state and private institutions that, led by the state, cooperate in achiev- ing social goals. Whether solid theory will come out of this line of thinking remains to be seen. APPENDIX. LIST OF AUTHORS AND CHAPTER TITLES FOR THE THREE VOLUMES OF THE HANDBOOK Volume 1. Chapter 1. Amartya Sen, "The Concept of Development" Chapter 2. W. Arthur Lewis, "The Roots of Development Theory" Chapter 3. Pranab Bardhan, "Alternative Approaches to Development Economics" Chapter 4. Gustav Ranis, "Analytics of Development: Dualism" Chapter 5. Joseph E. Stiglitz, "Economic Organization, Information, and Development Chapter 6. Lance Taylor and Persio Arida, "Long-run Income Distribution and Growth" Chapter 7. Moshe Syrquin, "Patterns of Structural Change" Chapter 8. C. Peter Timmer, "The Agricultural Transformation" Chapter 9. Howard Pack, "Industrialization and Trade" Chapter 10. Mark Gersovitz, "Saving and Development" Chapter 11. Jeffrey G. Williamson, "Migration and Urbanization" Chapter 12. Nancy Birdsall, "Economic Approaches to Population Growth" Chapter 13. T. Paul Schultz, "Education Investments and Returns" Chapter 14. Jere R. Behrman and Anil B. Deolalikar, "Health and Nutrition" Chapter 15. Mark R. Rosenzweig, "Labor Markets in Low-Income Countries" Chapter 16. Clive Bell, "Credit Markets and Interlinked Transactions" Waelbroeck 349 Volume 2. Chapter 17. Persio Arida and Lance Taylor, "Short-Run Macroeconomics" Chapter 18. Sherman Robinson, "Multisectoral Models" Chapter 19. Irma Adelman and Sherman Robinson, "Income Distribution and Development" Chapter 20. Ehtisham Ahmad and Nicholas Stern, "Taxation for Developing Countries" Chapter 21. Lyn Squire, "Project Evaluation in Theory and Practice" Chapter 22. Paul P. Streeten, "International Cooperation" Chapter 23. Christopher Bliss, "Trade and Development" Chapter 24. David Evans, "Alternative Perspectives on Trade and Develop- ment" Chapter 25. Jonathan Eaton, "Foreign Public Capital Flows" Chapter 26. Eliana A. Cardoso and Rudiger Dornbusch, "Foreign Private Capi- tal Flows" Chapter 27. G. K. Helleiner, "Transnational Corporations and Direct Foreign Investment" Chapter 28. Sebastian Edwards and Sweder van Wijnbergen, "Disequilibrium and Structural Adjustment" Chapter 29. Stephen R. Lewis, Jr., "Primary Exporting Countries" Chapter 30. Henry Bruton, "Import Substitution" Chapter 31. Bela Balassa, "Outward Orientation" Chapter 32. Dwight H. Perkins and Moshe Syrquin, "Large Countries: The Influence of Size" Volume 3A. Chapter 33. Angus Deaton, "Data and Econometric Tools for Development Analysis" Chapter 34. John Strauss and Duncan Thomas, "Human Resources: Empiri- cal Modeling of Household and Family Decisions" Chapter 35. Jan Willem Gunning and Michiel A. Keyzer, "Applied General Equilibrium Models for Policy Analysis" Chapter 36. Timothy Besley, "Savings, Credit, and Insurance" Chapter 37. Robert E. Evenson and Larry E. Westphal, "Technological Change and Technology Strategy" Chapter 38. Justin Yifu Lin and Jeffrey B. Nugent, "Institutions and Economic Development" Chapter 39. Partha Dasgupta and Karl-Goran Maler, "Poverty, Institutions, and Environmental Resource Base" Volume 3B. Chapter 40. Anne 0. Krueger, "Policy Lessons from Development Experience since the Second World War" Chapter 41. Michael Lipton and Martin Ravallion, "Poverty and Policy" 350 THE WORLD BANK ECONOMIC REVIEW, VOL. 12, NO. 2 Chapter 42. Hans P. Binswanger, Klaus Deininger, and Gershon Feder, "Power, Distortions, Revolt, and Reform in Agricultural Land Relations" Chapter 43. Emmanuel Jimenez, "Human and Physical Infrastructure: Public Investment and Pricing Policies in Developing Countries" Chapter 44. Vittorio Corbo and Stanley Fischer, "Structural Adjustment, Sta- bilization and Policy Reform: Domestic and International Finance" Chapter 45. Dani Rodrik, "Trade and Industrial Policy Reform" Chapter 46. Pranab Bardhan, "The Contributions of Endogenous Growth Theory to the Analysis of Development Problems: An Assessment" REFERENCES The word "processed" describes informally reproduced works that may not be com- monly available through library systems. Arrow, Kenneth J. 1973. "Higher Education as a Filter." Journal of Public Economics 2:193-316. Bairoch, Paul. 1975. The Economic Development of the Third World since 1900. Berke- ley: University of California Press. Balassa, Bela. 1971. The Structure of Protection in Developing Countries. Baltimore, Md.: Johns Hopkins University Press. Basu, Kaushik. 1989. "Rural Credit Markets, the Structure of Interest Rates, Exploita- tion, and Efficiency." In Pranab Bardhan, ed., The Economic Theory of Agrarian Institutions. Oxford: Oxford University Press. Behrman, Jere, and T. N. Srinivasan, eds. 1995. Handbook of Development Economics. Vols. 3A and 3B. Amsterdam: Elsevier. Boserup, Ester. 1981. Population and Technological Change. Oxford: Basil Blackwell. Caves, R. E. 1983. Multinational Enterprises and Economic Analysis. Cambridge, Mass.: Harvard University Press. Chenery, Hollis. 1974. Redistribution with Growth. Oxford: Oxford University Press. Chenery, Hollis B., and Michael Bruno. 1962. "Development Alternatives in an Open Economy: The Case of Israel." Economic Journal 72(March):79-103. Chenery, Hollis B., and T. N. Srinivasan, eds. 1988. Handbook of Development Eco- nomics. Vol. 1. Amsterdam: North-Holland. 1989. Handbook of Development Economics. Vol. 2. Amsterdam: North-Holland. Clark, P. B. 1975. "Intersectoral Consistency and Macroeconomic Planning." In Charles Blitzer, P. B. Clark, and Lance Taylor, eds., Economy-Wide Models and Develop- ment Planning. New York: Oxford University Press. Corden, Max. 1974. Trade Policy and Economic Welfare. Oxford: Oxford University Press. Domar, Evsey D. 1950. "The Effects of Foreign Investment on the Balance of Payments." American Economic Review 40(December):805-26. Dreze, Jean, and Amartya Sen. 1989. Hunger and Public Action. Oxford: Clarendon Press. Edwards, Sebastian. 1989. "The International Monetary Fund and the Developing Coun- tries." Carnegie-Rochester Conference on Public Policy 31:7-68. Waelbroeck 351 Fei, J. H. C., and George Ranis. 1961. "A Theory of Economic Development." Ameri- can Economic Review 51(September):533-65. Gershenkron, Alexander. 1951. A Dollar Index of Soviet Machinery Output. Santa Monica, Calif.: Rand Corporation. . 1962. Agricultural Backwardness in Historical Perspective. Cambridge, Mass.: Bellknap Press. Goldsmith, R. W. 1969. Financial Structure and Economic Development. New Haven, Conn.: Yale University Press. Grossman, Gene M., and Elhanan Helpman. 1991. Innovation and Growth in the Glo- bal Economy. Cambridge, Mass.: MIT Press. Harris, John R., and Michael P. Todaro. 1970. "Migration, Unemployment, and Devel- opment: A Two-Sector Analysis." American Economic Review 60(1):126-42. Hirschman, Albert 0. 1970. Exit, Voice, and Loyalty. Cambridge, Mass.: Harvard Uni- versity Press. Hoselitz, B. F. 1953. "The Role of Cities in the Economic Growth of Underdeveloped Countries." Journal of Political Economy 61(2, June):195-208. Johnston, B. F., and John Mellor. 1961. "The Role of Agriculture in Economic Develop- ment." American Economic Review 51(4):566-93. Kornai, Janos. 1990. The Road to a Free Society. New York: Norton. Krugman, Paul. 1979. "A Model of Balance of Payments Crises and Devaluations." Journal of Money, Credit, and Banking 11(August):311-25. . 1994. "The Myth of Asia's Miracle." Foreign Affairs 76(November-Decem- ber):62-78. Kuhn, Thomas S. 1962. The Structure of Scientific Revolutions (2nd ed.). Chicago: Chi- cago University Press. Kuznets, Simon. 1955. "Economic Growth and Income Inequality." American Economic Review 45:1-28. Lewis, W. Arthur. 1969. Aspects of Tropical Trade 1883-1965. Stockholm: Almqvist and Wiksel. Little, I. M. D., Tibor Scitovsky, and Maurice Scott. 1970. Industry and Trade in Some Developing Countries. London: Oxford University Press. McNamara, Robert S. 1973. "Address to the Board of Governors at the Nairobi Meet- ing." World Bank, Washington, D.C. Processed. Nehru, Jawaharal. 1989. The Discovery of India. Oxford: Oxford University Press. North, Douglas C., and Robert Paul Thomas. 1973. The Rise of the Western World, A New Economic History. Cambridge, U.K.: Cambridge University Press. Nurkse, Ragnar. 1953. Problems of Capital Formation in Underdeveloped Countries. Oxford: Basil Blackwell. Olson, Mancur. 1965. The Logic of Collective Action. Cambridge, Mass.: Harvard Uni- versity Press. Preobrazhensky, Evgeniaei Alexseevich. 1924. The New Economics. Oxford: Clarendon Press. Roemer, J. E. 1982. A General Theory of Exploitation and Class. Cambridge, Mass.: Harvard University. Rostow, Walter. 1965. "The Take-Off into Self-Sustaining Growth." Economic Journal 66(March):25-48. 352 THE WORLD BANK ECONOMIC REVIEW, VOL. 12, NO. 2 Salter, W. E. G. 1959. "Internal and External Balance: The Role of Price and Expendi- ture Effects." Economic Record 35:226-38. Samuelson, Paul A. 1947. Foundations of Economic Analysis. Cambridge, Mass.: Harvard University Press. Schulz, Theodore W. 1961. "Investment in Human Capital." American Economic Review 51. 1964. Transforming Traditional Agriculture. New Haven, Conn.: McGraw-Hill. 1975. "The Value of the Ability to Deal with Disequilibrium." Journal of Eco- nomic Literature 13:827-46. Schumpeter, Joseph A. 1934. The Theory of Economic Development. Cambridge, Mass.: Harvard University Press. Sen, Amartya. 1963. "Neo-classical and Neo-Keynesian Theories of Income Distribu- tion." Economic Record 39:54-64. Srinivasan, T. N. 1994. "Data Base for Development Analysis: An Overview." Journal of Development Economics 44(1 ):3-28. Srinivasan, T. N., ed., with contributions by Justin Yifu and Yun-Wing Sung. 1993. Agriculture and Trade in China and India: Policies and Performance since 1950. San Francisco: International Center for Economic Growth. Srinivasan, T. N., and John Whalley. 1986. General Equilibrium Trade Modelling. Cam- bridge, Mass.: MIT Press. Taylor, Lance, and Edmar Bacha. 1976. "The Unequalizing Spiral: A First Growth Model for Belindia." Quarterly Journal of Economics 90(2):197-218. Timberg, Thomas A. 1997. "Big and Free Is Beautiful: China and India, the Past 40 Years and the Next." Economic Development and Cultural Change 45(2).435-42. Townsend, Robert M. 1994. "Risk and Insurance in a Village in India." Econometrica 63(3, May):539-92. Vogel, Ezra. 1991. The Four Little Dragons: The Spread of Industrialization in Asia. Cambridge, Mass.: Harvard University Press. Waelbroeck, Jean. 1998. "Half a Century of Development Economics: A Review Based on the Handbook of Development Economics." Working Paper 1925. Research Ad- visory Staff, World Bank, Washington, D.C. Processed. Williamson, John. 1990. "The Progress of Policy Reform in Latin America." Institute for International Economics, Washington, D.C. Processed. . 1997. "The Washington Consensus Revisited." In Louis Emmerij, ed., Eco- nomic and Social Developments into the Twenty-First Century. Baltimore, Md.: Johns Hopkins University Press. Wilson, Edward 0. 197S. Socio-Biology, the New Synthesis. Cambridge, Mass.: Harvard University Press. Young, Alwyn. 1994. "Lessons from the East Asian NICS: A Contrarian View." Euro- pean Economic Review 38(3-4):964-74. Coming in the next issue of THE WORLD BANK ECONOMIC REVIEW September 1998 Volume 12, Nunmber 3 Demographic Transitions and Economic Miracles in Emerging Asia by David E. Bloom and Jeffrey G. Williamson Credibility of Rules and Economic Growth: Evidence from a Worldwide Survey of the Private Sector by Aymo Brunetti, Gregory Kisunko, and Beatrice Weder * A Database of World Infrastructure Stocks, 1950-95 by David Canning * Does Economic Analysis Improve the Quality of Foreign Assistance? by Klaus Deininger, Lyn Squire, and Szuati Basit * International Evidence on the Determinants of Private Saving by Pauil R. Masson, Tamim Bayouimi, and Hossein Samiei * The Bolivian Social Investment Fund: An Analysis of Baseline Data for Impact Evaluation by Menno Pradlia7n, Laura Rawlings, and Geert Ridder * Unfair Trade? The Increasing Gap between World and Domestic Prices in Commodity Markets over the Past 25 Years by Jacques Morisset