The Bank, the Fund, and the GATT: Which Institution Most Supported Developing-Country Trade Reform?*

The 1980s and 1990s saw a policy revolution in developing countries in which many highly protected (if not closed) economies were opened to world trade. These reforms were largely undertaken unilaterally, but international economic institutions such as the World Bank, the International Monetary Fund


Introduction
Alan Winters has had a distinguished career in academia at the University of Sussex and at institutions such as the World Bank and the United Kingdom's Department for International Development (DFID).He began his career in the 1970s and 1980s with insightful articles on the costs of trade restrictions on footwear and other goods.As globalization unfolded in the 1990s and 2000s, he focused on developing countries, writing about trade and development (Winters, 2003), trade and economic growth (Winters and Masters, 2013), and trade liberalization and poverty reduction (Winters and Martuscelli, 2014), among many other topics.
Over the course of Winters' career, the world economy has experienced dramatic changes, many of which he was directly involved with at the Bank and DFID.In particular, the decade from 1985 to 1995 was a period of dramatic trade policy reform by developing countries.Many of them shed import substitution policies that had been in place especially since the 1950s and embraced exchange rate and trade reforms that opened their economies to the world (Dean et al., 1994;Irwin, 2022).In doing so, previously closed economies such as China and India became open to world trade and investment, and other emerging markets in Latin America, Asia, and Africa reduced their trade barriers and increased their participation in global trade.These policy changes reshaped the world economy, enabled the emergence of global supply chains, and produced the high level of interdependence that we see today.
Most countries opened their economies by performing the trade policy three-step: (i) devaluing their currencies and establishing competitive exchange rates, (ii) abolishing import controls and converting quantitative import restrictions into tariffs, and (iii) gradually reducing the dispersion and level of those tariffs.In most cases, these reforms were undertaken unilaterally, often under the pressure of an economic crisis.The lessons of experience, such as the success that Taiwan and Korea enjoyed after opening their economies in the 1960s, along with changing ideas about economic policy, contributed to the decision to change trade policies (Krueger, 1997).
The World Bank, the International Monetary Fund (IMF), and the General Agreement on Tariffs and Trade (GATT)then the World Trade Organization (WTO) after 1995supported and encouraged the reform efforts.These institutions play an influential role in shaping international economic policy and their charters gave them a common purpose in promoting world trade. 1 Although these organizations may not have been the driving force behind the reform efforts, what impact did they have in promoting the trade reforms of the 1980s and 1990s? 2  Evaluating the contribution of these institutions to trade reform in developing countries is challenging because they approached the goal of expanding trade in very different ways.The GATT established trade rules and facilitated multilateral negotiations to reduce tariff and nontariff barriers to trade.The World Bank made loans to countries conditional on their making changes to their trade policies.The IMF sought 'exchange stability' to help 'in the elimination of foreign exchange restrictions which hamper the growth of world trade'.The institutions also differed in their ability to influence country policies.The GATT/WTO was the weakest of the three in having virtually no leverage over sovereign governments.The World Bank and IMF had financial resources that they could use to win compliance with the policies that they deemed desirable.
Empirical assessments of their impact on government policies and economic outcomes are plagued with difficulties.Studies based on observational data suffer from sample selection problems: the countries that choose to join the GATT/WTO, accept a World Bank loan, or enter into an IMF program are not randomly selected.These institutions dealt with different countries at different times and in different ways.The degree of compliance with loan conditionality is hard to observe.And it is not possible to know the counterfactual of whether a country's policies would have changed even in the absence of those actions.
That said, it is possible to reach some tentative if impressionistic judgments, perhaps even surprising ones, about the contribution of these institutions to the trade reform process.One might suspect that the GATT/WTO, which of the three institutions focuses most directly on trade, had the biggest impact on developing-country policies, but on closer examination its impact was limited.The World Bank provided billions of dollars in trade policy loans, but this may not have had a decisive influence on a country's decision to undertake trade reforms.Of the three, the IMF's role in promoting trade reform may be the most underrated.The IMF focused more on stabilization and macroeconomic stability and yet it provided critical ingredients to trade reforms by encouraging countries to devalue overvalued currencies and start the process of eliminating exchange controls and import restrictions.

The GATT/WTO
Over the postwar period, the GATT helped facilitate the reduction in trade barriers in the United States, Western Europe, and Japan (Bown and Irwin, 2017).Given this experience, the GATT would seem ideally placed to help developing countries reform their trade regimes.The GATT was to promote trade 'by entering into reciprocal and mutually advantageous arrangements directed to the substantial reduction of tariffs and other barriers to trade and to the elimination of discriminatory treatment in international commerce' (Preamble).The World Bank was 'to promote the long-range balanced growth of international trade and the maintenance of equilibrium in balances of payments by encouraging international investment for the development of the productive resources of members' (Articles of Agreement I:3).The IMF was 'to promote international monetary cooperation … to facilitate the expansion and balanced growth of international trade' (Articles of Agreement I:1-2). 2 As Rodrik (1994, 79) observed at the time, 'external actors have played at best a modest role in initiating recent [trade] reforms'.In a series of case studies, Devarajan et al. (2001, 34) find that all 'agree that economic policy is primarily driven by domestic politics, not by outside agents'.
As already noted, however, the GATT is a weak institution.It is mainly a forum for countries to discuss trade rules and negotiate tariff reductions.It did not have any independent power to affect policy in participating countries.It could not offer any financial incentives to promote policy changes.It did not advise countries as to what their policies should be or even promote research findings that might encourage policy reforms.It did not conduct any monitoring or surveillance of country policies, at least until the Trade Policy Review Mechanism was established in 1989 in the WTO.It had difficulty addressing violations of trade rules until the establishment of the dispute settlement system in 1995.
Despite these weaknesses, the GATT could have promoted reform through its rules on trade policy, multilateral negotiations to reduce import restrictions, and accession agreements.With some notable exceptions, none of these mechanisms played an important role in the trade reform wave of 1985-1995.
The GATT rules were ineffective because developing countries were exempt from key disciplines. 3For example, Articles XII and XVIII(b) permitted these countries to impose quantitative restrictions (QRs) on imports to safeguard the balance of payments and promote economic development.These were enormous loopholes that allowed developing countries to justify almost any restriction on imports.As a result, developing countries maintained extensive nontariff barriersincluding foreign exchange controls, import licensing, and other QRsostensibly on balance of payments grounds.Such measures were supposed to be temporary or transitional, but the GATT provided very little oversight and allowed them to persist for decades without challenge. 4 Developing countries were also exempt from reciprocity in trade negotiations. 5They did not participate in the tariff reductions negotiated during the Kennedy Round of the 1960s and the Tokyo Round of the 1970s.Perhaps no deal could have been struck, given developed countries' reluctance to open markets in agriculture and labor-intensive manufactures.But developing countries also insisted on special and differential treatment; that is, nonreciprocal and preferential treatment by developed countries. 6Even if they had participated in the negotiations, tariffs were not the principal barrier to trade in developing countries, given the QRs they imposed.
When developing countries began to dismantle their import control regimes and reduce their tariffs in the late 1980s and early 1990s, they did so unilaterally without the GATT being much involved in the process.Two-thirds of the reduction in the weighted average tariff of developing countries, which fell from 29.9% in 1983 to 11.3% in 2003, came from unilateral action (Figure 1).Just one quarter came from multilateral negotiations, most of which was due to China, and one tenth from regional trade agreements, according to a striking calculation by Martin and Ng  (2004). 7 Developing countries did make tariff concessions in the Uruguay Round, agreeing to reduce their bound tariffs and increase the share of their bound tariff lines.However, the bound tariffs 3 Finger and Winters (1998) concluded that GATT rules often amount to 'not good policy advice' for countries seeking to undertake policy reforms.

4
The IMF (1992) reported that some 80% of quantitative restrictions notified to the GATT by developing countries were justified for balance of payments reasons under Article XVIII(b).As Eglin (1987, 23) notes: 'the concept of the temporary application of restrictions has been lost.As a result, many developing countries have applied restrictions and invoked Article XVIII(b) over long periods'.One of the goals of the Uruguay Round was to introduce some disciplines on the invocation of these articles (Anjaria, 1987).As Article XXXVI (8) states: 'The developed contracting parties do not expect reciprocity for commitments made by them in trade negotiations to reduce or remove tariffs and other barriers to the trade of less-developed contracting parties.' 6 This spawned the creation of the United Nations Committee on Trade and Development in 1964.This special and differential treatment was formalized in 1979, when the GATT adopted the 'enabling clause' that permitted trade preferences for developing countries that would otherwise violate Article I of the GATT.
were much higher than the applied tariffs, which were left largely untouched in the negotiations.Figure 2 shows that the slight cuts in bound tariffs still left those rates considerably higher than the applied tariffs.Consequently, the reduction in bound tariffs by developing countries negotiated in the Uruguay Round had little impact on their trade flows.
Countries acceding to the GATT in the 1970s and 1980s did so without many demands placed on them.Mexico (1986), Morocco (1987), and Costa Rica (1990) joined while they were making, or after they had made, unilateral changes to their trade policy (Pastor and Wise, 1994).By contrast, countries seeking to join the WTO from 1995 on, mainly transition economies, were required to make significant concessions as part of the accession process, although this was largely after the reform wave of the late 1980s and early 1990s.The WTO accession process was vigorous and demanding in the cases of Bulgaria (1996), China (2001), Cambodia (2004), Saudi Arabia (2005), Vietnam (2007), and Russia (2012).China, for example, made extensive changes to its trade regime as part of a long, drawn-out process of gaining admission to the WTO (Lardy, 2002).These requirements were imposed so that these countries could not free ride on the previous tariff reductions made by others.
This accession process created a two-tiered system of insiders and outsiders.Nothing was asked of the insiders, the developing countries that were already part of the GATT, whereas much was asked of the outsiders seeking to join.They had to undertake far-reaching trade reforms to gain admission to the club.Subramanian and Wei (2007) and Dutt (2020) provide empirical support for this insider-outsider view.They find that developing countries already in the GATT did not see their imports increase significantly as a result of the Uruguay Round because they did not liberalize their trade within the GATT.Countries that joined the WTO saw a significant increase in their imports because they had to undertake substantial trade reforms.
In sum, the GATT was not the driving force behind the trade reforms of 1985-1995.But it was not entirely irrelevant to the liberalization process either.Countries outside the system, such as China, had a strong incentive to join the organization as the benefits of receiving most favored nation status in the global institution were growing.Given the importance of China to the Note: Developing countries defined using World Bank classification.Source: World Bank (2005, 42), based in Martin and Ng (2004).world economy, this example alone suggests that the GATT/WTO played an important role in the reform process of developing countries.

The World Bank
The World Bank was not always a champion of trade reform.In its first few decades, it focused on project loans and infrastructure lending, with economists taking a back seat to engineers.The Bank was inclined to support open markets, but it also sought to promote industrialization, which sometimes entailed financing projects whose goal was import substitution. 8 Economists began to play a larger role in the Bank in the 1970s under President Robert McNamara.The world economy was rocked by two oil shocks and other macroeconomic turbulence during that decade.There was also a growing sense that Bank-sponsored projects would not generate high rates of return or help increase growth and reduce poverty when recipient countries had poor economic policies. 9In 1979, McNamara (1979, 20) proposed new loans to assist developing countries in making 'structural adjustments for export promotion in line with their longterm comparative advantage'.In 1980, the Bank began offering structural and sectoral adjustment loans conditional on policy changes, including trade policy.Sectoral adjustment loans for trade policy amounted to $7.4 billion over the 1980s (mostly 1986-1989), about 5% of total Bank lending (World Bank, 1992, 18). 10 The Bank also provided technical assistance and advice to countries seeking to reform their trade policy.As Kapur et al. (1997, 451) put it: 'The early Bank favored international trade; it inclined toward openness.But "development" was preoccupied with industrializing, which tended to mean import substitution; and because of the infant-industry problem, import-substitution industrialization (ISI) entailed protection.Thus the Bank's trade and industrialization recipe in the 1950s and 1960s was to seek both import-substitution and export-promotion industrialization via a liberalized price rationing of inputs, protected by tariffs and partly ordered by national economic planning.'9 Isham and Kaufmann (1999) found that overvalued currencies, trade restrictions, and distortions in the prices of tradable goods all significantly reduced the rate of return of investment projects in developing countries.
The results of the trade policy loans were mixed, according to assessments inside and outside the Bank.Examining 40 countries that received trade policy loans in 1980-1987, Thomas (1991, 57) , 57) concluded that 'direct examination of the conditions in trade adjustment loans and their implementation records suggests that import protection on average has fallen only modestly in most of these countries'.Countries may have adjusted their exchange rates and removed export restrictions, but the reduction of import protection 'has been modest at best'. 11The report found that weak macroeconomic performance and inadequate government commitment to reform were responsible for the failure to fully implement or sustain the changes.Even strong reform advocates (e.g., Edwards, 1997;Nogués, 1998;Krueger and Rajapartirana, 1999) concluded that Bank lending did not produce much change in the trade policies of loan recipients. 12 Why was the track record so mixed?The Bank identified the lack of country 'ownership' of the reformsi.e., the reluctance of policymakers to undertake reformsas the key problem.If a country's political leaders did not want to reform, offering them financing was not going to persuade them to do so.Studies found that domestic political considerations usually determined whether a country was willing to reform, not the presence or absence of Bank-provided loans. 13However, in some cases, loans gave a political boost to reformers in governments that were divided over whether to reform or not, tipping the balance in favor of making an attempt at reform. 14 The Bank was particularly disappointed in its efforts to promote policy reform in Africa.The basic problem was that many countries were happy to take the loans but not change their policies.In a damning report, Devarajan et al. (2001, 34-35) wrote: 'In the pre-reform phase in which the government is not committed to reform, conditional loans have generally been a farce in which the government agrees to measures it does not believe in as a way to get funding, fails to carry them out, and then receives the funding from donors anyway.'This led to a repeated cycle of lending and noncompliance, in part because of institutional incentives in the Bank to continue providing loans.Breaking this cycle by withholding funds sometimes proved more effective at promoting reform because it forced countries to face a tighter budget constraint.As Michael Bruno, former chief economist at the World Bank, said: 'We did more for [reform in] Kenya by cutting off aid for one year, than by giving them aid for the previous three decades' (Devarajan and Khemani, 2018, 216).
The failure of conditionality to guarantee policy reform was apparent by the late 1980s.By the mid-1990s, the World Bank shifted its financial support for trade policy from subsidizing policy reform to supporting trade facilitation and infrastructure.
The Bank also provided technical assistance to help with the process of trade reform, particularly in dealing with sequencing and implementation issues (Thomas and Nash, 1991).The Trade Expansion Program, funded by the United Nations Development Programme and administered by the World Bank, began in 1987 to provide independent technical expertise in designing and 11 These qualitative assessments focus on whether the loan conditions were met or not.In one of the few empirical analyses of the loans, Jinjarak et al. (2013) compare countries that received a trade adjustment loan in 1987-2004 with a nonrecipient group.They find that loans are associated with slightly higher economic growth and growth in trade.
12 Krueger and Rajapartirana (1999, 739) concluded that the strongest reformers did not undertake reforms with Bank loans, rather they undertook reform on their own and sought assistance later'.Nogués (1998, 86-87) avers that the Bank policies failed 'mainly because they did not support … a strong trade liberalization program backed by a clear political commitment'.
13 Dollar and Svensson (2000) found 'no evidence that any of the variables under the World Bank's control affect the probability of success of an adjustment loan'.Rather, domestic political economy factorssuch as ethnic fractionalizationstrongly influence the success or failure of reform programs.14 As Morrissey (2004, 163) pointed out: 'Almost all studies reveal that conditionality has encouraged reform efforts that may not otherwise have been attempted, even if reform achievements are less than donors had expected.' Heckelman and Knack (2008, 526) note that 'several World Bank and IMF studies have concluded that policy reform is driven by domestic political economy considerations, and that conditionality is likely to be effective only in the early stages of reform, when it can bolster the position of reform advocates in government'.
implementing reform programs without conditionality.The goal was to have independent assessment teams (rather than World Bank staff) provide advice and thereby increase local ownership of reform.But even this provision of technical assistance was not a clear success.Berg (1998, 318) offered the sobering conclusion that 'local officials, economic interest groups, and political classes saw these policy reform programs as an outside imposition, a perception reinforced of course by the dominance of conditionality in the dialogue, and by the limited extent of genuine bureaucratic and political participation in the process.Therefore, adoption of reforms was slow and local commitment to them lukewarm.' 15 Arguably more important to reform than conditional lending was the Bank's role in disseminating research on economic policies that contribute to growth and development. 16The Bank helped shift the intellectual consensus among policymakers toward an understanding of the benefits of a liberalized trade and payment regime (Irwin, 2020).The prodigious writings and travels of Bela Balassa, a long-serving consultant at the Bank in the 1970s and 1980s, helped change prevailing viewsinside and outside the Bankaway from import substitution toward openness and export orientation.He forcefully and consistently argued that exports were a source of growth, that a market-appropriate exchange rate was critical to export expansion, and that import restrictions were often arbitrary and costly.Anne Krueger's intellectual leadership as chief economist at the Bank (1982Bank ( -1986) ) put a spotlight on trade research, especially the distortions caused by excessive trade restrictions.This effort, including the famous 1987 World Development Report on trade, added to the growing body of evidence that a more liberal trade policy could improve economic performance.The change in intellectual mindset about trade and development is hard to measure or even explain to anyone who did not live through the 1970s and 1980s. 17 A more concrete way the Bank contributed to trade reform was by serving as a training ground for officials who would return to their home countries, take high-ranking positions in government, and become important players on reform teams that brought about policy changes. 18 The rise of Western-trained economists, many with World Bank experience, to high-ranking government positions has been tied to the spread of trade liberalization around the world (Weymouth and Macpherson, 2012).
In sum, the effectiveness of the World Bank's efforts to promote trade reform through conditional lending is generally viewed as mixed: countries that wanted to reform did not need the loans to do so and countries that had no intention of reforming were happy to take the loans, although some countries may have been persuaded to undertake reforms that otherwise might not have been done.The Bank may have had a bigger impact on trade policy through the dissemination of research findings about the benefits of a more liberal trade policy and its technical expertise about how to implement such reforms.
15 Not every analysis took such a dismal view.Jones et al. (2011) find that when African countries did adopt trade reforms in the 1990s, they did so along the lines suggested by the Bank. 16 See Gavin and Rodrik (1995) and Clemens and Kramer (2016, 59), who note that 'the Bank plays an important role in how ideas move into policy by collecting data, generating ideas, bringing ideas to a wider policy audience, and turning ideas into specific policies'.17 Edwards (1997, 47) contends that 'the Bank has made a contribution to the intellectual debate on the consequences of alternative trade regimes'.Krueger and Rajapartirana (1999, 739) state that 'the Bank's economic work and research have acted as a conveyor belt to take trade policy issues from academic and in-house research into the developing country policymakers' domain … .It is the authors' judgement that the Bank was influential in policy decisions that led to trade reform in developing countries.But the extent of that influence cannot be adequately measured or determined.' 18 As Clemens and Kramer (2016, 60) note: 'A main channel of World Bank influence, and a measure of the Bank's prestige, is the flow of Bank staff to senior policy-making positions in their home countries.Frequently mentioned recent examples include: Ngozi Okonjo-Iweala in Nigeria, Luisa Diogo in Mozambique, Montek Ahluwalia in India, Kemal Dervis in Turkey, Richard Webb and Luis Miguel Castilla in Peru, Ellen Johnson Sirleaf and Antoinette Sayeh in Liberia, Moeen Qureshi in Pakistan, Vittorio Corbo in Chile, Ashraf Ghani in Afghanistan, Benno Ndulu in Tanzania, and many others.'

The International Monetary Fund
At first glance, the IMF appears to be the institution with the least relevance to a country's trade policy.Its primary focus is macroeconomic stability, for example through inflation control and fiscal deficit reduction.But in its attention to exchange rates and foreign exchange controls, the Fund played a very important role in enabling countries to take the first step of the 'trade policy three-step'.
Trade controls and import restrictions often stem from balance of payments difficulties, which in turn arise from an overvalued currency (Shatz and Tarr, 2002).An overvalued exchange rate reflects a failure to adjust the nominal exchange rate when a country experiences more rapid inflation than its trading partners or a significant terms-of-trade shock.An overvalued currency depresses exports and leads to excessive imports, putting pressure on foreign exchange reserves.Governments that fear devaluation as a remedy are led to impose exchange controls and foreign exchange rationing as a way of limiting imports. 19 The Fund's work in this area has been important because exchange controls and foreign exchange restrictions have been a considerable impediment to trade (see Wei and Zhang, 2007;Eichengreen and Irwin, 2010).The problem with the license raj in India, for example, was not the high tariffs but the exchange control regime that produced a complex maze of bureaucratic regulation that prevented imports from entering the country, ostensibly to conserve foreign exchange.The GATT was considered so ineffective in this area that in the 1970s IMF officials considered (but rejected) a proposal giving itself the authority to approve or disapprove of restrictions on current account payments (de Vries, 1985, 704-705).
From the start, the Fund encouraged countries to establish currency convertibility for current account transactions.It opposed illiberal payment regimes that included multiple exchange rates, exchange controls, and foreign exchange rationing.Although it was founded to promote exchange stability, the Fund grew to see the merits of exchange rate flexibility to ensure that the exchange rate was set at a realistic level. 20 The IMF began to play a more important role in developing countries in the 1980s, advising them on how to reduce external deficits while keeping payments systems open (Boughton, 2001).IMF officials were in frequent contact with central bankers and finance ministers, pressing them to work toward a greater opening of the economy.By quietly encouraging countries to devalue, the Fund was laying the groundwork for dismantling exchange control regimes and establishing current account convertibility.As Paul Collier (1993, 510) put it: 'The heart of liberalization is the conversion from using trade policy for payments balance to using the exchange rate.'That is precisely what the Fund was trying to do.Countries could not begin to convert quantitative restrictions on imports to tariffs and then reduce those tariffs without first undertaking a devaluation to eliminate the excess demand for foreign exchange.
The Fund's work in this area has been unheralded in part because it is very hard to measure or even document because every country's experience is unique.As a result, there is no systematic evidence on its impact on policy (Johnston, 1999).What is known is that the share of countries with nonunified exchange rates fell from about 50% in the mid-1980s to about 15% by 2000 (Figure 3).During the 1993 review of surveillance, the managing director of the IMF, Michel Camdessus, told the executive board that the Fund was not doing enough to get countries to 19 A key obligation under Article VIII of the Articles of Agreement is 'no member shall, without the approval of the Fund, impose restrictions on the making of payments and transfers for current international transactions'.Current international transactions here mean current account transactions, payments for imports of goods and services, not capital flows.20 Krueger (2014, 245) tells the story of Ernest Sturc, who had been on the Czech delegation to the 1944 Bretton Woods conference and was later an important official at the IMF.Whereas during the conference delegates toasted to the end of 'competitive devaluations', Sturc later said, 'Little did I dream that I would spend the rest of my life persuading countries to devalue.' eliminate exchange restrictions and accept the obligations of Article VIII (Boughton, 2001, 128-129).An increasing number of countries accepted those obligations throughout the 1990s.
While the IMF was deeply involved in exchange rate discussions with countries, it was less concerned about trade policy.In the 1980s, in cooperation with the World Bank, the Fund began to incorporate trade policy reforms in its lending programs.IMF conditionality is generally much tougher than the Bank's and therefore might be more effective at inducing policy change.But the Fund's trade reform conditions were never a key part of its structural adjustment programs. 21In a self-assessment, the IMF (1998, 6) found 'programs that targeted a reduction in restrictiveness generally succeeded in achieving their trade reform objectives' but 'medium-term trade reform targets although mentioned often were generally insufficiently specific and comprehensive' to be effective.Therefore, 'longer-term Fund involvement in countries during 1990-1996 does not seem to have promoted greater trade liberalization in the programs reviewed'. 22 Many countries joined the IMF in the early 1990s, but stringent entry conditions were not applied, unlike the WTO accession process.However, members that fell into arrears with the IMF (i.e., they did not repay its loans) were cut off from international capital markets and had to negotiate a tough package to escape from purgatory.The Fund helped several countriesnotably Peru, Cambodia, Vietnam, and Zambiamove out of arrears by adopting more realistic exchange rates and greater openness to trade.
In sum, the IMF has played a quiet and somewhat underappreciated role in the reform of developing-country trade policy.The IMF historian Margaret de Vries (1994, 229) once chided economists for failing to appreciate 'how the gradual lifting of exchange controls enabled world trade and services to increase, and how real exchange rate depreciation by developing countries Source: Ilzetzki, Reinhart, and Rogoff (2019).

21
As the IMF (1998, 33) noted: 'In most of the programs, trade liberalization measures were not used as performance criteria but were monitored in the context of overall reviews of performance, reflecting partly the difficulties in quantifying trade liberalization as well as attempts to minimize the number of performance criteria.'22 Wei and Zhang (2010) report that during the period 1993-2003, 99 countries had IMF programs and 77 had trade reform conditions in at least one of the programs.Trade conditionality is associated with an increase in openness on average, but the effect comes mostly from countries with a high willingness to reform, indicating selection biascountries willing to take on those obligations.helped integrate them into the world economy'.The IMF, she points out, 'was instrumental in persuading countries of the need for noninflationary macroeconomic policies and in pushing countries to lift trade restrictions in line with their improving external payments situations.'

Conclusion
The driving force behind the trade reforms of the 1980s and 1990s is not to be found at 1818 H Street NW or 700 19th Street NW in Washington, DC, or Rue de Lausanne, 154 in Geneva.Rather, finance ministries and central banks were the central actors in all the countries that undertook extensive reforms to open their economies during that period.
The Bank, the Fund, and the GATT/WTO contributed in some way and to some countries in this effort, often more indirectly than directly.The GATT played little role in the reform process, although countries that joined the WTO after 1995 were forced to engage in substantial trade liberalization.The World Bank provided a training ground for reformers and armed them with ideas about the rationale for and implementation of policy change.The IMF played an important role in reminding country officials of the value of a realistic exchange rate and the desirability of removing obstacles to current account payments. 1 5

Figure 1 .
Figure 1.Most developing economy tariff reductions between 1983 and 2003 were undertaken unilaterally.

Figure 2 .
Figure 2. Developing economies set tariffs below their obligations under the Uruguay round Note: Developing countries defined using World Bank classification.Source: Finger, Ingco, and Reinchke (1996). 8

Figure 3 .
Figure 3.The number of countries with nonunified exchange rates plummeted after the mid 1980s.