Speeches, Testimony, Papers

The Future Course of Trade Liberalization

by Gary Clyde Hufbauer, Peterson Institute for International Economics
and Barbara Kotschwar, Peterson Institute for International Economics

Paper for The World Bank
"Challenges for the Twenty-First Century"
October 1998

© Peterson Institute for International Economics

A defining feature of the post-war international economic system has been sustained growth of world trade. During the past quarter century the growth of world trade has outpaced that of world output, as seen in Figure 1. Global trade's share of world GDP rose from 13 percent in 1970 to 21 percent in 1995 (World Bank, 1998.)

As trade flows have grown over the past quarter century, the composition of trade has changed substantially, notably with the expansion of services trade and the larger role of developing countries. Driven by the activities of transnational corporations, various stages of the production process are located in different countries—a phenomenon termed by Michael Mussa the slicing of the value-added chain (Mussa, 1998). The resulting intra-industry trade and the decentralization of production activities have opened new opportunities for developing countries. Correspondingly, developing country exports have changed, evolving from a composition dominated by primary commodities to increasingly include manufactures and services. In fact, between 1985 and 1995, the share of manufactured goods in developing country exports rose from 47 percent to 83 percent. Correspondingly, developing countries play a much larger role in world manufactures trade than they did before—making up 20 percent of the world manufactures market in 1995, double their 10 percent share in 1970. (World Trade Organization (WTO), 1996).

In the past decade, greater participation of developing countries in the international trading system has helped maintain the fast growth of world merchandise trade: in value terms, developing countries have nearly doubled their exports of goods and services in the space of only five years—from $860 billion in 1990 to $1.5 trillion in 1995 (World Bank, 1998). One reason for this increase is the rapid trade growth of the East Asian countries, whose share of world manufacturing output grew from 6.8 percent in 1980 to 11.1 percent in 1995 (UNCTAD, 1997) Another reason for increased participation in the global trading system is the greater openness of developing countries to trade. Although imperfect indicators, openness ratios, which measure a country's trade as a percentage of output, can paint a picture of relative participation in international trade. These ratios give a rough indication of the economy's dependence on international trade and its participation in the international trading arena. All things being equal, a country with a higher openness ratio trades more with the world than a country with a lower ratio. Even when adjusting for size, geography, and other factors, the relative movement of a country with respect to this indicator will demonstrate a movement toward or away from more open trade. Developing country ratios have generally risen: from 1980 to 1995, the developing country share of exports of goods and services in GDP increased from 20 to 26. This compares with an overall increase in world openness from 19 to 21 (World Bank, 1998). This increase has not been uniform across developing countries. The greatest increase has been in the trade ratios for East Asian countries: collectively, they increased their openness from a low 9 percent in 1970 to 29 percent in 1995. Latin America and the Caribbean also rose, but more modestly—from 11 percent in 1970 to nearly 18 percent in 1995.

All of these facts have broad implications for the evolution of the world trading system. In this paper we discuss major factors that greatly influence the international trading arena. The first section outlines the importance of liberalization, identifying five gains from trade and investment liberalization. The second section gives an overview of international trade since the conclusion of the Tokyo Round, taking into account unilateral and regional, as well as multilateral trade measures. In the third section, the issue of trade and income inequality is discussed, while the final section speculates on the shape of future developments.

Section I: Gains from Trade liberalization

Five major kinds of gains from freer trade can be identified; however, there is considerable overlap among them. These gains come from unilateral policy shifts as well as from the liberalization that takes place through regional and multilateral negotiations. Quantitative estimates and further elaboration of the gains from liberalization will be found in a recent OECD book (1998b).

Static Gains

Free trade allows countries to export those goods and services that they can make efficiently, and to import those goods and services that they make inefficiently. "Efficiency" is measured in relative terms, not in absolute terms. For example, while the United States has an absolute advantage in the production of both aircraft and apparel by comparison with China—in the sense that the United States requires fewer man years to produce either a civilian airplane or a container load of blue jeans than China—both countries gain when the United States concentrates its higher-skilled workforce on producing and exporting aircraft, and China specializes in producing and exporting apparel.

The static gains from free trade—the gains that would come from the free exchange of goods and services when each country uses the same technology before and after freer trade—are typically estimated at less than two percent of GNP. In the context of the Uruguay Round, it was roughly calculated that the static gains from the negotiated reduction in industrial tariffs and agricultural barriers would grow the global economy by about $40 billion, of which about $30 billion would be realized by advanced countries and $10 billion by developing countries.¹ While they are not trivial, static gains alone do not measure the full benefits of liberalization.

Consumer Savings

Trade results in lower prices, enabling consumers to buy more with their paychecks. Estimates published by the Institute for International Economics suggest that totally free merchandise trade would save American consumers about 1.2% of U.S. GNP; Japanese consumers about 5% of Japanese GNP; and Korean consumers about 4% of Korean GNP (Hufbauer and Elliott, 1994, Sazanami et al, 1995, Kim, 1996). These savings occur even when markets are competitive without free trade. These measures are another way of computing static gains.

Trade also introduces more competition into domestic markets, undermining anti-competitive practices that usually prevail when domestic producers face few foreign challenges. Industries shielded from foreign competition often charge higher prices, have little incentive to hold down costs and often indulge slack management style. By injecting greater competition into sheltered domestic markets, international commerce forces national producers to live up to their potential. In recent years, the most dramatic example has been the fast-paced drop in international telecom rates as erstwhile monopolies have faced new competition. Simulation exercises carried out by Atje and Hufbauer (1996) suggest that when free trade converts a monopoly industry into a competitive industry, the efficiency gains are several times larger than the static gains resulting from the free trade among competitive industries.

Higher wages and more stable employment

Over the last decade, jobs supported by US exports rose four times faster than overall private-industry job creation. Consequently, 12 million Americans now owe their jobs to exports, out of a civilian labor force of 67 million people. The research of Richardson and Rindal (1995, 1996) shows that both production and non-production workers in US exporting firms receive on average about 14 percent higher pay than workers in firms that do not export. Export-oriented firms tend to have more stable employment patterns. Similar stories can probably be told for other countries that have increased the trade share of their economies.

Gains in total factor productivity

Free trade exposes countries to new production and management technologies that foster higher productivity at both the firm and industry level. Twenty years ago, American auto firms began learning from Japan how to produce cars more cheaply. Today, financial institutions in Tokyo are learning from the United States how to run banks and pension funds. These gains are closely linked to cross-border direct investment (foreign direct investment, or FDI). In fact, a major purpose of trade agreements is to promote investment, thereby simultaneously encouraging intra-firm trade by multinational corporations (MNCs) and higher productivity among all local firms that compete with MNCs.

Catch-up benefits

One of the biggest gains from freer trade in goods, services and capital is the opportunity for poorer countries to raise their productivity and income toward the levels already reached by richer countries. Research by Barro (1991, 1994), Sala-I-Martin (1991), Ben-David (1995), and others cited later suggest that freer trade helps close the income gap between poor and rich countries at a rate of about one-half of one percentage point per year faster than might otherwise occur. The catch-up largely reflects the acquisition of better technology, and improved management, and the spur to productivity that accompanies keen competition and foreign direct investment.

Section II: From Tokyo to the Year 2000

Tariffs have been cut substantially in the post-Second World War period (hereafter the post-war period)—both through multilateral and regional negotiations and through unilateral decisions. Throughout the 1960s and 1970s, the advanced countries (essentially the industrialized OECD countries) increasingly opened their markets to external competition by cutting tariffs. This was largely accomplished through successive rounds of GATT negotiations and the formation of the European Common Market. In the 1970s and 1980s, the advanced countries also relaxed their quantitative restrictions, with the exception of those protecting agricultural products and textiles and apparel. The developing countries initially lagged the advanced countries in cutting tariffs, but since 1980 many have started to catch up. Especially in the case of Latin America, countries have implemented trade-liberalizing measures within in a much shorter timeframe than the advanced countries.

The average tariff on manufactured goods for advanced countries was cut from about 40 percent at the beginning of the first Round of GATT talks, to about 6 percent at the end of the Tokyo Round, to just 4 percent when the Uruguay Round cuts are completely implemented (Schott, 1994). After the implementation of Tokyo Round cuts, the tariffs on manufactured goods averaged 6.0 in the European Community, 5.4 percent in Japan, and 4.5 percent in the United States (World Bank, 1987). The first seven rounds of tariff negotiations were conducted by contracting parties negotiating on a request-and-offer basis with their key trading partners, seeking concessions on specific items of interest. These concessions were then multilateralized to all GATT contracting parties, according to the nondiscrimination (most favored nation) principle.

These first seven rounds were basically conducted among OECD countries—the United States, Canada, the Western European countries, Australia and New Zealand, and, after the mid-1960s, Japan. The GATT was often seen as a rich man's club. While the advanced countries were reducing their tariffs within the GATT framework, developing countries were often adopting restrictive, inward-oriented policies in attempts to protect and to modernize their economies. Import-substitution industrialization (ISI) was the favored policy of the 1960s and 1970s.

To overcome the inherent scale limitations of these policies, some developing countries extended the ISI framework to a regional level by forging trading pacts with their neighbors. Indeed, south-south regional trading arrangements flourished, many of them inspired by ISI concepts. The 1960s and 1970s saw the creation of the Latin American Free Trade Association, the Central American Common Market, the African Common Market, the Central African Customs and Economic Union, the East African Economic Community, the Association of East Asian Nations, the Caribbean Community, the Andean Group, and the Economic Community of West African States. Often based on production-sharing arrangements and incorporating special payments mechanisms, with special provisions for the least-developed members, these regional arrangements erected high tariff barriers to third-party competitors and usually restricted foreign investment.

In the early 1960s, Charles Kindleberger (1962), among other academics, challenged the prevailing ISI strategy. However, the benefits of open markets to growth in developing countries were accorded little recognition in policy circles until Bela Balassa published a series of persuasive empirical studies, beginning in the late 1970s (Balassa, 1977). Influenced by Balassa, the link between trade policy and growth became the major theme in the World Development Report 1987. Broadly speaking, this literature established the proposition that developing countries that adopted outward-oriented as opposed to import substitution policies both enjoyed faster export growth and had higher GDP growth.

A few years later, policy analysts turned their attention to long-run per capita income convergence between states, regions and countries that started with different initial levels of income. Pioneering work was carried out by Barro (1991, pp. 407-443) and Barro and Sala-I-Martin (1991, pp. 107-158). The findings from these and subsequent papers were summarized by Barro in 1994. In their work, Barro and Sala-I-Martin established four broad propositions:

• Among states of the United States and regions of Japan and Europe, poorer geographic areas exhibited a long-run tendency to catch up with the richer areas at a rate of about 2 percent per annum.
• Between a large cross-section of countries, poorer countries tend to catch up with richer countries at a rate of about 1.5 percent per annum.
• As between countries, higher investment shares of GDP and more schooling accelerate income growth; while higher government consumption (excluding education and defense) and more frequent revolutions and coups reduce growth.
• More open trade and investment policies—proxied by lower tariff rates and lesser black market premiums foreign exchange—accelerate growth.

The work of Barro and Sala-I-Martin contained two findings of special relevance to the future trade policy agenda. The first is their observation that long-run income convergence has been about 0.5 percent per year faster within highly integrated economic areas—the United States, Japan and Europe—by comparison with convergence between countries that have a much lower degree of integration. This observation suggests a bonus payoff for poorer regions that join integrated economic areas. The second supporting observation is their statistical finding that measures of openness—low tariffs and low black market premiums—are significantly associated with faster growth among a large panel of countries. This second observation was recently confirmed in a highly demanding statistical test involving two million regressions performed by Sala-I-Martin (1997, pp. 178-183). Additional supporting evidence is reported by the IMF (May 1997) and the OECD (1998b).

Of course association does not prove causation. It can be and has been argued that growth is the cause of openness, rather than the other way around (Bradford and Chakwin, 1993; Bradford, 1994; Rodrik, 1994). However, in a joint paper, Jeffrey A. Frankel, David Romer and Teresa Cyrus (1996) use an instrumental variables approach to convincingly show that causation runs from trade openness to growth among East Asian countries. And Ben-David (1995) demonstrates that income convergence within artificial groups of countries created on the basis of their trade intensity is significantly greater than convergence within artificial groups created without regard to trade intensity.

Finally, there is the question whether policy—as opposed to geography, history and other factors—is all that important in determining the extent of a country's integration with the world economy. Considerable evidence shows that policy matters. We have the statistical findings of Barro and Sala-I-Martin, already cited, showing the significance of tariffs and black market premiums in the growth equation. And we have the work of Georgios Karras (1997), Dan Ben-David (1995) and Michael B. Loewy (1995), among others, which detects faster income convergence in regions with a stronger policy commitment to economic integration. These scholars detect strong convergence in Europe (where the European Union has been the driving force), mild convergence in Latin America (where free trade areas did not amount to much before the 1990s), and practically no convergence in Southeast Asia (where ASEAN has played a larger role politically than economically).

Supporting these findings, Frankel, Romer and Cyrus (1996) find that openness in East Asia is mainly a matter of geography, not policy. The two East Asian countries that exhibit policy-driven openness are Singapore and Malaysia. In most of East Asia, fortuitous geography created open economies and in turn open economies stimulated income growth; but integration policies were not strong enough to promote income convergence within the region. However, the region as a whole sharply narrowed its income gap with the industrial nations.

Policy Consequences

Lessons summarized in the foregoing section were not lost on developing countries. Protective domestic economic policies and closed regional arrangements ran out of steam by the late 1970s. Faced with economic stagnation and high external debt levels in the 1980s, many developing countries started reforming both their macroeconomic and commercial policies. During the 1980s and 1990s, developing countries sharply and unilaterally reduced their tariff barriers. As seen in Table 1, tariffs were more than halved between 1986 and 1996 in the Latin American countries, and substantially cut elsewhere.

In parallel with trade reforms, many of these countries also privatized state-run companies, reduced fiscal deficits and cut inflation. Several developing countries are now taking steps to improve the environment for competition by controlling restrictive business practices. Still, a great deal of pro-competitive work remains to be accomplished, especially in upgrading transportation and communications infrastructure and in building legal and judicial institutions.

As development models have changed, so have approaches towards regionalism and multilateral trade. As unilateral trade liberalization took hold, far more countries participated in the multilateral trading system. At its inception, in the late 1940s, the GATT comprised 23 signatories; by the 1970s, 102 countries participated in the Tokyo Round of negotiations; in 1995 there were 108 members of the new WTO, and in late-1998 there were 133. Since the Tokyo Round, the standards of GATT/WTO accession have become progressively more rigorous. As a condition of membership, countries must now liberalize to a significant degree. Hence, the rising membership figures reflect a genuine expansion of more liberal commercial practices.

In the Uruguay Round, unlike the previous seven GATT rounds, developing countries were pressured to extend greater market access to their trading partners. Developing countries, in turn, came to realize that by opening their markets, they would be taken more seriously as negotiating partners. In addition, liberalization was no longer seen as a exercise in masochistic pain; politicians came to recognize that liberalization would confer benefits on their own economies. After the formal launching of the Uruguay Round in 1986, more than 60 developing countries unilaterally lowered their trade barriers.² In the Uruguay Round negotiations, certain sectors—notably agriculture, where the Cairns Group played a leading role—benefited from joint coalitions of advanced and developing countries.

Results of the Uruguay Round

The completion of the Uruguay Round and the creation of the WTO broadened and deepened the scope of the multilateral trading system. The WTO provides the "constitution" for multilateral trade—setting out the do's and don'ts of accepted commercial practice, establishing a forum for continued negotiations and, importantly, creating a mechanism for settling disputes among members.

Among advanced countries, tariffs were largely reduced in previous Rounds, but the Uruguay Round established significant tariff commitments on the part of developing countries. Developing countries as a whole agreed to upper limits—or " bindings"—for 72 percent of their tariff line items, contrasted with 22 percent before the Uruguay Round. Sectors that were previously considered too sensitive, such as agriculture, textiles and apparel were brought into the folds of the multilateral framework, both by converting quotas to tariffs, and by expanding the quotas. New areas—notably trade-related intellectual property rights and services—were also brought into the system.

Services are particularly important. It is difficult to measure the amount of services traded—but the figure now probably exceeds $2 trillion annually, when services provided by FDI are counted. The service sector now accounts for about 70 percent of GDP in advanced countries and 50 percent in developing countries in the late 1990s.³ On the whole, service trade is probably less sensitive to geography than merchandise trade. The new General Agreement on Trade in Services (GATS) created a framework for future service trade liberalization, much like the original GATT created a framework for merchandise trade liberalization in 1947. Subsequent to the conclusion of the Uruguay Round, the Agreement on Basic Telecommunications Services, the Information Technology Agreement and the Financial Services Agreement (signed in February, March and December 1997, respectively) each advanced the liberalization of services.

Services are part of the so-called "built-in agenda" of future WTO talks. Negotiations on labor mobility issues, sea and air transport services, government procurement and subsidies remain to be addressed in the next round of negotiations on services, one that will be launched on or before January 1, 2000. While not yet a formal item on the WTO agenda, electronic commerce will be a big part of the new services agenda. Issues of privacy, encryption, taxation and registration are all at stake. The same forces driving globalization are enabling more and more commerce to be conducted electronically.

The expansion of trade in services is changing the character of international commerce. With the rapidly falling cost of telecommunications and the rapid development and dissemination of increasingly sophisticated software, a wide range of professional services may soon be provided far from the point of purchase. Professionals—engineers, professors, accountants, medical professionals, lawyers—located in Bombay will sell their services in Toronto without setting foot outside of their country—and vice versa. In the future global trading system physical migration may continue to remain tightly controlled—but the migration of minds will be practically unlimited.

Agriculture. The Uruguay Round agreements marked a turning point in agricultural protection. For almost the first time in GATT negotiating history, agricultural barriers were squarely placed on the negotiating table. As pointed out by Timothy Josling (1998), the Uruguay Round Agreement on Agriculture establishes a set of rules for future agricultural liberalization. The Agreement made three notable achievements (Schott, 1994, 43-54):

• WTO members agreed to cap their farm export subsidies, starting from a high 1986-90 base, and committed to reduce the extent of export subsidization by 36 percent in value, and 21 percent in volume coverage, over 6 years (starting in 1995);
• WTO members agreed to reduce their domestic support for agriculture by 20 percent over 6 years.
• WTO members promised to convert their nontariff agricultural barriers into tariff equivalents, and provide minimum access quotas for farm imports that were blocked by previous protection.

These commitments are part of the "single undertaking"—the mutual exchange of identical promises by all WTO members to one another—and therefore bind developing countries as well as advanced countries. Various quantitative estimates made near the conclusion of the Uruguay Round (General Agreement on Tariffs and Trade, 1994) indicated that one-third or more of the total income gains from the entire negotiations can be attributed to agricultural liberalization. Generally speaking, the countries which do the most to liberalize their farm imports will gain the most from the Uruguay Round Agreement. For example, countries like Argentina that maintained low protection for their field crops (wheat, coarse grains) prior to the Uruguay Round will enjoy some gains from increased exports; but countries with high field crop protection, like Chile, will enjoy considerably larger gains.

Agricultural trade remains far from liberalized. Export subsidies continue to exist. Developed countries, which have been reluctant to reform their agricultural policies, continue to protect their domestic farms. Tariffication of nontariff import measures has resulted in tariffs bound at high peaks supplemented by tariff rate quotas in a number of products. Further multilateral liberalization attempts can complement and support domestic moves taken to increase transparency and access in this sector. Tim Josling (1998) cites a few additional reasons to keep moving forward on the multilateral front. A sound international framework could "lock in" the reforms that have slowly been implemented in developed countries and improved international trade rules in this area could offer developing countries restraint and guidance to assist with their domestic policy choices. As such, agricultural talks will continue to be a mainstay of the multilateral trading system.

Dispute Settlement. As of October 1998, the WTO's new dispute settlement mechanism (DSM) has received 143 consultation requests. Developing countries have been involved in more than one quarter of these requests—29 requests were made by developing countries alone, while 10 requests were made by advanced and developing countries acting together.⁴ Developing countries have brought 19 cases against advanced countries and eight against other developing countries. Twenty-seven of the 77 cases brought by advanced countries targeted developing countries (WTO, 1998).

The issues addressed in the DSM context have varied as much as the members participating. They have included agricultural commodity cases, including a U.S. case against Korea; a case in which the United States, Honduras, Guatemala and Mexico questioned the EU banana regime; and cases involving EU beef imports, Hungarian export subsidies, and U.S. shrimp rules. The United States and the European Union have brought cases concerning intellectual property rights. Automobiles have featured in cases brought by the United States, Japan and Brazil. Further WTO disputes have included issues relating to underwear, photographic film, gasoline and liquor.

Given the range and complexity of issues addressed in the WTO, and the differences in the size and development of the parties in dispute, developing countries may feel at a disadvantage when facing off against larger, richer countries. However, WTO dispute settlement decisions do not seem to support this fear. Of the panels that have been decided and implemented to date, several have been won by developing countries. This includes the first case heard by the WTO dispute settlement body (DSB), a case brought by Venezuela and Brazil against the United States concerning discrimination in the U.S. gasoline market. The Standards for Reformulated and Conventional Gasoline case included all steps of the DSM: recourse to the WTO mechanism, appeal of the findings of the DSB, a period of bilateral consultations, an ultimately implementation of the WTO decision within the stipulated 15 months (Rodriguez, 1998). Other developing countries that have achieved panel rulings in their favor include Costa Rica, and India, and Guatemala, Honduras and Mexico, acting with the United States. While developing countries are certainly at a disadvantage vis-à-vis advanced countries in terms of financial, administrative and technical resources, they do have access to technical assistance provided by the WTO and other international bodies. Their lawyers and negotiators seem to be learning how to maneuver the waters of international trade law. One way developing countries could cut their legal WTO-related costs would be to use multilateral and national technical assistance funds to set up a special "pool" of trade lawyers, trained in the use of the WTO law for use in developing country DSB cases. This pool of legal knowledge could be further tapped in the sometimes complex and cumbersome task of implementing WTO decisions.

Antidumping and countervailing duties. A less auspicious trend must also be noted, in parallel with the increased use of WTO dispute settlement mechanisms by developing countries; antidumping and countervailing duty measures are increasingly used as instruments of episodic protection by developing countries. Until 1986, 95 percent of all cases were brought in the United States, Europe, Canada and Australia; however, over the past three years, more than half of antidumping cases have been brought in "new user" countries (Horlick, 1998). In this respect, they are following the lead of advanced countries, and building on their increased knowledge of permissible safeguards in the multilateral trading system. This has the potential to create an environment in which WTO obligations are filled—yet countries are able, during difficult economic times, to erect protectionist barriers against imports from their trading partners through the use of trade remedy laws.

Intellectual property rights. Trade Related Intellectual Property Rights (TRIPs) agreement incorporates fundamental IPR concepts from existing treaties, but expands the minimum standards of protection, especially for many developing countries. Intellectual property protection has been, and continues to be a contentious issue in international trade. The inclusion of intellectual property within the WTO trading system arose through pressure by the advanced countries to protect their investments in high-tech and creative industries. Many developing countries have questioned the assertion that stronger IPR protection will stimulate local innovators; instead they worry that TRIPS may retard their economies by raising the price of necessities such as medicines and computer software.

TRIPs set out a comprehensive global standard for the protection of copyrights, patents, trademarks, industrial designs, and trade secrets. The issue of intellectual property protection and its impact on welfare, especially for developing countries, is discussed at length elsewhere (Maskus, 1998; Abbott, 1998, for example) , and will not be treated here. It is worth mentioning, however, that this is an area of growing complexity, linked to other issues on the future agenda of the WTO. Complexity derives not only from the legal and economic intricacies of the covered subjects—for example trademarks, pharmaceutical patents and parallel imports—but also from issues on the agenda of new communities that now participate in the WTO, issues such as biological diversity, herbal remedies and traditional knowledge. Future WTO negotiations that will have links to the IP arena include electronic commerce, biotechnology and genetics, and competition policy. The challenge for the multilateral trading system will be to address these issues in a balanced and comprehensive manner, taking into account the varying interests and legal systems of the participants.

Regionalism

A fair characterization of regionalism in the 1990s is that three super-regional groups are in existence or forming—the European Union and associates, NAFTA and associates, and Mercosur and associates—plus a few strong regional arrangements such as the Australia-New Zealand CER and ASEAN. The subject of regional arrangements is comprehensively covered in a draft World Bank report (1998 ) as well as in a recent volume by Frankel (1997) and the extensive bibliography cited there. Here, regionalism requires only brief notice.

The number of regional agreements notified to the GATT/WTO seems staggering—about 180 as of 1998. Another 50 or so have not been notified. At this juncture, only a handful of WTO members are not also members of at least one regional group—Japan, Korea, and Hong Kong being the most prominent non-regionalists.

However, two observations qualify the impression that regionalism is running amok. Of the 180-odd notified agreements, at least half are defunct. Among the rest, about half are FTA arrangements among the European countries.⁵ To a large extent, these are staging arrangements for ultimate membership in the EU itself or a EU-centered customs union. Likewise, in the Western Hemisphere, many FTAs have been launched in the hopes that they will ultimately merge into a hemispheric FTAA, or as a fall-back into a South American FTA or customs union centered on Mercosur.

The post-1980 regional arrangements tend to be more pragmatic than those of the 1960s and 1970s —both in opening markets to outside goods and services and in seeking to establish policy credibility as a component of broad economic reform. Common external tariffs in today's customs unions tend to be lower than were national barriers before the creation or consolidation of the arrangement. Members of some FTAs have seen an advantage in unilateral liberalization to improve their prospects of attracting MNCs and direct investment.

Modern regional arrangements tend to cover merchandise trade in all industrialized goods and some agricultural goods and provide for the elimination of quantitative restrictions and other nontariff barriers. Coverage of services and investment is considerably less extensive. For the subjects covered, implementation time frames have been shortened as the coverage of goods has increased. Despite the heated debate between those who see regional trade arrangements (RTAs) as inimical to the multilateral system, and those who see RTAs as the answer for less developed economies, it seems likely that in ten years this wave of regional arrangements will be seen as a way station in global economic integration. Many will fall into disuse, some will become obstacles to further liberalization, and a few will carry their internal liberalization initiatives into the WTO.

Role of Developing Countries in Regional Arrangements

As mentioned earlier, approximately 180 regional arrangements have been notified to the GATT/WTO, and another 50 or so have not been notified. At least on paper, nearly every WTO member (of which there are 133) belongs to at least one regional arrangement. However, at least 100 of the arrangements exist only on paper. Of those with some vitality, at least half are agreements of one kind or another between and among the European Union and its numerous candidate and associate members. Against this background, several stylized observations can be made about the role of developing countries in regional arrangements.

• Regional arrangements between the EU and its candidate countries are essentially staging arrangements prior to full EU membership. They serve as vehicles to acquaint the candidate with the numerous commercial and political obligations of EU membership. There is no question of renegotiating the basic EU acquis; the only question is the pace at which the candidate country fully opens its markets to the EU and vice versa. The phasing of agriculture liberalization is usually sensitive for both parties.
• Regional arrangements among the candidate countries serve three functions: to enlarge the intra-European web of trade, prior to enlargement of the EU; to ensure that early candidates become accustomed to freer trade with other candidates prior to becoming EU members; and finally to ensure that FTAs among candidates do not entail major concessions that will have to be renounced when one of the candidates becomes a EU member.
• Regional arrangements between the EU, Turkey and some North African states are intended to enlarge the zone of free trade without pointing to EU membership with its political commitments and its labor mobility requirements.
• Most regional arrangements in Africa are defunct. The major exception is the Southern Africa regional arrangement, anchored by South Africa with bordering states as members. South Africa has essentially set the terms of this arrangement.
• The South Asian regional agreement, anchored by India and intended to include most of its bordering countries, has little economic content. Until India and Pakistan can reach a minimal level of accommodation, this agreement has little promise.
• The ASEAN arrangement and the promised ASEAN Free Trade Area (AFTA) have yet to reach their economic potential. ASEAN was born as a political agreement, designed to quell fighting and proxy guerilla warfare among neighboring countries, and it has succeeded in that objective. However, it has not achieved free trade, especially not in sensitive manufactured items like automobiles and textiles, nor in agricultural products.
• In Latin America, one regional agreement dating from the 1970s, the Andean Pact, was reinvented and renamed in the 1990s. The original Andean Pact was a highly institutionalized arrangement, with production-sharing and investment-limiting features, and very little free trade. The Andean Community of the 1990s has genuine free trade in a few sectors, investment liberalization, and an approach to external trade that is half FTA and half customs union.
• Two other Latin regional agreements dating from the 1970s, the Caribbean Common Market (Caricom) and the Central American Common Market (CACM), were paper agreements for many years and even now are barely stirring.
• The most dynamic regional feature in Latin America is Mercosur, a customs union between Brazil, Argentina, Paraguay and Uruguay that entered into force in 1995. Like the original European Common Market, the Mercosur served an important political purpose, namely defusing tensions between Argentina and Brazil. Unlike earlier Latin American regional agreements, the Mercosur countries reduced their external barriers, liberalized investment, and tackled difficult sectors (autos, computers, electric power).
• In addition to these agreements, a number of Latin American free trade agreements have been negotiated as potential staging arrangements for the members to join NAFTA, Mercosur, or the promised Free Trade Area of the Americas (negotiations are to be completed by 2005). For example, the Chile-Canada FTA, the G-3 FTA between Mexico, Colombia and Venezuela, and the Bolivia-Mercosur FTA are in part staging arrangements. The difference between such agreements in the Western Hemisphere and their counterparts in Europe is that the end game in the Western Hemisphere is much less certain.
• Finally, a number of developing countries participate in one-way preference schemes with the European Union (the ACP countries) and the United States (the CBI countries and potentially several African countries). Since these are one-way arrangements, they are generally tailored by the advanced nation to exclude or limit imports of sensitive products—often products that would most benefit the developing country exporters.

From these stylized facts and other aspects of recent economic history, we distill several lessons that developing countries can draw from the post-1980 history of regional arrangements.

The first lesson is that small and medium-sized developing countries find it very hard to create a genuine FTA solely among themselves. They are hard because the potential loss of tariff revenue from third-country transshipments, coupled with high reliance on tariffs as a share of fiscal receipts (often more than 10 percent), can only be squared through devilishly tight rules of origin. On a practical level, a certain amount of pre-FTA trade is necessary so that business firms find it worthwhile lobbying to eliminate the remaining tariff and nontariff barriers.

The second lesson is that, when seeking accession to a larger group (i.e., the European Union or NAFTA), or preferences from the larger group, small and medium-sized countries must play the role of demandeur. The role of demandeur requires that small and medium-sized countries: (a) accept the political cost of free trade and investment (reserving very few, if any, sectors); (b) accept the regulatory framework of the larger group; and (c) field well qualified experts to prepare the ground for negotiations (a good place to start is by reviewing the WTO commitments of both parties). Moreover, small and medium-sized countries must be prepared to wait a long time and accept a quarter-loaf or less when seeking preferences. These were the experiences of Mexico in joining NAFTA, of Chile in trying to join NAFTA, of Caricom and CACM in seeking enlargement of CBI preferences, of the Andean countries in negotiating Andean preferences from the United States, and of North African states in seeking preferences form the European Union. The major counter-example is Mercosur: pursuing a South America vision, Brazil was anxious to sign FTA deals with Chile, Bolivia and others.

The third lesson is that the biggest economic gains to small and medium-sized developing countries are linked to cross-border investment and takeovers. The spread of direct investment and multinational companies (MNCs) boosts productivity, facilitates intra-firm trade, and enlarges the distribution possibilities for goods and services produced in the small and medium-sized countries. However, many local firms will not survive as independent companies. Only a few will expand their horizons to encompass the regional arrangement (e.g., Cimex, based in Mexico, has become a North American cement producer, and YPF, based in Argentina, could become an important second-tier energy company).

The fourth lesson is that an FTA without a single currency requires small and medium-sized members to orient their fiscal and monetary policies towards exchange rate stability—or suffer the consequences. A true FTA implies wide scope for portfolio capital flows, originating from hot money speculators, the intra-firm transactions of MNCs, and wealthy households. To buffer the effects on their exchange rates, the small and medium-sized members must be prepared to use fiscal and monetary policy in an aggressive, disciplined and often unpopular manner. Even with virtuous macro policies, the member can experience extreme exchange rate pressure (note the recent woes of Singapore, Mexico and Argentina).

Section III: Trade and Labor Standards

In the last five years, new themes have crowded onto the trade agenda: notably, labor standards, environment, competition policy. In this section, we discuss only one of the new themes, labor standards. There are obvious parallels with trade and environment, fewer with competition policy.

With developing countries now embracing the themes of globalization, labor advocates in advanced countries fear the impact on unskilled workers at home. Building on economic theory and recent wage and employment statistics, trade unions are calling for the adoption of international labor standards to improve working conditions in developing economies. Supporters of free markets, however, see the proposed links between labor standards and trade sanctions as a move towards protectionism. This section surveys the debate, starting with the connection between trade and wages. The World Development Report 1995 contains an excellent review of the literature.

The Impact of Trade on Wages

The most striking feature of labor markets in advanced countries is the widening gap between the fortunes of skilled and unskilled workers.⁶ In some OECD countries, such as the United States, Canada and Australia, the divide takes the form of increasing wage differentials.⁷ In Europe and Japan, the divide is visible in higher unemployment rates among young and unskilled workers (see Table 2). Since these changes of fortune have occurred at the same time that all countries have become increasingly intertwined with the world economy, many observers suspect that globalization is the culprit. Well-known economic theories seem to support this position, but a close reading of the data suggests otherwise.⁸

The celebrated Heckscher-Ohlin (HO) theory predicts that countries will export the products that use their abundant factors intensively—capital-intensive and skill-intensive goods in the case of advanced countries, and labor-intensive goods in the case of developing nations. Building on this theory, the Stolper-Samuelson (SS) theorem shows that, given several restrictive assumptions, freeing trade will reduce the returns to the scarce factor of production in each country. The related "factor-price equalization" theorem (FPE) leads to an even more dramatic conclusion: the wage structure in advanced and developing countries will eventually converge to a common level. In its most extreme form, the FPE theorem has led to fears of a worldwide "race to the bottom": advanced countries will lower wages and labor standards to levels not seen since the Great Depression.

Casual "post hoc ergo propter hoc" observations appear to support these fears. During the period from 1979 to 1988 in the United States, the ratio of the average wage of a college graduate to the average wage of a high school graduate increased by over 15 percent (Bound and Johnson 1992, 371). What is even more striking is that the widening gap occurred even as the share of college graduates in the work force increased from 19 percent to 26 percent, while the share of high school graduates remained constant at about 33 percent (Cline 1997, 28). Whereas in 1979, there were 4 high-school-educated workers for each college-educated worker in import-competing industries, there are now only 2; the ratio in export industries has also been cut in half (see Table 3). An oft-quoted statistic shows that the real average hourly earnings of production workers decreased by almost 15 percent between 1973 and 1994; however, the real hourly compensation of production workers (a measure which includes the self-employed and also fringe benefits) increased by 9 percent over the same period (Lawrence 1996, 17).

The shift in wage rates has also been mirrored by a shift in employment. Whereas the manufacturing sector accounted for 27 percent of US employment in 1970, its share fell to 17 percent by 1990; concurrently, imports rose from 11 percent to 38 percent of manufacturing value-added (Krugman and Lawrence 1993, 3). Many see a causal relationship: free trade has adversely affected unskilled workers. In other words, the logic of the HO, SS and FPE theories seems to be borne out by recent experience.

However, several counterintuitive facts must be added to the story. The overwhelming share—about 70 percent—of advanced country trade is between advanced countries themselves (see Tables 4 and 5). In fact, imports as a percentage of gross national product (GNP) in 1996 were on average only 16 percent in advanced countries, of which only 5 percent came from developing countries (see Table 6). More than half the imports from developing countries consist of petroleum and other natural resource products. It is difficult to believe that non-resource imports from developing countries, which amount to less than 3 percent of GDP in advanced countries, could alone cause massive changes in the fortunes of skilled and unskilled workers.

Other factors also indicate that trade is not the main reason for the rising disparity in wages between different skill levels in advanced countries. The most important qualification to the FPE theorem is labor productivity: even though wage rates in developing countries are a fraction of those in advanced countries, so is labor productivity. Plant-level technology and managerial skills are far lower in developing countries, but a key assumption underlying the FPE theorem is equivalent technology between trading partners. Because technological differences are so great, it turns out that unit labor costs are approximately the same in countries with vastly different wage levels, leaving firms in advanced countries with only weak incentives to relocate production to developing countries (Golub 1998). In fact, the great bulk of foreign direct investment is between advanced countries, not from advanced to developing countries.

As a consequence, the most attractive feature of many developing countries is simply their lower wage rates, and that advantage alone is not enough to entice a flood of investment. A better way for developing countries to make themselves more attractive investment locations is to embrace the free trade and investment agenda of the super-regional groups (the European Union and NAFTA). So far, few developing countries have followed that route (Singapore, Hong Kong, and Mexico are exceptions; so were Ireland, Spain, Portugal and Greece two decades ago).

Despite the stringent assumptions that are necessary to fully realize the predictions of the HO, SS and FPE theorems, most economists see some substance to fears that trade has reduced the demand for unskilled workers in advanced countries. However, opinions differ as to the magnitude of the impact. Adrian Wood calculates that the composition of trade between advanced and developing countries has reduced the relative demand for unskilled workers in advanced countries by a stunning 22 percent (1995). This is more than ten times the effect that Robert Lawrence attributes to North-South trade—under 2 percent (1996). William R. Cline calculates that North-South trade added 6 percentage points to the wage differential in advanced countries over the period 1973-1993, while other globalization factors—outsourcing and immigration—contributed respectively an additional 1 percentage point and 2 percentage points of the differential (1997). Cline's estimate is somewhat larger than Lawrence's, but significantly smaller than Wood's. Cline and Lawrence, among other economists, find that while North-South trade has contributed to the increasing inequality between skilled and unskilled workers, it is not the largest factor.⁹

A factor that many scholars regard as more important than trade in explaining rising wage inequality in advanced countries is skill-biased technological change. Workers in high import-share industries tend to be less educated (see Table 7) and are therefore more vulnerable to skill-biased technological change. Cline finds that technological changes explain 29 percent of the rising disparity in wages between skilled and unskilled workers (1997, 264), while Lawrence attributes 7.5 percent of the rising differential to this explanation (1996, 69). In fact, Krugman and Lawrence conclude that "the old-fashioned concern about loss of manufacturing jobs because of automation is closer to the truth than the current preoccupation with loss of manufacturing jobs because of foreign competition" (1993, 8).

Strides in technological progress may however be related to globalization. In a global economy, characterized by abundant unskilled labor in developing countries, firms based in advanced countries may find that it makes more sense to enhance mind-saving technologies than to spend money creating the sort of labor-saving devices that hallmarked the industrial age. Consequently, technological progress may be biased towards increasing the productivity of skilled, rather than unskilled, workers.

Developing Countries

The Stolper-Samuelson theorem (SS) implies that free trade would reduce the wages of skilled workers while increasing the wages of unskilled workers in low income countries. To the contrary, recent studies in Latin America and Asia show rising wages of skilled workers relative to their unskilled counterparts. (Asian Development Bank 1998, and Lusting and Deutsch 1998). In most developing countries, manufacturing employment is a relatively skilled activity. During the period 1970-1990, manufacturing wages grew fastest in countries that were most open to trade (World Bank 1995, 55). This implies two things. First, skill-biased technological growth is a worldwide, not just an advanced country, phenomenon. Second, with the spread of multinational companies (MNCs), and the rapidly falling costs of computers and telecommunications, it is getting easier for manufacturing workers and skilled professionals to sell their goods and services into the markets of Europe and the United States.

International Labor Standards

While the call for international labor standards is not new, with globalization it has acquired new urgency. In order to lower their costs, the argument runs, MNCs move from country to country in search of lower wages. Poor, capital-hungry nations court these MNCs by touting their cheap labor force and minimal monitoring of working conditions. Thus, skeptics say, globalization "…pulls the law [on workers' rights] downward in search of the lowest common denominator" (Greider 1997, 34). They argue that this vicious circle can only be countered by international labor standards, adopted through the International Labour Organisation (ILO) and enforced through trade agreements and the World Trade Organisation (WTO).

Most developing countries disagree. In a report presented to the ILO in 1997, 113 developing countries asserted that standards should be recognized as "benchmarks in the process of development" (quoted in Financial Times, 12 June 1997, 6); moreover, they fear that trade sanctions could easily be hijacked by protectionists. They insist that the ILO, and not the WTO, is the only body qualified to deal with labor standards, and that there should be no link to trade sanctions.

In reality, the ILO is limited in its ability to influence working conditions. Apart from its technical assistance program, the ILO's main activity is the establishment of non-binding Recommendations and Conventions. Member countries can choose whether or not to adopt these provisions, but they are required to report progress (or lack thereof) on Conventions they have adopted. Countries are subject to investigation by the ILO when other countries complain about the violations of standards, but offending nations do not face ILO punishment beyond moral censure.

Supporters of stricter international labor standards, notably trade unions in the United States and France, have grown impatient with the ILO's inability to enforce standards, and have pressed for a WTO link between standards and trade. Developing countries object. After much debate during the WTO's first ministerial level meeting in Singapore in December 1996, the WTO tried to bow out of the labor standards debate, stating the ILO was the "competent body to set and deal with these standards" (quoted in Charnovitz 1997, 156).

However, when Michael Hansenne, the then ILO Director-General, called for increased standards-setting activity by the ILO only seven months later, there were immediate cries of outrage among developing countries. Although Hansenne limited his proposals to targeted standards, greater adoption of ILO recommendations, more stringent evaluations, and perhaps the creation of an "overall social label", 113 non-aligned countries plus five observers (including China) accused the ILO of "…introduc[ing] an untenable link between labor standards and trade which we do not accept. The ILO has no role with regard to the multilateral trading system nor is it mandated to promote or impede globalization" (quoted in Financial Times, 12 June 1997, 6).

Developing countries have since tempered their stand. At the 86th International Labour Conference, held in 1998, the 174 members of the ILO adopted the "Declaration on Fundamental Principles and Rights at Work" which commits them to respect the principles inherent in seven core labor standards. The Declaration emphasizes that "labour standards should not be used for protectionist trade purposes".

To that end, supporters of labor standards (and of free trade) differentiate between standards that are centered on human rights (such as freedom of association) and those that are essentially economic (such as minimum wages). In other words, a distinction is shaping up between "core" and "cash" labor standards. Advocates of this distinction argue that "Adherence to core standards will not substantially affect the comparative advantage of developing countries nor have more than a minimal effect on trade" (Freeman 1996, 87).

Rather than forcing governments to regulate standards, some propose a market-based solution: create labels that would attest to the working conditions under which a good is made. US consumer groups are pushing this approach, and some labels already exist, a famous example being the Rugmark label for carpets which guarantees that the carpet was not made with child labor. A labeling program can only function on a wide scale, however, if each business purchaser is responsible for monitoring the working conditions of its suppliers.

A companion free market approach to promoting labor standards is the adoption of Codes of Conduct, which typically list minimum labor standards. The firm that adopts the code then advises its sub-contractors that if they are found in violation of the Code, they will lose the contract. The language in Codes is often vague and enforcement can be lax, but steps have been taken to address those problems, starting with the establishment of a universal standard for ethical sourcing. That initiative has just begun, and it is uncertain how many companies will agree to be audited, and how tough the audits will be.

Codes and similar free market approaches to improving labor standards can meaningfully impact only the export sectors of developing countries. These approaches exert little effect on other industries and sectors that are even more prone to labor abuses—such as agriculture, mining for local use, the informal services industry, and prostitution. Developing countries themselves must correct the abuses in these sectors, which typically account for more than three-quarters of all workers.

Future Developments

Pessimists project a continuation of rising wage differentials and long-term unemployment in advanced countries; however, we expect the trend to change in favor of unskilled workers in advanced countries. Although one year's data does not make a new trend, figures assembled by the Economic Policy Institute suggest a modest narrowing of U.S. wage differentials between 1996 and mid-1998 (Wall Street Journal, 17 July 1998, A2). Meanwhile, European unemployment rates are edging down. Several basic factors that could presage a trend change can be listed.

• The information revolution is transforming lower-paid service jobs by simplifying skill requirements; high speed checkout counters with scanners that record prices and keep track of inventory illustrate this trend. This technology enables some unskilled workers to occupy jobs that once required far more training.
• Furthermore, faced with large wage differentials and prolonged layoffs, unskilled workers are strongly motivated to acquire new skills. Faced with tight labor markets, as in the United States and selected parts of Europe, Australia and Canada, firms are motivated to offer training programs. To some extent, governments are also aiding individual efforts with public training programs.
• Finally, the aging population of advanced countries will create a huge demand for personal services that are largely provided by less skilled workers.

Trade will also play a role, specifically trade in services. In 1991, international cross-border trade in services (measured on a balance-of-payments basis) was worth about $900 billion; at least another $900 billion of services business was done through overseas affiliates of service providers, which fall outside the scope of balance of payments accounting. With future technological advances, a multitude of services once regarded as nontradeables will become far more tradeable than petroleum or semiconductors, and trade flows will grow rapidly.

One effect of technological advances and the growth in services trade will be to trim the salary premiums enjoyed by highly-skilled workers in advanced countries and to boost the earnings of similar workers located in developing nations. Indeed, over the next few decades, this trend could provoke a defense of economic rents in advanced countries. For example, university professors may try to stop the television classroom, pension-fund managers may attempt to erect barriers to foreign money managers; U.S. and European actuaries may question the professional credentials of their Indian counterparts, and U.S. and European airline pilots may try to block the use of foreign pilots or cabotage routes. The policy debate will center on appropriate shields for a wide range of service activities that face competition imported electronically from abroad. Specific