International Trade

Helen Milner. Handbook of International Relations. Editor: Walter Carlsnaes, Thomas Risse, Beth A Simmons. Sage Publication. 2002.

International trade has become one of the most potent issues in both domestic and international politics these days. Under the rubric of globalization, international trade has become a contentious issue in domestic politics, as the recent WTO conference in Seattle showed. In international politics, trade is today a premiere instrument of statecraft, as witnessed by the US–China trade agreement and the EU’s accession negotiations with the countries of East and Central Europe. How can we explain the trade policy choices that states make? What theories do we possess that illuminate the nature of countries’ trade relations?

It is important to note that trade has become such a critical issue largely because countries’ economies are now more than ever open to trade flows. This has occurred both because of technological changes as well as government policies. Since the 1970s countries across the globe have adopted freer trade policies. Many lesser developed countries, like Mexico, India, Poland, Turkey, Ghana and Morocco, have chosen to unilaterally liberalize their trade policies. In addition, the successful conclusion in 1994 of the multilateral trade negotiations under the GATT (the Uruguay Round) further liberalized trade among many developed countries and between them and developing ones. This global ‘rush to free trade,’ as Rodrik (1994) has called it, is important because it has helped further integrate countries into the world economy. But it has also increased their exposure to the pressures of quickly changing global markets, thus upsetting domestic politics at times.

The scholarly literature on international trade is vast. Both economists and political scientists have contributed much to it, as recent surveys by economists such as Reizman and Wilson (1995) and Rodrik (1995) and political scientists such as Cohen (1990) and Lake (1993) demonstrate. This chapter will focus more on the contributions of political scientists, but will include the research of economists where it is particularly important.

Much of the research by economists in international trade has dealt with topics that political scientists have not examined. By and large economists remain very interested in the three issues: the composition and direction of trade flows and the welfare effects of trade. Why certain countries import and export particular goods or services to certain other countries has been a central question for them. Much theory in international trade addresses this question; for instance, one of the central theorems in trade theory, the Heckscher–Ohlin one, explains trade flows. Economists have also devoted attention to the issue of trade barriers. The central theoretical conclusion of the field, of course, has been that free trade is the best policy for most countries most of the time. Thus economists have puzzled over why, given this finding, countries invariably employ at least some protectionist policies. They have tended to ask why countries protect certain of their industries, when free trade would be better economically. By and large their answer has focused on the preferences of domestic actors for protection. Using the Stolper–Samuelson theorem and other economic theories, they have explored why certain domestic groups would prefer protection and why they would expend resources to lobby for it. A large part of this debate involves whether specific-factors models of trade perform better than Stolper–Samuelson type models depending on factor endowments. This has resulted in a large empirical literature examining levels of protection across industries and recently in the development of models of such protection. Ultimately, then, economists have been pushed into studying the politics of trade. But a great deal of this literature explains why protectionist policies should never change, which is anomalous given the dramatic changes in trade we have seen (e.g., Drazen, 1996; Fernandez and Rodrik, 1991).

In contrast, political scientists have rarely focused on explaining the composition or pattern of trade flows, and they have been less concerned with the welfare consequences of trade. Only some recent work has explored the political roots of import and export flows among countries. But like economists they have been interested in the issue of protectionism. However, they have tended to see trade protection as more of the norm and have puzzled more over why a country would ever liberalize its trade policy or adopt free trade. Politically, protectionism seems eminently reasonable. Explaining both protectionist and free trade policies and changes in them over time have occupied political scientists.

There are at least four sets of factors that political scientists refer to when trying to understand trade politics. In this chapter I will survey how these four factors have been discussed in the literature. First, some focus on the preferences of domestic groups for protection or free trade. These scholars see trade policy as ultimately being shaped by the preferences of strongest groups in domestic politics. The questions of central import here are why do some groups favor protection, and some free trade. Do these preferences change over time? And if so, why? Which groups have a greater ability to have their preferences heard and translated into policy?

Second, political institutions may affect the formation of trade policy. Much as in the macroeconomic issue area where independent central banks are an important factor, political institutions may matter for trade. They may shape the ways in which the preferences of actors are translated into policy. They may affect which domestic groups have the most access and voice in policy-making. Generally, institutions may be important for aggregating preferences and implementing policy. Changes in institutions may provide a natural way to examine their impact.

Third, some claim that factors at the international level shape trade policy choices. The nature of relations among countries and the structure of the international system may affect domestic choices about trade. Hegemonic stability theory was an early structural theory of trade. Finally, some scholars have asked whether and how international trade itself affects states and the international political system. They use trade as an independent variable. The debate on globalization is especially relevant here. Some claim that rising trade flows produce important changes in domestic preferences, institutions and policies. The rest of this chapter asks how political scientists have addressed these four central issues about trade politics.

What do we know about Trade and Trade Policy?

Since the Second World War, the main instrument of trade policy, tariffs (which are taxes on imports), among advanced industrial countries have been reduced to insignificant levels. After the latest round of international trade negotiations sponsored by the GATT–the Uruguay Round, completed in 1994–the average tariff for the developed countries was reduced from 6.3 per cent to 3.8 per cent (World Trade Organization, 1996: 31). Non-tariff barriers (NTBs), which include quantitative restrictions, price controls, subsidies, voluntary export restraints (VERs), etc., on the other hand, have proliferated, in part countering the decline in tariffs. But again the Uruguay Round slowed or reversed this, helping to reduce quotas, subsidies and VERs across a wide range of industries and to convert these barriers into more transparent tariffs (World Trade Organization, 1996: 32). Nevertheless, while tariffs have declined for advanced industrial countries, NTBs still make up an important arsenal of barriers to trade. For these countries close to 20 per cent of all categories of imports are subject to some form of NTBs (Laird and Yeats, 1990).

For most of the postwar period, LDCs have used trade barriers extensively, many for the explicit purpose of import-substituting industrialization (ISI). But from the late 1970s, many developing countries began to liberalize trade and adopt an outward, export-orientation (International Monetary Fund, 1992). The conclusion of the Uruguay Round promoted this by reducing trade barriers in many areas of key interest to the LDCs, such as textiles and agriculture; it also brought many new developing countries into the international trade organization, the WTO (World Trade Organization), inducing them to follow its rules. In addition, the transition from command or communist economies to market-based ones by many countries in the 1990s further accelerated the trend toward trade liberalization globally. All of these changes have resulted in one striking fact about the period since 1980: there has been a far-reaching liberalization of trade barriers across the globe (Rodrik, 1994, World Trade Organization, 1996).

Concomitantly and in part a product of this, the growth of world trade has surged. For most of postwar period, the growth of trade has outpaced growth in world output. Also important are changes in the nature of global trade: there has been tremendous growth in intra-industry trade (IIT) and in intra-firm trade (IFT). IIT, which involves the exchange of goods from within the same industry, say Toyotas for BMWs, now accounts for between 55 and 75 per cent of trade in advanced industrial countries (Greenaway and Milner, 1986: Table 5–3); for the United States, this figure was 83 per cent in 1990 (Bergsten and Noland, 1993: 66). IFT, which involves transfers of goods within one company across national boundaries, has also grown; it now accounts for over 40 per cent of total US imports and 30 per cent of US exports (Encarnation, 1992: 28). These two types of trade are important because they tend to have different effects than standard, inter-industry trade. Generally, they are associated with fewer displacement effects and less conflict. As Lipson (1982: 453) argues, ‘intra-industry trade provides a powerful new source of multilateral interest in the liberal trade regime: diminished adjustment costs in some sectors, and higher net gains from trade as a result.’

Finally, there has been a significant regionalization of trade. Intraregional trade flows within the European Union, East Asia, North America and Latin America especially have become more important as a share of total trade. This is partially a result of the regional integration agreements signed by these countries in the past two decades–for example, the single market in Europe, NAFTA, ASEAN, APEC and Mercosur (WTO, 1996: 17–22). Also indicative of regionalism’s growth are the increasing rates at which preferential trade agreements (PTAs) formed and states joined them throughout the post-Second World War period (Mansfield, 1998). The number of regional agreements notified to the General Agreement on Tariffs and Trade (GATT) from 1948 to 1994 has waxed and waned. Few PTAs were established during the 1940s and 1950s; then a surge in preferential agreements occurred in the 1960s and 1970s, and the incidence of PTA creation again tailed off in the 1980s (de Melo and Panagariya, 1993: 3). But there has been a significant rise in such agreements during the 1990s; and more than 50 per cent of all world commerce is currently conducted within regional trade arrangements (Serra et al., 1997: 8). Indeed, PTAs have become so pervasive that all but a few parties to the WTO now belong to at least one (World Trade Organization, 1996: 38). This regionalization of the trading system has been treated as evidence both of increasing protectionism and of increasing liberalization. The key issue is whether these agreements, which lower barriers between participants, also lower barriers with non-members. If so, then they might foster greater trade liberalization globally; if not, then they may be a force for undermining the integrated world economy, creating exclusive trading blocs.

Trade Policy Preferences and Domestic Politics

Some of the earliest models explaining trade policy have focused on ‘pressure group politics.’ That is, they explain the trade policy choices by governments as a function of the demands made by domestic interest groups. Domestic groups seek protection or liberalization because such policies increase their incomes. The distributional consequences of trade policy thus become the explanation for its causes. Adam Smith ([1776] 1937) may have been one of the first to recognize this, when he noted that the subversion of the national interest in free trade is the frequent outcome of collusion among businessmen. Schattschneider (1935) was another early proponent of the view that special economic interests were mainly responsible for the choice of protectionism; he showed how these pressure groups hijacked the American Congress in 1929–30 and via a logroll produced one of the highest tariffs ever in American history, the Smoot–Hawley tariff.

Since then, development of the pressure group model has attempted to delineate more specifically the groups who should favor and oppose protection and the conditions under which they may be most influential. One motive for this has been the observation that the extent of protection and the demands for it vary both across industries and across countries. If all domestic groups always favored protection, then such variance should not exist. Explaining this variance has been a key feature of the literature. It has depended on theories about two factors: the sources of trade policy preferences and the nature of political influence of these interest groups.

The former area has been a more prolific research topic and a highly divisive one. The main divide over the sources of trade policy preferences has been between so-called factoral versus sectoral (or firm-based) theories of preferences. In both cases, preferences are deduced as a result of the changes in income that accrue to different actors when policy changes from free trade to protection or vice versa. These types of theories focus on the distributional effects of trade; they associate preferences for protection with those who lose (income or assets) from greater trade flows and preferences for liberalization with those who gain. Factoral theories rely on the Stopler–Samuelson theorem, which shows that when factors of production, like labor and capital, can move freely among sectors, a change from free trade to protection will raise the income of factors in which a country is relatively scarce and lower it for factors that are relatively abundant. Thus scarce factors will support protection, while abundant ones will oppose it. Rogowski (1989) has developed one of the most interesting political extensions of this, claiming that increasing (decreasing) exposure to trade sets off either increasing class conflict or urban–rural conflict according to the factor endowments of different countries.

In contrast, sectoral and firm-based theories of trade preferences follow from the Ricardo–Viner model of trade–also called the specific-factors model. This model claims that because at least one factor is immobile, all factors attached to import-competing sectors lose from trade liberalization while those in export-oriented sectors gain. Conflict over trade policy thus pits labor, capital and landowners in sectors besieged by imports against those who export their production. How tied factors are to their sectors–that is, the degree of factor specificity–is the key difference between these two models (Alt et al., 1996).

A number of studies have tested these two models, sometimes singly and sometimes simultaneously. Frieden (1990), Irwin (1994, 1996), Magee, Brock and Young (1989), have found evidence in support of the specific-factors model; in contrast, Balestreri (1997), Beaulieu (1996), Midford (1993), Rogowski (1989) and Scheve and Slaughter (1998) find support for the Stolper–Samuelson type factoral models. Despite such differences, most agree that domestic pressures from industry and labor play a significant role in both increasing protection and in preventing it (e.g., Milner, 1988).

In addition to these models of trade preferences, others have looked at how particular characteristics of industries affect patterns of protection. Anderson (1980), Baldwin (1986), Caves (1976), Marvel and Ray (1983), Pincus (1975), Ray (1981) and Trefler (1993) have shown how specific characteristics make an industry more likely not only to desire protection but also to be able to induce policymakers to provide it. These regression analyses tend to straddle the debate between sectoral and factoral models of trade politics. Their comparison across industries suggests a sectoral type of model, but many of their findings do not disagree with those resulting from a more factoral view of the world. For example, they tend to demonstrate that in advanced industrial countries low-skill, labor-intensive industries with high and rising import penetration are frequently associated with high protection. In addition, many have shown that export-oriented industries and multinationals tend to favor freer trade and be associated with less protection (Milner, 1988). This attention to anti-protectionist groups is particularly interesting given the global move toward trade liberalization; one question is whether this movement has been due to the growth in importance of these types of groups domestically.

Can these models of societal preferences explain trade policy? As noted above, many of these theories are fairly good at explaining variance across industries in any one country. But in terms of explaining overall directions in national trade policy and crossnational differences these theories have a number of weaknesses. First, they have no theory of how preferences are aggregated at any level, let alone the national one. If firms in an industry are divided over trade policy, how does the sector choose a policy to advocate? If some industries are opposed to liberalization and some support it, how can we predict whether political leaders will agree to international negotiations to reduce trade barriers? The issue of which groups are able to influence policy and which are not depends much on political factors, such as the clout of the industry or how institutions shape its access to policymakers. Second, might not these differences in preferences give policy-makers a great deal of leeway to implement their own preferred policies, thus weakening the influence of interest groups? Political leaders could simply pick and choose the groups that they wanted to ‘represent’ and then build coalitions around their own preferences, rather than being driven by industry pressures. The literature on interest groups in trade policy-making continues to wrestle with these issues.

The preferences of other domestic actors have also been the focus of some attention. Many assume that individual voters take their preferences from their role as consumers. Since consumers gain from free trade, they should favor it (e.g., Grossman and Helpman, 1994). Other models of individual preferences contradict this. Mayer (1984), for example, introduces an electoral component into the determination of trade policy. Trade policy is determined by the median voter’s preferences, which depend on that voter’s factor endowments. The more well endowed he (or she) is in the factor used intensively for production of import-competing goods, the more protectionist he will be. Scheve and Slaughter (1998) add a new component by asking how asset-ownership is affected by trade policy. They show that the preferences of individual voters will depend on how trade affects their assets. Individuals living in regions with a high concentration of import-competing industries will be more favorable to protection because as imports rise economic activity in the region will fall causing their housing assets to fall in value. Some surveys have also shown that voters respond positively toward protection out of sympathy for workers who lose their jobs because of import competition. Thus, whether individual voters favor protection or free trade is an area demanding further research, especially in democracies where elections are often linked to trade policy decisions. Moreover, understanding changes in these preferences may help us account for the recent push to liberalize trade.

A number of scholars have argued that the preferences of interest groups and voters are less important in determining trade policy than are those of the policy-makers themselves. Bauer, Pool and Dexter (1972) were among the first to make this point. From their surveys, they showed that constituents rarely had strong preferences about trade policy and even more rarely communicated these to their political representatives. Trade policy depended much on the personal preferences and ideas of politicians. Baldwin (1986) and Goldstein (1988) have also argued that it is the ideas that policy-makers have about trade policy that matter most. Rather than material factors determining preferences, ideational ones are paramount. Interestingly, Krueger (1997), an economist, claims that it is ideas that have mattered most in trade policy-making in the lesser developed countries lately. She argues that it is ‘ideas with regard to trade policy and economic development [that] are among those [factors] that have changed most radically’ from 1950 to the 1990s, helping to explain the recent rush to free trade. Many suggest that the failures of ISI policy and the glaring success of the export-oriented newly industrializing Asian countries in the 1980s forced policy-makers to adopt new ideas about trade policy. A key example of this is Fernando Henrique Cardoso, who co-authored one of the most important books in dependency theory in the 1970s, arguing for the continuation of ISI policies to shelter LDCs from the capitalist world economy (Cardoso and Faletto, 1979). In the 1990s, of course, Cardoso was elected president of Brazil and initiated a major economic reform program, including extensive trade liberalization. Changes in the ideas that policy-makers have about trade policy may then, as this example suggests, play a large role in affecting trade policy choices.

Economic conditions may also affect the preferences of actors and lead to changes in trade policies. While Krueger and others, such as Bates and Krueger (1993), Haggard and Kaufman (1995), and Rodrik (1995), attribute leaders’ decisions to initiate trade policy reform to crises and economic downturns, another strand of literature on the macroeconomics of trade policy concludes in the opposite direction. For many scholars, bad economic times are a prelude to rising demands for protection and increasing levels of protection. Cassing, McKeown and Ochs (1986), Gallarotti (1985), Magee and Young (1987), Takacs (1981) and Wallerstein (1987) all find that declines in economic growth or capacity utilization and/or increases in unemployment and imports tend to increase the demand and supply of protection. This earlier literature then sees policy-makers responding increasingly to the rising demands for protection from domestic groups in bad economic times.

The more recent literature, however, implies that bad economic times allow policy-makers more freedom to maneuver, so that they can overturn existing protectionist policies by blaming them for the bad times. For example, Rodrik (1992: 88–9) notes that

It is paradoxical that the 1980s should have become the decade of trade liberalization in the developing countries. Thanks to the debt crisis, the 1980s have also been a decade of intense macroeconomic instability. Common sense would suggest that the conventional benefits of liberalization become muted, if not completely offset, under conditions of macro instability.

But he (1992: 89) claims that ‘a time of crisis occasionally enables radical reforms that would have been unthinkable in calmer times.’ He argues that the prolonged macroeconomic crises of the 1980s were so bad that ‘the overall gain from restoring the economy’s health [in part via trade liberalization] became so large that it swamped distributional considerations [raised by such reforms]’ (1994: 79). On the other hand, others, especially Haggard (1995), have argued that crises reduce the room for maneuver of political leaders. They suggest that in the 1980s these leaders were almost forced to liberalize trade (and make other reforms) because of the lack of options and international pressures. Noting the difference between the 1930s and 1980s crises, Haggard (1995: 16–19) points out that

why external shocks and corresponding macroeconomic policy adjustments might also be associated with trade and investment liberalization … is puzzling. In the 1930s, balance of payments and debt crises spurred the substitution of imports … and gave rise to a more autarchic and interventionist policy stance. In the 1980s, by contrast, an inward-looking policy seemed foreclosed … The opportunities for continued import substitution were limited, and ties to the world economy had become more varied, complex and difficult to sever.

The effect of economic crises on a country’s decisions to liberalize trade thus seems contingent on a number of other factors, such as the prevailing ideas about trade, the extent of openness existing at the time and the influence of international factors.

A similar debate exists concerning the impact of the exchange rate on trade policy. Appreciation of the exchange rate may increase protectionist pressures because it increases imports and decreases exports, thus affecting the balance of trade preferences domestically (Mansfield and Busch, 1995). Others suggest that the effects of an exchange rate change may have little impact. For instance, Rodrik (1994: 73) shows that a devaluation, which is the opposite of an appreciation, increases the domestic price of all tradables–imports and exports, thereby allowing both import-competing and export-oriented sectors to benefit. But under certain conditions–for example, when foreign exchange is rationed, devaluations can work just like trade liberalization, prompting demands for new protection from import-competing sectors. Some studies reveal such an association between periods of currency devaluations and rising tariffs; Simmons (1994) points out that many of the same conditions–but not all–that drove states to devalue also pushed them to increase tariffs in the interwar period. Both policies were intended to increase demand for domestic output, thus counteracting the effects of the depression. Much debate continues over the macroeconomic conditions that produce increasing domestic pressures for protection and/or that induce policymakers to relent to or resist such pressures.

Can these preference-based theories explain trade policy? These theories seem best at explaining the domestic sources of opposition to and support for trade liberalization. The role of interest groups, voters and policy-makers are obviously important in sketching the domestic politics of trade. But how far can these theories go? Without a concomitant theory of which groups are able to organize and exert influence, theories about interest groups and voters are best able to explain the demand for trade policy domestically. The preferences of policymakers may play a different role. They may be more likely to explain the supply-side of trade policy; that is, they may indicate the willingness of political leaders to supply protection or liberalization, as separate from the demand for it. But our models of such preferences seem the most under-specified and post hoc. Why are some policymakers more favorable to protectionism than others? Why and when do their preferences change? Theories about the conditions under which policymakers will abandon ideas that produce ‘bad’ results and what ideas they will adopt in their stead are largely unavailable. In sum, theories of trade preferences seem to provide an initial level of explanation for the supply and demand for trade policy. But they cannot as of yet provide a complete explanation of this process.

Political Institutions

Can theories that focus on political institutions do better at explaining trade policy-making? A number of scholars have argued that political institutions, rather than preferences, play a major role in explaining trade policy. While preferences play a role in these arguments, the main claim is that institutions aggregate such preferences. Different institutions do so differently, thus leading to distinct policies. Understanding institutions is necessary to explain the actual supply of protection, rather than simply its demand (Nelson, 1988). On the domestic side, different institutions empower different actors. Some institutions, for example, tend to give special interest groups greater access to policy-makers, rendering their demands harder to resist. For example, many believe that the fact that the US Congress controlled trade policy exclusively before 1934 made it very susceptible to protectionist pressures from interest groups (Baldwin, 1986; Destler, 1986; Goldstein, 1993; Haggard, 1988).

Other institutions insulate policy-makers from these demands, allowing them more leeway in setting policy. Thus, some argue that giving the executive branch greater control over trade after the Reciprocal Trade Act of 1934 made trade policy less susceptible to these influences and more free trade-oriented. In general, concentrating trade policy-making capabilities in the executive’s hands seems to be associated with the adoption of trade liberalization in a wide variety of countries (e.g., Haggard and Kaufman, 1995: 199). As Haggard and Webb (1994: 13) have noted about trade liberalization in numerous LDCs, ‘In every successful reform effort, politicians delegated decisionmaking authority to units within the government that were insulated from routine bureaucratic processes, from legislative and interest group pressures, and even from executive pressure.’

Other aspects of political regimes may make them more or less insulated from societal pressures. Rogowski (1987), for example, has argued that policy-makers should be most insulated from domestic pressures for protection in countries having large electoral districts and proportional representation (PR) systems. Mansfield and Busch (1995), however, find that such institutional insulation does indeed matter, but often in exactly the opposite direction: greater insulation (that is, larger districts and a PR system) leads to more protection. Similarly, Rodrik (1998) shows that ‘political regimes with lower executive autonomy and more participatory institutions handle exogenous shocks better,’ and this may include their response to shocks via trade policy. Thus it is not clear that greater insulation of policy-makers always produces policies that promote trade liberalization; the preferences of those policy-makers also matter.

The administrative capacity of the state is also seen as an important factor shaping trade policy. It is well-established that developed countries tend to have fewer trade barriers than do lesser developed countries (Conybeare, 1982, 1983; International Monetary Fund, 1992; Magee et al., 1989: 230–41; Rodrik, 1995: 1483). Part of the reason is that taxes on trade are fairly easy to collect and thus in LDCs where the apparatus of the state is less well developed such taxes may account for a substantial portion of total state revenues (between a quarter and a half, according to Rodrik, 1994: 77). As countries develop, their institutional capacity may also grow, thus reducing their need to depend on import taxes for revenue. Thus the introduction of the personal income tax in 1913 in the United States made trade taxes much less important for the government, thereby permitting their later reduction. Hence political institutions and changes in them may help explain trade policy.

Large institutional differences in countries’ political regime types also may be associated with different trade policy profiles. Some have argued that democratic countries are less likely to be able to pursue protectionist policies. Wintrobe (1998) claims that autocratic countries will be more rent-seeking, and protection is simply one form of rent-seeking. Mansfield, Milner and Rosendorff (1998, 2000) also show that democratic pairs of countries tend to be less protectionist and more likely to sign trade liberalizing agreements than are autocratic ones. Many of the countries that have embraced trade liberalization have also democratized. Mexico is a prime case here. The growth of political competition and the decline of the hegemonic status of the governing party, the PRI, seem to have gone hand-in-hand with the liberalization of trade policy beginning in the 1980s.

In contrast, others point out that trade reform in many LDCs occurred before the transition to democracy and was often more successful when it did occur this way (Haggard and Webb, 1994). Chile, Turkey, Taiwan and South Korea all began their trade liberalization processes before their democratic transitions. Rodrik argues more generally that any change in political regime is likely to induce trade reforms: ‘Historically sharp changes in trade policy have almost always been preceded (or accompanied) by changes in the political regime … Not all political transformations result in trade reform, but sharp changes in trade policy are typically the result of such transformations’ (Rodrik, 1994: 69). None the less, Milner and Kubota (2001) find evidence that democracy in general and democratization have contributed to the lowering of trade barriers in a number of LDCs since the 1970s. Thus it may be that the character of political regimes has a direct effect on trade policy choices.

On the other hand, Verdier (1998) argues that because of the political conflict engendered by trade, democracies may be less likely to pursue free trade and more likely to adopt protection against each other, except when intra-industry trade dominates their trade flows.

The postwar democratic convergence among OECD countries did not hurt trade because similarity in endowments, combined with the presence of scale economies, allowed these countries to engage in intra-industry trade–a form of trade with few, if any, wealth effects … The current wave of democratization endangers trade. Only in the presence of scale economies [and thus intra-industry trade] can democratic convergence sustain trade. (Verdier, 1998: 18–19)

Haggard and Kaufmann (1995) are more circumspect, arguing that the presence of crises and the form of autocracy may have more to do with the ability to adopt economic reforms, like trade liberalization, than does regime type alone. Debates over the impact of regime type on trade policy have just begun.

The structure of the government and the nature of the party system have also been seen as an important institutional factor shaping trade policy. Parties very often take specific stands on trade policy, and their movement in and out of government may explain trade policy changes, as many have contended about the United States (e.g., Epstein and O’Halloran, 1996). In general, partisanship as a source of trade policy had been unexplored except in the United States. But theory suggests that partisanship and the nature of the political party system may matter greatly. For example, countries with highly polarized party systems, in which the main parties are separated by large ideological differences, may experience dramatic swings in policy and generally produce unsustainable trade reforms. On the other hand, countries with large numbers of parties may experience coalition governments frequently, which may be unable to change the status quo. Haggard and Kaufman (1995: 170) predict that countries with fragmented and/or polarized party systems will be unable to initiate economic policy reforms, including trade liberalization, let alone to sustain them. In general, these perspectives suggest that fragmented political systems are similar to ones with many veto players, and like them are resistant to change (Tsebelis, 1995).

Party systems also interact with the structure of the government. For example, Lohmann and O’Halloran (1994) and O’Halloran (1994) have argued that when government in presidential systems, like the United States, is divided–that is, one party controls the legislature and the other controls the executive branch–protectionism is likely to be higher. Milner and Rosendorff (1996) also argue that divided government in any country is likely to make the lowering of trade barriers either domestically or internationally harder in most cases. In sum, ‘political systems with weak executives and fragmented party systems, divided government, and decentralized political structures responded poorly to crises’ and were unable to mobilize the support necessary for the initiation of economic reforms, like trade liberalization (Haggard and Kaufman, 1995: 378). In all of these cases, however, the trade policy preferences of the parties matter for the outcome. Political institutions tend to affect whose preferences will become dominant in policy-making.

Many of these institutional arguments thus depend on prior claims about actors’ preferences. For instance, many of the arguments about insulation assume that the policy-makers (usually executives) who are insulated from societal demands are free traders. But as Mansfield and Busch (1995) show they may actually be protectionist, in which case insulation allows greater protection than otherwise. The arguments about divided government, party systems and democracies also rest to some extent on assumptions about each actor’s preferences. Divided government matters most when preferences of the parties differ, and differences in the preferences of autocratic leaders and democratic ones may be important for the implications of different regime types. Thus having theories that bring together both preferences and institutions seems most valuable, since we know that both matter. Very few studies, however, try to bring together theories of both preference formation and institutional influence; Gilligan (1997) and Milner (1997) are examples. Moreover, the matter of which comes first, preferences or institutions, is far from settled. Those who focus on preferences tend to argue that institutions are often shaped by the preferences of those in power; in contrast, those who emphasize institutions argue that they may actually shape actors’ preferences. The growing consensus is that both matter and are jointly determined, but parsimoniously modeling and testing this is an area for future research.

International Politics

Trade policy is not just affected by domestic forces. A number of factors about the international system have been connected to a country’s trade policy choices. A favored argument among realists has been that the distribution of capabilities in the international system has a fundamental effect on trade. The so-called theory of hegemonic stability (HST) posited that when the international system or economy was dominated by one country, a hegemon, then free trade would be most likely (Gilpin, 1987; Gowa, 1994; Krasner, 1976; Lake, 1988). A large number of critics have challenged this claim both theoretically and empirically (Lake, 1993; Keohane, 1998). Conybeare (1984) has shown that large countries should favor optimal tariffs, not free trade, even if others retaliate; Snidal (1985) and others have claimed that small numbers of powerful countries could maintain an open system, just as well as a single hegemon could. The theory has also faced empirical challenges that imply that a hegemon is neither necessary nor sufficient for an open trading system (e.g., Krasner, 1976; Mansfield, 1994). In light of these results, the theory has been modified as scholars examine more closely the dynamics of interaction among countries in the trading system.

Perhaps the most interesting point about this theory is that it has tried to explain change over time in the overall level of openness in the trading system; that is, it looks at the sum of countries’ trade policy choices. The main claim of this theory is that changes in the distribution of capabilities over time should provide clues to changes in the openness of the international trading system. In the 1980s, many argued that the decline of American hegemony from its zenith after the Second World War would lead to a rise in protectionism and perhaps the fragmentation of the international economy into rival blocs (e.g., Gilpin, 1987). This prediction, however, would not seem to explain well the rush to free trade witnessed since the mid-1980s.

One possible retort, however, is that US hegemony has risen, not declined, since 1980, as Russett (1985) and Strange (1987) have argued. Thus the renewal of American preeminence in the international system explains the turn away from protectionism. This type of argument would fit well with a broader claim about the dominance of American ideas about free markets and trade, and their impact on other countries’ trade policy choices. After all, the package of market-oriented reforms including trade liberalization that have been proposed for the LDCs and ex-communist countries is called the ‘Washington consensus.’ Others have also argued that American hegemony matters, but more through the direct pressure it may exert on lesser developed countries. Haggard (1995), for example, argues that changes in United States trade policy in the 1980s help explain the move toward free trade. The United States began exerting strong bilateral pressure on LDCs to liberalize their economies or face closure of the American market to their exports. American hegemony and its renewed will to exert influence may help explain recent changes in trade policy.

Other scholars have felt that aspects of the international security environment best explain the pattern of trade. Gowa (1994) has argued that countries which are military allies trade more with each other, and that is especially the case for those within the same alliance in the bipolar system. That is, when countries are allies in a system featuring one other major opposing alliance group, such as that during the Cold War, they will tend to trade the most freely among themselves. The security externalities of trade will drive their behavior, inducing them to help their allies while also punishing their enemies. Gowa and Mansfield (1993) and Mansfield and Bronson (1997) provide strong evidence for this effect. In terms of this argument, there should be a direct link between trade policy and the end of the Cold War and the dissolution of the Eastern bloc. Predictions from this model seem to be incomplete. The argument appears to suggest that protectionism should rise, not decline, with the demise of bipolarity and the emergence of multipolarity. A description of the current structure of the international system might be one of either multipolarity, or unipolarity, in which case the theory seems to have no single prediction.

Another aspect of the international system that scholars have noted for its effect on trade policy is the presence and influence of international institutions. Although a long debate has occurred over whether international institutions matter, many scholars now conclude that the fact that countries have been willing to set up and participate in such institutions would seem to imply that states feel that they matter (e.g., Keohane, 1984; Ruggie, 1983). In the trade area, a number of institutions provide support for an open, multilateral trading system; these include the GATT and its successor the WTO, as well as the International Monetary Fund (IMF) and World Bank. While regional trade institutions may have a more ambiguous effect on the multilateral system (Mansfield and Milner, 1999), some of them, including the EU, NAFTA and ASEAN, seem to have positive effects on lowering trade barriers and reinforcing unilateral moves toward freer trade.

These institutions are postulated to have a number of different effects on countries’ trade policy choices. Some suggest that their main role is to provide information about other countries’ behavior and compliance with the rules of the game (e.g., Keohane, 1984). Others see these institutions as providing a forum for dispute resolution so that partners in trade can feel more secure and thus more likely to trade (e.g., Yarbrough and Yarbrough, 1992). Others view such international institutions as encapsulating the norms by which countries agree to play the trading game, which again provides a common framework for sustaining trade flows (e.g., Ruggie, 1983). All of these arguments hypothesize that the presence of these institutions should be associated with a freer trade environment; moreover, they imply that the depth and breadth of these institutions should be positively related to trade liberalization and the expansion of trade.

Certainly the presence of institutions like the GATT and IMF have added leverage to arguments for trade liberalization; the IMF and World Bank have for instance often made loans conditional on trade policy reform. Some have argued that when countries are in severe economic crisis and need external financing, then these institutions may be especially powerful. As Rodrik (1992: 89) points out, ‘The 1980s were a decade of great leverage for these institutions [i.e., the IMF and World Bank] vis-à-vis debtor governments, especially where poorer African governments are concerned. The trade policy recommendations of the World Bank were adopted by cash-starved governments frequently with little conviction of their ultimate benefits.’ Others tend to argue that international institutions help lock in such domestic reforms. For example, Mexican unilateral trade liberalization seems much more secure now that Mexico is part of NAFTA and the WTO.

Finally, the creation of the WTO out of the GATT Uruguay Round represents a step toward the deeper institutionalization of an open trading system. The influence of these international institutions may depend either on the economic condition of debtors or on changing domestic preferences and ideas about trade. While there is little doubt that these institutions helped support trade liberalization globally, it seems likely that their influence varies over time and across countries (Haggard and Kaufman, 1995: 199). But, as with domestic political factors, these institutions may be an important element of the trade policy-making environment.

Effect of Trade on Countries and the International System

A final area of interest in the political economy of trade policy is the reciprocal effect of international trade on domestic and international politics. Once countries have liberalized or protected their economies, what might be the subsequent effects of such choices? Scholars have examined this question with attention to at least three aspects of the domestic political economy. First, some have argued that trade liberalization can in its wake change domestic preferences about trade. As countries liberalize, the tradables sector of the economy should grow in size along with exposure to international economic pressures. Rogowski (1989) has argued that this should lead to heightened or new political cleavages and conflicts between scarce and abundant factors domestically. These new cleavages in turn will alter domestic politics, as for example new parties arise to represent these groups or new coalitions form. Milner (1988) also argues that increasing openness to trade changes preferences domestically. Openness raises the potential number of supporters of free trade as exporters and multinational firms multiply; it may also reduce import-competing firms as they succumb to foreign competition. Hathaway (1998) presents a dynamic model that shows that trade liberalization changes industry structure in ways so that future demands for protection are reduced. ‘Trade liberalization has a positive feedback effect on policy preferences and political strategies of domestic producer groups. As industries adjust to more competitive market conditions, their characteristics change in ways that reduce the likelihood that they will demand protection in the future’ (1998: 606). James and Lake (1989) suggest an ingenious argument about how repeal of the protectionist Corn Laws in the UK created the necessary conditions for the creation of a successful coalition for free trade in the United States. Each of these arguments in distinct ways suggests that increasing exposure to trade leads to increasing pressure against protection, thus creating a virtuous cycle of rising demand for freer trade. As an explanation for trade policy in the advanced industrial countries over the past few decades, this type of argument seems plausible since these countries were increasingly exposed to trade. For the developing countries, their abrupt rejection of ISI and protectionism seems less explicable in these terms.

A second aspect of domestic politics that may be affected by increased trade flows involves the character of national political institutions. Among the advanced industrial countries, Cameron (1978) long ago noted the relationship between those that were very open to international trade and those with large governments. He and Katzenstein (1985) attributed this to the need for governments with open economies to provide extensive domestic compensation to the losers from trade and to employ flexible adjustment strategies for their industries. Rodrik (1997) has found strong evidence of this relationship around the globe. He claims that greater exposure to external risk, which trade promotes, increases the volatility of the domestic economy and thus that ‘societies that expose themselves to greater amounts of external risk demand (and receive) a larger government role as shelter from the vicissitudes of global markets’ (1997: 53). Increasing exposure to international trade may thus create demands for more government intervention and a larger welfare state, which in turn are necessary to sustain public support for an open economy.

Rogowski (1987: 212) has argued that as countries become more open to trade, they will find it increasingly advantageous to devise institutions that maximize ‘the state’s insulation, autonomy and stability.’ For him, this implies parliamentary systems with strong parties, proportional representation (PR) and large districts. He finds a strong relationship especially between openness and PR systems. Hadenius (1992) also finds that trade may have effects on domestic institutions. He argues that exposure to international trade brings higher rates of economic growth, which through the development process may translate into better conditions for the emergence of democracy. Thus trade liberalization may over time foster conditions conducive to political liberalization. This again suggests a virtuous cycle: trade liberalization fosters democratization and democracy in turn may promote more trade liberalization, and so on.

Besides its effects on preferences and institutions, trade may constrain the policy choices available to decision-makers. The recent literature on internationalization, or globalization, suggests this constraining influence. Rodrik (1997) provides some of the most direct evidence of how greater openness may force governments to relinquish the use of various policy instruments. In particular, he notes that openness often makes governments cut spending on social programs and reduce taxes on capital. In order to maintain competitiveness, governments are prevented from using many of the fiscal policy measures they once could.7 Whether such constraints are good or bad depends on the value one places on government intervention in the economy. For some, like Rodrik (1997), this constraint is worrisome since it reduces the government’s ability to shelter its citizens from external volatility and thus may erode the public’s support for openness. Here the impact of trade liberalization may not be benign. It may produce a backlash, undermining societal support for openness and creating pressures for protection and closure.

In terms of international politics, trade liberalization may also have important effects. As countries become more open to the international economy, it may affect their political relations with other countries. In particular, scholars have asked whether increased trade promotes peace between countries or increases their chances of conflict. A number of scholars, such as Polachek (1980), Gasioworski (1986) and Russett, Oneal and Davis (1998), have found that increases in trade flows among countries (or between pairs of them) decrease the chances that those countries will be involved in political or military conflicts with each other. Others such as Waltz (1979) and Barbieri (1996) argue that increased trade and the interdependence it creates either increase conflict or have little effect on it. One way that trade policy might affect the international political system then is by increasing or decreasing the level of political-military conflicts. There are a variety of different feedback mechanisms that might exist. For instance, if trade promotes pacific relations among trading nations, then such a pacific environment is likely to stimulate further trade liberalization and flows; on the other hand, if increasing trade produces more conflict, then we might expect more protectionism and less openness as a result.

These more dynamic models of how international trade and domestic politics interact are an important area of research. They may tell us a good deal about what affects trade policy choices. For example, will the global liberalization process bring increasing pressures for more openness and for democracy? Or will it undermine itself and breed demands for closure and a backlash against the governments and international institutions which support openness, as O’Rourke and Williamson (1999) have shown happened in the early twentieth century? Will openness produce a peaceful international system or one prone to increasing conflict? The answers to these questions will in turn tell us much about the future direction of trade policy globally.

Conclusion

I have examined preeminent theories of trade policy and see how they explain trade policy and changes in it. The point of this conclusion is to assess how well they have done and where future research might be useful.

What factors drive trade policy and changes in it? Existing theories suggest several answers to this question. The first involves preferences about trade policy among domestic actors. Economic theory suggests that domestic groups may have clear preferences about trade policy. If groups are rational and prefer profit maximization, then policies that increase profits should be favored. Whether factors endowments or sectors or firms are the correct unit of analysis, these models suggest that the demand for trade policy should follow clear patterns domestically. It is probable that these groups recognize their interests as well, since they are more likely to be organized and to receive large, concentrated benefits from policy. For voters the question is more difficult. Voters are consumers but they may also be workers and asset-owners as well; hence their preferences for trade may be torn in different directions. Moreover, voters’ capacity to organize is not well developed, as collective action theory suggests to us.

Political leaders, on the other hand, may be able to take action, but it is harder to deductively derive their preferences for trade policy. Should we conceive of them as benign leaders intent on maximizing national social welfare or as politically motivated leaders dependent on special interests for support and often maximizing their own personal interests? The former might lead us to attribute to them preferences for free trade, while the latter view would incline us to see them as protectionists. Under what circumstances, should we expect which behavior?

This question leads to a discussion of political institutions and their role in shaping trade policy. Both the influence of domestic groups and the preferences of leaders may depend on the political institutions in place. Some argue that democracy makes a difference. Political leaders may be forced to concern themselves more with the national interest than with just special interests, according to one theory; others claim the opposite: that in democracies the relentless search for political support makes politicians more sensitive to the needs of and subservient to the demands of special interest groups. We need more research on these issues before we can conclude.

Other features of political institutions may also matter. Whether institutions–democratic or not–insulate policy-makers from special interest pressures may matter. After all, some non-democratic countries in Asia, such as Hong Kong and Singapore, have long had fairly liberal trade policies. Electoral rules, the nature of the party system and other institutional features may also affect which interest groups can exert the most influence. Institutions that can internalize the costs of protection so that all members bear them can make protection much more difficult for political leaders to choose. The role of political institutions is underexplored. We need more cross-national studies of trade policy.

As for international factors, they have also only received a small amount of attention. Hegemonic Stability theory remains the central explanation for trade policy at the systemic level. Security concerns also seem important. The gains from trade do pose security externalities. How in the new international system do such factors affect states? How should we expect trade policy to change in the wake of the collapse of socialist and communist economies and the end of the Cold War? One might anticipate that the lack of direct threats globally should lead all to adopt freer trade. On the other hand, some suggest that the end of bipolarity and the decline of American hegemony might lead to the fragmentation of the world economy into rival trading blocs, centered on the United States, Europe and Japan. Some claim that this is what the growing role of regional trade agreements is leading to. Others see such PTAs as fostering the extension of a multilateral trading system. The impact of security concerns and the balance of capabilities and threats on trade policy is another area demanding empirical research.

The role of international institutions is also of importance. The creation of the GATT/WTO, the EU and a slew of regional organizations like Mercosur make it plain that such institutions play an important role in trade. But what exactly is this role? Can they alter states’ behavior or preferences? Do they just provide information and hence help prevent cheating? Or are they simply the instruments of the most powerful states in them? And how do such institutions act to change the behavior of the weaker states in the grouping? Research on the role of international institutions in trade is a growing topic for both economists and political scientists (e.g., Bagwell and Staiger, 1999).

None of our existing theories appears to do very well in explaining trade policy. A better understanding of how political leaders form their trade preferences and how these preferences are connected to societal ones is essential. Moreover, theories about the relationship between democracy and trade are in their infancy. And knowledge of the conditions under which international institutions are able to exert greater (or lesser) influence over countries is necessary. Hence, although we have many theories of trade policy, none can provide a complete explanation of the trade policy process. Most are also too centered on the American political system and economy to provide convincing explanations of other countries’ policies. Moreover, more empirical studies examining these theories are needed.

Finally, we need to be able to theorize about the possible future direction of trade policy. Will the recent moves toward freer trade around the globe be sustained or reversed? Will trade barriers remain as low as they are and keep declining, or will protectionism return? Again, the factors discussed above should give us some bearing on this issue. If leaders’ or social groups’ preferences for free trade are maintained or grow, then we might expect liberalization to remain in place. Factors, such as economic crises, that cause actors to question these preferences will limit their sustainability. We might also expect that the return of authoritarian governments would be associated with the return to protection, but democracy itself is not a sufficient condition for liberalization. Finally, the role of international institutions seems to be heightened by the severity of domestic economic crises. This suggests that as good times return political leaders who do not favor free trade may reject the policies forced on them by their lenders and turn protectionist. These and other factors will be important for understanding the sustainability of trade liberalization. But again these factors only give us some preliminary clues about where to look for the forces that may influence trade policy in the future.