If there is a nightmare scenario for the trading system of the late 1990's it is that the regionalism that has recently come into the system leads to increased friction among trade blocs and ultimately to full trade hostilities. It may seem farfetched to suggest that postwar cooperation in trade, personified for four decades by the GATT, and since the end of the Uruguay Round by the World Trade Organization (WTO), could almost overnight be replaced by an all-out interbloc noncooperative trade war. Yet, history shows that cooperation is the handmaiden of subsequent conflict. John Coneybeare (1987) examined trade relations back to the 12th century and suggested that hostility can rapidly follow cooperation, citing repeated and extreme examples of conflict among former trading partners, such as the burning of ships during the Anglo-Hanse trade wars. Thus, the possibility of regionalism triggering a breakdown of the multilateral trading system cannot be dismissed out of hand.
The idea that strengthened regional blocs may lead to global trade conflict has recently gained currency in both academic and policy circles. Jagdish Bhagwati (1992) has characterized regional trade arrangements as a threat to the multilateral system. Paul Krugman (1991) has pointed out that enlargement of trading blocs results in more retaliatory power for each, which, in a noncooperative environment, can lead to higher interbloc trade barriers.
These analyses have generated considerable resonance in policy circles, because they have reinforced widespread fears, motivated in part by the experience of the 1930's, of what might follow a progressive erosion of GATT/WTO disciplines. In one scenario, the major powers would each further develop their networks of regional arrangements with neighboring groups of countries that were already in place at the end of the 1980's (EU/Lome/EFTA; United States/Canada). They would then enter new and more expansive arrangements with other regional partners (NAFTA/Latin America; EU/ Eastern Europe; Japan-United States/APEC). What Ronald Wonnacott (1990) has termed a "hub and spoke" structure would come to dominate, with an increased potential for trade conflict between the major powers.
These same fears, in turn, have fueled the recent growth of regional blocs, as smaller countries have rushed to join larger blocs to insure themselves against the risk of being squeezed out of vital foreign markets. They have also been instrumental in larger countries' obtaining asymmetric concessions from smaller countries in the resulting regional negotiations.
Are fears that growing regionalism may trigger a return to the 1930's justified in the late 1990's? History seems to be of only limited help. The 1930's are by now a distant and clouded event, in which exactly what happened with retaliatory trade policy is less clear with every passing year. But a closer scrutiny of those events suggests that repeated trade retaliation did not occur; nor was the collapse in trade flows in the 1930's simply the result of protectionism. Recent experience, furthermore, such as the EU-U.S. trade conflicts in the 1980's, has shown that large powers are capable of mediating their disputes quickly before the damage becomes significant.
Our conclusion is that the idea that growing regionalism risks taking the trading system into a retaliatory trade implosion seems, on the basis of present literature, a little farfetched. It reflects a romanticized view of destructive noncooperative behavior which entered the trade literature in the 1950's and has been progressively embellished since. It is based, at best, on a partial picture of the 1930's, which we believe bears even less resemblance to the risks that regionalism actually poses for the trading system today.
I. Theoretical and Analytical Considerations
Here, we draw on analytical literature and on recent modeling work of our own and others to review critically the argument that regionalism increases the risk of breakdown of multilateral trade cooperation. We identify two distinct elements of retaliation risk. The first involves the consequences that could follow for global economic performance if unconstrained interbloc trade retaliation were actually to occur (the cost of a trade conflict). The second is the probability that retaliation may occur (the likelihood of a trade conflict). We argue that recent theoretical literature offers some insight on the first but offers little guidance on the second.
Early analyses of Nash tariff wars (Harry Johnson, 1953; Terence Gorman, 1958) showed that postretaliation tariffs will typically be higher for larger countries. The implication then seemed to be that as trade blocs form, a smaller number of larger blocs should result in higher trade barriers postretaliation, with larger negative impacts on trade. This possibility has also recently been explored by Krugman ( 1991 ), who has shown that the simultaneous formation of regional blocs can lead to higher external barriers and lower world welfare. Krugman's work has also been interpreted as implying that recent regionalism could pose new difficulties for multilateral trade relations.
This conclusion, however, is not fully warranted. First, Krugman's analysis implicitly assumes that regional agreements would survive a global trade war. This begs the question of why a regional agreement should be less fragile than a multilateral agreement, when explicit mechanisms for enforcing both types of arrangement are weak.
Second, the situation examined by Krugman (a reduction in the number of blocs and an increase in their size) corresponds to a scenario where bloc enlargements take the form of customs unions, in which members adopt common external barriers. But many of the new regional agreements are free-trade areas. Since, external policies are uncoordinated under this arrangement, an enlargement of a free-trade area does not enhance the retaliatory power of its members. John Kennan and Raymond Riezman (1990) show that if regional agreements take the form of a free-trade area, they need not imply an escalation in retaliation.
In a related study, Perroni and Whalley (1994), we perform numerical simulation analysis on the effects of regional trade agreements, comparing their effects not to free trade but to an unconstrained Nash outcome in the global economy. In their model, a simultaneous formation of North American and European customs unions under a global retaliation scenario leads to a fall in world welfare. However, when the enlargements take the form of free-trade areas, world welfare rises.
Perroni and Whalley's simulations also show that the formation of a free-trade area between a large and a small country can actually undermine the retaliatory power of the large country. With zero bilateral barriers between the two regional partners, the small country's lower optimal external tariff acts as a constraint on retaliation by the large country against third parties. Thus, one could argue that the formation of hub-and-spoke arrangements between large countries and smaller satellites who maintain independent trade policies can reduce the potential cost of global retaliation.
Recent literature also shows that when trade policies reflect a political-economy equilibrium, the formation of preferential trading areas can lead to lower external tariffs (Martin Richardson, 1993). Other strategic aspects of trade policies may also affect conclusions. David Collie ( 1995 ) shows that, in an export subsidy game with oligopolistic industries, bloc enlargement results in lower, not higher, subsidies.
Third, implicit in this kind of analysis is the assumption that once retaliation gets under way, it will continue until the Nash outcome prevails. No intermediate form of limited cooperation beyond that associated with free trade is considered. In reality, we observe that countries seem to be able to act cooperatively, maintaining low tariffs for prolonged periods of time. Trade retaliation is episodic; when it occurs, it is usually limited in severity, and cooperation is eventually restored. Coneybeare (1987), for example, points out that the experience of the 1930's was closer to a single retaliatory exchange than a Nash tariff war. The analysis of single-shot tariff games can shed light on how bloc enlargements can affect the potential costs of retaliation but has little bearing on the question of whether regional agreements make retaliatory episodes more or less likely.
To explain how countries manage to maintain a cooperative trade regime, even when trade agreements are not easily enforceable, recent theoretical literature has resorted to infinitely repeated game constructions. In these, cooperation can be sustained noncooperatively by an appropriate supporting penalty structure. This explanation of how cooperation is preserved relies on a well-known property of infinitely repeated games: namely, that, provided the discount rates applied to future payoffs are sufficiently low, any outcome that is Pareto superior to the disagreement payoffs can be supported by suitable strategies as a subgame-perfect Nash equilibrium of the repeated game (the so-called "folk theorem" ) . New regional trade arrangements may thus move the global trading system from one cooperative equilibrium to another, rather than propel it to the disagreement point. However, the basic insight afforded by this construction, while powerful, offers limited predictive power as to what form of global cooperation in trade will actually occur under growing regionalism.
Furthermore, in such a framework, trade retaliation should never be observed. Ray Riezman ( 1991 ) attempts to explain how limited retaliation can nonetheless occur in this framework, by offering a noncooperative characterization of cooperative tariff equilibria in a repeated-game setting, where countries cannot perfectly observe other countries' barriers. In this framework, random shocks can trigger temporary retaliatory episodes, which simply represent mistakes that are necessary to support cooperation. Regional trade arrangements may disturb a preexisting equilibrium, but the notion that this acts as a trigger propelling the global economy into an all-out trade war again does not follow.
Two recent papers which are relevant for the question of how the formation of regional free trade agreements affects multilateral trade cooperation are those by Kyle Bagwell and Robert Staiger ( 1993) and by Eric Bond and Constantinos Syropoulos ( 1993 ) . Bagwell and Staiger show that during the transition to a free-trade arrangement tariffs may temporarily rise in anticipation of the trade-diversion effects stemming from the agreement, although their analysis does not explicitly deal with tariff retaliation as such. Bond and Syropoulos show that an increase in bloc size raises the incentive to deviate from cooperation and thus lowers the maximum discount rate under which free trade can be sustained. Neither study directly addresses the question of how regional agreements affect the likelihood of temporary retaliatory episodes, which, as experience has shown, do not imply a permanent breakdown of cooperation.
In summary, then, theoretical underpinnings for the notion that growing regionalism may trigger an outbreak of large-power retaliation in the trading system have been largely propositions relating to the impact of customsunions formation on one-shot Nash tariff games. These, however, do not extend to freetrade areas and break down when alternative trade policy-making mechanisms are considered. Furthermore, recent work in game theory stresses how a range of cooperative equilibria can be supported by penalty structures, and a new regional arrangement could simply move the global economy from one cooperative equilibrium to another. Indeed one might speculate that, as the number of blocs falls, the cost of communication among trading partners will fall, which might result in improved observability, reduced communication costs, and a lower chance of retaliatory mistakes, making cooperation easier.
II. Retaliation Risk in the Later 1990's Seen from the 1930's
Parallels are also commonly drawn between the risks of large-power trade conflicts in the later 1990's and the events of the 1930's. But despite extensive discussion of the 1930's in the literature, as is noted in Colleen Hamilton and John Whalley (1996), there is a surprising degree of ambiguity as to what actually happened in this period, and what was the role of trade policies.
Charles Kindleberger (1973) has documented the dramatic collapse of world trade in the late 1920's. By the end of 1931, world trade was less than half of what it had been at the beginning of 1929; and by 1933 it was only about 33 percent of what it had been in 1929. This collapse coincided with a rise in trade barriers and a surge in regionalism. Great Britain and France negotiated preferential arrangements with their colonies in an effort to maintain market shares, and several smaller European countries also negotiated regional trade arrangements.
However, the link between regionalism and increased trade barriers, on the one hand, and sharply reduced trade performance, on the other, is unclear, and a comparison of the mid1990's and the 1930's reveals both similarities and differences. The increasing resort today to regional and preferential trade arrangements (EU, Canada-United States-Mexico, APEC, and others) parallels what happened in the 1930's; but in the 1930's regionalism was to a large degree a response to the collapse of the system, not one of the initial factors causing its collapse. Also, in Kindleberger's view, a central factor explaining what happened in the 1930's was a lack of leadership, and the same could be claimed today. The system today seems inherently more unstable than at any time since 1947, as the leadership role of the United States has weakened. The EU is caught up in further integration with Eastern Europe and with other concerns and shows no willingness to lead. Nor does Japan, who largely participated passively in the system throughout the 1970's and 1980's and remains in a similar stance today.
While the 1930's saw retaliatory responses by all European countries to higher U.S. tariffs, the collapse in world trade which followed cannot be accounted for solely by increases in trade barriers. The Depression was already under way when the United States implemented the Smoot-Hawley Tariff (raising the average U.S. tariff rate to 53 percent in June 1930); and the situation was further complicated by a series of competitive devaluations. Thus, the trade-policy experience of the Depression does not fully match the beggarthy-neighbour trade-policy characterization so often portrayed.
Furthermore, trade performance in the period since the late 1940's also clearly stands in sharp contrast to the events of the 1930's. The largest players, the United States and the EU have consistently displayed a determination to mediate their trade disputes in the 1980's, triggered by EU enlargement. And today's global economy is much more interdependent than it was in the 1930's. Firms and industries have become more reliant on export markets, and there is more interindustry trade. There is also the major difference of the presence of the GATT/WTO, accompanied by bindings on tariffs achieved in eight rounds of negotiations; and, despite its weaknesses, a GATT/WTO dispute-settlement procedure has continued to function.
Thus, while there may seem disturbing similarities between today and the 1930's, to argue that a repeat of the 1930's is possible in the current environment, and that it will be triggered by growing regionalism, seems to us an overstatement.
III. Concluding Remarks
In summary, the view that regional arrangements could spark a global trade war is not supported by current analytical trade literature, nor by observation of events in the 1980's, when large-trade-power conflict was successfully mediated outside of the GATT. In part, this possibility has been given prominence by the popularity in the analytical literature of the 1950's of Nash tariff games in which no cooperation was possible, and by allusions to the events of the 1930's. Since then, game theory has shown that a range of cooperative equilibrium outcomes are possible even in the absence of explicit coordination mechanisms, suggesting that new regional agreements may merely move the global trading system from one equilibrium to another.
What impact bloc enlargement may have on the stability of cooperative global trade arrangements is thus still an open question. Using the 1930's as a whipping boy for global trade policy to keep the system open may have its uses. But drawing on what happened then does not necessarily suggest acute danger ahead, triggered by new regionalism.
REFERENCES
Bagwell, Kyle and Staiger, Robert W. "Multilateral Tariff Cooperation During the Formation of Regional Free Trade Areas." National Bureau of Economic Research (Cambridge, MA ) Working Paper No. 4364, May 1993. Bhagwati, Jagdish. "The Threats to the World Trading System." World Economy, July 1992, 15(4), pp. 443-56. Bond, Eric and Syropoulos, Constantinos. "Trading Blocs and the Sustainability of InterRegional Cooperation." Department of Economics Discussion Paper No. 93-17, University of Birmingham, U.K., March 1993. Collie, David R. "Bilateralism is Good: Trade Blocs and Strategic Export Subsidies." Mimeo, University of Warwick, August 1995. Coneybeare, John A. C. Trade wars: The theory and practice of international commercial rivalry. New York: Columbia University Press, 1987. Gorman, W. M. "Tariffs, Retaliation and the Elasticity of Demand for Imports." Review of Economic Studies, June 1958, 25(3), pp. 133-62.
Hamilton, Colleen and Whalley, John. The world trading system after the Uruguay Round. Washington, DC: Institute for International Economics, 1996 (forthcoming). Johnson, Harry G. "Optimum Tariffs and Retaliation." Review of Economic Studies, 1953-54, 21(2), pp. 142-43. Kennan, John and Riezman, Raymond. "Optimal Tariff Equilibrium with Customs Unions." Canadian Journal of Economics, February 1990, 23(1), pp. 70-83. Kindleberger, Charles P. The world in depression 1929-1939. London: Allen Lane Penguin, 1973. Krugman, Paul R. "Is Bilateralism Bad?" in Elthanan Helpman and Assaf Razin, eds., International trade and trade policy. Cambridge, MA: MIT Press, 1991, pp. 9-23. Perroni, Carlo and Whalley, John. "The New Regionalism: Trade Liberalization or Insurance?" National Bureau of Economic Research (Cambridge, MA) Working Paper No. 4626, January 1994. Richardson, Martin. "Endogenous Protection and Trade Diversion." Journal of International Economics, May 1993, 34(3-4), pp. 309-24. Riezman, Raymond G. "Dynamic Tariffs with Asymmetric Information." Journal of International Economics, May 1991, 30(34), pp. 267-83. Wonnacott, Ronald J. U.S. Hub-and-Spoke Bilaterals and the Multilateral Trading System. Toronto: Howe Institute, 1990.
By CARLO PERRONI AND JOHN WHALLEY *
* Perroni: University of Warwick, Coventry, CV4 7AL, United Kingdom; Whalley: University of Western Ontario, London, Canada N6A 5C2, and University of Warwick, and National Bureau of Economic Research.
Copyright American Economic Association May 1996
©1996 UMI Company; All Rights Reserved. Only fair use, as provided by the United States copyright law, is permitted. UMI Company makes no warranty regarding the accuracy, completeness or timelines of the Publications or the records they contain, or any warranty, express or implied, including any warranty of merchantability or fitness for a particular purpose, and shall not be liable for damages of any kind or lost profits or other claims related to them or their use.