¹ Keohane, Robert O., “The Demand for International Regimes,” International Organization 36 (Spring 1982)CrossRefGoogle Scholar.

² Krasner, Stephen D., “Global Communications and National Power: Life on the Pareto Frontier,” World Politics 43 (April 1991)CrossRefGoogle Scholar. See also Fearon, James, “Bargaining, Enforcement, and International Cooperation,” International Organization 52 (Spring 1998)CrossRefGoogle Scholar.

³ Welfare maximization is defined in standard trade theory as the efficiency gains that accrue when patterns of comparative advantage drive trade. Defining benefits here as increased trade is consistent with observed behavior, as GATT members sought to increase their exports. See Wolf, Martin, “Differential and More Favorable Treatment of Developing Countries and the International Trading System,” World Bank Economic Review 1 (May 1987)CrossRefGoogle Scholar.

⁴ Rose, Andrew K., “Do We Really Know That the WTO Increases Trade?” American Economic Review 94 (March 2004)CrossRefGoogle Scholar; Subramanian, Arvind and Wei, Shang-Jin, “The WTO Promotes Trade Strongly but Unevenly,” NBER Working Paper 10024 (2003)Google Scholar; and Tomz, Michael, Goldstein, Judith, and Rivers, Douglas, “Membership Has Its Privileges: The Impact of GATT on International Trade” (Manuscript, Stanford University, Stanford, Calif, 2004a)Google Scholar. See also idem, “How Do Institutions Affect International Relations? Standing, Embeddedness, and the GATT/WTO” (Manuscript, Stanford University, Stanford, Calif, 2004b)Google Scholar.

⁵ Strange, Susan, “International Monetary Relations,” in Shonfield, Andrew, ed., International Economic Relations of the Western World, 1959—1971, vol. 2, International Monetary Relations (London: Oxford University Press, 1976)Google Scholar.

⁶ Negotiations to establish the ITO, intended to govern trade, began in 1946. The GATT was an interim agreement concluded in 1947 to guide the trade round then under way. The Truman administration chose not to resubmit the ITO charter to Congress for ratification in 1950. Thus, the GATT remained in effect until the WTO replaced it in 1995. See Bagwell, Kyle and Staiger, Robert W., The Economics of the World Trading System (Cambridge: MIT Press, 2002), 46;Google Scholar and Aaronson, Susan Ariel, Trade and the American Dream:A Social History of Postwar Trade Policy (Lexington: University Press of Kentucky, 1996)Google Scholar.

⁷ Dam, Kenneth M., The GATT: Law and International Economic Organization (Chicago: University of Chicago Press, 1970)Google Scholar,

⁸ We identified the rankings of trading partners we report here using 1928 data, described below.

⁹ Pomfret, Richard, Unequal Trade: The Economics of Discriminatory Trade Policy (New York: Basil Blackwell, 1988), 53;Google ScholarTussie, Diane, The Less Developed Countries and the World Trading System: A Challenge to the GATT (London: Frances Pinter. 1987), 15Google Scholar. Even after the Article XXXV exemption was lifted, member countries “sought to isolate” Japan, leaving it “at the periphery” of the GATT even as late as the Tokyo Round; Hart, Michael, Fifty Years of Canadian Tradecraft: Canada at the GATT, 1947-1997 (Ottawa: Centre for Trade Policy and Law, 1998), 126.Google Scholar Patterson agrees, noting that the MFN obligations nations assumed with respect to Japan “were qualified and the principle of ‘orderly marketing’ was endorsed”; Patterson, Gardner, Discrimination in International Trade, the Policy Issues: 1945—1965 (Princeton: Princeton University Press, 1965), 295.Google Scholar

¹⁰ Zamagni, Vera, The Economic History of Italy, 1860-1990 (Oxford: Clarendon Press, 1993), 31Google Scholar, 368.

¹¹ Milward, Alan S., The Reconstruction of Western Europe, 1945-51, 2nd ed. (Berkeley: University of California Press, 2000), 441.Google Scholar

¹² Gardner, Richard, Sterling-Dollar Diplomacy: Anglo-American Collaboration in the Reconstruction of Multilateral Trade (Oxford: Clarendon Press, 1956), 8.Google Scholar

¹³ The source of the quotations in this paragraph is Chap. 474 U.S. code, Sec. 354, June 12, 1934, 945.

¹⁴ Goldstein, Judith L., “Creating the GATT Rules: Policy, Institutions, and American Policy,” in Ruggie, John Gerard, ed., Multilateralism Matters: The Theory and Praxis of Institutional Form (New York: Columbia University Press, 1993)Google Scholar. The Interdepartmental Committee on Foreign Trade Agreements and the Committee on Reciprocity Information were “responsible for identifying likely negotiating partners, designating concessions to be offered and demanded during negotiations, and supervising the negotiations as they were undertaken by the Department of State”; Butler, Michael A., Cautious Visionary: Cordell Hull and Trade Reform, 1933-1937 (Kent, Ohio: Kent State University Press, 1998), 104Google Scholar. Thus, before negotiations began, “the public is given notice of the Government's intention to negotiate, is advised of the list of articles on which the grant of import concessions will be considered, and is invited to supply the Government with any information that might be useful to it in the conduct of the negotiations.” U.S. Tariff Commission, Operation of the Trade Agreements Program, July 1934 to April 1948, pt. 2, History of the Trade Agreements Program (Washington, D.C.: Government Printing Office, 1948), 53.Google Scholar

¹⁵ Oye, Kenneth A. argues that this was in the self-interest of the United States, given the way in-terwar trade agreements were structured; Oye, Economic Discrimination and Political Exchange: World Political Economy in the 1930s and 1980s (Princeton: Princeton University Press, 1992), 95–96.Google Scholar

¹⁶ Hawkins, Harry C., Commercial Treaties and Agreement: Principles and Practice (New York: Rinehart and Co., 1951), 81.Google Scholar

¹⁷ John M. Leddy, staff member of the Trade Agreements Division, Department of State, Harry S. Truman Library, Oral History, http://trumanlibrary.org/oralhist/leddyj, 8-9. First quotation in Hawkins (fn. 16), 81. “Hull believed that “if the United States signed [a treaty] with the low-cost producer, no industry would be hurt” by imports from other producers. Goldstein (fn. 14), 208. “Butler (fn. 14), 135.

²⁰ Hawkins (fn. 16), 92.

²¹ Pomfret (fn. 9), 6.

²² Bidwell, Percy, What the Tariff Means to American Industries (New York: Harper and Brothers, 1956), 52–54.Google Scholar

²³ See Gardner (fn. 12), 150; Goldstein (fn. 14), 209.

²⁴ Hoekman, Bernard M. and Kostecki, Michael M., The Political Economy of the World Trading System: The WTO and Beyond, 2nd ed. (Oxford: Oxford University Press, 2001), 128.CrossRefGoogle Scholar

²⁵ Republican Party preferences were similar. Its 1944 platform asserted that because “the domestic market is American's greatest market,” the tariff “which protects it against foreign competition should be modified only by reciprocal bilateral trade agreements”; Goldstein (fn. 14), 227n13.

²⁶ Leddy (fn. 17), 37. A heated debate exists about the RTAA and the trade preferences of presidents relative to Congress. See, e.g., Lohmann, Susanne and O'Halloran, Sharyn, “Divided Government and U.S. Trade Policy: Theory and Evidence,” International Organization 48 (Autumn 1994)CrossRefGoogle Scholar; Hiscox, Michael, “The Magic Bullet? The RTAA, Institutional Reform, and Trade Liberalization,” International Organization 53 (Autumn 1999)CrossRefGoogle Scholar; Bailey, Michael A., Goldstein, Judith, and Weingast, Barry R., “The Institutional Roots of American Trade Policy: Politics, Coalitions, and International Trade,” World Politics 49 (April 1997)CrossRefGoogle Scholar. We do not engage this debate here, since we are concerned about only this one instance of executive-congressional interaction.

²⁷ Gardner (fn. 12); Zeiler, Thomas W., Free Trade Free World: The Advent of GATT (Chapel Hill: University of North Carolina Press, 1999), 45Google Scholar. An appealing by-product of the approach was that it gave exporters an interest in lowering tariffs abroad. See Bagwell, and Staiger, (fn. 6); Gilligan, Michael J., Empowering Exporters: Reciprocity Delegation and CollectiveAction in American Trade Policy (Ann Arbor: University of Michigan Press, 1997)Google Scholar; and Irwin, Douglas A. and Kroszner, Randall S., “Interests, Institutions, and Ideology in Securing Policy Change: The Republican Conversion to Trade Liberalization after Smoot-Hawley” Journal of Law and Economics 42 (October 1999)CrossRefGoogle Scholar.

²⁸ Pomfret (fn. 9), 30; Ruggie, John Gerard, “International Regimes, Transactions and Change: Embedded Liberalism in the Postwar Economic Order,” International Organization 36 (Spring 1982), 397CrossRefGoogle Scholar. Preserving its bargaining power and maintaining the Commonwealth also, of course, motivated British efforts to preserve imperial preferences. Milward (fn. 11), 351.

²⁹ Leddy (fn. 17), 21.

³⁰ More generally, Article XXIV sanctioned preferential trading agreements (PTAS) as long as they were not “more restrictive” than they had been previously. For the GATT text, see http://pacific.commerce.ubc.ca/trade/GATT. ³¹Snnivasan, T. N., Developing Countries and the Multilateral Trading System: From the GATT to the Uruguay Round and the Future (Boulder, Colo.: Westview, 1998), 10.Google Scholar

³² U.S. Tariff Commission (fn. 14), 65; U.S. Tariff Commission, Operation of the Trade Agreements Program, 9^th Report, July 1955-June 1956 (Washington, D.C.: Government Printing Office, 1957), 65.

³³ Havana Charter, chap. 4, Art. 17.2(a). For the Havana Charter text, see www.globefield.com/havana.

³⁴ Ibid., Article 28.2(a) (bis).

³⁵ In the 1950s, for example, several countries sought to impose maximum tariffs on broad product groups instead of continuing with reciprocal trade concessions. About the same time, a majority of GATT contracting parties, including Belgium, Denmark, France, Germany, and the Netherlands, proposed roughly 30 percent cuts on average tariffs in ten sectors. U.S. allegiance to the “product-by-product and country-by-country” approach, however, left the status quo intact. Kock, Karen, Intemational Trade Policy and the GATT, 1947-1967 (Stockholm: Almqvist and Wiksell, 1969), 98Google Scholar; see also Dam (fn. 7), 67.

³⁶ The Trade Expansion Act of 1962 authorized the use of a linear approach. For discussion, see Dam (fn. 7), 67.

³⁷ Ibid., 76. The sectors were chemicals, cotton textiles, iron and steel, aluminum, and pulp and paper.

³⁸ Curzon, Gerard and Curzon, V., “The Management of Trade Relations in the GATT,” in Shonfield, Andrew, ed., International Economic Relations of the Western World, 1959-1971, vol. 1, Politics and Trade (London: Oxford University, 1976), 185-86Google Scholar.

³⁹ Hoekman and Kostecki (fn. 24), 130.

⁴⁰ U.S. Tariff Commission (fn. 32), 140, 125. Of the 1,014 tariff reductions that the United States agreed to between the wars, almost half involved tariff reclassifications. Verdier, Daniel, Democracy and International Trade: Britain, France, and the United States, 1860-1990 (Princeton: Princeton University Press, 1994), 196.Google Scholar⁴¹Bidwell (fn. 22), 54, emphasis in original. The growth of the U.S. tariff code attests to the effort to specialize tariffs. In 1963 it included 6,421 lines. By 2000 it had grown to 10,175 lines, even though the U.S. trade-weighted average tariff rate fell from 11.9 percent to 7.4 percent ad valorem in the interim. U.S. Tariff Commission, Simplification of the Harmonized Tariff Schedule of the United States (Washington, D.C.: Government Printing Office, 2000)Google Scholar.

⁴² Ibid., 125.

⁴³ Hoekman and Kostecki (fn. 24), 31.

⁴⁴ Ibid. For the internalization ratios for each GATT round though 1967, see Finger, J. M., “Trade Liberalization: A Public Choice Perspective,” in Amacher, Ryan C., Haberler, Gottfried, and Willett, Thomas D., eds., Challenges to a Liberal International Economic Order (Washington, D.C.: American Enterprise Institute, 1979), 424.Google Scholar

⁴⁵ Ibid., 42.

⁴⁶ CPI for Urban Consumers, ail items; 1982-84=100; www.economy.com/freelunch.

⁴⁷ Tomz, Goldstein, and Rivers (fn. 4, 2004a).

⁴⁸ That is, a state is a system member if it is a UN member or its population exceedsfivehundred thousand and it receives diplomatic missions from at least two major powers. Correlates of War 2 Project. 2003. State System Membership List, v2002.1, http://cow2.la.psu.edu.

⁴⁹ The analysis utilizes directed dyads composed of these countries, for various years depending on system membership and data availability.

⁵⁰ See fn.4.

⁵¹ Tomz, Goldstein, and Rivers (fn. 4,2004a), 4.

⁵² To construct their roster of “nonmember participants,”Tomz, Goldstein, and Rivers (fn. 4,2004a) use GATT archival material that is not yet publicly available. They have not responded to our request to allow us to use their data. We cannot be certain, therefore, that our roster of nonmember participants exactly replicates theirs. This is so, because, for example, while they note that “the maximum allowable duration of de facto status changed over time,” their papers do not define exactly how it did so. Similarly, they state that not all members accorded provisional members MFN treatment, but, with one exception, they do not identify these countries (pp. 7-8).

⁵³ The quotation is from Rose (fn. 4).

⁵⁴ For details, see www.faculty.haas.berkeley.edu/arose.

⁵⁵ Complete results for any finding we report in the paper are available from the authors.

⁵⁶ To be more specific, the GATT coefficient in Rose's OLS analysis is a statistically insignificant -0.04, (fn. 4), 104. Including dyadic fixed effects produces a statistically significant estimate of 0.15 (fn. 4), 104. We used Rose's data and variable definitions to estimate GATT'S impact between 1950 and 1994, the years we analyze here. An OLS analysis with year fixed effects produces a coefficient on GATT of 0.116 (p-value = 0.004). An analysis with dyadic and year fixed effects produces a coefficient on GATT of 0.046 (p-value = 0.150). In Subramanian and Wei (fn. 4), the GATT coefficient that an analysis with country and year fixed effects produces is -0.113, which is statistically significant at less than the 0.05 level (p. 28). Their analysis is difficult to compare directly to ours, as their regressand is the log of real imports of each state and they include data through 2000. Goldstein, Rivers, and Tomz use data from 1948 through 2001 and report a statistically significant estimate on GATT, using a model with dyadic and year fixed effects, of 0.37 (fn. 4,2004b), 26.

⁵⁷ Green, Donald P., Kim, Soo Yeon, and Yoon, David H., “Dirty Pool,” International Organization 55 (Spring 2001)CrossRefGoogle Scholar.

⁵⁸ Anderson, James and Wincoop, Eric Van, “Gravity with Gravitas: A Solution to the Border Puzzle,” American Economic Review 93 (March 2003)CrossRefGoogle Scholar; idem, , “Trade Costs,” NBER Working Paper 10480 (2004)Google Scholar; Matyas, L., “Proper Econometric Specification of the Gravity Model,” WorldEconomy 20 (May 1997)Google Scholar.

⁵⁹ The results in Table 1 are robust to the inclusion of country fixed effects, however.

⁶⁰ Since the data are in the form of directed dyads, dyadic fixed effects are specified separately for each direction. Based on a likelihood ratio test, Cheng and Wall find that a symmetry restriction on the dyadicfixedeffects rejects the null (that is, that fixed effects do not differ significantly between the dyad that reflects the imports of a from b and the dyad reflecting imports of b from a). Thus, they argue for including a separate term for each direction of the dyad. See I-Hui Cheng and Howard J. Wall, “Controlling for Heterogeneity in Gravity Models of Trade and Integration,” Federal Reserve Bank of St. Louis Working Paper 1999-010E (2004).

⁶¹ Barbieri, Katherine and Levy, Jack S., “Sleeping with the Enemy: The Impact of War on Trade,” Journal of Peace Research 36 (July 1999)CrossRefGoogle Scholar; Gowa, Joanne and Mansfield, Edward D., “Power Politics and International Trade,” American Political Science Review 87 (June 1993)CrossRefGoogle Scholar; Li, Quan and Sacko, David, “The (Ir)Relevance of Militarized Interstate Disputes for International Trade,” International Studies Quarterly 46 (March 1999)Google Scholar; Polacheclc, William Solomon, “Dyadic Disputes: An Economic Perspective,” Papers, Peace Science Society (International) 28 (1978)Google Scholar; Pollins, Brian M., “Does Trade Still Follow the Flag?” American Journal of Political Science 33, no. 3 (1989)Google Scholar.

⁶² In comparing these results with those in the existing literature, however, it is important to note that our analyses do not include Soviet bloc members.

⁶³ As reported in Subramanian and Wei (fn. 4), 9. Belgium and Luxembourg are not in our analysis, however, because the IMF reports their trade separately only in 1997.

⁶⁴ These include dyads where both members are nonindustrial countries or where one is an industrial country and the other a nonindustrial country.

⁶⁵ The coefficient on industrial-country trade that Rose (fn. 4) reports is a statistically significant 0.47 (p. 108). The corresponding statistic in Subramanian and Wei (fn. 4) is 0.322, also statistically significant (p. 29),

⁶⁶ We report Newey-West standard errors applied to panel data, which are robust to heteroskedas-ticity and autocorrelation. Newey, W. and West, K., “A Simple Positive Semi-Definite, Hetersoscedasticity and Autocorrelation Consistent Covariance Matrix,” Econometrica 55 (May 1987)CrossRefGoogle Scholar.

⁶⁷ This is also true, of course, of members of other dyads that either joined the GATT at its inception or were allies or democracies throughout the sample period.

⁶⁸ The thirteen remaining industrial countries—Austria, Denmark, Finland, Germany, Greece, Iceland, Ireland, Italy, Japan, Portugal, Spain, Sweden, and Switzerland—acceded to GATT between 1951 and 1967.

⁶⁹ We are grateful to Albrecht Ritschl and Nikolaus Wolf for making the data available to us. Iceland is not included in the Ritschl and Wolf data, so we exclude it from the analysis. The following states are in the Ritschl-Wolf data set but not in ours: Argentina, Belgium, Turkey, and several members of the former Eastern bloc (that is, Bulgaria, Czechoslovakia, Hungary, Poland, Romania, the Soviet Union, and Yugoslavia). Albrecht Ritschl and Nikolaus Wolf, “Endogeneity of Currency Areas and Trade Blocs: Evidence from the Inter-War Period” (Manuscript, Humboldt University 2003).

⁷⁰ Panel data analysis can accommodate time-invariant variables using a technique that Hausman and Taylor developed; Hausman, Jerry A.. and Taylor, William E., “Panel Data and Unobservable Effects,” Econometrica 49 (November 1981)CrossRefGoogle Scholar. However, their method assumes that not all explanatory variables are correlated with the fixed effects, an assumption that is not tenable here. The time-invariant nature of these variables also precludes an estimation based on first-differencing the data to account for the possible endogeneity of the GATT regime. Baier, Scott L. and Berg-strand, Jeffrey H., “Do Free Trade Agreements Actually Increase Members’ International Trade?” (Manuscript, 2005)Google Scholar.

⁷¹ Ritschl, and Wolf, (fn. 69); Nitsch, Volker, “Honey, I Shrunk the Currency Union Effect on Trade,” World Economy 25, no. 4 (2002)Google Scholar; Persson, Torsten, “Currency Unions and Trade: How Large Is the Treatment Effect?” Economic Policy 33 (October 2001)Google Scholar.

⁷² Blundell, Richard and Dias, Monica Costa, “Evaluation Methods for Non-Experimental Data,” Fiscal Studies 21, no. 4 (2000)Google Scholar.

⁷³ The privileged group and the founding-member group are mutually exclusive. The former consists of pairs of privileged-group members (Britain, Canada, France, Germany, and the United States). The latter consists of country pairs formed among founding members that were not in the privileged group, as well as founding members paired with each member of the privileged group.

⁷⁴ Ritschl and Wolf (fn. 69), 9.

⁷⁵ Our post-World War II data begin in 1950; hence, we use this as the first of the GATT years.

⁷⁶ An F-test for all the interaction terms (for GDP and group terms) is F(10, 17356) = 20.04 (Prob > F = 0.0000), which shows that all the interaction terms are jointly statistically significant. An F-test for all the interwar bloc variables (main and interaction effects) is F(1O,17356) = 29.80 (Prob > F = 0.0000), which indicates that the effects of the control groups and their corresponding interaction terms are jointly statistically significant.

⁷⁷ The MID variable is insignificant. Joint democracy exerts a statistically significant impact on trade, increasing it by about 17 percent. ⁷⁸These results are robust to the inclusion of country or dyadic fixed effects for countries that are not members of any of the groups in the analysis. In our sample, these countries are Greece and Japan.

⁷⁹ We did not include the European Community in this analysis, because it did not come into existence until 1958 and because its membership did not remain stable over the postwar period. We have, however, tested whether our results are robust to the inclusion of the EC6, since it formed relatively early in the postwar period. We find that the EC had a large positive (0.84) and significant (p-value = 0.000) impact on its members’ trade. Our results are robust to including the EC with these exceptions: the treatment effect for the gold bloc becomes insignificant (p-value = 0.772); the coefficient on the interwar alliance term becomes significant (p-value = 0.008), while its treatment effect becomes insignificant.

⁸⁰ Eichengreen and Irwin used interwar and early postwar trade to predict trade in 1949,1954, and 1964; Eichengreen, Barry and Irwin, Douglas A., “Trade Blocs, Currency Blocs and the Reorientation of World Trade in the 1930s,” Journal of International Economics 38 (February 1995)CrossRefGoogle Scholar; and idem, , “The Role of History in Bilateral Trade Flows,” in Frankel, Jeffrey, ed., The Regionalization of the World Economy (Chicago: University of Chicago Press, 1998)Google Scholar.

⁸¹ Our sample includes, however, only two Reichsmark bloc members (Austria and Germany). ⁸²To calculate these values we sum the coefficients of the control and treatment effect variables (that is, the group dummy and its corresponding interaction term). The corresponding significance level is based on F-tests for the joint significance of each group's control and treatment effect estimates. They are statistically significant for all groups, except for founding members: F(2,17356) = 1.85 (Prob > F = 0.152); privileged group: F(2,17356) = 121.52 (Prob > F = 0.000); exchange-control bloc: F(2,17356) = 5.78 (Prob > F = 0.003); gold bloc: F(2,17356) = 11.85 (Prob > F = 0.000); Commonwealth bloc: F(2,17356) = 62.00 (Prob > F = 0.0000); sterling bloc: F(2,17356) = 35.71 (Prob>F =0.000); and Reichsmark bloc: F(2,17356) = 107.71 (Prob > F = 0.000).

⁸³ To ensure that these results are not due to the abnormally low levels of trade during the Great Depression, we reanalyzed the data using only 1928 trade as the reference level. The results are robust, with these exceptions: the coefficient on alliances becomes positive and statistically significant for the interwar period (p-value = 0.048); its treatment effect is insignificant (p-value = 0.960).

⁸⁴ Ritschl and Wolf (fn. 69), 11.

⁸⁵ The Ritschl and Wolf results (fn. 69) are sensitive to model specification. When they control for trade diversion, they find that trade within the Commonwealth bloc, for example, increases significantly after its formation (p. 24). Eichengreen and Irwin (fn. 80,1995) find that the formation of the Commonwealth and Reichsmark blocs exerted a positive and significant effect on the trade of its members, while the formation of the currency blocs did not (p. 15).