We thank Bob Keohane, Marc Busch, Eric Anderson, James Fearon, Erica Gould, Barry Weingast, Simon Jackman, Brian Hanson, Richard Steinberg, an anonymous reviewer, and the editors of IO for comments on a previous version of this article.

1. Legalization refers to three aspects of international law: obligation, precision, and level of delegation to a centralized authority. Abbott et al., this issue.

2. The number and variety of groups participating in the politics of trade has grown in the last decades. Where the classic models assumed three groups with trade-related interests—consumers, import-competing groups, and exporters—other groups, whose interests span from human rights to a clean environment, have come to believe that their interests are influenced by trade negotiations. The logic of this article, explaining the interaction among international regimes, social mobilization, and domestic politics, applies to any interest that groups perceive to be influenced by international trade agreements.

3. Collective-action problems have been central to the literature on endogenous tariff formation. See, for example, Magee, Brock, and Young 1989; and Mayer 1984.

4. Gilligan 1997.

5. Milner 1988.

6. Young 1989.

7. Gilligan 1997.

8. Schattschneider 1935.

9. Numerous empirical studies document the importance of groups in setting trade policy. For a cross-national, cross-sectional examination of groups' involvement, see, for example, Verdier 1994.

10. On trade and interest groups, see Destler 1995; and Lohmann and O'Halloran 1994.

11. Whether it was a change in the balance of group interests or a shift in trade policymaking that explains the ability of governments to lower barriers to trade is difficult to determine in the early years of the GATT regime. Certainly, in the United States interest-group activity was muted because the costs of organizing increased when the president obtained increased control of trade policymaking. Still, the shift toward openness would not have occurred without underlying social support. For an analysis of the relationship between institutional and underlying social variables, see Bailey, Goldstein, and Weingast 1997.

12. One simple metric capturing the continued involvement of proprotection groups in the United States is the number of bills entered in Congress pertaining to imports. In the 93rd–97th Congress, 1973–81, over 2,200 bills were entered; in the 98th–101st Congress, 1982–90, almost 2,300 bills were entered; and in the 102nd–105th Congress, 1991–98, over 1,600 bills were considered. Although few of these bills become law, the data suggest that liberal policy exists in the shadow of growing resistance. Numbers were obtained from Thomas Legislative Information on the Internet located at <http://thomas.loc.gov/home/thomas.html>.

13. Keesing 1998.

14. Ibid.

15. The GATT's move to the Trade Policy Review Mechanism was motivated by the perception that information was key in negotiations but that it was available only to the larger countries. Ibid.

16. More applications were received from groups asking for admission to the talks than at any previous time in the GATT/WTO's history (private correspondence). See also the New York Times, 13 October 1999, A12Google ScholarPubMed.

17. New York Times, 4 December 1999, A6Google ScholarPubMed.

18. Richard Blackhurst interviews.

19. Under 1951 and 1958 laws, U.S. negotiators were forced to consider data on which industries would be hurt by a trade agreement. Under peril point legislation, the U.S. Tariff Commission provided public information on the effect of tariff reductions on particular industries. This list told U.S. trading partners exactly who could and could not be included in a trade deal. This generated more information to groups than is even the case under the WTO. One effect, consistent with this analysis, is that access to this information explains the relatively modest tariff changes made in the 1950s. Goldstein 1993.

20. Shonfield 1976, 175.

21. Ibid., 183.

22. Goldstein 1993, 166.

23. Evans 1971, 281.

24. American Selling Price was a valuation system used for chemical imports. Ibid., 300–303.

25. Winham 1986, 17–18. A reviewer suggested the Swiss cheese metaphor, particularly applicable here since the adopted formula was Swiss.

26. The three major players were all plagued by antitrade pressures during the round. In the EU, differences among member countries developed over the Common Agricultural Policy. In Japan, the issue of rice protection undercut the ruling Liberal party, and in the United States, pressure from groups polarized on fast-track legislation extension. Secchi 1997, 81.

27. Ibid., 79.

28. Goldstein 1993.

29. Josling 1999.

30. See Federal Register 62 (7 November 1997)Google Scholar, Docket No. 301–116 (62 FR 60299).

31. Federal Register 60 (27 October 1995)Google ScholarPubMed, Docket No. 301–101 (60 FR 55076).

32. Federal Register 56 (2 December 1991)Google Scholar, Docket No. 301–86 (56 FR 61278); see also 57 FR 38912 and 61 FR 25000.

33. Krueger 1998.

34. New York Times, 9 June 1995, D3Google Scholar.

35. See, for example, the history of agricultural trade in Josling 1999.

36. We use the term bindingness where the term obligation would seem appropriate to a political scientist. The reason is that obligation has taken on a particular legal meaning, and that meaning has been adopted in this issue. By bindingness we mean the political obligation created by international rules. It is a positive rather than a normative term, meaning the degree to which rules are binding, practically speaking, on governments. Rules with higher probability of enforcement, for example, are more binding (or obligatory) in this political sense.

37. Keohane 1984.

38. Krasner 1983.

39. Downs and Rocke 1995, chap. 4.

40. Contract law recognizes the same dynamic of uncertainty requiring flexibility in contracts, under the heading of efficient breach. See Roessler, Schwartz, and Sykes 1997, 7.

41. The idea is similar to that behind the Coase theorem: efficient agreements are reached through the mechanism of one party compensating another.

42. As we argue later, the safeguard reforms are counterintuitive for two reasons. First, they may be too difficult to invoke, undercutting their purpose. Second, since retaliation is limited, the stability evoked by activating export groups may have been undermined.

43. See also Sykes 1991, 259Google ScholarPubMed.

44. Hufbauer and Schott 1993.

45. Roessler, Schwartz, and Sykes 1997, 13.

46. See Sykes 1992; and Hudec 1999.

47. On the extent of changes in the WTO, see Krueger 1998.

48. Goldstein 1993.

49. “Tariff concessions and unforeseen developments must have caused an absolute or relative increase in imports which in turn causes or threatens serious injury to domestic producers … of like or directly competitive goods.” Although the invoking party is not saddled with the burden of proving that it has met these requirements, the requirements nonetheless have deterred countries from invoking the escape clause.

50. This often leads to a situation where the producers causing the problems in the first place could remain in a competitive position with the higher-cost home producer. The producers who get penalized are the middle-price traders who were not the problem. Shonfield 1976, 224.

51. Baldwin 1998.

52. GATT Analytical Index, various issues.

53. The United States invoked Article XIX fourteen times between 1950 and 1969. Of these they used nontariff barriers alone in only one case.

54. We assume a goal of reducing opportunism on theoretical grounds, without claiming that all negotiators had precisely this goal in mind. Certainly the agendas of negotiators were diverse, and reducing opportunism was only one goal among many.

55. Jackson 1998.

56. For a more thorough examination of patterns of disputes in the GATT and the WTO, see Hudec 1999; and Sevilla 1998.

57. Hudec 1993, 297.

58. Hudec 1999.

59. Hudec 1993, 296.

60. Sevilla 1998.

61. Reinhardt 1995.

62. Franck 1995.