Abstract
Why would some legislators alter their votes on trade agreements in return for environmental side agreements that may be hard to enforce? While numerous studies have examined the effects of side agreements, few have evaluated their impact on legislators’ positions on a trade agreement over time. This paper examines the effects of the environmental side deal attached to NAFTA, with novel time-series survey data that captures the evolution of House members’ positions on NAFTA during discussion and finalization of the environmental side of the free trade agreement. I find that pro-environmental legislators in safe districts tended to withdraw their support for NAFTA once the side deal was agreed upon, whereas those in competitive districts stood their ground and increased their support in the final stage of voting. Contrary to the conventional wisdom, I find little evidence that the side deal assuaged legislators in import-competing districts. This article shows how the effectiveness of international institutions is moderated in important ways by electoral considerations.
Acknowledgements
I want to thank Jeffry Frieden, Anne Sartori, Beth Simmons, Stephen Chaudoin, Mark Copelovitch, Bobby Gulotty, Connor Huff, Casey Kearney, Yoon Jin Lee, Francesca Parente, Yon Soo Park, and Gabriele Spilker for comments. All errors are my own.
Disclosure statement
No potential conflict of interest was reported by the author.
Notes
1 The Bush administration in particular negotiated extensive environmental side agreements starting in 2007 when the House was dominated by Democrats.
2 For a more comprehensive analysis of the relationship between presidents and legislators in the realm of international trade, see Milner and Tingley (Citation2015).
3 Environmental issues have been typically non-salient during legislative elections, especially at subnational levels. For example, when the Clinton administration attached an environmental side agreement to the North American Free Trade Agreement (NAFTA) to gain support from House Democrats, only 2% of the American voters listed environmental issues as the most decisive factor in the 1992 House election. See “The American National Election Studies, 1992 Time Series Study.” VAR 900228: Did you tend to prefer one of the House of Representatives candidates because of this issue?
4 Of course, there are numerous instances of cooperation (e.g., the alliance of Teamsters and Turtles at the 1999 WTO protest in Seattle). Given the long-running animosity, however, the formation of such coalitions should not be taken for granted.
5 See various issues in Inside U.S. Trade that cover business groups' opinions about the environmental side agreement attached to NAFTA. Among others, see “TEXT: Industry Letter to Kantor on NAFTA Side Accords.” (June 4, 1993). Retrieved from the Inside U.S. Trade online archive.
6 Also see Cinar and Gulotty Citationn.p. for the September survey.
7 See Special Report on NAFTA, Inside US Trade, June 11, 1993. S6-S7.
8 I use the data from Conconi et al. (Citation2014). Drawing from the Census data, the authors compute the number of employees of each constituency in export and import-competing industries.
9 I present the results using legislators’ lifetime scores measured in 1993, not the annual scores, because the cumulative data better capture their genuine commitment to environmental protection. Although the correlation between members’ lifetime and annual scores is relatively high (Pearson correlation = 0.89), I also report the results based on legislators’ annual environmental scores in Appendix E.
10 Inside US Trade records that the Clinton administration was thirty votes short as of August 1, and thirty-one votes short as of August 27, 1993. See Special Report in Inside US Trade, October 1, 1993.
11 The s in my models are relatively low. Although the alternative models with more control variables (i.e., legislators’ ideology scores, education, state fixed effects) report higher s (Appendix B), the adjusted s in those models are still low. Despite the limitation, the models provide important insights into how legislators update their attitudes on an important trade agreement. The low s do not change the interpretations of the significant variables in the models. Because this paper’s goal is to test the validity of important theories in the literature on trade agreements, not to predict legislative positions over time, the results indicate a real relationship between electoral competition and legislators’ increased support for NAFTA. Relatedly, the same models (with static legislative positions in each survey period) report higher s (on average 0.4). This means that the standard variables considered important in the trade literature (i.e., partisanship, voter characteristics, and legislator ideology) do a good job explaining legislative positions at a specific moment but do a poor job accounting for how legislators “change” their positions over time. Although it would be difficult to get systematic data, including unobservable exchanges between the executive and legislators could enhance the models’ explanatory power.
12 Campaign finance data is drawn from the Federal Election Commission at www.fec.gov. I identified pro-NAFTA companies based on membership information regarding the USA*NAFTA coalition. Also see Osgood (Citation2017).
13 In total, there are 46 missing observations in the June survey: 24 members in New York, 18 in Texas, 3 in Tennessee, and one in California. Except for California, the missing observations are due to the random omission of two pages of the survey in the Presidential Records. The seat for California’s 17th district was vacant in this time period.
Additional information
Notes on contributors
Boram Lee
Boram Lee ([email protected] | http://www.boramlee.org) is a postdoctoral fellow at the Browne Center for International Politics at the University of Pennsylvania.