Does the U.S. Congress Respond to Public Opinion on Trade?

• Published date: 01 November 2023
• DOI: http://doi.org/10.1177/1532673X231174854
• Author: Boram Lee,Michael Pomirchy,Bryan Schonfeld
• Date: 01 November 2023
• Subject Matter: Articles

Article

American Politics Research

2023, Vol. 51(6) 731–748

© The Author(s) 2023

Article reuse guidelines:

sagepub.com/journals-permissions

DOI: 10.1177/1532673X231174854

journals.sagepub.com/home/apr

Does the U.S. Congress Respond to Public

Opinion on Trade?

Boram Lee

1,2

, Michael Pomirchy

3

, and Bryan Schonfeld

4

Abstract

Are U.S. legislators responsive to public opinion on trade? Despite the prevalence of preference-based approaches to in-

ternational trade, not much work has directly assessed the relationship between constituency opinion and positioning by

members of Congress on trade bills. We assess dynamic responsiveness (whether shifting constituency opinio n on trade yields

corresponding changes among legislators) by exploiting an original dataset on the positions of members of Congress on the

North American Free Trade Agreement at various points leading up to the November 1993 roll-call vote. We f‌ind no evidence

of dynamic re-sponsiveness to shifting constituency opinion on even a highly salient piece of trade legislation. We provide

qualitative evidence that interest group inf‌luence may instead be the predominant source of shifting legislator positioning on

trade.

Keywords

congress, representation, dynamic responsiveness, trade

Introduction

Do American legislators adhere to public opinion when

voting on trade issues? While there is already substantial

evidence that American legislators are responsive to their

constituencies on many policy issues (Canes-Wrone, 2015;

Erikson et al., 2002;Stimson et al., 1995;Trounstine, 2010),

trade policy is still a relatively understudied area for exam-

ining legislative responsiveness. In most policy areas, in-

cluding defense spending, the literature generally f‌inds that

legislators are more likely to approve a policy when there is

greater constituency support (Bartels, 1991;Erikson et al.,

1993;Lax & Phillips, 2009;2012). However, others have

suggested that there is a representational inertia, such that

there might be persistent mismatches between districts and

representatives (Anzia, 2011;Bafumi & Herron, 2010;Lax &

Phillips, 2012;Lee et al., 2004;McCarty et al., 2009).

The idea that shifting constituency preferences should

yield corresponding changes in government policy is a

fundamental tenet of normative democratic theory. Most

existing studies of dynamic responsiveness measure the as-

sociation between constituency opinion and policy passage or

roll-call voting over time (Caughey & Warshaw, 2018;

Fowler, 2005;Lee et al., 2004;Stimson et al., 1995;Warshaw,

2016). However, there are some challenges researchers face

when making temporal comparisons. In particular, it is hard to

control for the issue agenda across time. In these contexts, it is

unclear whether an increase in responsiveness can be

attributed to actual movement of legislator positioning on the

same issue over time, rather than changes in which issues are

on the legislative agenda. Furthermore, some studies examine

responsiveness by treating different bills as pertaining to the

same policy area (e.g., pro-LGBT bills). Scaling different

bills requires making assumptions about whether the bills

belong to the same dimension.

We resolve these concerns–in the trade context–by ex-

ploiting time-series data on legislator positioning throughout

1993 on the North American Free Trade Agreement

(NAFTA), one of the most important trade agreements in US

history. In particular, we use a unique original set of legislator

surveys conducted by a group called USA-NAFTA in March,

June and September 1993, in addition to October and No-

vember positioning data from Congress Daily and the As-

sociated Press. An interesting aspect of the data is that we

1

Department of International Relations, London School of Economics and

Political Science, UK

2

Department of Government, Wesleyan University, Middletown, CT, USA

3

Institution for Social and Policy Studies, Yale University, New Haven, CT,

USA

4

Department of Politics, Princeton University, Princeton, NJ, USA

Corresponding Author:

Boram Lee, Department of International Relations, London School of

Economics and Political Science.

Email: boramette@gmail.com

have access to concerns about the bill that legislators spelled

out in an open-ended section of the survey, allowing us to

account for whether shifts in legislator positioning are at-

tributable to changes in the NAFTA legislation itself. For

example, we code whether legislators had concerns about

labor, sugar, and the environment, which directly relate to and

may have been later addressed by the side agreements to

NAFTA. Using these legislator surveys and contemporane-

ous surveys of public opinion, we look at the relationships

between constituency opinion, the content of the bill, and

legislator positioning.

This design contributes to the literature in a few ways.

First, we are able to look at just one dimension, since all the

legislator positions were recorded on NAFTA. Second, to the

extent that the bill had undergone changes over the course of

the year, we can utilize the aforementioned data on legislator

concerns (on policy components that changed over time) to

investigate whether legislators most concerned about issues

that were addressed in the side agreements were most likely to

change their positions. Third, we provide an unusual and

relatively rare look at changing legislator positions on a

specif‌ic bill over time, instead of examining f‌inal roll-call

votes, and we are able to isolate the role of a few possible

determinants, including constituency opinion, of legislative

behavior. As we illustrate below, analyzing interim legislator

positions can help us understand how legislators gain targeted

benef‌its by changing their positions on legislation and

whether changes in legislation, like the side agreements to

NAFTA, increased the size of the winning coalition.

There are a few important limitations to our analysis. First,

we should note here that there are multiple determinants of

roll-call voting (or legislator positioning), and we are spe-

cif‌ically zeroing in on a few possible factors that may be

present here. Second, our analysis is naturally limited by the

granularity of our data in the year of 1993. To this end, we

have collected as much data on public opinion and legislator

positioning as we can f‌ind on NAFTA. Wehave 14 surveys of

mass opinion that in total include more than 10,000 re-

spondents, and we have f‌ive separate detailed surveys of

legislators in the House over time. Third, we might be

concerned that legislator responses to surveys sponsored by

pro-NAFTA organizations might be motivated by social

desirability bias; to address these concerns, we assess the

relationship between the legislator scores on the

USA*NAFTA surveys and legislator signing of public pro-

and anti- NAFTA letters.

Our main result is that we f‌ind no evidence of legislator

adaptation to shifting constituency opinion over time. We

compare changes in legislator positioning with changes in

contemporaneous public opinion and f‌ind no evidence of

responsiveness. We highlight one particular reason that this

might be the case—the inf‌luence of interest groups. Studies

pinpoint the disproportionate inf‌luence of interest groups on

trade legislation (Ehrlich, 2008). Scholars have cited interest

group behavior as a key inf‌luence on legislative decision-

making (Fordham & McKeown, 2003;Grossman &

Helpman, 2002), and often, interest groups are perceived

as altering the direction of responsiveness away from the

majority of the public (Anzia, 2011;Gilens, 2012). Given

this, it is possible that on bills where there is a lot of interest

group activity, like trade bills, we might expect district

opinion to matter less than on other issues. Following our

analysis, we provide qualitative evidence that interest group

pressure is likely a source of the non-responsiveness of

legislators to shifting opinion. It is hard to conclusively

determine how generalizable our results are to other pieces of

legislation, though we suggest some factors that may point in

either direction. On the one hand, NAFTA was unusually

highly salient, which might suggest that members of Con-

gress may be less adaptive on lesser-known pieces of trade

legislation. On the other hand, trade policy is also unlike

some other policy areas in that citizens know very little about

the complexities of trade agreements, and as mentioned

before, interest groups are pivotal players.

Related Literature

Do US legislators adjust their policy positions in response to

public opinion on trade? A strand of recent studies critically

examines the validity of the assumed link between public

preferences and legislative policy stances: Guisinger (2009,

2017) shows that American voters, even those with a high

stake in trade, often fail to hold politicians accountable for

their trade policy decisions; Bailey (2001) demonstrates that

the American public’s diffuse interests in trade liberalization

exert little inf‌luence on congressional trade voting; and

Feigenbaum and Hall (2015)’sf‌indings suggest that U.S.

legislators’ability to gain reelection is not affected by lo-

calized trade shocks, although they respond to these shocks

by voting in a more protectionist direction.

It is crucial to study why legislators are unlikely to respond

to public opinion in their trade policy-making process.

Studies show that American voters have a weak under-

standing of the economic consequences of trade (Betz &

Pond, 2019;Bearce & Moya, 2020). Specif‌ically, Rho and

Tomz (2017) show that most American voters fail to predict

the economic consequences of protectionist policies accu-

rately. For example, their survey f‌inds that less than one-third

of the respondents correctly identif‌ied whether low- or high-

educated Americans would benef‌it from restrictions on im-

ports of low-skill intensive products. They further demon-

strate that public attitudes on trade policy change easily

depending on information about the winners and losers of

trade policy. If public preferences on trade policy are mal-

leable, politicians may have a weak incentive to adjust their

policy position due to shifting public opinion.

Alternatively, it is plausible that politicians form their

policy positions based on the information provided by interest

groups who have better knowledge and concentrated interests

in trade policy. In line with this line of reasoning, studies

732 American Politics Research 51(6)