Affecting the Unaffected:
Priming opposition to free trade among students in Canada, Mexico, and the U.S.
John H. Aldrich
Department of Political Science
Duke University
Durham, NC
Jennifer L. Merolla
Department of Politics & Policy
Claremont Graduate University
Claremont, CA
Laura B. Stephenson
Department of Political Science
University of Western Ontario
London, Ontario
Elizabeth J. Zechmeister
Department of Political Science
University of California – Davis
Davis, CA
This project was funded by DIPE, ASRC, the Centers for North American Studies and Latin American and Caribbean Studies at Duke, and the Trent Foundation. Early versions of these results and this paper were presented at the Globalization, Equity and Democratic Governance Seminar at Duke, the Juan March Institute, Madrid, the Midwest Political Science Association 2001 Conference, Chicago, IL, and the 2001 Liberty Fund Conference on Liberty and Hierarchy, Park City, Utah. We thank the participants at all these functions for their helpful comments. We also thank Andrew Sancton for his assistance in facilitating the experiments in London, Ontario, and Alejandro Poiré, Gerardo Domínguez González and María Fernanda Vergara for their great help with the experiments at ITAM, in Mexico City.
Abstract
This paper investigates a mechanism by which information can be translated into policy preferences. Specifically, it speaks to claims made by some political economists that economic liberalization results in welfare state growth by inducing insecurity and changing policy preferences, and assesses whether information on trade policy can directly persuade and/or make certain considerations more relevant for individuals’ preferences over trade and over government provision of goods and services, regardless of whether these individuals are directly affected by globalization. Further, we test whether individuals respond differently to positive versus negative messages about trade policy. Using data generated by a cross-national experiment with student populations, we find that positive information about globalization can influence preferences over trade policy, even among a population that is not directly harmed by free trade, but that negative information does so even more. At the same time, we find that only negative information translates into preferences over government welfare spending.
The welfare state has grown markedly in OECD countries over the last decade. Some political economists argue this growth is caused by citizens’ demand for greater government provisions and protection in the face of losses generated by economic liberalization (Garrett 1995, 1998b, Rodrik 1997). The logic behind this argument is clear. A plant shuts down because a firm can make goods cheaper in another country. The workers from this plant ask the government for assistance. The assertion of the “giant sucking sound” pulling jobs overseas is compelling and the government responds by extending its safety net in the form of increased welfare. Others, however, argue that these claims and the causal story are exaggerated even if visually compelling, because the losses from liberalization are concentrated while the benefits are diffuse: most individuals, including the median voter, will benefit from increased liberalization. The argument that because all consumers receive benefits from free trade, while only a minority are hurt due to concentrated costs, draws from central tenants of economic theory tracing back to David Ricardo (see Garrett 1998a).
These conclusions, even though starkly opposing, are based on common assumptions about how individuals receive, assimilate, and use information to develop policy preferences: Individuals’ objective economic circumstances influence their perceived economic (in)security, which in turn shapes their preferences over public policy. But one might ask why, if those assumptions are true, were not the vast majority of protesters at the WTO meetings from Seattle on those who had lost their jobs from globalization? Or, why was there so much anti-NAFTA sentiment on the part of the PRD in Mexico, if many of their supporters stood to gain from globalization? Or, finally, why did the Liberal and NDP parties campaign so vehemently against the Canadian Free Trade Agreement (CUFTA) when so many of their stood to benefit from these closer ties with the U.S.?
Perhaps these actions demonstrate that objective economic conditions are not the only determinants of preferences abut globalization. Perhaps, that is, politics, as well as economics, matters in understanding why citizens desire more or less from their government. Politicians and interests groups are constantly bombarding the public with messages about national policy in an effort to lure voters to their side. While such messages may be aimed at those most affected by trade policies, the messages reach all citizens, including those who, even though not be directly affected by such policies, might nonetheless be persuaded to support freer trade or greater government protections.
In this paper, we ask how information about trade policies shapes peoples’ preferences not only over such policies themselves but also over policies affecting the government’s provision of goods and services, regardless of whether or not these individuals are directly affected by globalization. We ask whether information about trade policy directly persuades individuals to hold particular preferences or whether the information primes relevant considerations thereby making them more important in reaching political judgements. Finally, we test whether individuals respond differently to positive as compared to negative messages about trade policy. Using data generated by a cross-national experiment with student populations, we find that positive information about globalization influences preferences over trade policy, but that negative information does so even more, even among a population that is not directly harmed by free trade. In contrast, we find that only negative information translates into changes in preferences over government welfare spending.
Political Economic Conjectures and Micro-Level Evidence
As we noted above, some scholars argue that globalization creates insecurities and dislocations that increase demand for government services and for trade protections (Garrett 1995, 1998; Rodrik 1997, 1998).[1] They further argue that government responsiveness to these consequent changes in policy preferences account for much of state growth in this era of globalization. The models used by these scholars posit a direct relationship between individuals’ objective economic circumstances and their preferences over government policy. However, the studies rely on macro data and therefore cannot test whether micro-level evidence supports their assumptions. We seek here to place some of these aggregate inferences onto firmer micro-foundations.
There is a nearly direct linkage between those helped [harmed] by trade and their support for [opposition to] freer trade. The connection between objective economic conditions and preferences over government growth in the provision of goods and services is less immediate. Those harmed by current macro-economic conditions are likely to feel those effects subjectively and translate feelings of insecurity into preferences for greater government provision of items that help “insure” them against economic loss or that provide for retraining or other compensatory assistance in recovering from economic loss. Conversely, those who benefit from increased trade liberalization are assumed to desire a decrease in the size of the government and its role in people’s lives, and therefore desire a decrease in the government’s provision of goods and services.
Testing of these presumed microlinkages has only begun (important initial studies include Iversen and Soskice, 2001; Mughan, Bean, and McAllister, 2003; and Scheve and Slaughter, 2001a, 2001b, 2004) . Aldrich et al. (1999a, 1999b) examined NES and Eurobarometer survey data and found that these microlinkages were real, but relatively weak, while Kramer, Stephenson and Lange (2000) found evidence of somewhat stronger connections.[2] While such refinements are likely to continue, even these stronger results do not appear to be sufficient to induce the magnitude of political changes that the political economists have detected at the macro level.
Survey evidence may be relatively weak for two reasons. First, these survey respondents are typically not primed by the current political context and surely are not primed by the survey itself in ways at all similar to how the protestors and displaced workers who comprise the dramatis personae in the political economists’ stories have been primed by real events in real time. Second, survey analysis has mostly focused on models assuming direct effects of globalization. Such citizens who are directly affected may be too few in number and too concentrated in location to be able to yield effects in a cross-sectional survey sufficient to account for major governmental activity. We hardly need add that this group almost certainly does not include the median voter to whom policy is aimed.
Those with little direct effects of trade policies nonetheless encounter news media and their messages about such policies. If they also have developed attitudes on the subject, such messages should also serve to prime those attitudes. However, for the unaffected, such attitudes are likely to be more weakly held and therefore, such message could (also) be persuasive to them, leading them to revise or change their preferences. Such persuasion is more likely for negative messages, precisely because the costs are concentrated and therefore likely to impinge heavily on what are often also the most vulnerable citizens. The resulting accounts can therefore be more dramatic and thus potentially more persuasive. The benefits of freer trade, while real, are more diffuse and therefore make less compelling media stories and campaign advertisements.
A Model of Persuasion over Preferences for Economic Liberalization
To begin, let us consider what sorts of messages about globalization might the public receive? Two of the best known are illustrated below.
You implement that NAFTA, the Mexican trade agreement, where they pay people a dollar an hour, have no health care, no retirement, no pollution controls, et cetera, et cetera, et cetera, and you're going to hear a giant sucking sound of jobs being pulled out of this country right at a time when we need the tax base to pay the debt and pay down the interest on the debt and get our house back in order. - Ross Perot (Third Presidential Debate 1992, http://www.debates.org/pages/trans92c.html, 6/29/01).
The NAFTA-GATT trade deals sold out America’s workers, ravaged our manufacturing base & caused disruption in our small towns & farming communities. … Our once mighty steel industry now begs relief from the World Trade Organization for problems that were caused by the Clinton-Gore headlong march into the New World Order. - Pat Buchanan http://www.issues2000.org/Celeb/Pat_Buchanan_Free_Trade.htm#Globalization, 6/29/01).
We noted above that survey evidence might find relatively weak linkages in part because only a minority of the population are directly affected (especially negatively) by trade policies. These citizens, however, include those who are expected already to hold preferences over trade policies and government insurance and protection, and it is quite possible that they hold preferences sufficiently strong to “counter-argue” any political communication about trade. If so, such messages are likely to prime already held beliefs but relatively unlikely to be persuasive messages themselves.
There are two ways in which such messages might influence people’s opinions. First, the messages may transmit facts or emotional feelings, to some individuals, whether it is about trade policy itself or about who supports or opposes such policies. What would we expect their reaction to be to a political message about trade? Not only might those directly affected be expected to hold beliefs about trade, but it is also likely that many hold preferences sufficiently strong to “counter-argue” any communication about trade. If so such messages are likely to prime already held beliefs but relatively unlikely to be persuasive emssages themselves. If so, such messages will increase access to such attitudes without changing them, that is, bring the issue and already held beliefs to mind when answering other survey questions (Iyengar and Kinder 1987, Krosnick and Kinder 1990; Miller and Krosnick 2000). For example, an individual who has seen or read a number of news stories about Iraq may weight Bush’s performance on Iraq more heavily in her overall assessment of his performance relative to other issues, such as the economy. Feldman and Zaller (1992) and Zaller (1992) argue that this occurs because when answering survey questions, people sample from the considerations that are at the “top of their heads” (i.e., are accessible). Thus, when a given message highlights particular considerations, those become more salient to individuals and are thus more accessible.[3]
What about others? Those relatively unaffected directly by trade policies nonetheless encounter news media and their messages about such policies. If they also have developed attitudes on the subject, such messages should also serve to prime those attitudes. However, for the unaffected, such attitudes are likely to be more weakly held and therefore, such message could (also) be persuasive to them, leading them to revise or change their preferences. Such persuasion is more likely for negative messages, precisely because the costs are concentrated and therefore likely to impinge heavily on what are often also the most vulnerable citizens. The resulting accounts can therefore be more dramatic and thus potentially more persuasive. The benefits of freer trade, while real, are more diffuse and therefore make less compelling media stories and campaign advertisements.
We also argue that information about trade may be especially important in priming attitudes that make sociotropic (national level) economic considerations more salient. The balance of evidence from the sociotropic versus egoistic effects suggests that individuals base their evaluations and choices on national conditions rather than their own personal economic situation (Kiewiet 1983; Kinder, Adams and Gronke 1989; MacKuen, Erikson, and Stimson 1992; McGraw, Best and Timpone 1995; Sears and Funk 1990). The mechanism behind this effect is that the public receives information from the media about sociotropic conditions, not their personal situation (Mutz 1992). After all, free trade is a national level economic policy, and such messages typically highlight its effects to a number of people rather than “merely” to the individual listener. A similar mechanism might be at work with respect to globalization. Negative or positive arguments about globalization may make citizens’ perceptions of the national economy more salient, which in turn influence their preferences about trade and government protection.
However, not all messages will necessarily carry the same weight in influencing individual opinion. Many strands of scholarship find that negative information is a) more likely to draw one’s attention, which in turn b) makes this information more effective in the formation of impressions of people or things, and in one’s decision-making. There are several different causal explanations for the negativity bias. Expectancy-contrast theories posit that since the experiences most people have are positive, negative information stands out as being extreme; thus, the latter is more effective in grabbing our attention and in our decision-making (e.g., Helson 1964; Lau 1982, 1985). Another theoretical explanation is frequency-weight theories, which posit that negative cues are more effective in that they are more informative or novel (e.g., Fiske 1980; Kahneman and Tversky 1979; Lupia and McCubbins 1998). Finally, more recent models draw from neurobiology and argue that the negativity bias arises because negative and positive information are processed in different ways (e.g., Brader 2006; Marcus, Neumann, and MacKuen 2000). When things are going well, the disposition system causes people to rely on habit when making decisions. In contrast, the surveillance system scans the environment for the sudden intrusion of threat, and in the presence of threat causes individuals to stop their behavior and process the information to avoid potential harm. Thus, negative cues are more likely to be recognized and processed than are positive cues, even when people are not cognizant of the negative content of the cues (Ito, Larsen, Kyle and Cacioppo, 1998; Pratto and John, 1991). Strong effects of negative information have been found in the domain of studies on globalization; for example, Cobb and Kuklinski (1997) found that negative arguments were the most persuasive when used to influence opinions regarding NAFTA in particular. Thus, individuals exposed to negative information might be more likely to absorb information that leads to changes in preferences over government policy.