Going Nativist: How Nativism and Economic Ideology Interact to Shape Beliefs about Global Trade *

Existing research explains variation in trade attitudes by pitting explana- tions rooted in the foreign part of foreign economic policy, like nativism, against economic beliefs like a commitment to free market principles. But what happens when these factors create signiﬁcant cross-pressures—how do free market-oriented nativists think about trade? We argue that na- tivism is a higher-order belief that constrains the relationship between lower-order economic attitudes and beliefs about international trade. We test our argument using representative samples from the United States and United Kingdom. First, we analyze observational data and ﬁnd a sig- niﬁcant interaction whereby nativism moderates the relationship between free market attitudes and beliefs that trade provides national and global beneﬁts. Second, we report results from a survey experiment to show that a message about the long-term beneﬁts from free trade increases support for free trade in both samples. Importantly, we also ﬁnd that nativist values weaken the treatment effect in the US sample. As long as international re- lations scholars focus on cultural or economic antecedents on their own, we miss much about how elements in belief systems interact. effects on the nation and the world. Hypothesis 1b : Free market beliefs will be associated with positive beliefs about trade’s effects on the nation and the world. Hypothesis 2 : Nativism will moderate the relationship between free market beliefs and beliefs about trade. The positive relationship between belief in the free market and trade attitudes will weaken as nativism increases.


Introduction
Political scientists have accumulated substantial individual-level evidence that citizens in Western democracies, like the United Kingdom and United States, often hold protectionist views that contradict economists' pro-globalization consensus (Herrmann, Tetlock, and Diascro 2001;Scheve and Slaughter 2001; on the other. 1 Each approach contributes to how we understand public attitudes toward trade. Broadly, economic approaches contend that trade attitudes depend on trade's real or perceived material effects. One version of this approach relies on the "open economic model" to test whether trade attitudes follow from an individual's position in the labor market (e.g., Lake , 2009a). Trade agreements create domestic winners and losers, and people whose pocketbook benefits from access to foreign markets should support free trade (Scheve and Slaughter 2001;Beaulieu 2002;Mayda and Rodrik 2005;Fordham and Kleinberg 2012). In Western democracies where unskilled labor is scarce whereas human capital is abundant, citizens without a college education support protectionist policies at greater rates than their degree-holding counterparts (Scheve and Slaughter 2001). 2 Models that link economic self-interest to trade attitudes have "come under fire" in response to evidence that (1) proxies for skill level lose their predictive power when we include other individual-level factors in our models (Mansfield and Mutz 2009) and (2) most people lack accurate information about how trade policy affects them (Rho and Tomz 2017, S88; see also Hainmueller and Hiscox 2006;Guisinger 2017). On the first point, recent work has moved beyond rough proxies to show that measurement makes a difference-when scholars account for occupation characteristics (Owen and Johnston 2017) or put a "magnifying glass" on policies with clear and concentrated distributional effects (Malhotra, Margalit, and Mo 2013), they find stronger evidence that material interests play a role in economic attitudes. 3 However, on the second point, researchers find that neither the pocketbook effects implicated by trade's distributional consequences nor the pro-trade "Econ 101" arguments that academics embrace percolate throughout the public. Indeed, many Americans seem to spend little time thinking about trade policy (Guisinger 2009(Guisinger , 2017. Experiments demonstrate that people can and do update their preferences in response to new information (Hiscox 2006;Rho and Tomz 2017;Bearce and Moya 2020) or pro-trade political messages (Guisinger 2017). But the results also reveal important nuances: Some people respond with altruism rather than egoism (Rho and Tomz 2017) and treatment effects are weaker among some demographic groups (Guisinger 2017). Average effects may mask meaningful heterogeneity-a point we return to below.
Material interests matter sometimes for some people, but without an experimental prompt or a "crash course" in economic theory (Kertzer et al. 2019, 6), many individuals draw from other economic cues and ideas to form their opinions about trade (Fordham and Kleinberg 2012;Kuo and Naoi 2015). First, people can turn to their beliefs about how trade affects the national or local community economy. Mansfield and Mutz (2009, 427), for example, show that Americans' trade attitudes are sociotropic-related to "perceptions of how trade affects the country as a whole." 4 Colantone and Stanig (2018) similarly claim that the "Chinese import shock," which caused economic distress in some parts of the United Kingdom, helps explain Brexit support. Although a thriving community can affect individual 1 Scholars often treat explanations rooted in material self-interest, like labor market position, as separate from economic predispositions or sociotropic perceptions. We discuss these elements together due to their shared attention to free trade's economic implications, though different research designs are better suited to testing propositions about material self-interest in particular (Owen and Walter 2017). 2 Although much of the research in this program tests hypotheses based on the Stolper-Samuelson factor endowments model, others like Beaulieu (2002) test the Ricardo-Viner specific factors model. Consistent with the specificfactors hypothesis, Beaulieu find evidence that trade supporters disproportionately work in export-oriented markets. 3 See Owen and Walter (2017) for a review and discussion of how material interests have more explanatory power when they are measured correctly, especially pages 188-191. 4 Mansfield and Mutz (2009) further argue that trade attitudes are shaped by out-group anxiety (though cf. Fordham and Kleinberg 2012), a factor we return to below.
Downloaded from https://academic.oup.com/fpa/article/17/3/orab015/6281490 by University of Glasgow user on 02 June 2021 prosperity, 5 the authors find that people living in negatively affected regions were more likely to vote "Leave" regardless of whether their individual circumstances had changed. As an empirical matter, Guisinger (2017) finds a substantial degree of overlap between sociotropic and egoistic perceptions; Americans who think that trade helps others also believe that trade bolsters their own prospects (and vice versa). Second, people can use their broader economic ideology as a guide for their specific attitudes toward trade-including as a precursor to their perceptions of trade's effects. Free market beliefs provide a useful shortcut to evaluate a complicated set of policies. Economic conservatism and commitment to the the free market predict support for free trade in the United States and Canada (Wolfe and Mendelsohn 2005;Rathbun 2016). In East Asia, Wu (2019, 4) also finds that people who think that the government should bear responsibility for citizens' well-being "favor protectionism over laissez-faire." In short, people project their beliefs about domestic markets to the international sphere. Importantly, theories about economic ideology as an antecedent for trade attitudes only require that members of the public align their views with what they believe will produce material benefits-they do not require that people have accurate ideas about the global economy. These free market beliefs represent one key component of the cross-cutting pressures that motivate our puzzle.

Nativism and Protectionism
Faith in the free market joins other economics-oriented factors to explain some of the variance in trade attitudes-but opening domestic markets to international competition carries symbolic implications too. British consumers might appreciate cheaper clothing, but scoff at a "made in Vietnam" label because it means that an out-group produced the product. And, indeed, many scholars argue that trade attitudes stem from psychological and symbolic predispositions rather than material concerns or economic ideas (Kaltenthaler and Miller 2013, see Kuo and Naoi 2015 for a review). In line with this emphasis on the foreign in 'foreign competition', we focus on nativism's role in shaping public support for protectionism. Although other values might play a role in public opinion about trade (Mayda and Rodrik 2005), we turn to nativism-"the individual-level attitude that a uniquely [national] culture and way of life needs to be protected against foreign influence" (Knoll and Shewmaker 2015, 88)-to help consolidate the many studies that implicate outgroup attitudes under a unified banner and test a key source of partisan crosspressures (e.g., Mansfield and Mutz 2009;Margalit 2012;Mutz and Kim 2017).
Nativism is a value that shapes how people respond to problems in the political domain, and in particular to foreign policy. Like other domain relevant values (Hurwitz and Peffley 1987), nativism sits atop a hierarchy that ranges from abstract commitments to concrete policy preferences. In such a hierarchy, abstract, higherorder guides constrain lower-order beliefs like economic ideology and trade policy preferences. Specifically, Schwartz (1994, 20) argues, "A value is a (1) belief (2) pertaining to desirable end states or modes of conduct, that (3) transcends specific situations, (4) guides selection or evaluation of behavior, people, and events." Values provide an intuitive shortcut: Faced with a complicated situation or policy choice, an individual can make a quick judgment based on her "gut response" (Haidt 2001). Nativism serves this function by providing a strict definition for who belongs to the national in-group and evoking a person's desire to protect that group. There are few constructs more fundamental than in-group/out-group distinctions when it comes to political judgment (Kinder and Kam 2010), such that nativism serves as an important guide to the complex world of international trade. 6 To maintain a homogeneous nation-state, nativists oppose "foreign cultural influence" and are wary of non-native ideas and customs that undermine their traditional way of life (Margalit 2012). New arrivals bring undesirable variety to the nativist's home country. Mansfield and Mutz (2009) and Van der Waal and De Koster (2018) find associations between ethnocentrism and protectionism in the United States and Netherlands, respectively, while Margalit (2012) presents cross-national evidence that people who fear foreign cultural influence also believe that they are personally harmed by trade liberalization. Antipathy to others can in some instances be so great that Americans will only support trade policies that offer hardly any benefit to a trading partner (Mutz and Kim 2017).
At the same time, nativism has analytical purchase beyond trade-a fact that underscores its centrality within political belief systems. Sides and Citrin (2007) show that opposition to immigration correlates with a broader preference for cultural unity, and Schildkraut (2005) similarly links ethnoculturalism to support for English-only laws. Nativism predicts support for radical right parties in Europe (Zhirkov 2014), and-in conjunction with anti-elitism-Brexit support (Iakhnis et al. 2018). McConnaughy et al. (2010) find that nativism shapes coethnic candidate evaluations in the United States, and Knoll and Shewmaker (2015) show that nativists oppose domestic policies like the Affordable Care Act when they are framed as "un-American." In short, the collective empirical record suggests that when a policy or candidate arouses concerns about mixing foreign people, ideas, or products domestically, nativists react with opposition.

Theoretical Expectations
We expect that nativism and free market attitudes separately shape beliefs about international trade and support for specific policies. Nativists conclude that trade is harmful and oppose lowering tariffs, whereas free market proponents believe that free trade brings important benefits. But when a higher-order value clashes with economic beliefs, the former will constrain the latter (Prather 2017)-nativism will 6 We are confident in characterizing nativism as a value and therefore causally prior to trade attitudes for several reinforcing reasons. First, there is an overwhelming amount of evidence that emphasizes the importance of in-group/outgroup distinctions in evaluation and decision-making contexts (Kinder and Kam 2010). Out-group intolerance has both historical and neurological roots (Sapolsky 2019;Homola, Pereira, and Tavits 2020), and group distinctions affect political beliefs from partisanship and polarization (Mason 2015;Huddy and Bankert 2017) to militarism (Herrmann, Isernia, and Segatti 2009) and social welfare (Kinder and Kam 2010). As noted in the text, research demonstrates that nativism predicts attitudes toward immigration and candidate evaluation in addition to beliefs about trade-consistent with Schwartz's (1994) assertion that values transcend specific situations. Second, these concepts are clearly embedded in leading comprehensive theories of values. Opposition to change and outside influences are encompassed in Schwartz's (1994) tradition and security dimensions and Haidt and Graham's (2007) loyalty and purity dimensions (with the latter in each case a slightly more abstract concern). Third, as an empirical matter, our nativism scale shows significant stability over time. Data used in this paper were part of a larger panel study where a select set of the nativism questions were asked at multiple timepoints, three for the US study (spring 2012, summer, 2013, late 2018) and four in the United Kingdom (spring 2012, summer 2013, spring 2014, late 2018). First-order autoregressive paths connecting nativism across timepoints indicated high over-time stability: Lagged 2012 nativism values explain 92 percent of the 2013 variance on nativism in the United States and 94 percent in the United Kingdom, for example. We also observe striking stability between 2014 and 2018, a time period that saw both the election of Donald Trump and acrimonious debate over immigration in the United States and the Brexit vote in the United Kingdom. Nativism from 2014 explains 69 percent of the 2018 variance in the United States and 78 percent of the variance among UK respondents. These results are consistent with what we would expect from a value that lies at the center of an individual's belief system and constrains lower order attitudes (Goren 2005; see Online Appendix Section 1.2 for additional details). Although theory and evidence therefore lead us to conclude that nativism constitutes a value, a skeptical reader might disagree due to the policy implications associated with nativist beliefs. This logic would move nativism one step down in a hierarchical belief structure-akin to the mid-level "general postures" like militarism that mediate the relationship between values and specific policies in Hurwitz and Peffley's (1987) influential model-but it would remain causally prior to beliefs about whether tariffs improve the national economy. blunt the influence of economics on how members of the public think about trade. We test these expectations using both observational and experimental data.
To evaluate the relationship between nativism, free market attitudes, and public opinion about trade, we first turn to their associations with two sets of beliefs about international trade.
On one hand, we know that people tend to hold consistent and reinforcing views about trade's domestic effects-it either helps or harms the nation in general (Guisinger 2017). Those who believe that trade encourages national economic growth will also believe that it bears fruit for domestic consumers and helps factory workers, for example. Nativists, whose apprehension about openness leads them to prefer a closed homeland (Margalit 2012), conclude that trade harms the nation's economy whereas free market proponents tout the benefits from lower tariffs (Rathbun 2016;Wu 2019).
Although previous research has provided consistent support for an association between our key independent variables and beliefs about trade's domestic effects, we also measure participants' beliefs about how open trading affects the world. We expect that nativists will reject claims that trade is a global good. 7 The relationship between nativism and beliefs about trade's global effects provides an important test for our theory: If nativist opposition to trade depended on rationally weighing the costs and benefits from trade, nativist values should be unrelated to beliefs about broader, global effects that do not carry domestic consequences. Instead, we argue that nativist values will automatically arouse a negative reaction to the prospect of mixing foreign goods and cultures anywhere. Values lead us to evaluate situations based on our emotional "gut" reactions (Haidt 2001). Driven by their abstract commitment to cultural homogeneity and a summary assessment that globalization is "bad," nativists, in turn, believe that trade brings global harm. We also expect those who oppose government intervention in the economy to apply their intuitions to trade: For economic liberals, free trade offers the best path to global development just as it facilitates growth at home. This logic leads to our first two hypotheses.
Hypothesis 1a: Nativism will be associated with negative beliefs about trade's effects on the nation and the world.
Hypothesis 1b: Free market beliefs will be associated with positive beliefs about trade's effects on the nation and the world.
Evidence to support hypotheses 1a and 1b will largely confirm the conventional wisdom on both sides of the divide about what shapes trade attitudes. However, if core values entail an automatic reaction and constrain the effects of lower-order beliefs (Hurwitz and Peffley 1987), then people high in nativism may support protectionist policies despite the fact that they endorse the free market domestically. Mutz and Kim (2017) hint at this relationship when they argue that nationalism and social dominance orientation, characteristics that are correlated with Republican partisanship, enhance support for trade agreements in which the United States maximizes relative-rather than joint-gains. We expect that nativists, who prioritize in-group protection over government deregulation, will have a weaker relationship between free market attitudes and positive beliefs about trade. This moderating effect has important implications for contemporary politics. If nativism attenuates the relationship between faith in markets and pro-trade beliefs, it could help to explain the protectionist turn in the US GOP (Grand Old Party) and, to a lesser extent, right-wing parties in the United Kingdom.
Hypothesis 2: Nativism will moderate the relationship between free market beliefs and beliefs about trade. The positive relationship between belief in the free market and trade attitudes will weaken as nativism increases.
Finally, if nativism constrains the effect of beliefs about the economy, committed nativists will maintain their opposition to trade even when they are exposed to a message that lays out the prospects for prosperity under an open economic system. Most people who learn about national-level gains from trade advocate for lower tariffs (Mutz and Kim 2017), and people update their policy preferences in response to information and elite cues (Guisinger 2017;Rho and Tomz 2017;Bearce and Moya 2020). At the same time, we know that people do not respond to persuasive messages in a vacuum (Lodge and Taber 2013). Values work in the background to shape how we interpret messages and how-or whether-we update our attitudes (Hurwitz and Peffley 1987; see Kertzer and Powers 2019 for a review). If any foreign influence is "fundamentally threatening" (Mudde 2007), nativists will not update their support for tariffs in response to a short lesson about the free market. Economic benefits do not outweigh the cultural costs. 8 To test our proposition that nativism moderates the effect of pro-market messages about trade, we embed an experiment in both surveys that randomly assigns a message about how open trading improves the national economy in the long run. Whereas the survey measures beliefs about whether people are better off in a free market economy, the experiment targets those beliefs directly by providing a common macroeconomic argument about how open trading redounds to citizens' material benefit. We build on recent experimental work showing that Americans support trade when the United States stands to gain jobs (Mutz and Kim 2017; Bearce and Moya 2020), but we focus on long-term trends rather than the effects of a particular trade agreement. One political argument in favor of trade depends on the idea that liberal trade policies have long time horizons-industries that do not have a comparative advantage may suffer in the short term, but stronger economies abroad eventually open new markets to goods produced at home. 9 When nations break down trade barriers, economies thrive in the long run. We expect that exposure to this message increases support for free trade, but also that values matter: Long-term benefits from a free market approach to trade may not move nativists to support lower tariffs.
Hypothesis 3: Exposure to a political message touting trade's long-term national benefits will increase support for free trade.
Hypothesis 4: Nativism will moderate the effect of a political message touting trade's long-term national benefits. The effect of the treatment will weaken as nativism increases.

Nativism, Free Market, and Beliefs about Trade
We conducted nationally representative surveys of 2,346 American and 2,349 British participants in 2012. We selected the United States and United Kingdom for two reasons. First, our theory engages research that heavily relies on American samplesperhaps because the United States is the world's largest economy-and we aim to test our expectations about cross-pressures using the same population as past work. Second, although the United States and United Kingdom are both independently interesting given their positions in the globalized economy, the paired samples allow us to probe the theory's generalizability. The United States and United Kingdom share memberships in both NATO (North Atlantic Treaty Organization) and the G-7 and cultural similarities. And although some research finds evidence for similar foreign policy attitude structures (Reifler, Scotto, and Clarke 2011), others report key differences in the cross-national correlates of these attitudes .
While the primary independent and dependent variables are drawn from a single cross-sectional survey, the data are part of a larger panel study. We include independent variables from a prior wave when available.

Dependent Variable: Latent Beliefs about Trade
Scholars often model trade attitudes using single items available on prominent survey instruments. Mayda and Rodrik (2005) and Hainmueller and Hiscox (2006) rely on data from the International Social Survey Program (ISSP), for example, which asks participants whether they agree that their home country "should limit the import of foreign products in order to protect its national economy." 10 An item on the ANES (American National Election Studies) describes pro-and anti-trade positions before asking participants if they favor new limits on imports (see, e.g., Scheve and Slaughter 2001;Hainmueller and Hiscox 2006). The Latinobarometro survey (Beaulieu, Yatawara, and Wang 2005), National Annenberg Election Survey (NAES; Mansfield and Mutz 2009), and World Values Survey (Kaltenthaler and Miller 2013) ask about support for particular protectionist or liberal policies like restricting imports to protect domestic jobs, and the Chicago Council surveys ask if participants "believe that globalization, especially the increasing connections of our economy with others around the world, is good or bad for the United States?" (Mansfield and Mutz 2009, 435). 11 These measures "allow scholars to solicit citizens' gut-based, general reactions to trade" and have been instrumental in advancing our knowledge (Kuo and Naoi 2015, 102). They are appropriate for research on policy preferences, and are available on national and cross-national surveys. However, as Guisinger (2017, 4) points out, trade policy is interesting in part because "its implications are neither primarily domestic nor international," and we want to understand how values and economics relate to beliefs about both sets of beliefs. In particular, we need to test whether nativist values shape beliefs about trade's global effects to demonstrate its centrality in individuals' belief systems. Moreover, beliefs about trade's effects correlate with attitudes toward specific policies (Mansfield and Mutz 2009), and understanding where these beliefs come from remains an important research question (Guisinger 2017). We therefore include a prompt that asks participants, "Do you think globalizationthe increased trade between countries in goods, services, and investments-has had a negative effect, no effect, or a positive effect on each of the following?" Eight target items, listed in table 1, tap national and global effects (see, e.g., Woodward, Skrbis, and Bean 2008 and Bechtel, Bernauer, and Meyer 2012 for similar approaches). 12 Four items ask about the domestic effects of trade. Following sociotropic theories, we ask about the American (British) economy in general. The other three items target specific groups-businesses, consumers, and factory workers. Each of these categories has a different relationship with the globalized economy (Lake , 2009b), but we expect that they comprise a single latent dimension that taps perceptions of trade's national effects.
10 See the documentation for ISSP Research Group (1995). 11 See Kuo and Naoi (2015) for a review of survey measures of trade attitudes. 12 Although the eight items do not exhaust the myriad ways that trade affects national or global politics-and do not account for individual or community-level factors (Guisinger 2017)-they represent a range of important outcomes that we expect to reflect the latent national and global trade factors that interest us. Notes. The table reports the standardized results of the CFA, which generates factor scores for the "Global" and "National" benefits of trade utilized as dependent variables in the multivariate analyses to follow. Final fit statistics of the model are as follows: United States: χ 2 = 37.14(df = 15), RMSEA = 0.03, CFI = 1.00; United Kingdom: χ 2 = 89.68 (df = 15), RMSEA = 0.05, CFI = 0.99. Correlation between "Global" and "National" factors: US = 0.65; UK = 0.50. Error correlation between "Improving women's rights…" and "Improving democracy and human rights…": US = 0.41; UK = 0.48 (not used in the derivation of factor scores utilized below).
Second, we measure whether people believe that trade has negative or positive international consequences with four items. One draws inspiration from theories that connect economic interdependence to peace (Russett and Oneal 2001) and asks whether participants think that trade reduces prospects for war. The other three draw from political economy research on the relationship between trade and development and democratization and ask about whether the liberal international economy is good for developing countries (Milner 1999;Milner and Kubota 2005).
A confirmatory factor analysis (CFA) on these eight items revealed a two-factor solution in both the United States and United Kingdom. Factor scores estimated from the two factors-"national benefits" and "global benefits" of globalized tradecomprise our dependent variables. The first factor is positively correlated with beliefs that globalization has a positive effect on women's rights, poor countries' economies, the prospect of interstate war, and human rights abroad. The second factor is positively correlated with beliefs that economic globalization has positive sociotropic effects-it helps the national economy in general alongside specific groups in society. 13

Independent Variables
To evaluate hypotheses 1a, 1b, and 2, we include two key independent variables: nativism and support for free markets. Additive scales for nativism based on eight (United States, α = 0.85) or nine items (United Kingdom, α = 0.90) tap participants' antipathy to immigrants and preference for a homogeneous homeland (Iakhnis et al. 2018). 14 The nativism items comprise a single latent factor and tap 13 See Online Appendix Section 3 for a thorough discussion of measurement validity. One possible concern is that our primary independent variable, nativism, and dependent variables, beliefs about globalization, are co-constitutive. In Online Appendix Section 3.1, we show that nativism and trade attitudes form distinct latent factors.
14 Survey items and sample characteristics are included in Online Appendix Sections 1 and 2. The nativism scale was not included in the 2011 wave, but preceded the globalization questions in the survey instrument for the 2012 wave. We also estimate our models using a shorter, two-item version of the nativism scale that excludes any immigration references. Though we lose precision with this shorter scale, the results in Online Appendix Section 6 are largely consistent with the findings we report here and give us additional confidence that latent nativism values play an important role in forming beliefs about trade.
the cultural fears and group-based distinctions at the heart of the construct. To measure our other key predictor variable, free market attitudes, we include a single item from the 2011 survey wave that asks participants if they agree or disagree that "people are better off in a free market economy." This item separates economic liberalism from other elements of political ideology, since people "keep orientations across policy domains … largely distinct" (Feldman and Johnston 2014, 338). Its precision also lends cross-national construct validity to our analysis, because abstract measures of liberalism/conservatism imply a different set of priorities to citizens in the United Kingdom compared to those in the United States. Additional independent variables control for economic perceptions, other values, and demographic characteristics. A four-item additive index measures retrospective and prospective evaluations of participants' national and household economic situations (α UK = 0.80, α US = 0.85). 15 Next, we include measures of universalism and conformity values to account for the possibility that other values confound the relationships that we expect between nativism, free market beliefs, and trade attitudes (Schwartz 1994;Caprara et al. 2006;Bayram 2015;Goren et al. 2016). Universalism is a "self-transcendance" value that suggests a "belief in the interdependent nature of international relations" (Rathbun et al. 2016, 128). Conformity, in contrast, is a "conservation" value that stresses "social solidarity" as a way to improve prospects for the in-group (Rathbun et al. 2016, 128). In Rathbun et al.'s research on personal values and foreign policy attitudes, universalism is the strongest predictor of cooperative internationalism, while the conservation dimension is the strongest predictor for militant internationalism and isolationism. Since beliefs about trade capture an important if oft-overlooked aspect of internationalism (Gravelle, Reifler, and Scotto 2017), our analyses account for these theoretically relevant controls.
Finally, we account for relevant demographic factors (Hainmueller and Hiscox 2006;Guisinger 2017): race (a dummy variable coded 1 for white), gender (coded 1 for male), age (in years), education (dummy coded 1 for university graduate), and employment status (dummy variables for full time, part time, unemployed, student, retired, and others). An eleven-point scale captures left/right ideology, and we include a seven-point scale-measured in the previous survey wave-to measure attitudes about whether "the government has an essential role to play in regulating the market." Table 2 presents the results of 12 ordinary least-squares (OLS) models that regress participants' belief in the global (models 1-6) and national (models 7-12) benefits of globalized trade on nativism, belief in the free market, and the interaction. The table indicates which models include the full complement of control variables we described above. The dependent variable and all continuous measures, except age, are rescaled from 0 to 1. All models include survey weights based on population parameters. 16 15 An exploratory factor analysis revealed that evaluations of the economy loaded on a single dimension despite questions that separately targeted the country, household, and prospective/retrospective assessments for each. 16 Results in table 2 are from a "two-step" approach to modeling (James, Mulaik, and Brett 1982;Anderson and Gerbing 1988), whereby factor scores are first derived from a measurement model, with the regression models estimated in a second step. This is done for practical purposes-the weighted least-squares with adjusted means and variances (WLSMV) estimator appropriate for estimating a CFA when indicators are ordinal only is available in the Mplus software package, while Hainmueller, Mummolo, and Xu's (2019) kernel estimation procedure to plot the marginal effects from the interaction is implemented in R. However, the "two-step" approach has strident critics, both from a model fitting perspective (Hayduk 1996) and due to concerns about biases in the means, variances, and relations with other variables when using derived factor scores in a two-step procedure (Skrondal and Laake 2001). For this reason, we perform a robustness check by simultaneously estimating both the measurement and structural models in Mplus and present the

Results
(3) (8) (10)  Notes: Models report OLS coefficients with standard errors in parentheses. The dependent variables and continuous independent variables, except age, have been rescaled from 0 to 1. Models incorporate population weights, and models 3, 6, 9, and 12 include controls for race, gender, age, university education, and employment status, omitted for space. Full model results are reported in the Online Appendix Section 4.
Models 1, 4, 7, and 10 provide support for hypotheses 1a and 1b, and confirm our expectations that both economic and cultural beliefs independently shape trade attitudes. In both the United States and United Kingdom, significant coefficients show a negative association between nativism and the belief that trade benefits the world (b US = −0.42, b UK = −0.31, both p < 0.01) or the nation (b US = −0.37, b UK = −0.26, both p < 0.01). The same people who believe in closing their borders to immigrants are less likely to believe that globalized trade has positive consequences. This finding complements existing research on the relationship between out-group anxiety (Mansfield and Mutz 2009) or cultural fears (Margalit 2012) and protectionism, but with a key addition: Nativists are not just concerned about damage at home-they are unpersuaded that trade is good for developing countries, either. This finding supports our theory about nativism as a core value that guides how people think about the consequences from open trading in general.
We find that people who embrace free markets extend their commitment to the international realm-lowering barriers to trade brings global benefits (b US = 0.07, b UK = 0.17, both p < 0.01) just as the invisible hand improves living standards at home (b US = 0.05, b UK = 0.26, both p < 0.01). The effect is positive and statistically significant in both samples, though nearly three times larger in the United Kingdom than in the United States. The substantive effect of a one-unit change in nativism on beliefs about global benefits is consistently large compared to a one-unit change in belief in the free market-5.6 times larger in the US sample and 1.8 times in the UK sample.
The main effects of nativism and belief in the free market are consistent with our expectations, but the remaining models in table 2 provide evidence for an important interaction: Nativism moderates the relationship between free market beliefs and beliefs about trade's positive effects on the nation and world. The interaction coefficient is negative and statistically significant across countries and outcomes, indicating that the effect of free market beliefs weakens among more nativist respondents.
To illustrate the relationship belief in the free market and beliefs about trade conditional on nativism, we follow Hainmueller, Mummolo, and Xu's (2019) advice and use their kernel estimation procedure to plot marginal effects without assuming a linear functional form. This method uses a five-fold least-squares cross-validation procedure to estimate local effects across the range of nativism. Estimates include the full complement of controls included in models 3, 6, 9, and 12. Panels (a) and (b) in figure 1 plot the marginal effect of free market endorsement on beliefs about the global benefits of globalized trade, conditional on nativism, while panels (c) and (d) plot the same quantity of interest on beliefs about trade's effects on the nation.
All four panels display a clear downward trend: The marginal effect of free market decreases as nativism increases. Among nativists, people who support the free market at home are less likely to translate that support into their beliefs about the effects of international trade. Panel (a) shows that in the United States, the estimated effect of belief in the free market on beliefs that globalized trade has global benefits approaches 0 when nativism reaches 0.55 (the 70th percentile). Panel (c) shows a similarly large decrease in the marginal effect on beliefs about the national consequences of economic globalization. Indeed, the effect of free market beliefs is negative among the most ardent American nativists. For Americans who are animated by cultural homogeneity, economic principles do not usurp their belief that trade barriers create important benefits. The invisible hand reigns at home, but not in foreign interactions.
In the United Kingdom (panel b), the interaction is notably weaker albeit still significant. Whereas marginal effect estimates dip below zero among the most nativist results in Online Appendix Section 5. Overall model fit statistics remain adequate across all specifications, and the signs and magnitude of the coefficients of the nativism, belief in the free market, and interaction predictors are comparable to what we see in table 2. Americans, the conditional effect of belief in the free market on both the global and national goods of trade is positive across the range of nativism in the United Kingdom. Although our research design does not directly test contextual moderators, this result comports with contemporary cross-national differences in partisan connections between anti-immigration platforms and global economic policy-the US GOP has turned more strongly against trade than nativist parties in the United Kingdom. This result suggests that national context can shape the relationship between values and policy preferences, a point we return to in the conclusion. Turning to the control variables, the results from models 3, 6, 9, and 12 show that economic sentiment is positively associated with beliefs that global trade has widespread benefits. This brings cross-national evidence to bear on a theory that has been primarily tested in American samples. Our evidence adds external validity by showing that in at least one additional Western democracy, economic perceptions are associated with positive attitudes toward open trading.
University education is associated with positive beliefs about trade's effects in the United States (b global = 0.05, b national = 0.04, both p < 0.01) (Hainmueller and Hiscox 2006;Mansfield and Mutz 2009), but this does not hold in the United Kingdom, where university-educated respondents are slightly less likely to believe that trade benefits the nation (b = −0.02, p < 0.01). 17 This finding suggests an important avenue for future research into how variation in university curricula might affect the relationship between education and trade attitudes, but we are wary of making strong claims given the goals of the present study. 18 Full-time students in the United Kingdom appear to hold more positive beliefs about trade than their unemployed counterparts, but the data do not reveal other consistent and substantively meaningful relationships between employment status and our dependent variables.
Previous work finds that the personal values from Shalom Schwartz's (1994) circumplex are related to foreign policy preferences . That work implicates universalism and conservatism in particular, and we included those two values to more clearly separate the effect of nativism from other values. Neither universalism nor conformity values are related to beliefs about trade's global effects, though universalist Americans are less likely to believe that globalized trade is good for the nation (b = −0.06, p < 0.05). Conformity has a similarly small but positive coefficient in the American sample (b = 0.04, p < 0.05). Variation in the effect of these core personal values would benefit from additional research into how the antecedents of trade attitudes might differ from other foreign policy orientations (Gravelle, Reifler, and Scotto 2017).

Nativism and the Limits of Economic Messages
Our survey results provide support for our theory that nativism is a core value that shapes beliefs about trade and constrains the effect of free market ideology. The observational evidence supports hypotheses 1a, 1b, and 2-nativists tend to believe the worst about global trade, and foreign antipathy outweighs free market attitudes. In this section, we present the results of an experiment that builds on the correlational evidence to test hypotheses 3 and 4. Given their commitment to the homeland, will nativists update their beliefs and support free trade if they are exposed to a message that describes the economic benefits from an open economy? Previous work demonstrates that trade attitudes are malleable (Guisinger 2017), but also that people respond differently to economic information-some participants in Rho and Tomz's (2017) experiment reacted to distributional cues with altruism rather than egoism, for example. We argue that on average people will respond to a message about how the free market improves domestic job prospects with stronger opposition to tariffs, but that people who hold nativist values will be less susceptible to persuasion on the basis of trade's economic effects. We analyze the results of a survey experiment to test whether nativism moderates the effect of an important macroeconomic message.

Methods and Results
We embedded a between-subjects experiment in wave 2 of the survey to assess the effect of a message that describes how open trading leads to long-term economic growth on support for lowering tariffs. All respondents read one message about possible short-term tradeoffs from tariff reductions: Free trade reduces poverty abroad and lowers the cost of consumer goods, but also leads to domestic job losses in specific industries. Via random assignment, half of the sample received an additional message (in italics below) about long-term economic growth prospects, noting that when poor countries develop, they become a new market for goods and services, which ultimately improves the developed countries' economies and encourages job growth. 19 One way that rich countries could reduce poverty in the developing world is to lower trade restrictions, allowing developing countries to sell more goods like food and clothing products. Lowering these trade restrictions would lower prices in rich countries for food and clothing, but would also cause significant job losses in these industries. However, as poor countries continue to grow their economies through exports, they will be able to buy more products and services from developed countries, resulting in economic growth and new jobs for the developed countries. Overall, would you support or oppose rich countries lowering trade barriers allowing developing countries to sell more goods like food and clothing?
We focus on these long-run, downstream advantages because they demonstrate a potentially persuasive political argument in favor of trade even as many economists focus their research on short-and medium-term consequences from opening export markets. 20 Per this logic, trade makes groceries more affordable, but also redounds to the benefit of American or British workers by driving growth and creating future employment opportunities. The Obama administration implored the public to support the Trans-Pacific Partnership using this logic, for example: "The more we sell abroad, the more higher-paying jobs we support here at home." 21 The dependent variable is a five-point scale from "strongly oppose" to "strongly support," rescaled to range from 0 to 1. We measure nativism with the same additive scale as reported above, rescaled to range from 0 to 1. 22 We find that this simple message about how trade generates long-term gains for developed countries increases support for lowering trade barriers, relative to the condition where participants learned only about developing country gains and national losses. Figure 2 plots means and 95 percent confidence intervals by country and condition, where higher values represent greater support for free trade. While UK residents are more supportive of removing trade barriers than their US counterparts, the treatment effect is comparable across samples: Mean support is 0.044 units higher among both Americans (t = −4.49, p < 0.01) and Britons (t = −4.81, p < 0.01) assigned to the treatment group.
These results provide cross-national support for hypothesis 3 and evidence that messages touting the benefits of free markets for national job gains may counteract protectionism in Western democracies (Guisinger 2017;Mutz and Kim 2017). Importantly, results presented in Online Appendix Section 8 show that the treatment effect is not limited to the most educated citizens. We find no interaction between 19 Importantly, our vignette and treatment focus on one plausible argument about costs and benefits from trade, but the reality of trade's effects on jobs entails much more nuance and debate than we present in the experimental prompt. For example, economists point out that protectionist policies sometimes cause immediate domestic job losses-tariffs might help keep jobs in "upstream" industries like steel producers, but lead to cuts in "downstream" producers like tractor companies who rely on steel to make their products and immediately must pay more for the raw materials (Irwin 2017). The net effect of trade in goods on jobs in the short term, then, might be neutral insofar as lower tariffs reduce jobs in inefficient industries but increase available jobs in other industries. In the medium term, however, research on the Chinese import shock shows that import competition caused widespread job losses in the United States, and that these job losses were not offset by new jobs in industries that were not exposed to new competition (Acemoglu et al. 2016, S183). Our view of the political debate around trade-perhaps because of the Chinese import shock-is that those advocating against trade have successfully framed the issue as generating immediate costs. Our experimental prompt was designed to offer an argument about long-term benefits as a competing consideration rather than a direct refutation. 20 We additionally chose to write the experimental prompt as a form of intertemporal choice or delay discounting problem-will an expression of long-term benefits overcome an expression of near-or medium-term losses? While some public figures may advocate for trade by emphasizing nearer term gains (perhaps even concentrating on gains for certain sectors), we chose not to focus on those types of messages here. 21 The White House, President Barack Obama."What's at Stake: Writing the Rules to Support American Jobs." Ac-  To test hypothesis 4, which states nativism moderates the effect of the economic message, we estimate OLS models that regress support for free trade on the treatment, nativism, and the interaction. 23 In both countries, negative coefficients on nativism (b US = −0.24, b UK = −0.26, both p < 0.01) comport with our expectation that nativism and protectionism go hand in hand among participants in the control group.
We find mixed support for our hypothesis that nativism moderates the effect of the pro-market message. In the United States, we observe a clear interaction (b = −0.089, p = 0.052). 24 Participants who reject nativist sentiments respond more positively to the treatment compared to their high nativism counterparts. They express more support for free trade in general, and respond positively to the prospect of new jobs in developed countries. By contrast, the OLS estimate for the interaction coefficient is non-significant in the UK sample. Figure 3 plots the effect of the treatment conditional on nativism in both samples. Panel (a) shows that nativist Americans do not change their attitudes in response to a message designed to shift their macroeconomic beliefs, consistent with our expectations that values limit the effect of lower-order beliefs. Panel (b) shows that in the United Kingdom, the treatment effect is positive across the range of our moderating variables. We find some evidence that the treatment effect is smaller at higher levels of nativism, but the confidence interval indicates substantial uncertainty. This 23 Full regression results are presented in Online Appendix Section 7. 24 To further evaluate the difference in treatment effect among the most nativist and least nativist parts of our sample, we conducted a post hoc analysis in which we split the sample into terciles on the nativism moderator. We find evidence for a significant interaction between an indicator for low nativism-people who scored in the bottom 33 percent on the scale-and the information treatment, relative to people in the top tercile on nativism (b = 0.054, p = 0.019). finding suggests that nativism's constraining effect on trade attitudes may depend on national context. In a national environment rhetorically tilted more toward free trade, nativist and non-nativist Britons alike respond to a positive economic message by increasing their support for tariff reductions. In Online Appendix Section 11, we show that nativism is uniquely important-neither ideology nor partisanship interacts with the economic message treatment in the United States or United Kingdom. Although our results show that nativism moderates the treatment in the United States, our approach does not allow us to adjudicate exactly what participants glean from the prompt or the extent to which they update their broader economic belief system. The treatment message describes how the free market benefits national and global economies, but it potentially implicates both self-and national-interests by targeting jobs. We would encourage future work to explore whether one account is better than the other. Moreover, both the treatment and control groups include information about how developing countries also win from free trade. This makes relative gain concerns salient, and nativists might join people with high social dominance orientations in eschewing relative losses (Mutz and Kim 2017). This concern over relative gains suggests a further avenue for research exploring whether nativists oppose free trade even when politicians frame it as a win-win proposition; those concerned with relative gains would perhaps only support free trade if domestic gains far exceed foreign gains. Similarly, American nativists might update their beliefs in response to an economic message that only targets the United States.

Conclusion
Research on trade attitudes is divided about whether members of the public prioritize economics or values related to in-group protection and out-group antipathy when they think about trade. Our findings, summarized in table 3, affirm existing theories and evidence about symbolic values and economics, but add important new wrinkles. First, in the case of free market beliefs and nativism, our theoretical and empirical frameworks suggest that the latter constrains the former. Those higher in nativism are more likely to believe that trade has negative domestic and international consequences, and nativism weakens the connection between beliefs about what makes a domestic economy thrive and beliefs about trade. This finding has important implications for the debate between symbolic and economic explanations, but also for a growing body of research on how competing predispositions interact with each other to shape foreign policy public opinion (Hurwitz and Peffley 1987;Prather 2017;Rathbun, Powers, and Anders 2019). Moreover, it suggests fruitful paths for engagement with the recent wave of research on the individual-and community-level effects from import shocks. The influx of competition created a rightward and authoritarian shift in both the United States and the United Kingdom (Ballard-Rosa, Jensen, and Scheve 2018; Colantone and Stanig 2018), and our research suggests that preexisting values might moderate the effects of such exogenous shocks on attitudes and economic policy preferences. We encourage future research to further explore this possibility.
Second, our observational evidence reveals a similar pattern in both the United States and the United Kingdom, but our experiment suggests that national context might shape the degree to which nativist values constrain support for open economics. This finding is striking in light of the fact that much existing research on foreign policy public opinion aims to describe fundamental relationships, and past work provides evidence that belief systems have similar structures in the United States and United Kingdom (Jenkins-Smith, Mitchell, and Herron 2004;Reifler, Scotto, and Clarke 2011;Margalit 2012;Gravelle, Reifler, and Scotto 2017). Yet if context and elite messaging influence how tightly people tie their nativist values to specific policy preferences (Knoll and Shewmaker 2015), structural relationships can differ even in two otherwise similar Western populations : For example, the UK Independence Party (UKIP) morphed from a H1a: Nativism will be associated with negative beliefs about trade's effects on the nation and the world. H1b: Free market beliefs will be associated with positive beliefs about trade's effects on the nation and the world. H2: Nativism will moderate the relationship between belief in the free market and beliefs about trade. The positive relationship between belief in the free market and trade attitudes will weaken as nativism increases.
(though weaker) H3: Exposure to a political message touting trade's long-term national benefits will increase support for free trade. H4:Nativism will moderate the effect of a political message touting trade's long-term national benefits. The effect of the treatment will weaken as nativism increases.
X single-issue anti-EU party to a nativist party whose 2012 manifesto read "Britain for the British first" 25 -but the party's message also embraced free trade as a route to prosperity. 26 And recent research shows that polarization shapes the effect of elite cues (Guisinger and Saunders 2017). If free trade was less polarized in the United Kingdom compared to the United States, expert messages targeted at updating free market beliefs should be enough to move opinion, whereas nativist Americans might require a message from a party leader before they shift their attitudes. Alternatively, nativist citizens could influence party platforms, a possibility we return to below. Although our study was not designed to test important contextual moderators like elite messaging or polarization-we focus on bottom-up influences on public opinion rather than top-down factors-our findings highlight the added value in pursuing comparative research on foreign policy public opinion and suggest important avenues for future work (Hurwitz, Peffley, and Seligson 1993;Bjereld and Ekengren 1999;Fattore and Fitzpatrick 2016;Wu 2019). Our findings are especially interesting because nativism and free market beliefs are not only correlated with one another but historically tied to right-of-center parties. In the 1980s, Ronald Reagan and Margaret Thatcher each advocated for-and doggedly pursued-classically liberal economic policies. President Reagan was effusive in his praise for Milton Friedman, an economist renowned for his passionate defense of free markets. Prime Minister Thatcher made radical changes by dramatically reducing state control and support of industry in the United Kingdom.
The nativist tendencies of recent US and UK leaders have been plain to see. Donald Trump spent 2012 reviving the long discredited "birther" conspiracy against President Obama, and in 2016 ran on an anti-immigration platform. Former Prime Minister Theresa May pursued a policy designed to create a "hostile environment" toward illegal immigrants as Home Secretary. She paid for billboard vans to circulate in immigrant communities with the message "Go home or face arrest," moved to deport legal residents who migrated from the Caribbean in the 1960s (the "Windrush" scandal), tightened family migration and student visas, and imposed stricter limits on work visas for those from non-EU countries. Current Prime Minister Boris Johnson wrote in an August 2018 Daily Telegraph column that Muslim women wearing veils "go around looking like letter boxes" and compared them to bank robbers. 27 But Trump has also radically transformed the economic platform of the Republican Party-far from Reagan's embrace of the free markets, the 2020 GOP cancels free trade agreements and imposes tariffs. The Conservative Party's drive to undo a lasting part of their 1970s' legacy under Prime Minister Edward Heath and a cause championed by Thatcher early in her Prime Ministerial tenure-the UK's EU membership-is a staggeringly bold reversal of the Party's heritage. And, just recently, Prime Minister Boris Johnson has sealed Britain's exit from the EU. Despite his frequent rhetoric extolling free trade and a "global Britain," he also signals a turn away from a comprehensive free trade deal with the EU. Although observers discuss "Canada" or "Norway" style trading arrangements for the post-Brexit United Kingdom, Prime Minister Johnson suggested a much more distant relationship when he suggested that an "Australian" style deal-which is not a free trade arrangement-is an acceptable path forward.
What we find especially intriguing is the possibility that our data presaged these dramatic changes in the economic orientations of the Republican and Conservative parties. These transformations are perhaps especially noteworthy because one 25  could argue that they were largely bottom-up. In the United States, Trump won the Republican nomination in 2016 despite widespread elite opposition. In the United Kingdom, the story is more nuanced, but the Tories repositioned themselves to prevent voters from fleeing to Nigel Farage and UKIP. Our data suggest that these turns away from classical economic liberalism and economic globalization are rooted in the negative interaction we observe. In our data, right-of-center voters still endorse free markets (large majorities of Republicans and Tories agreed that "people are better off in a free market economy"), but the extent to which this endorsement results in greater support for free trade and economic globalization declines with greater nativist sentiment. As nativist rhetoric has taken center stage in these parties, perhaps the parties' turn away from their historic support from economic liberalism is not so surprising.