Free trade agreements are a growing phenomenon, with important consequences for national economies and political systems. Nowhere is that more true than in developing countries aligned with large developed markets. Developing economies can be altered significantly for good or ill as trade patterns and investment flows adjust to new openings, especially in sensitive sectors (Wals et al. Reference Wals, Theiss-Morse, González and Gosda2015), or in regions that are exposed to trade competition. Much depends on their economic structure and level of development.
Democratic accountability is also potentially affected by free trade agreements. On the one hand, national governments lose some policy autonomy under regional trade rules. Governments commit to common rules, and this can have an impact on their capacity to serve domestic interests. On the other hand, for developing countries in democratic transitions, the agreements can alter expectations and lock in democratic reforms, acting as a constraint on national governments (Pevehouse Reference Pevehouse2002; Seligson Reference Seligson1999; Davis et al. Reference Davis, Gabel and Coleman1998; Beliz and Chelala Reference Beliz and Chelala2016). Moreover, trade agreements are often associated with neoliberal governments, and their legitimacy is sometimes contested by opponents on ideological grounds (Cárdenas Gracia Reference Gracia2015).
The prevalence of free trade agreements between developed and developing countries is growing, not simply in the case of the United States and its many trade partners, but also between the European Union and its trade partners. Given this impact, it is important to know as much as possible about public opinion in developing partner countries, in particular what factors are associated with this opinion. While elites may enter into trade agreements as a means of promoting development (Fairbrother Reference Fairbrother2019), how do average citizens perceive them? Is trade integration seen as a way to create (or destroy) economic opportunities—a utilitarian rationale? Is it seen through a nationalist or ideological lens, as a means by which the dominant cultural identity may be weakened? Do citizens use cues from political elites as shortcuts to form opinions? Answers to these questions are important because the very legitimacy of these agreements is at stake.
This article examines Mexican attitudes toward the North American Free Trade Agreement (NAFTA) over the period 2008–2018. Mexico is one of Latin America’s two largest economies. It is a key player in North American trade, with dense networks of integrated supply chains, and is the second-largest US trading partner. How favorable were Mexicans to NAFTA? NAFTA has now been replaced by the U.S.-Mexico-Canada Agreement (USMCA), but our findings hold important lessons, and they are striking.
Three novel findings emerge, challenging the status quo regarding our understanding (and for that reason the title implores Mexicans to “wake up”). First, despite NAFTA’s overt economic purpose and its plausible differential material impact on Mexican society, we find that Mexicans’ opinion of NAFTA is not associated with their outlook on material or economic issues but instead is related consistently and strongly with their opinion of the United Nations. Assuming that this finding were confirmed in further research, this would mean that Mexicans essentially interpret NAFTA as though it were an international organization such as the UN, not a means to improve their own welfare or that of the country. Given that NAFTA and the UN have vastly different memberships, purposes, and obligations, this is a very important finding. NAFTA may have assumed a new role in the minds of Mexicans; namely, as a way to be part of the international framework of rules constraining and guiding state behavior, rather than an engine of development for the national economy.
Second, reinforcing this finding, location within subnational regions is an insignificant predictor of support, despite the clear fact that northern Mexico has benefited economically from NAFTA far more than other regions. This finding goes against expectations that people from the north of Mexico would be more favorable to NAFTA, given its economic importance for that part of the country. Additionally, while other cueing variables remain important correlates of Mexican opinion— namely, attitudes toward the Mexican president and toward the United States and Canada—they are far weaker than attitudes toward the UN, and in the case of the president, they have weakened considerably. This again is a departure from earlier findings, and it may herald at long last a separation in Mexicans’ minds between what they think of the president and what they think of NAFTA. The presidential cue, for so many years a key determinant of public opinion on NAFTA, appears to have weakened.
Third, the campaign and election of Donald Trump as president of the United States, with his virulent anti-Mexico and anti-NAFTA messages, put pressure on Mexico. When threatened by Trump, Mexicans rallied behind NAFTA as never before, including leftists who traditionally had opposed NAFTA as a neoliberal device designed to increase the wealth of a narrow elite. This is in sharp contrast to Mexico’s European counterparts, for whom external threats inevitably led to increases in nationalism. Mexican support for NAFTA is unconnected to nationalist ideology. In fact, explanatory variables that are closely associated with support for the EU in Europe, such as left-right ideology, nationalism, and material well-being, play no role in explaining Mexican support for NAFTA.
These findings are highly relevant, both for the Mexican case and as a potential harbinger for other developing countries in their own trade agreements. Public opinion in a democracy is an important measure of democratic legitimacy; when trade agreements are negotiated and entered, they have implications for sectors of the economy and regions of the country, which we should see reflected in public opinion. In the Mexican case, NAFTA was intended not simply to help foster economic development but also to lock in democratic transition (Cameron and Tomlin Reference Cameron and Tomlin2000). Trade agreements, including NAFTA, can be renegotiated or terminated, so public opinion is an important measure of their “appropriateness” at a given point in time. Moreover, public opinion on NAFTA is structured by factors that are very different from those that gave rise to the creation of the agreement in the first place; they include helping create economic opportunities and democratic consolidation in Mexico. This is a very significant finding. What this means for other developing countries that followed in Mexico’s wake is an important question.
Our source of data is the biennial waves of the Las Américas y el Mundo (LAYEM) survey from 2008 to 2018, which permit analysis of these questions. Our strategy is to draw from existing literature to construct models that enable us to better understand Mexican public opinion about NAFTA. We do not seek to evaluate perceptions of economic reform per se or foreign policy more generally, but rather NAFTA itself. The following section reviews the economic and political context around Mexico’s participation in NAFTA. Subsequent sections review the literature on public attitudes toward regional organizations and free trade agreements, draw out testable implications related to public attitudes toward NAFTA, introduce our data and methods, and discuss results.
Mexico and the North American Economy
Mexico is a critically important case study of public opinion for several reasons. As with other developing countries linked to large market economies, a regional trade agreement was a key element in Mexico’s strategy of modernizing and opening to the international economy. Mexico has since entered into many other free trade agreements, both in Latin America and beyond, but NAFTA remains the cornerstone of its external economic strategy, given the size of the US market and the dependence on that market for Mexican exports. In 2019, 80 percent of Mexican goods exported went to the United States, at a total value of $387.8 billion. This represented an increase of more than 700 percent from the pre-NAFTA level.Footnote 1 The main beneficiaries were manufacturing exports, such as electronics and automotive manufacturing (Economist 2014). Other sectors fared less well, especially parts of the agriculture sector, such as maize. Meanwhile, by 2019, Mexico had received $100.9 billion in foreign direct investment stock from the United States.
Gains and losses were felt not simply in different sectors but in different regions of the country. NAFTA benefited the north of Mexico more than the south, for reasons of geographic proximity in the north and poor infrastructure in the south. Moreover, economic problems in the United States had an outsized impact on Mexico, given the reliance on its big neighbor. The year 2008 marked a severe downturn in the Mexican economy, due to the Great Recession in the United States, and it underscored Mexico’s vulnerability in this relationship. In short, the economic relationship is both very important to Mexico and also very well known among Mexicans. By the time negotiations began in 2017 over reforms to NAFTA, most Mexicans had a favorable opinion of it (Zamora Saenz Reference Zamora Saenz2017). Yet because NAFTA is responsible for variations in the performance of sectors and regions, there is reason to believe that public opinion would reflect these variations.
As these economic developments unfolded, a number of political changes occurred that potentially affected public opinion. In 2012, 2016, and 2018, presidential elections took place in Mexico or the United States or both, often with statements addressing the trade agreement and the economic relationship. During the sexenio of Enrique Peña Nieto (2012–18), a series of corruption scandals and ineffective responses to crime caused the president’s approval rating to plummet to just 28 percent by 2017 (Vice and Chwe Reference Vice and Chwe2017). In 2014, the twentieth anniversary of NAFTA again brought renewed scrutiny of it, as did the renegotiations that culminated in 2018. Trump’s presidential campaign added to the upheaval and uncertainty.
Mexico was the first developing country to be linked to a major economy in a trade agreement—a full ten years before Eastern and Central European countries joined the European Union. NAFTA required Mexico to make many substantial changes to its legislation and to commit to enforceable obligations that were unusually strict for the time (Kose et al. Reference Kose, Meredith and Towe2004, 6–7). Thus, not only should we expect NAFTA to matter a great deal to Mexicans, but given its 25-year history, it is important to look for ways that public opinion evolved.
Mexico’s gradual (if imperfect) democratization means that the relevance of individual perceptions matters, especially given its economic dependence on its giant neighbor. Do respondents perceive NAFTA in simple material terms—how it benefits them economically? Do they think about the impact of NAFTA on Mexico as a whole? Do noneconomic factors, such as identity or attitudes toward other international organizations or countries, matter? Is it true that attitudes toward important politicians (i.e., the president) or other actors serve as a shortcut to enable respondents to form a judgment about NAFTA?
Because trade agreements such as NAFTA can be renegotiated, citizen attitudes can also help inform policymakers contemplating such changes. Nevertheless, while research on Mexican public opinion regarding NAFTA has made some important contributions, it is relatively scarce, confined to one-year snapshots, and dated. We therefore draw from public opinion research on regional trade agreements more widely to create a set of initial hypotheses on the Mexican case. We aim to draw conclusions about Mexico itself, but we believe that our findings will help set a baseline for comparative studies of other developing countries, such as Chile and Colombia, that have since entered into trade agreements with powerful market economies, including the United States.
Understanding Attitudes Toward Regional Trade Agreements
How do people view trade agreements? Are they simply economic tools in the minds of citizens, or do they assume some broader relevance, linking Mexico to the international system, like other international organizations? And should we expect opinion to remain unchanged over the years, given that the trade agreement itself did not change, or do changing circumstances also have an impact on opinion? If so, why would that be the case? This section presents findings from earlier research and hypotheses derived from them. Much of the literature is at odds in terms of how to predict the effect of various factors on opinion, although in our hypotheses we take a position consistent with the most plausible associations between explanatory factors and NAFTA opinion.
Trade agreements affect economic interactions between the member states, and so opinion about them is plausibly guided by how individuals perceive their economic well-being. Indeed, much research has shown that personal material or utilitarian considerations structure individual attitudes toward integration in Europe in important ways (Gabel Reference Gabel1998; Jackson et al. Reference Jackson, Mach and Miller-González2011). Moreover, Seligson (Reference Seligson1999) found that perceptions of individual and national economic circumstances positively predicted support for economic integration in Latin America. Others have reported similar findings in the region (Jara Ibarra Reference Ibarra2014; Merolla et al. Reference Merolla, Stephenson, Wilson and Zechmeister2005).
Jara Ibarra (Reference Ibarra2014) distinguished between individual-level variables, such as perception of self-interest, and national-level variables, such as perception of benefits to the country as a whole. Personal economic situation positively predicted support for both economic and political integration across Latin America. Levels of inequality, as measured by the Gini coefficient, had a negative impact on support for integration across all education levels, according to her findings. However, higher levels of inequality led to increased support for integration among those on the right but lower levels of support among those on the left.
Similar findings have emerged in Eastern and Central Europe, where perceptions of the EU have been driven by beliefs about utilitarian benefits. Eastern and Central European states are relevant for Mexico, given their status as less developed regional partners, takers of integration norms and rules, and dependence on larger markets in the core. In one study, positive perceptions of the EU were related to the magnitude of transfers from the EU, as well as to increases in personal income (Jackson et al. Reference Jackson, Mach and Miller-González2011). Other research on the 1991–2003 period, before Eastern and Central European countries became EU members, found that economic winners were more apt to back EU membership than economic losers (Herzog and Tucker Reference Herzog and Tucker2010). The same relationship was shown by Tucker et al. (Reference Tucker, Pacek and Berinsky2002); namely, that economic winners in Eastern and Central Europe were more supportive of membership (as were free market advocates).
Other research undermines the notion that material considerations drive support (Cichowski Reference Cichowski2000; Davis et al. Reference Davis, Gabel and Coleman1998; Davis Reference Davis1998). In Costa Rica and El Salvador, personal material utility mattered less to attitudes toward regional integration than it did in Europe (Davis et al. Reference Davis, Gabel and Coleman1998). In the first paper published about attitudes toward NAFTA, only four years after its entry into force, Davis (Reference Davis1998) found that utility considerations had very little direct effect on NAFTA opinion in Mexico.
In any case, the exact effect of material impacts is unclear, even if it were to exist. Specifically, the material consequences of NAFTA may vary according to the type of economic activity of the respondent, and therefore we feel it is appropriate to test for variation according to occupation or sector. Theoretically, the sector of employment could matter to opinion because as economic conditions change, favoring or impeding trade, certain sectors will benefit or be harmed. This assumption is based broadly on the Ricardo-Viner notion that labor mobility is restricted and that personal fortunes are tied to the sector of employment.
Changing economic and political conditions support this decision. For example, throughout the period of this study, the value of the peso declined by over 70 percent against the dollar—from 11.2 in 2008 to 19.2 in 2018—meaning that Mexican exports to the United States were progressively cheaper and imports from the United States into Mexico were more expensive. This helped Mexican export industries and those that compete against dollar-denominated imports. In addition, once Trump was on the scene as a relevant actor, his threat to withdraw from NAFTA would have left the export sector severely harmed, and in principle we would expect to see a change in the opinion of those in industries tied to export markets in the United States, at least in the 2016 and 2018 surveys. Relatedly, if sector matters to opinion, we should see variation in support according to the region of the respondent. We assume that those in northern Mexico are likely to be more favorable to NAFTA, given the region’s economic reliance on trade with the United States and the fact that NAFTA was directly responsible for the growth of manufacturing hubs in northern Mexico, linked to integrated production chains with US partners.
However, studies have also shown that skill levels drive opinion about trade agreements, given that those in higher-skilled occupations are more likely to be able to change jobs or find alternative employment if necessary. This assumes labor mobility across sectors and greater opportunities for those in high-skilled positions because of having greater knowledge of opportunities and how to take advantage of them, or more capacity to adapt to changing work conditions. Deutschmann and Minkus (Reference Deutschmann and Minkus2018) looked at attitudes toward political and economic integration in 17 Latin American countries and reported higher levels of support for economic integration among those in more skilled jobs. Thus, we cannot be sure whether it is better-paid, higher-skilled workers or those in successful export sectors (at whatever skill level) who are more supportive of integration. At any rate, material consequences for individuals are likely to vary depending on the precise opportunities provided by a given trade agreement.
H1. Material
1. Personal: Wealthier individuals will be more supportive of NAFTA because higher-income citizens are more likely to be engaged in (or prepared for) international economic activity.
2. Skills and sector: Those with higher levels of skills or those who work in export sectors will be more supportive of NAFTA.
3. Region: Those in the north of Mexico will be more supportive of NAFTA, given the region’s economic reliance on exports to the United States
Similarly, cognitive variation also potentially structures attitudes toward trade agreements. People with higher education levels sometimes show stronger support for regional accords, given their possible gains economically, their familiarity with economic ideas and information, their flexibility, and their knowledge of other cultures (Seligson Reference Seligson1999; Hainmueller and Hiscox Reference Hainmueller and Hiscox2006; Jara Ibarra Reference Ibarra2014; Deutschmann and Minkus Reference Deutschmann and Minkus2018). Nevertheless, the impact of educational attainment depends on the extent of exposure to the impact of integration. The more exposure, the more relevant education may be as a predictor of support (Elgün and Tillman 2007).
Hakhverdian et al. (Reference Hakhverdian, Erika van Elsas and Kuhn2013) found that the effect of educational levels on approval of the EU strengthened over time, possibly due to the way that globalization widened the economic gap between winners and losers. Factors such as human capital, convertible skills, networking, and mobility contribute to differences in attitudes toward the EU, and of course, these skills and attributes are affected by education. Hakhverdian et al. found a negative correlation between educational level and Euroskepticism. Others, looking at Central American countries, have found no relationship between support for regional integration and levels of education (Davis et al. Reference Davis, Gabel and Coleman1998).
On the other hand, if education is a proxy for skills, it is not clear that there should be a positive correlation with NAFTA attitudes in Mexico. As discussed above, a factor-based argument would suggest that low-skilled workers in countries with an abundance of low-skilled labor (such as Mexico) have more to gain from trade opening. Indeed, it is the lower-skilled workers who have benefited most from NAFTA, and consistent with this empirical reality, Fairbrother et al. (Reference Fairbrother, Maldonado, Pérez-Armendáriz and Long2019) showed that (among Mexicans) education is negatively associated with support for free trade and cooperation among North American countries. We add an additional cognitive hypothesis related to interest in foreign affairs news to test whether there is an association between that interest and NAFTA support. Literature shows, on the one hand, that interest in public affairs depends on cognitive abilities and, on the other hand, that knowledge and awareness determine citizens’ policy preferences (Althaus Reference Althaus1998; Gilens Reference Gilens2001; Hill Reference Hill2017; Stancheva Reference Stancheva2020).
H2. Cognitive
1. Education. Those with higher levels of education will be more supportive of NAFTA because their economic opportunities or cultural awareness will be higher.
2. Interest. Those with higher levels of interest in news about international issues should be more supportive of NAFTA because of the relationship between awareness and knowledge of the issues at stake, and policy preferences.
Attitudes toward trade agreements can also be structured by ideational factors. Those on the left may hold different positions from those on the right, depending on how they interpret the effects of the trade agreement. However, the impact of ideology is inconclusive and depends largely on existing domestic conditions. Garry and Tilley (Reference Garry and Tilley2015) showed that leftists who live in countries with high income inequality and low state ownership tend to support the EU, but leftists in countries with low income inequality and high state ownership are less supportive. They attribute this difference to the perceived economic consequences of European integration as compared to preexisting national economic conditions. Similarly, Brinegar and Jolly (2005) show that ideology as a predictor of support for EU integration is affected by varieties of capitalism, or the extent of the national welfare state in respondents’ countries. These sociotropic contextual factors color respondents’ evaluations of the EU, which may otherwise be similar across member states. In Latin America, results regarding the importance of ideology have been mixed: some studies show that leftism increases support for regionalism (Deutschmann and Minkus Reference Deutschmann and Minkus2018), and others do not find an effect (Davis et al. Reference Davis, Gabel and Coleman1998; Davis Reference Davis1998; Magaloni and Romero Reference Magaloni and Romero2008).
Identity also matters in a number of studies, especially in the EU. Those with an exclusively national identity are more opposed to European integration, as citizens come to interpret the EU as a threat to national identity or culture (Hooghe and Marks Reference Hooghe and Marks2005; Elgün and Tillman 2007). Nationalism and cultural intolerance generate Euroskepticism; but Euroskepticism is also related to education levels. Some scholars show that these relationships have become stronger over time, especially since the Maastricht Treaty in 1992, because of the increased salience and impact of the EU (Hakhverdian et al. Reference Hakhverdian, Erika van Elsas and Kuhn2013). Similar findings of the impact of nationalism on regional integration have been found in the Americas; namely, that as national pride increases, support for political and economic integration declines (Jara Ibarra Reference Ibarra2014). A related study showed that attitudes toward offshoring in the United States—the practice of placing portions of business operations overseas—were structured more by ethnocentrism and antiforeign beliefs than by income or employment (Mansfield and Mutz Reference Mansfield and Mutz2013, 572). Our ideational hypotheses take into account earlier findings on left-right ideology and nationalism, and we also add one hypothesis on attitudes toward economic globalization to test whether respondents associate NAFTA with globalization.
H3. Ideational
1. Left-right ideology: Those identifying as right will be more supportive of NAFTA, given its aim to liberalize trade and protect investors.
2. Nationalism: Those who hold nationalist attitudes will be less supportive of NAFTA because NAFTA erodes national economic protections and permits the entry of foreign products, services, practices, customs, and ideas.
3. Globalization: Those who favor economic globalization are expected to favor NAFTA because free trade agreements require economic openness.
Cueing also impacts support for trade agreements. Cuing refers to the phenomenon whereby the public follows the positions (or cues) of political leaders instead of learning the issues for themselves. Numerous scholars have shown that respondents’ attitudes toward integration or foreign affairs more broadly are structured by their positions on domestic conditions, such as democratic performance and ratings of domestic leaders, political parties, and governments (Gabel Reference Gabel1998; Anderson Reference Anderson1998; Davis Reference Davis1998; Davis et al. Reference Davis, Gabel and Coleman1998; Seligson Reference Seligson1999; Elgün and Tillman Reference Elgün and Tillman2007; Boomgaarden et al. Reference Boomgaarden, Schuck, Elenbaas and De Vreese2011; Hakhverdian et al. Reference Hakhverdian, Erika van Elsas and Kuhn2013; Harteveld et al. Reference Harteveld, van der Meer and De Vries2013; Schlipphak Reference Schlipphak2015; Beliz and Chelala Reference Beliz and Chelala2016; Deutschmann and Minkus Reference Deutschmann and Minkus2018). When citizens hold high opinions of domestic elites or trust them, they also tend to hold more favorable views of regional organizations and agreements of which the country is a member.
In older research on Mexico, scholars found a clear relationship between individual support for a particular party and support for policies such as economic reform (Domínguez and McCann Reference Domínguez and McCann1996; Kaufman and Zuckermann Reference Kaufman and Zuckermann1998). Yet it is conceivable that we may see a change over time in the relationship between support for the president and support for NAFTA. Given that both Presidents Felipe Calderón (2006–12) and Enrique Peña Nieto (2012–18) came under fire for domestic actions, their capacity to cue Mexicans on NAFTA could have eroded. Calderón organized a disastrous war on drug cartels; Peña Nieto’s administration suffered serious corruption allegations. Did Mexicans still trust them enough to use them as cues?
Cueing is not simply a domestic phenomenon. We also see evidence that external actors, such as powerful states and international organizations, can structure attitudes toward trade agreements. Given Mexico’s relatively weak status as a “taker” of international agreements such as NAFTA, it is not unreasonable to assume that citizens look to external models on which to base their views of NAFTA—most obviously the United States as a powerful partner, but conceivably other organizations of which Mexico is a member and has sought to play a role, such as the UN, the Pacific Alliance, and the OECD. Thus, support for regional integration may be tied to perceptions of domestic success and the prestige of external actors. Cueing may be especially important in a developing country, where citizens realize that much hinges on the success of their external agreements.
For example, positive attitudes toward the European Union and satisfaction with the state of democracy led to higher levels of support for economic integration in 17 Latin American countries (Seligson Reference Seligson1999; also Davis Reference Davis1998). Davis showed that approval of NAFTA among Mexicans was structured above all by approval of the government of Carlos Salinas and by positive attitudes toward the United States. Research by Davis and Bartilow (Reference Davis and Bartilow2007) indicated how cognitive and affective orientations toward the United States helped Mexicans to form opinions about economic agreements between the two countries, including NAFTA.
Positive attitudes toward the United States are also associated with positive attitudes toward regional organizations in sub-Saharan Africa in one study (Schlipphak Reference Schlipphak2015, 367). Cues are related to trust in political actors and to low levels of a priori knowledge of the organizations in question, so that well-known actors serve as a proxy. Wals et al. (Reference Wals, Theiss-Morse, González and Gosda2015) argue that trust specifically in foreigners (rather than trust in a more general sense) predicts positive attitudes toward NAFTA among Mexicans. Conversely, scholars have shown that anti-US attitudes in Mexico—stemming from the US failure to treat Mexico on equal terms—have prevented further deepening of NAFTA (Bow and Santa Cruz 2011). Many Mexicans have reservations about the intentions and trustworthiness of the United States, especially with regard to their sovereignty, as Bow and Santa Cruz point out (see also Carreras et al. forthcoming). At the same time, respect and status plausibly have an impact, too. In other words, those who have a higher regard for the United States (even if they do not trust it) may have more positive views of NAFTA. The election of Donald Trump during the period of our study gives us an opportunity to test this conjecture more directly. Given the low opinion of Trump among Mexicans and Trump’s harsh rhetoric about NAFTA and Mexicans in general, we would expect to see some change in public opinion following his election, to the effect that “if Trump thinks it’s bad, it must be good.” Our data enable us to test for that effect.
International organizations such as the United Nations could plausibly also serve as external cues, given that they provide rules in an otherwise chaotic, power-based international system. Both the UN and NAFTA introduce some clarity on state behavior and (theoretically) constrain nationalistic and self-interested actions by powerful member states. In this sense, the UN could be a proxy for the narrower economic realm of NAFTA. However, this relationship is unclear, given their very different purposes: the UN is not an economic agreement, nor is it regional in scope. It also entails a set of obligations that differ markedly from regional trade agreements like NAFTA. Little research tests this link directly.
Moreover, the direction of causality in these external cueing relationships is not always clear. Flores-Macías and Sánchez-Talanquer (Reference Flores-Macías and Sánchez-Talanquer2019) argue that NAFTA improved Mexicans’ opinions of the United States from 1993 to 2008, the reverse of the relationship assumed in many other studies. Other research shows that communication by national governmental and civil society elites can affect attitudes toward international organizations, thus potentially altering the exogenous role of cues (Dellmuth and Tallberg Reference Dellmuth and Tallberg2021). Moreover, Johnson (Reference Johnson2011) showed that negative public opinion about an influential country within an international organization will affect the latter’s perceived legitimacy. She calls this “guilt by association”—the institutional, financial, or ideational power of the strong state is assumed to bias or prejudice the neutrality and fairness of the international organization. The effect is especially strong with respect to the United States in relation to the UN, the World Bank, and the International Monetary Fund.
H4. Cues
1. Domestic: Those who have positive perceptions of the president are more likely to favor NAFTA.
2. International: Those with positive attitudes toward international organizations or NAFTA partner countries are more likely to favor NAFTA.
We also test for the effects of age. The age effect is unclear; some research has shown that younger respondents are more supportive of regionalism in Latin America (Deutschmann and Minkus Reference Deutschmann and Minkus2018) or the EU (Down and Wilson Reference Down and Wilson2017), but it could be argued that older respondents have more to gain economically, depending on their skills.
H5. Age
1. Respondents’ age is predicted to have an effect on support for NAFTA.
While the studies we review here help us frame our hypotheses and predict likely effects, many suffer from weaknesses. A large number of them are based on surveys taken in a single year, so they do not provide time series data enabling us to determine how relationships between variables may evolve (e.g., Dellmuth and Tallberg Reference Dellmuth and Tallberg2021; Schlipphak Reference Schlipphak2015; Johnson Reference Johnson2011; Boomgaarden et al. Reference Boomgaarden, Schuck, Elenbaas and De Vreese2011; Harteveld et al. Reference Harteveld, van der Meer and De Vries2013; Magaloni and Romero Reference Magaloni and Romero2008; Davis and Bartilow Reference Davis and Bartilow2007; Estrades Reference Estrades2006). Moreover, in Schlipphak’s important study, one of the Latin American regional organizations that he examines is UNASUR, which had been in existence for only a year when the survey was undertaken. UNASUR’s principal purpose was security- related. It provided for no real economic integration, so it was unsurprising that economic issues were unimportant in determining respondents’ positions on UNASUR. Additionally, the literature provides only a piecemeal, fragmented picture of Mexico—with divergent findings in relatively few studies over more than two decades.
Our data help us fill in some of these gaps. In terms of predicting how opinion changes over time, the existing literature gives us few clues. NAFTA itself did not change. Some sectors of the economy became enmeshed in dense networks of investment, production, and trade, particularly the automotive and electronics sectors. Overall trade grew rapidly. Those who benefited plausibly became more favorable to NAFTA over time, in line with the kinds of material effects we discussed earlier. On the other hand, given that much of the country failed to benefit in any significant way, it could be that opinion turned negative over time. As described above, this should be seen in certain sectors and in the south of the country.
Looked at a different way, it could be that NAFTA became disconnected from material considerations altogether, even though it was principally an economic instrument. Why this should occur is unclear. It could be that its technocratic, relatively uncontroversial nature and low salience pushed it from people’s minds, and it disappeared into the forest of international organizations—one among many, each with its well-defined aims and functions. NAFTA, after all, had no agency or “actorness”—it did not impose sanctions or close borders or invade countries. It was stable and static, in addition to being just another trade agreement for Mexico. The lack of day-to-day relevance perhaps impelled people to place NAFTA in the same category as any other international organization, neither posing a threat to Mexico nor offering anything new. On the other hand, the negative relationship between age and support for NAFTA in important anniversary years (2010 and 2014) provides another possible noneconomic clue about changes over time. In those years, Mexico celebrated the bicentennial of independence (2010) and in 2014, the twentieth anniversary of NAFTA’s implementation. It is possible that for older Mexicans, these milestones reactivated nationalist concerns about the effect of the agreement on Mexico.
Additional possible explanations for change over time come from the discussions above about the Trump effect and the declining support for Presidents Calderón and Peña Nieto. Plausibly, Trump’s campaign and election resulted in some differences to the structure of support for NAFTA. We could see this change in the levels of nationalism (i.e., we would see changing opinion among those Mexicans who are more nationalist). Alternatively, we could see shifts in opinion among certain segments of the population, such as those on the left, a group that has historically been skeptical. Furthermore, the serious problems Calderón faced with the drug cartels and Peña Nieto with security issues and corruption could result in a weakening of the longstanding presidential cue. In other words, we may find that the president was no longer taken seriously as a cue for support for NAFTA.
Data and Methodology
We take into account a wide variety of variables, in keeping with findings presented above on public opinion in Europe, Latin America, and elsewhere, in addition to Mexico. Both the time series and the breadth of potentially significant factors that we test should give us greater confidence in the findings about the Mexican case and enable us to theorize Mexican public opinion on NAFTA.
Our analysis extends existing research by examining Mexican opinion over the course of ten years, through six waves of opinion surveys. The Las Américas y el Mundo (LAYEM) survey questions inhabitants of various Latin American countries about their perspectives on domestic and international politics. It is the only survey in the region that gathers data on citizens’ attitudes, interest, and knowledge regarding foreign affairs. The project began in Mexico in 2004 and, in 2008, became a regional Latin American study. In Mexico, national surveys are conducted through a probabilistically representative sample that allows for the generation of estimations applicable to the totality of the population of interest, composed of Mexicans aged 18 and older, living in any entity of the national territory at the time the survey was conducted.Footnote 2
Our dependent variable is operationalized from the following survey question: “On a scale from 0 to 100, where 0 is very unfavorable, 100 very favorable, and 50 is neither favorable nor unfavorable, what is your opinion of the North American Free Trade Agreement? If you do not have an opinion or you do not know the organization, please tell me so.” The dependent variable is this numerical opinion, from 0 to 100. Data are complete for every survey from 2004 to 2018. However, in the 2004 and 2006 versions, attitudes toward NAFTA were measured in a different way from later surveys. Beginning in 2008, the survey began to use this 0–100 “thermometer scale” to measure overall approval or disapproval of NAFTA. Therefore, we use the 2008, 2010, 2012, 2014, 2016, and 2018 editions for the sake of consistency.
As for the explanatory variables, our models draw from other relevant LAYEM questions.Footnote 3 For hypothesis 1, Material, we use the number of lightbulbs in the household as a proxy for income, as well as information on sector of employment and job category and the region where the respondent lives.Footnote 4 For H2, Cognitive, we use questions about educational attainment, the level of interest in news about Mexican foreign relations, and internet use. For H3, Ideational, we include three variables: political ideology, attitudes toward nationalism, and attitudes toward economic globalization. For H4, Cues, we used four variables: opinion about the Mexican president, opinion about the United States and Canada, and opinion about the UN.Footnote 5
Results and Discussion
Figure 1 presents descriptive statistics of Mexican opinion.Footnote 6 Support for NAFTA has consistently been above the theoretical mean (50 points on scale from 0 to 100). In 2008, the mean was 65.5. Approval dropped slightly in each subsequent survey wave, so that by 2016, the mean was 59.8. However, opinion became more favorable again in 2018, with the mean reaching 63.7 points. This change is plausibly explained by the events and discourse emanating from the United States, including bellicose remarks about NAFTA and renegotiations undertaken by the three partner countries beginning in 2017.
Mexican Attitudes Toward NAFTA, 2008–2018
Source: Authors’ calculations based on CIDE, LAYEM.
Our statistical analyses tested the hypotheses with linear regression models. We developed six models, one for each year, and then one model including all the samples and the year fixed effects (see table 1). It is important to be clear that the biennial survey data and our statistical analyses of the data do not permit us to report causality per se, so we try to couch our reporting in appropriate terms.
Models Predicting Opinion About NAFTA, 2008–2018
*p < 0.05; *p < 0.01; ***p < 0.001.
Notes: Linear regression models. Method of estimation: ordinary least squared estimates with countries’ weights. Robust standard errors in parentheses.
Source: Authors’ calculations based on CIDE, LAYEM
What Accounts for Support for NAFTA?
Three striking findings emerge from the data. First, our observations indicate that Mexicans no longer associate NAFTA with economic outcomes, but rather with the UN, another international organization whose role is quite distinct from trade, investment, and other economic objectives. Attitudes toward the UN are highly significant and strong predictors of NAFTA opinion in every year. This is a very important finding because the UN provides Mexico with no immediate material benefit and is not a regional organization. If confirmed, this finding would suggest that Mexicans understand NAFTA as not solely the purview of the president or simply an economic partnership or a tool in Mexico’s economic toolbox, but rather as representing the emergence of Mexico as a partner in rules-based international organizations. In other words, to Mexicans, NAFTA may be about international order, the rules of the game, and perhaps Mexico’s role as an international partner, rather than about getting rich, either as an individual or as a nation.
Other cueing variables are also associated with NAFTA opinion. Attitudes toward the Mexican president and the North American partners are highly significant in most years but with lower effects than attitudes toward the UN. As in other studies of Latin America and Mexico, perspectives on the Mexican president are significant (positive) predictors (Davis Reference Davis1998; Davis et al. Reference Davis, Gabel and Coleman1998; Seligson Reference Seligson1999; Beliz and Chelala Reference Beliz and Chelala2016; Deutschmann and Minkus Reference Deutschmann and Minkus2018). This is the case for every year (and all years combined) except 2016, by which point President Peña Nieto was tainted by scandals and suffering from low approval ratings, even though approval of NAFTA itself remained high (Zamora Saenz Reference Zamora Saenz2017). Quite possibly, respondents dissociated the president from free trade policy, and he no longer served as a cue for public opinion.
In several years, attitudes about the United States and Canada are also related to opinion on NAFTA. In every polling year except 2012 and 2014 (plus the combined- years variable), US opinion is positively associated with opinion about NAFTA, as expected, and statistically significant. Opinion about Canada is also positively associated with NAFTA support in various years (but not in 2010 or 2016). Nevertheless, statistical results suggest that NAFTA is seen less as a project of the president or a proxy for opinion about regional partner countries and increasingly as another rules- based international organization whose purpose is now something other than simple economic benefits. In fact, the international organization cue has more explanatory power than the presidential, the US, and the Canadian cues combined.
Results from regions reinforce this initial finding. Despite common assumptions that NAFTA benefited the north of Mexico far more than other regions and widespread perceptions that northerners are more favorable to NAFTA, being from the north simply has no statistically significant association at all with support for NAFTA. Nor does being from the south in most years. Mexicans from the north appear to be much more opposed to NAFTA in 2016, but in other years this variable was not statistically significant. These regional results are very revealing, given received assumptions and the impact on tradable sectors, but perhaps they would be less surprising if NAFTA’s “economicness” were substituted in the minds of Mexicans with a sense in which NAFTA is another tool in the country’s international toolbox—predictable, ordinary, but not necessarily material in nature.
Second, we see a plausible “Trump effect” in 2016 and 2018, as the relevance of opinion about the United States became stronger and (in 2018) leftists were slightly more supportive of NAFTA (even though leftists have often expressed deep criticism of its neoliberal bias). This change could have been caused by Trump’s criticism of NAFTA (and of Mexicans), given that NAFTA itself did not change in a more progressive direction. Also, the pooled data combining all years show that in 2018, respondents were dramatically more positive about NAFTA in comparison to 2008. Ironically, Trump could have been more of a presidential cue for Mexicans (in a negative sense) than the Mexican president himself.
Third, variables that are associated with attitudes toward regionalism in other parts of the world (especially Europe) have very little or no impact in the case of NAFTA. Their statistical significance is patchy, and their impact is weak and inconsistent. This is especially the case with individual material well-being and nationalism, and to a lesser extent left-right ideology, age, and education. Mexicans’ views of NAFTA are not associated with nationalistic anti-integrationist opinion, for example, as often occurs in Europe. In the EU, threats and crises often provoke a nationalistic response, as we saw in the dual Euro and migration crises over the past decade. In fact, during the Trump assault, Mexicans’ opinion of NAFTA rose rapidly. The one ideational variable that was associated with support for NAFTA was opinion about economic globalization, which is (unsurprisingly) positively associated with NAFTA in some years (not 2012 or 2016), and the effect is very strong.
This globalization result is the only hint we get that utilitarian rationales may matter—albeit at a more abstract level.
On the other hand, individual characteristics appear to be weak explanatory variables. Age is significant only in 2010, 2014, and all years combined. In both years, the relation is negative—younger respondents are associated with more proNAFTA opinion. This is possibly because of the media attention around NAFTA’s twentieth anniversary in 2014 and the bicentennial in 2010, celebrating the initiation of the Mexican independence struggle. It is possible that old arguments about the value of NAFTA for Mexico were reactivated during these two anniversary moments and that older Mexicans voiced negative opinions. However, we do not know if media attention had any effect at all, as our results are inconclusive. Here, material well-being never explains opinion, unlike other studies.
Education is significant in 2014, with a negative relation. The twentieth anniversary of the agreement plausibly highlighted the positive impact of NAFTA on less-skilled workers, leading the less educated to voice more support, though this is conjecture. Interest in foreign affairs news is a highly significant and strong estimate of NAFTA support in 2010 and in all years combined. This variable may be measuring education levels as proxied by media interest, though we cannot be sure. We also measured internet use and found a negative relationship to NAFTA in 2008, but positive in 2018. Television use is insignificant as a predictor, and we have omitted it from table 1. In short, cognitive explanations appear to be weak and shifting.
Left-right ideational factors are generally not associated with NAFTA opinion, and the same is true for party affiliation. Right ideology is positively associated with NAFTA support in 2008 (and very significant and strong), but ideology is only significant in one other year (2018), and by this time the effect is negative, suggesting that leftists are more closely associated with favorable NAFTA opinion. This, too, is potentially due to the Trump effect, given that leftists have not historically been supportive of Mexico’s neoliberal policy of trade opening.
Considering the year effects, over the ten-year period of study, Mexican opinion on NAFTA only changed in a statistically significant way in 2014 (the 20th anniversary of the agreement), and in 2018 (during the renegotiation). The impact as measured by the coefficient was very strong in those years, particularly in 2018, plausibly as a result of the publicity surrounding the renegotiations and the threat posed to the agreement by Trump. It is possible that events such as the twentieth anniversary of NAFTA in 2014 reactivated some of the arguments about NAFTA among the older generations. Finally, we note that the explanatory value of the models increased over the ten-year period.
Political Economy Explanations
Given NAFTA’s significance for the Mexican economy and the plausible economic rationales for public opinion, we report these data separately here. Our datasets contain information in all years on the sector in which the respondent was employed. The categories are public sector, agriculture, fisheries and livestock, industry, commerce, construction, education, services, and other. However, we found virtually no association between levels of support for NAFTA and these occupational sectors (see online appendix 3). Those in industry were more positive in 2008, those in construction in 2014, and those in fishing and livestock in 2016. Why this occurred is unclear, though given the size of the coefficient for the fishing and livestock sector in 2016, it seems that those respondents may have seen a Trump threat to their exports. Commerce as an occupation predicted approval only in 2008. These results give us enough confidence that we can reject sectoral explanations as possible predictors of NAFTA support.
Next we examined the relationship between skill levels and NAFTA opinion (online appendix 4). Data are limited here, but we do have information in every year for those people engaged in activities outside formal employment: working in the home, student, retired, and unemployed. Employed persons are the reference category. This test assumes that categories such as home workers represent people in lower-skilled activities than those in employment. Home workers were more favorable to NAFTA in 2018, and the unemployed in 2010 (see online appendix). Being a retired person never predicted NAFTA support. Students were favorable to NAFTA only in 2016.
Finally, we report results on skills as represented by position of employment (see online appendix 5). We have data on several categories of occupation in 2008 and 2010, and in 2016, our data show more specific occupational categories, with differences in presumed skill levels. Taking “Manager or Director” as the reference category, we can see that differences in occupations make no statistically significant difference in support for NAFTA. In fact, the cueing variables—especially the UN variable—continue to be statistically very significant and to have a strong impact, even with these additional variables as controls. In summary, we feel confident in rejecting political economy variables as explanations.
Conclusions
The central findings of this study have very important implications for policymakers and researchers alike. Free trade agreements have significance for democratic accountability. The finding that Mexicans associate NAFTA with the UN—another international organization but with very different roles and attributes—suggests that NAFTA may play an important role in Mexican thinking. Indeed, we see a potential divergence between elite views of NAFTA as a development tool and citizen views of NAFTA as an international organization. In its guise as a rules-based organization, NAFTA may make it easier to hold public officials to account in Mexico, given their commitments. It also (in theory at least) makes it easier for Mexico to constrain its powerful neighbor. We do not know whether Mexicans will have a different attitude toward the new agreement, the USMCA.
This does not mean that NAFTA became less important for Mexicans; rather that its importance evolved from being a means to create wealth to a means to participate in the international system. The absence of a subnational regional effect in Mexican support lends additional credibility to the argument that NAFTA was not associated simply with economic outcomes in the minds of Mexicans. Moreover, when threatened, Mexicans rallied behind NAFTA, including those on the left, who traditionally had opposed NAFTA. The upshot here is that Mexicans are not like Europeans, who tend to see the EU through nationalist, materialist, and ideological lenses.
We suggest new avenues to address some of the limits of these findings. First, there is much more to learn about how cueing works. Are we witnessing the decline of the Mexican president cue (Davis Reference Davis1998) as citizens separate their views of the president from their opinion of NAFTA and other external agreements? Second, do we have the direction of effect right when we speak of the importance of the United States as a cue for support of NAFTA? Some research suggests that the relationship could be the reverse—that opinion about NAFTA is what leads to changing attitudes toward the United States (Flores-Macías and Sánchez-Talanquer Reference Flores-Macías and Sánchez-Talanquer2019).
Third, our data show the relevance of the United Nations as an important cue for Mexican opinion on NAFTA. But how (and why) does an international organization take the place of a president or a powerful trading partner as a guide for opinion on a regional economic agreement? We have posited that this occurs when respondents begin to see the regional agreement in a new light—one among many international organizations of which their country is part. Thus there could be a dawning realization that an agreement such as NAFTA serves purposes other than material wealth. Nevertheless, we need to know more about whether international organizations have fixed and exogenous effects on opinion over time, or whether a process of (de)legitimizing (through elite discursive practices or through actual events) takes place and affects opinion (Dellmuth and Tallberg Reference Dellmuth and Tallberg2021; Johnson Reference Johnson2011).
Fourth, the literature provides a much clearer guide to how people feel about regional organizations than whether they feel anything at all. Researchers should examine why the propensity to respond to survey questions changes over time. It may be due to changes in the structure of the economic relationship between Mexico and the United States or to specific economic circumstances, such as exogenous economic shocks. Or it could be due to the arrival of new political elites with alternative narratives about the relationship, such as Trump (Shiller Reference Shiller2019). We simply have no roadmap to answer this question from among existing published works. Answers to these questions will help us shed more light on these important relationships and could serve as a guide to examining public opinion in other developing countries.
Supporting Information
Additional supporting materials may be found with the online version of this article at the publisher’s website: Appendix. For replication data, see the authors’ file on the Harvard Dataverse website: https://dataverse.harvard.edu/dataverse/laps
