How, if at all, does self-interest bear on individuals’ economic policy preferences? Though conventional theories of preference formation usually assume a role for self-interest, the informational demands associated with understanding economic policies can prevent individuals from forming attitudes on an egocentric basis. Building on work showing that policy-specific information facilitates egocentrism, we argue that personal experience with alternative policy options is necessary for self-interested preferences to materialize. To test this argument, we leverage Ecuador’s adoption of the U.S. dollar (i.e., dollarization) and examine whether material-based preferences toward exchange rate policy are conditional on individuals experiencing the transition from the Ecuadorian sucre to the dollar. We find that lived experience with dollarization causes policy preferences to align more closely with citizens’ self-interest, as proxied by measures of capital ownership and skill level. In addition, personal experience with dollarization drives attitudes against a dollarized economy, but primarily among poor and low-skill workers—precisely the groups that benefit less from this policy shift. Rather than entirely discredit the role of economic self-interest, these findings suggest that scholars devote greater attention to how contextual factors can strengthen egocentric policy attitudes.
We thank Natália Bueno, Daniel Hidalgo, and David Bearce for their comments and suggestions. A previous version of this article was presented at the Southern Political Science Association’s (SPSA) Annual Meeting in January 2021. This research was funded in part by the University of Colorado Boulder’s Department of Political Science through a Graduate Student Research Grant.
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
The dominant approach among political economists in understanding voters’ policy preferences involves deducing individuals’ material self-interest from economic theorems (Lake, 2009; Oatley, 2011). Across a range of issue areas—including trade (Scheve and Slaughter, 2001b), immigration (Mayda, 2006; Scheve and Slaughter, 2001a), foreign direct investment (Pandya, 2010), and the welfare state (Hasenfeld and Rafferty, 1989)—scholars have demonstrated that policy preferences are driven by individuals’ position within the domestic economy as proxied by demographic traits, such as education level, wealth, and occupation. For this reason, egocentric preferences are professed to form the “hard core” of international political economy (IPE) research (Lake, 2009, p. 227). Indeed, seminal explanations of economic policy outcomes that emphasize democracy and political parties are anchored in these foundational preference models (e.g., Bearce and Hallerberg, 2011, Milner and Kubota, 2005).
The behavioral turn in political science, however, has cast doubt on self-interest-based explanations of policy preferences (Hafner-Burton et al., 2017). A number of studies demonstrate that policy preferences are driven less by material self-interest and more by non-economic factors, including personally held beliefs, identity, and partisanship (Citrin et al., 1997; Hainmueller and Hiscox, 2006, 2007; Mansfield and Mutz, 2009; Sears and Funk, 1991). Furthermore, even if individuals do form their attitudes based on self-interest, they may lack the knowledge to understand how government polices shape their well-being, especially in complex issue areas (Bearce and Tuxhorn, 2015; Rho and Tomz, 2017).
Anzeige
Does material self-interest play a significant role in driving attitudes toward economic policy? And if so, how do relatively uninformed individuals arrive at self-interested policy preferences? To tackle these questions, recent work has relied on experimental research designs to adjudicate between material and non-material drivers of policy attitudes. With some exceptions, this emerging research program has produced overwhelming evidence that economic policy preferences are rooted in non-material considerations.1 For instance, exposing individuals to different messages about trade (Mutz, 2021) and outsourcing (Mansfield and Mutz, 2013) tends to activate sociotropic or symbolic attitudes, with minimal evidence of economic self-interest. Such findings have led scholars to recommend against “inferring ordinary people’s preferences based on complex economic models” (Mutz, 2021, p. 5).
Yet, there are several issues with relying on experiments to assess the presence of egocentric attitudes. First, the distributional implications of economic policies are extremely difficult to portray in a laboratory setting. Vignettes and images often do not adequately capture the material effects of a policy, thus making self-interest hard to identify even for otherwise selfishly driven respondents (Naoi, 2020, pp. 345–347). In addition, while some survey experiments demonstrate that information provision can cause individuals’ attitudes to align more closely with their self-interest (Bearce and Tuxhorn, 2015; Rho and Tomz, 2017), there remains a lack of attention to how these findings translate to real-world settings. This critique should not be understated, as voters are unlikely to encounter situations in which they thoughtfully consume unbiased information about macroeconomic policies. The information-based defense of egocentrism therefore appears quite hollow without identifying an everyday mechanism by which individuals learn about policy.
To accomplish this task, we argue that personal experience with different policy regimes—specifically, living under both the status quo and alternative policies—bestows individuals with low-cost information that informs their preferences. Living through policy transitions allows individuals to compare their well-being under different policies while also exposing them to media reporting and intense political debates about policies’ distributional implications. This information then assists individuals in identifying policies that best serve their interests. By contrast, citizens who only experience the status quo are less informed and, consequently, less likely to form self-interested preferences. This experience-based argument offers a novel process by which individuals arrive at egocentric policy attitudes in a real-world setting.
We evaluate this argument in the context of exchange rate policy by analyzing public attitudes toward dollarization (i.e., adopting the U.S. dollar as an official currency) in Ecuador. Our choice to focus on exchange rate or monetary preferences—terms we use interchangeably throughout this article—is motivated by several factors. First, whereas select survey research finds income and sector of employment (interpreted as proxies for self-interest) to be relevant predictors of monetary preferences (Broz et al., 2008; Fernández-Albertos and Kuo, 2016), others fail to uncover these same relationships (Gabel and Hix, 2005; Jupille and Leblang, 2007). This has led some researchers to discount—perhaps prematurely—the relevance of economic self-interest when explaining monetary preferences (e.g., Hobolt and Leblond, 2009; Kaltenthaler and Anderson, 2001). Second, while exchange rate policies can be important in boosting cross-border commerce and preventing inflation, they are arguably not as salient to voters as more visible policies, such as immigration or social spending. For our purposes, then, monetary preferences approximate a “least likely” scenario for uncovering egocentric attitudes.
Anzeige
Ecuador also offers a unique opportunity to measure how an abrupt policy change shapes public opinion. Ecuador’s government formally replaced its national currency, the Sucre, with the U.S. dollar in January 2000. Because of this sudden transition, we are able to exploit variation in Ecuadorians’ experience with different exchange rate regimes. Using 2017 survey data from the Latin American Public Opinion Project (LAPOP), we find that monetary preferences are more likely to align with conventional measures of economic self-interest among individuals who experienced both monetary regimes. In particular, wealth and skill level are positively associated with support for dollarization but only among individuals that experienced both the Sucre and the U.S. dollar. These results corroborate our argument that experience with alternative policies is often necessary for self-interested preferences to materialize. Rather than completely discard the role of economic self-interest, we advocate that scholars devote greater attention to how contextual factors cause egocentrism to become more or less relevant.
Egocentrism and Monetary Preferences
Before turning to the role of experience, we briefly present the conventional wisdom surrounding monetary preferences. When explaining individual attitudes toward different monetary regimes, political economists typically emphasize the role of economic self-interest or “pocketbook” considerations.2 These models center on a policy trade-off outlined by Mundell (1962) and Fleming (1962): When capital is internationally mobile, governments must choose between a fixed exchange rate regime or control of domestic interest rates. For instance, a government seeking to stimulate their economy may choose to lower interest rates but, by doing so, cause its currency to depreciate as investors shift their money abroad to capitalize on higher foreign interest rates. Conversely, a government wishing to keep their exchange rate fixed must use their domestic interest rates to defend the currency peg rather than to serve domestic policy aims, such as fighting unemployment. While both of these policy options carry potential costs and benefits for the economy as a whole, they also elicit clear distributional implications that affect how individuals value one policy over the other.
In predicting these social cleavages, many posit differences in monetary preferences among different factors of production—namely, capital owners and labor (Bearce, 2003, p. 377; Fernández-Albertos and Kuo, 2016; Jupille and Leblang, 2007).3 Given that low-skill labor is primarily concerned about unemployment and cannot “exit” the economy in the same way that capital can, workers are more reluctant to sacrifice monetary policy for a fixed exchange rate regime. This contrasts with capital owners, who are likely to express more immediate worries about inflation. From the perspective of these individuals, government control of the money supply threatens to erode the value of their domestic financial assets (Shambaugh, 2004).4 For this reason, capital owners should favor fixed exchange rate regimes as a commitment device that precludes currency depreciation and inflationary policies (Canavan and Tommasi, 1997).
The literature on exchange rate politics therefore leverages the distributional implications of different exchange rate regimes to deduce individuals’ preferences. Nevertheless, efforts to directly test the accuracy of these preference models is still a relatively recent phenomenon yielding mixed empirical evidence. On the one hand, select research focusing on Europe’s experience with the Euro uses individual measures of wealth and skill level to validate egocentric preference models. For instance, Fernández-Albertos and Kuo (2016) find that income and education level—proxies for capital ownership and human capital endowment—are positively associated with support for the Euro. Broz et al. (2008) also show that employment in exporting sectors correlates with a preference for stable exchange rates which tend to promote international trade.
On the other hand, there is still a variety of survey evidence failing to pinpoint the observable implications predicted by conventional preference models. Using two separate U.S.-based samples, Bearce and Tuxhorn (2015) find that income and education fail to predict monetary preferences consistent with economic self-interest. Similarly, studies by Gabel and Hix (2005) and Jupille and Leblang (2007) find either weak or no evidence of a relationship between education and support for the Euro in the United Kingdom, Denmark and Sweden. Many also question whether these attitudes are instead formed based on non-economic considerations, including feelings of nationalism or isolationism (Hobolt and Leblond, 2009; Kaltenthaler and Anderson, 2001). This raises the question: Does egocentrism play any role in driving individuals’ exchange rate preferences?
The Role of Experience in Informing Preferences
We argue the formation of self-interested preferences is conditional on individuals becoming aware of exchange rate policies’ distributional effects through the experience of policy transitions. The idea that experience affects monetary preferences follows from evidence that lived experience shapes attitudes more generally. Examples of this come from retrospective voting studies showing that individuals use their personal circumstances to gauge the performance of elected officials. In lieu of extensive policy knowledge, personal experience with economic hardship (Benton, 2005; Fiorina, 1981; Lewis-Beck and Ratto, 2013), crime (Pérez, 2015), and natural disaster relief (Bechtel and Hainmueller, 2011; Healy and Malhotra, 2009) serve as heuristics by which voters judge incumbents’ policy performance. Similarly, personal economic difficulty can shape opinions about specific policies, including welfare spending (Margalit, 2013) and monetary integration (Fernández-Albertos and Kuo, 2016). In this sense, direct experience stands in as a useful, albeit imperfect, substitute for more detailed policy knowledge.
However, the idea that personal experience conditions the effect of existing predictors of monetary attitudes is less studied. A nascent research program within international relations emphasizes information as a requirement for self-interested preferences (Bearce and Tuxhorn, 2015; Jamal and Milner, 2019; Rho and Tomz, 2017). According to this view, the inherent problem of egocentric models stems not from a lack of selfishness but from the informational demands placed on citizens. Although predicting the “winners” and “losers” of an economic policy is relatively clear-cut from an economist’s standpoint, voters often do not possess adequate knowledge about how policies affect their material well-being (Citrin and Green, 1990; Sears and Funk, 1990). And even when individuals are aware of how status quo policies affect them, they may not hold enough knowledge about the alternatives to act in a self-interested manner (Fernandez and Rodrik, 1991; Jurado et al., 2020).
A corollary of this claim is that providing individuals with policy information should cause their preferences to better align with egocentric models. Consistent with this, scholars have employed experiments to demonstrate how informational vignettes enhance the role of self-interest (e.g., Bearce and Tuxhorn, 2015; Jamal and Milner, 2019; Rho and Tomz, 2017). Lack of information can therefore provide an answer for why monetary preferences sometimes diverge from economic self-interest. But this explanation remains unsatisfying given concerns about how experimental findings generalize to everyday life. Individuals seldom encounter real-world situations where they inform themselves about esoteric economic policies through an informational vignette. Moreover, if individuals must become informed in a laboratory setting for self-interest to “activate,” it remains questionable how much value scholars should continue to place in egocentric models.
Thus, research on policy attitudes needs to identify real-world sources of information and to determine whether these information sources actually strengthen the role of self-interest. We propose two experience-based mechanisms by which the transition between policy regimes accentuates egocentrism: (1) by enabling individuals to make yardstick policy comparisons across different points in time; and (2) by exposing individuals to media coverage and elite cues relating to a policy’s distributional effects.
Facilitating Policy Comparisons
Lived experience of policy transitions allows individuals to compare their economic well-being (including their well-being relative to other groups) before and after changes in exchange rate policy. Whereas macroeconomic policies almost always create domestic winners and losers, a major informational barrier for individuals when forming their policy attitudes stems from uncertainty about policy alternatives. Even when individuals stand to gain from a policy reform, they may not comprehend ex ante whether the future policy benefits or hurts them (Fernandez and Rodrik, 1991). In other words, the “devil” individuals know may still be better than the “devil” they do not. Without the ability to make these yardstick comparisons, voters are often unable to correctly sort out the policies they profit from most (e.g., Belleflamme and Hindriks, 2005; Carrillo-Viramontes, 2022).
By contrast, living under two different policies—in our case, fixed and floating exchange rates—provides individuals with information about the effects of alternative policy regimes. This information assists individuals in identifying their material self-interest by providing multiple reference points over which to compare their well-being. In this way, policy transitions allow individuals to “answer the easier question” (Kahneman, 2011, pp. 97–104) by relying on their first-hand experience instead of dense factual details about exchange rate policy. Specifically, this is a case of attribute substitution (see Kahneman and Frederick, 2002), where individuals substitute readily available information (i.e., their well-being before and after dollarization) for less accessible information (i.e., the precise trade-offs across exchange rate regimes).
Experienced-based judgments, in turn, facilitate egocentrism since personal experience is strongly shaped by individuals’ position within the domestic economy. Consider, for example, a state that abruptly shifts from a floating to a fixed exchange rate. As detailed above, capital owners and skilled workers stand the most to gain from this shift, as they benefit less from efforts to combat unemployment and more from the stable inflation rates generated by fixed exchange rates. While many are initially unaware of these distributional implications, experiencing the transition between alternative policies allows individuals to compare their economic well-being under each and, consequently, to form preferences that best reflect their material interests. On average, capital owners and high-skilled workers will experience the benefits of more stable inflation rates and learn to favor fixed exchange rates. Meanwhile, most poor and low-skilled workers will experience adverse effects on their wages and employment prospects, forming preferences for floating exchange rates. This hypothetical example mirrors real-world shifts in public opinion during the Cold War era where beneficiaries of trade liberalization in developing countries did not support free trade until after such policies were implemented (Fernandez and Rodrik, 1990).
Of course, substituting personal experience for factual information can still produce biased judgments about how policies match individuals’ interests; some individuals’ experiences of policy change may not reflect their self-interest as defined in the IPE literature. For this reason, heuristic judgments will not accentuate egocentric preferences for all individuals. Nevertheless, miscalculations about one’s self-interest are likely to be more egregious when individuals only experience the status quo, especially in complex issue areas where individuals are poorly informed. As Kahneman and Frederick (2002) argue, “biases of insufficient information cannot be described as errors of judgment, because there is no way to avoid them” (p. 54). Given the general lack of information about monetary policy, personal experience should work as a “good enough” substitute that promotes self-interested policy attitudes. We therefore expect experiences with different exchange rate regimes to strengthen the role of egocentrism on average.
Exposure to Media Coverage
A complementary mechanism involves increased exposure to elite cues among individuals who experience the transition between policy regimes. Citizens learn about policies through media that present inconsistent coverage of different policy issues (Jerit et al., 2006; Barabas and Jerit, 2009). Media coverage of abstruse economic policies is typically limited, leaving most individuals with little, if any, information about topics such as exchange rate policy. However, increased media scrutiny during policy transition—when policy differences become salient—should increase individuals’ knowledge of alternative policies, thereby facilitating egocentric preferences.
Media coverage during policy transitions can shape preferences both by providing new factual information and through coverage of opinion leaders. First, media coverage provides information (e.g., Barabas and Jerit, 2009) that directly shapes individuals’ policy views. For instance, news coverage may inform low-skill workers about how fixed exchange rates limit macroeconomic policies to reduce unemployment. These workers would then form preferences for floating exchange rates by considering that information in light of their (un)employment prospects. Experimental work showing that policy information facilitates egocentric preferences indicates the plausibility of this simple informational mechanism.
Yet, many individuals exposed to media coverage might not understand how the aggregate effects of exchange rate policy affect them personally. Work on elite cue taking suggests that these individuals can still develop egocentric preferences by following cues from elites who represent their interests. Citizens can overcome topical complexity and form self-interested preferences by adopting the views of opinion leaders, including peak business organizations and unions (Lupia et al., 1998; Lupia, 1994). In this case, individuals who experience the transition between exchange rate regimes are exposed to messages from elites, then develop self-interested preferences by adopting the views of leaders who best represent their pocketbook interests.
Hypotheses
The conventional wisdom posits that exchange rate preferences are a function of capital ownership and individual skill level. Predictions made by these models are formally stated as follows:
H1A
Wealthy individuals are more likely to favor fixed exchange rates than poor individuals.
H1B
High-skill workers are more likely to favor fixed exchange rates than low-skill workers.
We argue, however, that these egocentric predictors are conditional on lived experience. Through the two mechanisms outlined above, experiencing a policy transition provides information that individuals use to determine the policy that benefits them most. It follows that individuals who experience a transition between monetary regimes should be more likely to hold preferences that match their economic interests. We therefore expect egocentric monetary preferences to materialize only among these “experienced” individuals:
H2A
Increases in wealth are positively associated with support for fixed exchange rates but only among individuals that experienced a transition between policy regimes.
H2B
Increases in skill level are positively associated with support for fixed exchange rates but only among individuals that experienced a transition between policy regimes.
Research Design
Dollarization in Ecuador
We test our argument by leveraging Ecuador’s adoption of the U.S. dollar. Dollarization stands as a unique and particularly rigid type of fixed exchange rate regime. Whereas some Latin American countries have currencies that are pegged to the U.S. dollar, Ecuador currently uses the U.S. dollar as legal tender in lieu of a national currency. Since the supply of U.S. dollars is controlled exclusively by the U.S. Federal Reserve, Ecuador’s government effectively has no sovereignty over its monetary policy and is unable to manipulate the domestic money supply to achieve certain macroeconomic policy objectives, including combating unemployment. This is in stark contrast to Ecuador’s pre-2000 exchange rate regime which involved a freely floating currency, the Sucre.
Ecuador introduced the dollar in January 2000 as a “measure of last resort” intended to escape uncontrolled inflation (Beckerman and Solimano, 2002). Beginning in 1998, Ecuador underwent a “triple crisis” involving the near-collapse of the banking sector, the government’s inability to make foreign debt payments, and the inflationary spiral of the Sucre (Montiel, 2013, p. 198). These crises prompted high-profile discussions about the direction of Ecuador’s currency policy among international experts, business leaders, and various domestic groups (Jameson, 2003; Rohter, 2000). On January 9th, 2000, President Jamil Mahuad announced the country would abandon the Sucre and adopt the dollar in an attempt to escape hyperinflation and revive the economy. Though Mahuad was deposed later that month, his vice-president, Gustavo Noboa, oversaw Ecuador’s formal transition to the dollar with the Ley para la transformación económica del Ecuador (Law of Economic Transformation) in March (Beckerman and Solimano, 2002, pp. 2–3). Dollarization eventually succeeded in stemming inflation and generating monetary stability that fueled Ecuador’s economic recovery (Anderson, 2016; Quispe-Agnoli and Whisler, 2006). For this reason—and similar to other crisis-stricken countries in Latin America (Weyland, 1998)—Ecuador’s switch to dollarization and broader neoliberal reforms garnered popular support from voters. To this day, many Ecuadorians associate dollarization with a shift from economic crisis toward relative stability.5
But despite the success of dollarization in alleviating Ecuador’s inflation crisis, the benefits and costs of abandoning the Sucre have not been evenly distributed. In addition to the drawbacks of relinquishing control of monetary policy which exposes workers to economic shocks (Paredes, 2017), Ecuador’s chaotic transition to dollarization also generated clear domestic winners and losers that reinforce conventional predictors of monetary preferences. Whereas dollarization was largely welcomed by domestic banks and business interests, others decried that the policy shift would produce “massive impovershment” (Rohter, 2000). One reason for this was that Ecuador’s economy was already partially dollarized before the official transition in 2000, with many higher-end products already purchased in U.S. dollars. Poorer citizens, who were less likely to use or already own dollars, therefore faced higher transition costs in switching to the dollar for everyday transactions. Furthermore, Ecuadorians’ bank accounts—then largely still in Sucres—were frozen in March 1999 to prevent mass withdrawals and bank collapses. Since the Sucre continued to depreciate until January, the conversion of Sucres to dollars was harmful to workers holding pensions and savings accounts denominated in Sucres (de la Torre Muñoz, 2019; Roura, 2019).
With these distributional implications in mind, we exploit Ecuador’s shift to the U.S. dollar to compare the monetary preferences of individuals who were of working age before dollarization with those who were not. We utilize survey data from the 2016/2017 wave of the Latin American Public Opinion Project’s (LAPOP’s) AmericasBarometer.6 Survey data were collected through in-person interviews with individuals aged 18 or older in Ecuador between November 8th, 2016 and January 5th, 2017. While conventional models (H\(_{1A-1B}\)) predict that preferences toward dollarization will pit low-skill labor against capital owners and human capital,7 our argument is that these self-interested policy preferences are conditional on individuals experiencing both the pre-dollarized and dollarized economy (H\(_{2A-2B}\)).
Dependent Variable
To measure preferences toward dollarization, we construct our dependent variable from a question asking respondents: In general terms, do you think that having the dollar as a currency has been: very good; good; bad; or very bad for Ecuador? We collapse these response options into a dichotomous measure, Pro Dollar, where 1 is assigned to respondents believing dollarization is “good” or “very good” and 0 is assigned otherwise. While our main results use Pro Dollar as the dependent variable, we obtain similar findings when estimating an ordered probit with a four-category measure as the dependent variable. These models are provided in the appendix.8
Explanatory Variables
Wealth
To measure capital ownership, we construct a Wealth index by taking the first component from a principal component analysis (PCA) of items asking whether individuals own eleven different assets in their household (personal vehicle; home internet access; indoor plumbing; etc.). Asset-based wealth measures such as this are preferable to income-based measures which suffer from significant item non-response and may fluctuate over short periods of time, especially in lower income regions (Córdova, 2009; Filmer and Pritchett, 2001). We normalize Wealth to range from zero to one, with values closer to one denoting wealthier respondents.
Skill Level
To measure respondents’ skill level, we follow existing research which distinguishes between high- and low-skill workers using respondents’ level of education (e.g., Mayda, 2006, Scheve and Slaughter, 2001a; b).9 Specifically, we use a dummy variable, Education, indicating whether the respondent made it beyond the compulsory twelve years of primary and secondary schooling.10 Although there are valid concerns that education may also detect individual characteristics other than skill level (Hainmueller and Hiscox, 2006, 2007), we address this problem by controlling for different non-economic attitudes, interest in politics, and political ideology in our models. We discuss these controls in depth below.
Experience
H\(_{2A-2B}\) posits that the effects of wealth and skill level are moderated by one’s experience with different monetary regimes. Testing this claim requires establishing a cut-off that distinguishes between Ecuadorians with working experience under the Sucre and those without. We start by defining experience under the Sucre as being at least 18 years of age immediately before Ecuador’s dollarization. Thus, our variable, Monetary Experience, equals 1 if a respondent was 18 or older during the year 2000 and 0 if otherwise.11
We acknowledge that there are potential issues with operationalizing experience in this fashion. First, while we intend to use the cut-off age of 18 as a rough indicator of entrance into the labor market, many Ecuadorians begin working before the age of 18. This means we are incorrectly categorizing a small subset of respondents with “false” zeros for the Monetary Experience variable. Given that we do not possess information on the precise time when respondents began working, we address this issue by observing how different measures of monetary experience (i.e., with different cut-off points) affect our results. We discuss these sensitivity analyses in the Robustness Checks section, though these models do not alter our findings.12
Monetary Experience, by its nature, is also correlated with age which may affect monetary preferences. To be sure, it is not necessarily problematic if age is negatively associated with support for dollarization—a relationship that we indeed find. The more relevant question is whether age is moderating monetary preferences through mechanisms that are distinct from experience. To attend to this concern, we control for different individual characteristics likely associated with both age and monetary preferences. These controls are discussed below, and include nationalism, political ideology, partisanship, interest in politics, and retirement. In the appendix, we also estimate models that include Age as an additional covariate.13
Model Specification
We start by estimating the following probit regression model with standard errors clustered on geographic region:
where \(X_{i}\) denotes a vector of controls, and \(\delta _{p}\) and \(\eta _{r}\) denote party-affiliation fixed effects and region fixed effects, respectively. Note that traditional egocentric models (H\(_{1A-1B}\)) predict a positive coefficient for \(\beta _{1}\) and \(\beta _{2}\). However, since our argument posits that the effect of each of these explanatory variables are conditional on whether one experienced both the pre-dollarized and dollarized monetary regime, we interact Monetary Experience (denoted by Exp in the above equation) with wealth and skill level:
According to H\(_{2A-2B}\), we expect a positive interaction term but an insignificant constituent term for Wealth and Education.
Control Variables
The vector, \(X_{i}\), controls for individual-level characteristics likely associated with our main independent variables and policy attitudes. A major concern expressed by Hainmueller and Hiscox (2006, 2007) is that measures of material self-interest are correlated with non-material attitudes, including nationalism or isolationism, that also drive policy preferences. We address this concern by including several controls that capture important non-economic attitudes. We first include National Pride, which is an ordinal ranking of how proud an individual is to be an Ecuadorian. As several works on the Eurozone have shown (e.g., Hobolt and Leblond, 2009; Kaltenthaler and Anderson, 2001; Müller-Peters, 1998), feelings of nationalism correlate with attitudes toward monetary integration given the symbolic value of a national currency.
Additional research shows that trust in political institutions shapes attitudes about economic integration (Franklin et al., 1995; Jupille and Leblang, 2007). Because dollarization implies tying the hands of Ecuador’s government in the area of monetary policy, loss of trust in Ecuadorian politics could cause individuals to become more partial toward a foreign currency. We therefore include an ordinal measure of how much trust an individual places in Ecuador’s legislature. As a final set of measures to control for non-economic attitudes, we include a four-point scale of one’s interest in politics as well as dummy variables for a respondent’s political ideology, party affiliation, and vote choice in the previous presidential election.14 Past vote choice is a particularly important control given President Correa’s well-known opposition to dollarization. For instance, Correa dedicated a chapter of his book to dollarization, a policy he referred to as “monetary suicide” (Correa, 2009, 75). Generally, we expect that left-leaning respondents and Correa supporters will have more negative attitudes toward a dollarized economy.15
Employment status can also affect monetary preferences. Unemployed workers face an immediate cost of dollarization, as the Ecuadorian government does not have control over the supply of U.S. dollars to remedy domestic unemployment. Meanwhile, retirees are more likely to be dependent on government social expenditures which are restricted under dollarization due to Ecuador forgoing its printing press. In addition, certain segments of Ecuador’s population may be more vulnerable to unemployment while retirement is naturally associated with old age (and by connection, our Monetary Experience variable). We therefore include the two dummy variables, Unemployed and Retired, as controls.
We next incorporate a Female dummy variable given the well-documented gender differences in globalization attitudes (Mansfield et al., 2015). While the driving force behind the “gender gap” remains debated, men and women are nevertheless likely to differ in their income and educational attainment due to structural gender inequality. In combination with region fixed effects, our models include a Rural dummy variable to account for the varying nature of local economies. Lastly, we control for whether respondents receive migrant remittances. Not only do remittances mitigate concerns about relinquishing control of monetary policy by providing a stable income source during crisis times (Singer, 2010), but middle- and high-income families are more likely to receive remittances given the high financial barriers to emigration. Complete summary statistics are provided in Table A1 of the appendix.
The Evidence
Table 1 reports the main results. To gauge the relevancy of standard egocentric predictors, Model 1 estimates the unconditional effect of our measures of economic self-interest (i.e., wealth and education) and Monetary Experience on pro-dollarization attitudes. Consistent with H\(_{1A-1B}\), both Wealth and Education are positive and statistically significant at conventional levels, though the substantive effects are relatively small. When holding other covariates at their mean, individuals with a post-secondary education are roughly 4 percentage points more likely to favor dollarization than those without a post-secondary education. Similarly, 1 SD increase in the Wealth index from its mean elicits a roughly 3 percentage point increase in favoring dollarization. Together, these results suggest that capital and human capital endowment are associated with a preference toward dollarization.
Table 1
Attitudes Toward Dollarization in Ecuador
Dependent Variable: Pro Dollar
(1)
(2)
(3)
Monetary Experience
− 0.371\(^{***}\)
− 0.654\(^{***}\)
− 0.460\(^{***}\)
(0.047)
(0.170)
(0.059)
Wealth
0.448\(^{*}\)
0.064
0.445\(^{*}\)
(0.178)
(0.284)
(0.180)
Monetary Experience \(\times\) Wealth
0.739\(^{*}\)
(0.290)
Education
0.120\(^{**}\)
0.114\(^{**}\)
0.031
(0.038)
(0.036)
(0.057)
Monetary Experience \(\times\) Education
0.161\(^{*}\)
(0.068)
Control Variables
\(\checkmark\)
\(\checkmark\)
\(\checkmark\)
Party Dummies
\(\checkmark\)
\(\checkmark\)
\(\checkmark\)
Region Fixed Effects
\(\checkmark\)
\(\checkmark\)
\(\checkmark\)
Observations
1169
1169
1169
Controls are reported in Table A8 of the appendix
Standard errors are shown in parentheses
***, **, * and + indicate statistical significance levels of .1, 1, 5 and 10 percent
According to our argument, however, egocentric attitudes are likely to be muted without accounting for individuals’ disparate experiences. Measures of capital ownership and human capital endowment should exhibit a positive effect on support for dollarization, but only among individuals that experienced both monetary regimes. To test this claim, Models 2 and 3 interact Monetary Experience with our measures of economic self-interest. Consistent with H\(_{2A-2B}\), the interaction term is positive and statistically significant at conventional levels in both models. The constituent terms for Wealth and Education are also statistically insignificant.
To illustrate these results, Fig. 1 plots the marginal effects of Wealth and Education, depending on whether a respondent was of working age at the time of dollarization. Figure 2 also illustrates this relationship by plotting predicted support for dollarization according to individual levels of wealth and educational attainment. When looking at Ecuadorians who only have experience with Ecuador’s dollarized economy, attitudes toward dollarization are roughly identical between rich and poor respondents. However, among Ecuadorians that also worked under the Sucre, Wealth exhibits a sizeable positive effect. For instance, when Monetary Experience is equal to 1, the wealthiest Ecuadorians (i.e., Wealth\(=\) 1) are approximately 27 percentage points more likely to favor a dollarized economy than the poorest individuals (i.e., Wealth\(=\) 0). The relationship between capital ownership and support for dollarization is therefore restricted to individuals who experienced both regimes.
Fig. 1
Conditional marginal effects of economic self-interest (95% confidence intervals). All other covariates are held at their means
×
We obtain similar findings when analyzing the relationship between respondents’ skill level and monetary preferences. There is no statistically significant relationship between human capital endowment and support for dollarization among individuals who have only worked during Ecuador’s dollarized economy. Conversely, experience with the floating Sucre causes preferences to align more closely with economic self-interest. Among these individuals, Education elicits a roughly 7 percentage point increase in support for a dollarized economy.
Fig. 2
Predicted support for dollarization conditional on experience (95% confidence intervals). All other covariates are held at their means
×
The interaction effects in our models corroborate an experience-based story of egocentric preferences.16 These results are also hard to square with non-material factors as a competing explanation. For instance, in addition to being a proxy for skill level, education predicts economic policy preferences since schooling tends to socialize individuals to adopt more global-friendly attitudes, including favoring monetary integration (Hainmueller and Hiscox, 2006, 2007). However, if education or wealth were solely picking up these non-material beliefs, there is no obvious reason to believe that this mechanism would be restricted to individuals who have past experience with the Sucre. Thus, in addition to controlling for non-economic attitudes through nationalism, partisanship, and ideology, there are theoretical reasons to be confident that our results are indeed a story of economic self-interest.
Fig. 3
Conditional marginal effects of experience (95% confidence intervals). All other covariates held at their means
×
Another notable finding regards the effect of Monetary Experience. Although Monetary Experience is negative and statistically significant in Model 1, the interaction models suggest that the negative effect of Monetary Experience is concentrated primarily upon the economic losers of dollarization. The Monetary Experience constituent term is negative and statistically significant in Models 2 and 3, but the positive interaction term in these models attenuates this negative effect among wealthy and high-skill respondents. We illustrate this relationship in Fig. 3 by plotting the marginal effect of Monetary Experience at different values of wealth and education. Monetary Experience exhibits a negative effect on support for dollarization but only when the Wealth index is below 0.8. Similarly, experiencing the transition to dollarization is associated with a 15 percentage point decrease in support for dollarization among the less educated but only a 9 percentage point decrease among those with a post-secondary education. These results suggest that experience with the transition to the U.S. dollar has stained attitudes toward dollarization among poor to middle-class Ecuadorians and low-skill workers.
Robustness Checks
A remaining concern with the results involves the construction of the Monetary Experience measure. While we define monetary experience as being at least 18 years of age immediately before Ecuador’s dollarization, many Ecuadorians enter the labor force before turning 18 and therefore possess some working experience under the Sucre despite being coded as 0 by our measure. Given that the survey questionnaire does not permit us to precisely identify when respondents began working, we are unable to address this issue of measurement validity directly. However, we are able to alter our definition of monetary experience to see how, if at all, alternate measures of Monetary Experience affect our main results.
In Tables A3 and A4 of the appendix, we reevaluate support for H\(_{2A-2B}\) by interacting each of our measures of self-interest with modified measures of Monetary Experience. We sequentially expand our definition of monetary experience to include respondents who were at least 17, 16, and 15 before Ecuador’s dollarization. Figure 4a and b plot the interaction coefficients (with 95% confidence intervals) when interacting different iterations of Monetary Experience with Wealth and Education. Findings supporting H\(_{2A-2B}\) remain virtually unchanged when focusing on wealth and education: The interaction coefficients remain positive and statistically significant at the 95% level, while the constituent terms for Wealth and Education remain statistically insignificant.
Fig. 4
Sensitivity analysis (95% confidence intervals)
×
Another concern with our findings stems from the bundled nature of Ecuador’s economic reforms during its transition to the U.S. dollar. Given that dollarization was accompanied by other pivots in economic policy, including privatization and trade liberalization, it is possible that support for dollarization tracks with approval of the broader neoliberal paradigm. This raises an alternative interpretation of our results: Lived experience pushes wealthy and educated respondents to favor dollarization, not because it informs their interests vis-à-vis exchange rate policy, but because of their general acceptance of post-crisis reforms.
To rule out this story, we implement two different placebo checks. First, we analyze whether Monetary Experience conditions the attitudes of wealthy and high-skill respondents toward privatization. Whereas past research makes clear predictions about the association between wealth/skill level and exchange rate attitudes (Fernández-Albertos and Kuo, 2016; Jupille and Leblang, 2007), there is no obvious reason to think these traits would predict preferences about state-owned enterprises. Second, we assess opinions about state intervention in the economy to combat inequality. These models incorporate the same set of control variables and are reported in Table A7 of the appendix. The \(Wealth \times Experience\) and \(Education \times Experience\) interaction terms are statistically insignificant and often incorrectly signed. We therefore find no evidence that living through dollarization conditions the effect of Wealth and Education on attitudes toward these other policy initiatives. Thus, living through dollarization matters for exchange rate attitudes but does not appear to shape attitudes toward other neoliberal reforms.
Discussion
Recent arguments about citizens’ economic preferences posit a negligible role for material self-interest. Furthermore, the information-based defense of egocentrism has relied on arguably unrealistic tests where respondents are presented with clear policy information then immediately asked about their preferences (Bearce and Tuxhorn, 2015; Rho and Tomz, 2017). This article moves beyond this research by demonstrating how experience with policy transitions endows individuals with sufficient information, making self-interested preferences more likely to materialize. Experience with policy transitions enables individuals to make comparisons while also exposing them to different elite cues about policies’ distributional effects. Since our Monetary Experience variable is based on respondent age, we are unable to disentangle which of these mechanisms is most relevant. Future work might therefore further explore the mechanisms by which personal experience shapes policy attitudes.
Research should also examine other real-world sources of information that facilitate egocentrism. For example, burgeoning research on the social communication of information in Latin America (e.g., Baker et al., 2020; Smith, 2018) , suggests that individuals may develop selfish preferences through conversations with their family and peers.17 Additionally, the informational value of experience is likely to be conditional on the preexisting levels of policy knowledge. In models provided in the appendix, we confirm that Monetary Experience tightens the connection between wealth/skill level and exchange rate attitudes, but only among Ecuadorians who have a low interest in politics and are presumably less informed about dollarization.18
It is also worth exploring the generalizability of our findings outside of Ecuador. In other recently dollarized Latin American countries, such as El Salvador, we would expect that lived experience similarly strengthens the role of material self-interest. Although El Salvador’s former president, Francisco Flores, adopted the U.S. dollar under much more stable conditions than Ecuador, survey research in the years following El Salvador’s dollarization indicates that the primary opponents of the U.S. dollar became poor and low-skill workers (Towers and Borzutzky, 2004, pp. 45–46). Alternatively, scholars could examine countries with long-standing fixed exchange rate policies to see whether self-interest-based preferences fail to form when citizens only experience the status quo. It is perhaps for this reason that unpopular exchange rate policies in the West African CFA franc zone have nevertheless proven to be resilient (e.g., Wilson, 2021).
Broadly, our findings are consistent with bounded rationality as a model for understanding citizens’ preferences. Citizens’ preferences are more likely to be guided by self-interest when their experience provides them with low-cost information about their policy options. The abrupt transition to the dollar in Ecuador is just one example of how experience can facilitate self-interested preferences on the most complex policy issues. Lastly, these findings have significant implications for democratic accountability: Experience with only the status quo can leave individuals too poorly informed to articulate and advocate for their interests, making collective action problems harder to solve. Experience with different policy regimes, in comparison, leaves citizens better situated to champion policies that enhance their well-being. Rather than jettison egocentric preference models, scholars should focus on the conditions that make them more or less relevant.
Open AccessThis article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
In addition, government decisions to devalue their currency further stoke inflation and increase the real debt burden of capital owners with foreign-currency-denominated liabilities.
One might raise the concern that low-skill workers indirectly benefit from the U.S. dollar due to how dollarization promotes international trade. We do not expect this rationale to be salient among Ecuadorian workers since, despite being a relatively small economy, Ecuador is only moderately dependent on international trade with most of its exports concentrated in the capital-intensive oil sector and most of its labor force concentrated within the informal sector (Canelas, 2019).
In Table A6 of the appendix, we also substitute in a continuous measure of skill level, Years of Education. These models generate similar findings. While the Years of Education X Experience interaction term loses statistical significance in Model A16, it is positive and statistically significant in the ordered probit regression model (Model A18).
For vote choice, we specifically include two dummy variables for whether the respondent voted for Rafael Correa (Vote Correa) or another candidate (Vote Other), with non-voters forming the baseline.
Figure A1 of the appendix confirms that Correa supporters are more opposed to dollarization than supporters of other candidates. However, in Table A10, we find no evidence that this relationship is stronger among low-skill workers and the poor.
In Table A11 of the appendix, we also find evidence that unemployed respondents are more likely to oppose dollarization but only when they lack experience with both monetary regimes (see Figure A2). This could suggest that unemployment provides a negative heuristic for workers about the status quo when they otherwise lack information.
