Systematic scholarship on mass political behavior in Latin America is no longer a marginal enterprise, but ’twas not always thus. When I began graduate school in 1994, the study of public opinion and voting behavior in the region seemed poised for takeoff because of democratization and the emergence of a new generation of quantitatively trained Latin Americanists. However, scholarship in Latin American politics took an institutional turn, as political scientists focused on parties, legislatures, electoral systems, and other kinds of rules and organizations.
Fortunately, this turned out to be a temporary delay. In the last decade, Latin American survey data (that adhere to academic standards) have proliferated, thanks in part to the biannual efforts of the Latin American Public Opinion Project (LAPOP), which began in 2004. Moreover, in the last five years alone, numerous monographs and edited volumes on the topic have been published. The ground-breaking and award-winning Brokers, Voters, and Clientelism (Susan Stokes et al. 2013) was soon followed by a bevy of books on clientelism and vote buying in the region (e.g., Rebecca Weitz-Shapiro, Curbing Clientelism in Argentina, 2014) and the ambitious and wide-ranging edited volume The Latin American Voter (Carlin et al. 2015). In short, Latin American voting behavior is finally a hot topic in political science.
Latin American Elections wades into this booming territory with an important contribution that seeks to bring theoretical order to the newfound diversity of findings in the field. The authors import from U.S. voting studies the Michigan model and its “funnel of causality” theoretical framework. Developed in the 1950s, the funnel of causality groups the causes of vote choice into a set of categories that are ordered from long-term, slow-moving forces (such as race, gender, and class) to short-term ones that are idiosyncratic to each election (such as performance and candidate evaluations). The authors organize the book around these causal categories, with chapters on demographics, socioeconomic affiliations, partisanship and ideology, and positional and valence issues.
To test a variety of hypotheses in each category, the authors estimate regression models of vote choice using LAPOP data from 2008, 2010, and 2012. LAPOP typically does not poll during election campaigns, so the authors use as their dependent variable a measure of vote intention (in a hypothetical election to occur in the next few days) for the incumbent party or candidate. The authors pool respondents from 18 countries across all three years into statistical models that include various measures of the funnel of causality categories. Across the categories, the authors argue that demographic and socioeconomic factors have the weakest (but still nonzero) effects on vote intention, partisanship has the strongest effects, and ideology and issues yield effects that fall somewhere in between. The authors also report results from sets of single-country models and show how the weights of different factors vary cross-nationally.
The statistical significance of various Michigan model factors leads the authors to make their summary conclusion: “the behavior of voters in the region rests on the same foundations identified elsewhere in the world” (124). The weights of the different factors may vary across world region and countries, but the authors justify this conclusion on the grounds that the collective factors of the Michigan model account for a good deal of the individual-level variance in voting behavior. At the same time, one of the values of their approach is that the authors are able to determine that the impacts of some factors, especially partisanship and ideology, are not as strong in Latin America as they are in the United States and Western Europe. The authors do consider one factor that has received much scholarly attention in Latin America but that is not part of the Michigan model—vote buying—and conclude that vote buying is not all that prevalent.
The authors are to be commended for making Latin American Elections a short and readable volume (144 pages of text). Not surprisingly, a book of this brevity glosses over some things that would have made the argument more compelling and convincing. On the theoretical side, the argumentation is not always grounded in the Latin American context and literature. The authors instinctively begin each theoretical section by talking about the effect of the variable at hand (e.g., class, ideology) on voting behavior in the United States or Western Europe. They then transition to a discussion of how this variable may matter in Latin America, but this is typically brief and somewhat bereft of relevant information about the realities and contexts in Latin America.
For example, in one such paragraph, the authors write, “There is no clear reason to expect that this [positive] relationship between income and right-leaning vote choice ought to be different in the emerging economies of Latin America” (48). In actuality, this statement overlooks the large literature on the relative absence of class voting in Latin America. As another example, the authors derive some hypotheses under the assumption that left-of-center (right-of-center) parties in Latin America espouse the same kinds of policies and stable ideologies as do left-of-center (right-of-center) parties in the United States and Western Europe. However, when it comes to redistribution and welfare states, a recurring finding in Latin American political economy is that right-leaning parties have been, in recent years, almost as likely as left-leaning ones to implement progressive social policies. All told, the theoretical arguments and hypotheses sometimes fail to fully exit the shadow of the established democracy contexts from which they are derived.
On the empirical side, some readers may not entirely agree with the authors’ interpretations of their statistical results. The authors read their coefficients as sound estimates of causal effects and provide rather little discussion of the potential for reverse causation or confounding. For example, they conclude that “partisanship is the key determinant of voting behavior in Latin America” because their measures of partisanship correlate more strongly with vote intention than do their other independent variables (120). However, they do not explore whether partisanship is the exogenous unmoved mover (in their words, an “anchor variable”) that is implied by this statement. Indeed, the correlation between respondents’ expressed sympathy for a party and their hypothetical choice for the “party or candidate of the incumbent president” borders on the tautological, especially since survey interviews typically occurred outside the context of a campaign with actual candidates.
Similarly, the authors decline to wrestle with the motivated reasoning logic that would make respondents’ evaluations of corruption, democracy, and public safety under the incumbent endogenous to hypothetical vote intention for the incumbent. Finally (and admittedly nitpicky), the authors do not adjust for the violation of statistical independence that is introduced by pooling respondents for all countries and years. Because of this, their reported standard errors may be too small.
These issues notwithstanding, the value of Latin American Elections lies in the application of a unified but also flexible framework to a set of explanations and a literature that is now swimming in hypotheses and findings. The funnel of causality could provide a means of cross-national comparison—something that has been overlooked, given the concept’s single-country origins—by allowing researchers to explain the country-level variation in weights that voters apply to different factors. It also forces subsequent scholars of voting behavior to be more explicit and careful about where causal factors fall in a complex chain of events. (Given the availability of panel data from the region, this could be a fruitful area of subsequent research.) As they put it, the authors have taken Michigan south, providing a model of theoretical unity to the study of voting behavior in Latin America.
