Abstract
Realism in academic research is fundamentally about getting researchers outside their own heads. The key thing to realize is humility: others may see and engage the world differently from how we ourselves see and engage it. Christopher H. Achen and Larry M. Bartels’s Democracy for Realists is a compendium of evidence from the domain of democratic elections that promotes realism in this sense. Because of its breadth and ambition, it will command an audience that crosses multiple subfields in political science.
Yet empirical data alone is necessary rather than sufficient for political realism. Also required, arguably, are exercises of imagination. Counterfactuals and hypotheticals are crucial to realism’s ability to navigate the complexity and opacity of human affairs. Careful study of similarities and differences in social and political processes requires imaginative research design and, in turn, drives the imagination toward new insights. On this latter front, Achen and Bartels’s efforts are less conspicuous and less fruitful.
The book’s impressive and wide-ranging analysis of empirical data features in nine body chapters, bookended by introductory and concluding chapters that are primarily conceptual and polemical. There is one chapter on congruence between voters and policies; one chapter on direct democracy in American states; two chapters on retrospective accountability in general, using American evidence; one chapter on economic voting in particular, using American evidence; one chapter on economic voting around the world during the Great Depression; and three chapters on voter psychology in the United States, with a special focus on social identities.
Readers already familiar with these empirical literatures will find in each chapter much that suits their narrow professional interests. More importantly, though, the considerable effort that Achen and Bartels give to reviewing previous research in various fields could make this book an aid to intellectual cross-fertilization. Scholars of comparative politics and political theory, for instance, will find here vital perspectives on American voters and elections. Achen and Bartels do not just summarize and synthesize others’ findings; they also engage in admirably frank polemics with several research traditions. Their targets include median-voter theory (24–27), the theory that cues and heuristics enable uninformed voters to behave rationally (38–40), Condorcet’s jury theorem (40–41), studies that find congruence between voters and policies over time (45–48), and studies of economic voting (158–74), among others.
Achen’s and Bartels’s intended theoretic contribution in this book is summed up in an early passage: “if voting behavior primarily reflects and reinforces voters’ social loyalties, it is a mistake to suppose that elections result in popular control over public policy” (4). The argument aims first to demolish what they call “the folk theory of democracy” and then to introduce a group- and identity-based theory of voting behavior to fill the gap.
Some comparativists and theorists may struggle to identify either the rationale or the target of the demolition job that occupies most of the book. After all, the “mistake” that “elections result in popular control over public policy” may be pretty well understood in their own subfields. Adam Przeworski, Bernard Manin, and Susan Stokes coedited Democracy, Accountability, and Representation in 1999, the same year as the separate publication of an essay by Przeworski on “minimalist” democracy. Many democratic theorists had moved beyond the presumption that competitive elections are adequate vehicles of popular agency even earlier. With this already healthy tradition of electoral skepticism in mind, what else is there to learn?
This kind of reaction would be both eminently understandable and a little unfair. Achen and Bartels themselves admit that their work on this project began many years ago, that no particular scholar is cited as endorsing the folk theory, and that their critique “breaks little genuinely new ground” (298). All the same, they say, many working scholars have not shed their residual attachments to idealistic assumptions about electoral democracy. At first blush, this is an astonishing claim. The fact that Philip Converse’s 1964 essay on “The Nature of Belief Systems in Mass Publics” was cited over 700 times in 2013 alone (32n) would not normally square with Achen and Bartels’s premise that electoral realism is in need of retrieval. Yet they also make some concluding observations that add up to a troubling indictment of academic research on American politics in the last half-century. Achen and Bartels judge that knowledge of latent interest groups is “not nearly enough” (citing a 1951 study as the best effort) and that the evidence of how powerful interest groups operate is “uncomfortably thin” (citing studies from 1928, 1941, and 1949 as the best efforts). If these remarks on the challenges of building their new research agenda are fair, it is not inconceivable that many of their colleagues have simply failed to notice the ruins from which the new building must rise.
Achen and Bartels, then, deserve forbearance for revisiting some old themes in democratic theory. In prestigious generalist journals of political science, quantitative studies of electoral data are routinely published that affirm the benefits of “responsiveness,” “congruence,” and “accountability” on the basis of quick and easy interpretations of the evidence. The thinly veiled normative message of these studies that “the system works” may be more common in American journals, but that is precisely the milieu in which Achen and Bartels’s professional currency is strongest. Their intervention, furthering themes recently anticipated by Kirby Goidel’s America’s Failing Experiment (2013) and Bruce Cain’s Democracy More or Less (2014), is likely to have some nonnegligible impact.
At the same time, there is considerable danger lurking in this sort of critique. Though many scholars could probably benefit from embracing the same “intellectual conversion experience” (xiv) that carried Achen and Bartels on the road from idealism to realism, new converts sometimes take it too far, overshooting Damascus and ending up in the desert. The theoretic argument that emerges from Democracy for Realists is potentially misleading, incoherent, and self-defeating. It may not even be “realist,” exactly.
For example, in Chapter 2 the critique is that voting cannot enable citizens to control politicians or policy, whereas in Chapter 3 it seems to be that citizens use the vote to make poor choices (e.g., resisting the fluoridation of municipal water supplies). The implicit distinction here, between elections of personnel and referendums on policy, must reckon with the shared vulnerabilities to elite manipulation in both kinds of voting; they are beset by the same limitations on voters’ time, interest, and information, as well as the same campaigning by moneyed interests and activist cadres, and both typically (in the United States) offer only two ballot choices, neither of which necessarily reflects a rational assessment of important issues or desirable alternatives. The name-calling directed by the authors at initiative, referendum, recall, direct primaries, and term limits—“foolish reform movements” (302), “naive reformers” (321), “mind-numbing cliches about giving power to ordinary people” (327)—suggests that Achen and Bartels’s critique of the folk theory may need active reinterpretation by the attentive reader. The problem for them is less that voters are disempowered than that they are unworthy of power.
Thus the analysis of the first two-thirds of the book fortifies itself with evidence of durable pathologies in voter behavior across time and space. This kind of research design is most readily suited, of course, to a theory of essences or natures; in this case, the nature of what we might call homo elector, the human as voter. One problem with this approach, as Charles Merriam long ago pointed out—only to be dismissed as “shrill” by the authors (87)—is that it is difficult to confine the logic to some kinds of voting (say, referendums and primaries) but not others (say, general elections). Achen and Bartels mean to deal a devastating blow to idealistic visions of ordinary citizens’ conduct as voters while nonetheless asserting that universal suffrage is nonnegotiable (53)—without offering reasons to foreclose the deployment of their logic and evidence to contrary purposes. A coherent realist might believe (a) that voters deserve power and therefore must be skilled up, (b) that voters cannot be skilled up and therefore do not deserve power, or (c) that voters deserve power but voting is not empowering. Achen and Bartels must have found a fourth way, because they do not consistently defend any of these three.
Uncertainty clouds even their best moments, as when they acknowledge some oligarchic corruptions in American institutions (320–21) and then warn readers of “how far we have to go to become seriously democratic” and “how bitterly [] reforms will be opposed by the groups that profit from the current inequalities and the ensuing injustices” (322). Yet Achen and Bartels savage Charles Beard and John Dewey (22, 86) as hopeless idealists while aligning their realism with Joseph Schumpeter and William Riker (10, 23, 298), when their own sense of what counts as corruption would seem more readily addressed through the thought of the former two than of the latter two. It may be fair to ridicule proposals for changing human nature, but changing the structures and institutions within which humans operate is a different matter, perhaps deserving a greater presumption of respect. For Achen and Bartels to introduce economic inequality and the financing of federal campaigns as causes for concern after railing against state-level direct democracy for its effects on first-responder funding—with no thought given to the contribution of the federal constitution to national pathologies—is incongruous, to say the least.
To be clear, these may be puzzles and tensions rather than outright contradictions. An argument that creatively resolved such tensions would be a greater contribution than one that multiplies them without acknowledging them.
A deeper problem with Achen and Bartels’s account of homo elector is (alas) methodological. There was a time when college-level instruction in political theory invariably introduced students to Rousseau’s critique of Hobbes’s theory of human nature. Rousseau made a basic point that applies to many kinds of research: we cannot simply observe something and conclude that our observations give us true reflections of that thing’s nature; we must consider the possibility that intervening factors in the environment transform essences into outcomes. The logic behind Rousseau’s critique is classically comparative. Achen and Bartels are intent on documenting similar phenomena across time and space but show little awareness of intervening (and variable) institutional factors. Their consideration of evidence from outside the United States (41, 200–211), though welcome in itself, is limited to two of the nine empirical chapters. The authors show a special preoccupation with the 1930s in general and Nazi Germany in particular, which some readers may consider timely in a 2016 book. For the purposes of comparative analysis and democratic theory, however, this focus is tasty without being nourishing. Contemplating the rise of Hitler without mentioning the institutions of Weimar is little better than bemoaning American oligarchy (320–21) after celebrating the genius of 1787 (67, 86).
At bottom, the authors provide a powerful and salutary reminder that humans are generally only attentive to what they care about and what they can control. But those captivated by this narrow version of realism should reckon with the equally powerful proposition that the specific contents of what people care about and what they can control are not constants but variables, and that cultural and institutional factors may induce change in attitudes and behaviors. Perhaps those lamenting the irrationality of voters should spare a thought for the irrationality of ballots—whose structures and options are largely set by the very “professional politicians” that Achen and Bartels want to shield from the ravages of reform (50).
Similar considerations make Achen and Bartels’s critique of deliberative democracy (29, 302) less effective than they imagine. The standard realist response is to scoff at overpromising, and that is undoubtedly a fair reaction to much of deliberative theory. But empirical research on deliberation supplies some support for the hypothesis that ordinary citizens’ patterns of political thought and action change when formal processes and informal practices around them change. Achen and Bartels’s dismissal of the 2004 British Columbia Citizens’ Assembly as “an expensive, dramatic failure” (302) is not based on careful consideration of the subject; it is merely an academic cheap shot, and not the only one in this book. Ultimately Democracy for Realists does not set a conspicuous example for moving beyond the realism of the datum and the scoff.
Some excitement invariably accompanies a book like Achen and Bartels’s, which has been so heavily briefed in advance. Draft chapters were circulated in multidisciplinary and cross-subfield workshops at elite institutions for several years prior to publication. Multiple public presentations at professional conferences were made in the early stages of writing. These papers’ informal prestige became so compelling that peer reviewers of scholarly manuscripts sometimes demanded citation of Achen and Bartels’s unpublished material. In short, this book project became the subject of its own folk theory: it was supposed to be an authoritative statement of empirical political science on the subjects of electoral accountability and democratic theory.
Comparativists and theorists, at least, will find some morsels to relish here but not enough to sustain a lasting contribution. The importance of the questions asked and the variety of data analyzed will guarantee Democracy for Realists a large and attentive readership. Nonetheless, the shortcomings of such a thoroughly vetted and widely anticipated publication can only be discouraging. Perhaps theorists can make some difference, not only with interpreting empirical results but also with research design itself, by testing our mettle more often in collaborative projects with nontheorists. Academic political science might then be in better shape for understanding elections than elections are in for serving democracy.