In a year when voter dissatisfaction with electoral choices and outcomes has reached fever pitch, enter Christopher Achen and Larry Bartels with social-scientific evidence demonstrating that the act of voting reflects neither a forward-looking political vision held by individual voters nor a retrospective verdict on the past performance of parties and politicians. In short, people don’t seem to like or even understand the political choices they make. Democracy for Realists explains this counter-intuitive circumstance by showing that “liberalism is simply too shallow to bear the weight put on it in conventional [folk] democratic theorizing” (p. 216). Political behavioralists looking for evidence that voters vote in their rational interests or to punish leaders who fail to represent their interests have fallen into a trap: they cannot show that what democratic theorists told them ought to happen actually happens. Thus, Democracy for Realists exposes empirically a fundamental vulnerability inherent in democratic theory: many justifications for self-rule are predicated on an idea of the self, the rational individual, that is largely fictive.

In spite of the book’s title, readers are left to infer the definition of realism and what makes one a realist. I proceed based on the understanding that they use the word realism to mean that which is not rational and not individualist, specifically emotions and groups. ^{Footnote 1} Achen and Bartels are also not focused on what the implausibility of the folk theory of democracy says for citizens or for democratic theory. They want to know what this means for political scientists, and in particular, what it means for the foci and methods of behavioral political science. I proceed with the assumption that one cannot debunk the plausibility of a keystone idea in democratic theory without gesturing toward the larger implications for democracy.

Achen and Bartels recommend to political scientists two somewhat divergent approaches. First, they endorse the application of comparatively new methods for studying the psychological sources and effects of human behavior as they pertain to politics. Second, they urge a return to origins, and a reinvestment in political theory, classical sociology, and qualitative methods (p. 217).

Conceivably these psychological and sociological research methods could be combined in a way that would send political science down the path of behavioral economics. Behavioral economics considers the non-rational motivations of human behavior and from those extrapolates ways in which people’s decisions can be judiciously influenced in the direction of their rational self-interest ^{Footnote 2}

The research provides lenses through which the non-rational motivations of individuals can be identified and harnessed to achieve those rational goals. The behavioral turn has allowed economists to recover from their version of the very vulnerability Achen and Bartels describe in behavioral political science, that people are not consistently rational but their irrationality may still be both predictable and very much worth understanding. ^{Footnote 3}

However, democratic politics departs from economics, psychology, and sociology insofar as most justifications for democracy expect political subjects to be individuals who reason and choose rather than herdish groups who nearsightedly emote and identify. The behavioral turn is unproblematic for economists because recommending “libertarian paternalism” does not necessarily undermine the normative underpinnings of economics more generally. But it is another matter to suggest, as Achen and Bartels do, that liberalism is a mistake, and then forge ahead with a study of democratic procedures, whose justification is rooted in the existence of a rational individual. Liberalism and democracy are intertwined and cannot be cleanly severed from one another. Most justifications for the enfranchisement of demoi rely on the potential of individual citizens to think about and act based on some conception of the good. ^{Footnote 4} If individual citizens are granted the franchise with the expectation that they will behave rationally, but in reality they behave emotionally and herdishly, there is a mismatch between where political agency is vested (the rational individual) and where political motivations are generated (emotive identity group). This compels us to ask whether a politics produced by group affiliation and identities such as that described by Achen and Bartels can rightly be called democratic.

There are schools of thought that propose that the politics of group affiliation can be called democratic. For example, in a loose sense the politics of social class relies on an expectation that commonly shared rational interests will create group identification through class consciousness. Proponents of liberal nationalism argue that “liberals can acknowledge the importance of belonging and cultural affiliation” while “nationalists can appreciate the value of personal autonomy and individual rights.” ^{Footnote 5} And multicultural citizenship also allows a powerful political presence for competing groups, some of which can govern themselves semi-independently as long as core norms of liberal egalitarianism are met. ^{Footnote 6}

On the other hand, there are schools of thought that propose that emotions and group-based identities are unsuitable bases for political action. Utilitarianism and Rawlsian liberalism both rely on forms of reason-giving that preclude purely emotional appeals whose logic cannot be generalized.

From a distance the difference of opinion between the two may be more a matter of which emotions and affiliations motivate people’s decisions rather than a matter of whether people are motivated by emotions and affiliations. Some emotions and affiliations can be cast in terms that any member of a society could accept (e.g., what Rawls might call “reasonableness”).

If this is the case, three roads are suggested by Democracy for Realists. First, empirical scholars must catalog and classify the most evident triggers for unacceptable clannishness and emotional politics that cannot rightly be called democratic. Second, to stay true to the spirit of Achen and Bartels, realist scholars must identify procedurally democratic nudges or barriers to counteract such forms of clannishness and emotion. ^{Footnote 7} Third, normative theorists must revise democratic justificatory theses based on what can conclusively be known about political behavior, if such knowledge is agreed to be reliable. This takes Achen and Bartels on their own terms in all but one way. It puts the burden of proof on behavioralist researchers to come to consensus about the limits within which citizens act politically. One cannot deliberately do away with the basis for democracy and simultaneously lay claim to the study of democratic political behavior. Empiricists must remain attuned to the relationship of their work and normative theory.