Over the last twenty years or so, a vast academic literature has arisen, attempting to justify various paternalistic policies on the grounds that cognitive biases lead consumers to make systematic mistakes. Even before then, such policies as the War on Drugs and restrictions on smoking were justified in part by claims that shortsighted irrationality would otherwise lead people to make foolish choices they would later regret. In a notable new article, political philosophers Jason Brennan and Christopher Freiman argue that supporters of such policies must, if they wish to be consistent, also take a paternalistic approach towards voting decisions. After all, voters are prone to cognitive biases as much as consumers – perhaps even more. Here is their summary of their thesis:
Recent findings from psychology and behavioral economics suggest that we are “predictably irrational” in the pursuit of our interests.…Paternalists from both the social sciences and philosophy use these findings to defend interfering with people’s consumption choices for their own good.We should tax soda, ban cigarettes, and mandate retirement savings to make people healthier and wealthier than they would be on their own.
While there is an extensive literature arguing for paternalistic interference
with people’s consumption choices, little has been said on behalf of paternalistic interference with people’s voting choices. Brennan’s work in defense of epistocracy, for instance, focuses on the ways in which incompetent voters wrongly harm others.Our thesis is instead that the standard arguments offered in support of restricting someone’s consumption choices for their own good also imply support for restricting someone’s voting choices for their own good. Indeed, the case for paternalistic restrictions on voting choices is in many waysstrongerthan the case for restricting personal consumption choices. So, paternalists face a dilemma: either endorse less interference with consumption choices or more interference with voting choices….
We begin with a sketch of the social scientific research on cognitive bias and its effects on decision making (section I). From there we explore how this research informs recent philosophical defenses of paternalism: due to the pervasiveness of cognitive bias, paternalists claim, the state will frequently be positioned to better advance the aims of citizens than citizens themselves (section II). Next, we show that the same considerations that purportedly count in favor of paternalistic interference with citizens’ consumption choices also count in favor of paternalistic interference with citizens’ voting choices (section III). We then consider a variety of objections, including the claim that political liberties occupy a special status that shields them from coercive restriction (section IV). In closing, we acknowledge that the extent to which paternalists ought to endorse interference with the vote is an empirical question but insist that they are committed to such interference in principle (section V).
Brennan and Freiman do not necessarily support paternalistic constraints on either voters or consumers. Their point is that consistency requires those who support the former to also back the latter – or at the very least be open to it.
The form of voter paternalism Brennan and Freiman seem most partial to is Brennan’s own theory of “epistocracy,” outlined in his important book Against Democracy. The idea is to give better-informed members of the electorate extra influence relative to others. I have expressed various reservations about epistocracy here and here.
But, as the authors recognize, there are also many other ways to implement paternalism in the voting sphere. For example, we can give greater power over government policy to scientists and other bureaucratic experts, or use “sortition” to transfer some decisions to randomly selected subsets of voters who are then given incentives to become better-informed about the issues at stake.
We actually already have some paternalistic constraints on voting. For example, we deny the vote to children, largely because we think they are too ignorant and otherwise incompetent to be good voters. The same goes for denying the franchise to some of the mentally ill (as many states do). Immigrants are not allowed to become citizens with voting rights unless they pass a civics test most native-born Americans would fail.
We also, of course, turn over control of many public policy decisions to experts at least partly insulated from electoral constraints. Among the most significant examples is the Federal Reserve’s control over monetary policy.
A consistently paternalistic approach to voting rights might systematize these restrictions, and in some cases expand them. For example, if it is acceptable to exclude immigrants who can’t pass a civics test from the franchise, why not apply the same standard to natives? If children are barred from voting because they are likely to be ignorant, irresponsible, or immature, why not take the same approach with comparably ignorant and immature adults? If monetary policy is too important and too complicated to be entrusted to officials directly accountable to voters, perhaps the same is true of other areas of policy.
Brennan and Freiman explain why voter ignorance and bias are actually likely to be much greater than that of consumers:
[T]he assumption of voter competence is even more doubtful than the assumption of consumer competence.A priori, we would expect that every flaw in consumers to be worse in voters because the expected cost of an uninformed and biased consumption choice is higher than an uninformed and biased voting choice. A consumer bears most of the cost of their decision to smoke.
But unlike consumers, voters never have unilateral decision-making power. Their votes are thrown in with everyone else’s. Except in very tight elections, how individual voters vote (or whether they vote at all) has almost no effect; the expected utility of voting one way is the same as voting the other….
A massive body of evidence, collected over seventy years, indicates that the majority of voters are uninformed. We will spare you the details, but voters tend to be ignorant of political matters ranging from their local representative, which party controls Congress, or changes in economic performance, to changes in social indicators such as unemployment, recent changes in legislation, or the branches of government. They are not simply ignorant; rather, voters many have systematically mistaken beliefs about both basic political facts as well as basic social-scientific issues.
Voters, like consumers, are also subject to a variety of biases. Some biases are the same as those at play in the marketplace. Take motivated reasoning. Plenty of studies show that political partisans are selectively skeptical—they will accept evidence that confirms their preexisting policy commitments and reject evidence that threatens them. Just as a consumer may be motivated to rationalize their preference for an expensive luxury car, voters are motivated to rationalize their preference for the platform of their favorite party. So even when they are presented with relevant information, these voters will not update their beliefs appropriately….
Experimental evidence suggests that both political leaders and ordinary citizens are much more biased in their evaluation of political information than consumer information. Political ignorance is also likely to be deeper and more severe than consumer ignorance (I cover some of the relevant evidence in my book Democracy and Political Ignorance). To the extent that is true, the case for voting paternalism is actually stronger than that for consumer paternalism.
Brennan and Freiman effectively rebut a number of arguments suggesting that paternalism with respect to voting is inherently worse than consumer paternalism. One they do not consider is the danger that voter paternalism will be used in ways that discriminate against racial, ethnic, and other minorities.
There is, of course, a long history of invidious exclusion from the franchise on such grounds. But consumer paternalism has a comparably awful record of bias. The history of the War on Drugs is suffused with racism. Paternalistic regulation of sexual activity was for decades heavily influenced by homophobia. Perhaps we can cleanse consumer paternalism of such prejudices. But, if so, maybe we can do the same with voting paternalism. At the very least, we cannot just assume that the latter is inherently more tainted by bigotry than the former.
My own view is that the cognitive biases used to justify consumer paternalism are overblown, and the dangers of paternalistic regulation in this field often underestimated. Mario Rizzo and Glen Whitman’s recent book Escaping Paternalism: Rationality, Behavioral Economics, and Public Policy, provides a compelling exposition of both points (I reviewed it here).
For many of the same reasons as those outlined by Brennan and Freiman, I think voter ignorance and bias is a far more serious danger than consumer error. But I am skeptical of paternalistic solutions to the problem. I would instead prefer to shift more decisions to frameworks in which people can “vote with their feet,” and thereby have stronger incentives to seek out relevant information and assess it in an unbiased way. Expanding foot voting opportunities can diminish the danger of voter ignorance and bias without concentrating vast power in the hands of a small elite, and without giving the government broader authority to determine who is and is not a competent voter. This approach could also mitigate some of the partisan hatred and irrationality that currently poison the political process.
Of course, there are a variety of other possible strategies for addressing voter incompetence without resorting to paternalism. For example, some scholars are more optimistic than I am that we can greatly increase voter knowledge through education or by reforming media coverage of politics. But, as Brennan and Freiman point out, believers in the efficacy of such strategies should also be committed to using them as a substitute for consumer paternalism, as well.