‘Epistocracy’ is the name of a type of political power structure in which the power is held by the knowledgable—for example, by restricting the right to vote to those who can demonstrate sufficient knowledge. Though Plato and Mill defended epistocratic views, it has found few contemporary advocates. In a recent book, however, Jason Brennan argues that epistocratic power structures are capable of outperforming democratic ones. His argument is two-pronged: first, he argues that democratic procedures with universal suffrage allow poorly-informed voters to pollute the electorate, and that doing so has negative policy-related consequences that are easily avoidable. Second, he argues that voting does not possess any non-intrinsic value, and so restricting suffrage to the educated does not result in a loss of status or standing for less well-educated persons in any meaningful way. I argue that epistocratic techniques are (a) impossible to implement fairly, and (b) represent an ineffective solution for the problems they are designed to solve. On these bases, I recommend rejecting it.
Epistocracy is a type of political power structure in which the more knowledgeable or highly educated persons possess political power. The term is a neologism coined by Estlund (2008, p. 206) as a name for the kind of system Plato proposes in the Republic, and for systems that share that defining element. John Stuart Mill proposed a system with epistocratic components in Considerations on Representative Government, (Mill 2015, pp. 290–295) but it has found few contemporary defenders. However, in a pair of recent books, Jason Brennan first flirts with and then advocates the view that the right to vote should be limited along epistocratic lines.Footnote 1
There are various ways of spelling out how an epistocratic system might work. These include restricted suffrage, in which the right to vote is restricted to those who can demonstrate sufficient levels of politically relevant knowledge; plural voting, in which additional votes are granted to those who can demonstrate sufficient levels of politically relevant knowledge; epistocratic lottery, in which “pre-voters,” who are selected by lottery, can earn the right to vote by completing a training course; epistocratic veto, in which an epistocratic council, composed of those who can demonstrate sufficient levels of politically relevant knowledge, has veto power over otherwise democratically-selected policy; weighted voting, in which each vote is weighted according to the score the voter earns on an accompanying quiz (Brennan 2016, p. 15). Although Brennan does not take a firm stand about which epistocratic technique he favors, he conducts his discussion as though some form of the restricted voting scheme is the epistocratic archetype; Estlund’s discussion privileges Mill’s proposal to grant additional votes to the better-educated (Estlund does not endorse this proposal). Because Brennan’s presentation is the more thoroughgoing, and because only Brennan’s treatment of epistocracy is particularly favorable, I will presume that the epistocrat is proposing restricted suffrage unless otherwise specified.
In what follows, I will present and explain Estlund’s reconstruction of Mill’s argument for plural voting, as well as Brennan’s two-pronged argument for epistocracy. Mill’s argument is based on the idea that we should avoid giving equal political influence to persons who are unequal with regard to merit and intelligence. For Mill, this problem is exacerbated by the fact that in Victorian-era England the least-educated class was also the most numerous. The first prong of Brennan’s argument is an attempt to show that some citizens are not competent to vote, and that a prima facie case therefore exists for restricting suffrage. The second prong is an attempt to show that restricted suffrage does not deprive any individual of anything of any non-intrinsic value. The net effect is that (a) it would be better for everyone if an epistocratic regime of restricted suffrage were implemented, and that (b) doing so would not harm those who would be disenfranchised in any concrete or even merely symbolic way. This, according to Brennan, favors epistocracy and disfavors democracy.
I will then present two main criticisms of these arguments. The first is based on the fact that Brennan’s proposal does not adequately contend with the possibility that epistocratic restrictions will both intentionally and unintentionally be put to misuse; the second argues that epistocracy will not effectively resolve the problems it is designed to remedy.
I will assume but not argue for several potentially contentious principles and views. First, that we should have a government in the first place, and that it is possible somehow for that government to possess legitimate political authority, although perhaps in fact it does not. Second, that there are such things as moral and political rights. Finally, I will follow Brennan in doing non-ideal theory. The question Brennan addresses is whether we, in our non-ideal and non-idealized world, should adopt an epistocratic restriction. I will address this same question.
The Case for Epistocracy
Estlund’s reconstruction of Mill’s epistocratic argument is simple and intuitive. Mill’s view is that everyone should be granted the right to vote, but that persons who satisfy certain educational criteria, such as the possession of a university degree, be granted additional votes. Mill makes this claim against a background belief that democratic procedures can be justified in terms of the quality of the decisions they produce, but that a wiser electorate would produce better decisions. The purpose of Mill’s focus on concrete educational requirements is designed to circumvent a possible objection relating to the subjectivity and controversiality of judgments concerning wisdom—since it is controversial what wisdom consists in and it is unclear how we might measure it directly, we should instead measure it indirectly via educational criteria. It seems undeniable that persons with greater degrees of knowledge will make better, more effective decisions, which supports the view that the opinions of persons who have greater degrees of knowledge should be given greater political weight. Mill suggests that this should be accomplished via a plural voting scheme (Estlund 2008, pp. 209–11). See also (Mill 2015, chapter II, pp. 290–5).
Brennan’s argument is more developed, so we will dwell on it at some length. The first prong of his argument is an attempt to make a prima facie case in favor of epistocratic restricted suffrage based on voter incompetence. The second prong is an attempt to neutralize an obvious counterargument. For it seems obvious that the right to vote is extraordinarily valuable, and that this value outweighs virtually any potential political utility that might be gained by revoking it. Brennan must rebut this suggestion. He attempts to do so by arguing that the right to vote lacks two important kinds of value that it is widely presumed to possess: instrumental value, in that the right to vote benefits us in concrete ways; and symbolic value, in that the right to vote is a symbolic representation of full and equal dignity, autonomy, and humanity of all citizens.
Evidence of Democratic Incompetence
The argument for epistocracy begins with the claim, echoing Mill, that systems of government are tools whose function is to help their citizens lead prosperous lives, and should be evaluated in terms of their performance in this regard (Brennan 2016, p. 16). Brennan’s claim is that epistocracy would perform better than democracy. “Institutions are tools. Institutions that help us live together in peace and prosperity are good. Institutions that, compared to the alternatives, hinder us in doing so, give us little reason to support them, regardless of what they symbolize.” (Brennan 2016, p. 139).Footnote 2 This is important, according to Brennan, because it supports the view that there is a right to a competent, effective government, and this right outweighs the rights of uneducated, uninformed, or apathetic citizens to participate in the electoral process. He writes:
Democracy with unconditional universal suffrage grants political power in promiscuous ways. When [incompetent voters] vote, they exercise political power over others, and this cries out for justification. It needs to be justified against alternative systems—in particular, against epistocratic systems that try to reduce the damage [incompetent voters] might do. …Why should I be subject to the rule of [incompetent voters]? (Brennan 2016, p. 140)
Most of my fellow citizens are incompetent, ignorant, irrational, and morally unreasonable about politics. Despite that, they hold political power over me. These people can …wield the coercive power of the state against me. …They wield their power in ways that they cannot justify, and impose policies on me that they would not support if they were informed or processed political information in a rational way. (Brennan 2016, p. 142)
Additionally, Brennan stresses that the typical voter is subject to various cognitive biases and systematic patterns of error. These include confirmation bias, in which one tends to ignore evidence that disconfirms one’s prior beliefs, and tends to attend to evidence that confirms them; availability heuristic, which is a mental shortcut in which one relies on immediate examples that come readily to mind, in a way that favors recently acquired or memorable information despite the fact that such information is not necessarily particularly important; affective contagion, in which one’s emotional attitudes are influenced by the emotional attitudes of others around them—e.g. fear or anger towards racial and religious minority groups; framing effects, in which one tends to react to a particular choice or outcome differently depending on whether it is presented as a loss or a gain; among others (Brennan 2016, pp. 38, 43–8; Caplan 2007, pp. 2, 7–9, 43–8; Estlund 2008, p. 16).
Making a similar point, Estlund writes that “people have more or less systematic views about many issues. If their system is bad, so to speak, then they could easily be wrong all the time. If, for example, people in some time or place were systematically racist, or sexist, or both, it would not be surprising if their political decisions were worse than the performance of a coin flip would be on political matters involving race or sex.” (Estlund 2008, p. 16). Bryan Caplan also emphasizes that it is not simply that voters are ignorant, but that they are irrational, and that this irrationality is systematic rather than piecemeal. If the mistakes of voters were piecemeal, in that everyone was making mistakes but not generally the same ones, individual errors would not be repeated often enough to show up in the electoral results. But because the errors are systematic, the same mistakes are being made by many voters, and this leads to errors being reproduced in the electoral results (Caplan 2007, p. 7–9, passim).
For Brennan, this tendency toward systematic irrationality, as evidenced by our susceptibility to latent cognitive bias, is by far the most important source of voter error. Because the errors are systematic, various ways in which groups can make correct decisions more reliably than individuals fail to function.Footnote 3 Although he mentions voter apathy as another significant source of voter error, the systematic nature and universal scope of our latent cognitive biases makes these biases particularly dangerous.
Against the Non-intrinsic Value of Democracy
Brennan then considers whether voting or the right to vote possesses some possible nonintrinsic value that makes democracy worthwhile despite its shortcomings as an effective decision procedure. He first addresses the claim that voting is important because of its instrumental value, either in deciding the outcome of elections, or in encouraging the government to attend to one’s interests (Brennan 2016, ch. 4). Brennan denies that voting empowers us to advance our own interests in either of these ways. He suggests that whatever value an individual vote might possess must be linked with the probability that it will play a deciding role in determining the outcome of a decision, and points out that this probability is negligible in any practical application. The presence of the vast number of other voters dilutes the power of any one individual’s vote to near nothingness (Brennan 2016, p. 214).Footnote 4 Because of this, according to Brennan, individual votes are incapable of causing or motivating a government to act so as to advance the voter’s interests. There are so many voters that the government cannot be sensitive to the content of any individual’s vote, or even to its presence or absence (Brennan 2016, p. 85ff).Footnote 5
Second, Brennan claims that semiotic arguments, which trade on the symbolic value of suffrage, also fail (Brennan 2016, ch. 5).Footnote 6 According to such arguments, the right to vote is a symbolic representation of a citizen’s full and equal status as an autonomous agent with rights and dignity—suffrage expresses or symbolizes or represents this moral and civic status in the way a statue or a poem might. Brennan cites several prominent philosophers as advancing instances of this type of argument. For example, Robert Nozick argues that democratic institutions “themselves express and symbolize …our equal human dignity, our autonomy and powers of self-direction.” (Nozick 1989, p. 286) Similarly, Elizabeth Anderson writes that “pressure toward universal inclusion [in the franchise] follows from the demands of equality …whereby each adult actively recognizes everyone else’s equal authority to make claims concerning the rules under which we all shall live.” (Anderson 2009, p. 215)Footnote 7 Pablo Gilabert argues that non-democratic power-structures insult and injure the dignity of the disenfranchised (Gilabert 2012). (See also Griffin 2003, p. 120 and Estlund 2008, p. 37.) Brennan argues that this is inadequate. According to Brennan, it is an open question whether universal suffrage is the best way to symbolically recognize this equality, autonomy, and dignity, and that whatever symbolic (dis)value disenfranchisement may possess is outweighed by the value of ensuring that incompetent voters are not permitted to pollute the decision-making process with their participation. For there are ways other than universal suffrage to symbolize this human dignity, some of which would avoid the negative effects of allowing less-knowledgeable citizens to vote. He concludes that we should symbolize the equality, dignity and autonomy of such voters in those other, less destructive ways.Footnote 8
Epistocracy is Ripe for Abuse
Epistocratic restrictions on voting rights invite abusive implementation in both intentional and unintentional forms. It’s not merely that the potential for abuse exists in epistocratic systems, but that there is a high degree of certainty that if such a system were to be adopted in e.g. the United States, the country in which Brennan resides, it would be used specifically to prevent members of ethnic minority groups and other vulnerable people from voting.
One reason for thinking that epistocratic power structures would be subject to abuse is that they always have been in the past. The American South, for example, utilized epistocratic techniques, including restricted suffrage, as part of the program of racial segregation that was enacted during the period between Reconstruction and the Civil Rights Era, for the specific purpose of marginalizing Black Americans. Whenever epistocratic restricted suffrage has been implemented in the United States, it has been part and parcel of a larger concerted program whose purpose was to target specific minority groups and exploit them, ghettoize them, exclude them from public life, deprive them of a plethora of important political, civil, and due-process rights to the point where they could be murdered with impunity, and to prevent them from ever gaining the power necessary to do anything about any of this. This set of policies had a catastrophic impact for its victims, and these effects are why literacy tests and other epistocratic tools were prohibited in the United States by the Voting Rights Act of 1965. It would be a tremendous mistake to adopt a system that includes or might include epistocratic restricted suffrage that fails to acknowledge or cope with this fact.
This is particularly true in light of the fact that, in the South and elsewhere in America, contemporary governments at the state and local levels continue to use various means to limit the political power that is available to minority groups. Although epistocratic restricted suffrage was prohibited in the United States by the Voting Rights Act, the US Supreme Court’s 2013 decision in Shelby County v. Holder rendered key components of the law unenforceable.
Several states, not all of which are Southern, reacted by proposing strict voter identification laws that Federal Courts have consistently found are specifically designed, with discriminatory intent, to make it harder specifically for African American citizens to vote (Barnes and Marimow 2016; Weber 2017; Editorial Board 2017; Hajnal et al. 2017). In their ruling on North Carolina’s post-Shelby voter-ID law, the 4th circuit of the US Court of Appeals found that “the new provisions target African Americans with almost surgical precision. …Thus the asserted justifications cannot and do not conceal the State’s true motivation.” Epistocratic restricted suffrage represents a vastly more effective tool for the disenfranchisement of minority voters than voter-ID laws, which is why organizations who wish to disenfranchise minority voters generally tended to utilize epistocratic restricted suffrage for that purpose whenever it was available, and developed the use of voter-ID laws only after the epistocratic technique was prohibited by law. If epistocratic restricted suffrage were to be restored, it will be again be used as a tool of injustice by many of these same states, and will be used more effectively for that purpose than Voter-ID laws can be.
To take a specific example, Alabama’s strict voter-ID law went into effect in 2014 (Astor 2018). Early the following year, Alabama’s governor announced a plan to close most of the state’s Department of Licensing offices. These closures would have left 28 out of Alabama’s 67 counties without a Department of Licensing office, in a way that would disproportionately affect counties with high percentages of African American citizens. The closures would have left 8 of the 10 counties with the highest percentages of non-white registered voters without a Department of Licensing office. It would have left every county whose population of registered voters is at least 75% black without a Department of Licensing office. It would also have disproportionately affected counties with the highest levels of poverty.
This plan was eventually abandoned; all of Alabama’s Department of Licensing offices that were initially closed have been re-opened. But many of these offices have extremely limited hours—many are open only one or two days per month. Perry County, for example, is a rural county in central Alabama. Its demographics are: 68.7% Black, 30.3% White, 0.3% Asian, 0.2% Native American, 0.2% of some other race, 0.4% of two or more races, and 1.1% are Hispanic or Latino of any race. Its poverty rate is over 35%; by household income, it is the poorest county in Alabama. It voted for Clinton over Trump in the 2016 U.S. presidential election by a margin of 72.7–26.7%. Its Department of Licensing office is open only on the third Tuesday of each month from 8:00 AM to 2:30 PM with a half-hour lunch break at noon. That means that it is not open before the standard workday begins, or after it ends, or during the standard lunch break.Footnote 9
It should be emphasized that these attempts to enact discriminatory policies that disenfranchise minority voters are taking place in a legal environment that contains many protections and safeguards against the most blatant types of discrimination. Although some of these discriminatory policies are being overturned or weakened by the courts, some are being successfully defended and are doing significant harm (Hajnal et al. 2017). If they were permitted to exclude voters based on education level or ability to pass a test, whether it be a test for literacy (as in the Jim Crow era), high-school level civics, or college-level economics (as Brennan occasionally suggests), the harmful effects would be much worse. Although Brennan’s proposal is racially neutral on its face, there are many state and local governments in the United States that would be eager to implement it in racially non-neutral ways. Epistocratic restricted suffrage is a highly useful tool for this purpose, which is why literacy tests were prohibited in the United States by the Voting Rights Act. Brennan’s proposal contains no provisions that would serve to prevent this, or which are designed to address this possibility. Epistocratic restricted suffrage would be used by these governments as a weapon against the most vulnerable people who reside there, including ethnic minorities, women, and the poor.
In the Absence of Malice
Although there is good reason to suspect that Brennan’s proposed epistocratic restricted suffrage would be implemented in malicious ways, it would have negative effects even if this were not so. Poverty is inversely related to quality of education, is often geographically concentrated, and is correlated with racial and other demographic characteristics. This means that the effects of epistocratic restricted suffrage would disproportionately affect the poor, would serve to create entire neighborhoods that are largely disenfranchised, and would disproportionately affect vulnerable minority groups.
This is, of course, not to suggest that people who are affected by poverty or members of racial or ethnic minority groups are inherently less likely to possess or capable of gaining the relevant political knowledge. There is no in-principle reason why members of such groups could not pass an epistocratic exam. But because poverty is often geographically concentrated, and because public schools in the United States are generally organized by neighborhood, and because poverty disproportionately affects members of ethnic and racial minority groups, the public educational system in the United States is set up in such a way that members of ethnic and racial minority groups disproportionately lack access to high-quality educational opportunities, and therefore lack access to the relevant political knowledge.
For this reason, epistocratic restrictions on voting would also cause problems at the level of local government. Because poor educational outcomes are correlated with socioeconomic status, and because of the geographical concentration of poverty, epistocracy would disenfranchise entire neighborhoods at once. For example, the West Side of Tuscaloosa, Alabama is almost entirely Black, and contains neighborhoods in which, as of 2017, the median annual household income is in the $25,000 range, and in some areas is as low as $17,652 (U.S. Census Bureau 2017). The schools that serve these neighborhoods are under-funded, under-staffed, and poorly equipped, and the educational outcomes reflect this lack of resources. Of course there is the occasional student who is able to thrive academically in such an environment, but this is the exception rather than the rule. Epistocratic voting restrictions are likely to result in the disenfranchisement of almost everyone who lives in that neighborhood, which means that the interests of the neighborhood will lack representation in meetings of the city council, school board, state legislature, and US Congress. Iterations of this pattern would result in the underrepresentation of black and Hispanic citizens, and an overrepresentation of white and Asian citizens in the electorate.Footnote 10
These largely poor members of vulnerable ethnic minority groups would then have no voice in how their government operates, or on what their political and legal rights, responsibilities, and obligations are. They would have no hand in decisions regarding the schools to which they will send their children, and will have no access to the kinds of educational opportunities and resources that they would require in order to join the enfranchised class. With no influence at the local school board, the schools that serve those neighborhoods will not improve, and those neighborhoods will not gain political representation. They will become a permanent uneducated underclass.
It should be emphasized that these results do not require epistocratic techniques to be maliciously misused and abused in order to arise. Rather, these results will arise from a perfectly neutral, even-handed application of epistocratic restricted suffrage to the American electorate as it currently exists.
The Mill/Estlund proposal, according to which citizens who have achieved greater levels of education are granted additional votes, would have slightly less dramatic effects.Footnote 11 As we saw in §2, Mill’s proposal falls into the plural voting category, and as such does not take voting rights away from the less educated; it merely grants additional voting rights to the more educated. But although it would not result in the wholesale disenfranchisement of vulnerable minorities, it would nevertheless result in their comparatively reduced enfranchisement. And this reduced enfranchisement would conform to the same racial, ethnic, and socioeconomic patterns as Brennan’s proposed restricted suffrage scheme. It would therefore substantially reduce and limit the degree to which black and hispanic citizens and the poor are represented in their governments, including city council, the school board, and at the county, state, and federal levels, and increase the influence of white and Asian citizens, as well as the economically well-off, at their expense.
Brennan on the “Demographic Objection”
Brennan discusses the disenfranchisement of disadvantaged minorities in The Ethics of Voting (Brennan 2011, pp. 105–7). He writes that his “theory seems to suggest that some already disadvantaged people have an (unenforceable)Footnote 12 moral obligation to refrain from voting. On top of the mistreatment they have received from the social system, I am now claiming that their votes are a kind of pollution. How mean! Or so the objection goes.” (Brennan 2011, p. 105)Footnote 13 He goes on to consider the hypothetical case of the rich and powerful Blues and the systematically oppressed Greens, and claims that Greens ought to refrain from voting because their oppression renders them unqualified to vote correctly or responsibly.
Brennan considers an obvious counterargument: “One might argue that if the Greens vote en masse, then the Blues will be forced to reckon with them and will be forced to improve the lot of the Greens.” (Brennan 2011, p. 106) Brennan responds to this worry by claiming that
…This is true only if the Greens are sufficiently well informed, well meaning, and adequately rational in how they vote. That is, this is true only if the Greens are pretty good voters. If the Greens are bad voters, then the Blues will only need to pretend to care about the Greens’ problems but can get away with doing nothing to solve them. …if it turns out that poor minorities overwhelmingly qualify as bad voters on my theory, this does not mean my theory is wrong. Rather, it probably means that there is something wrong with our society, and we should try to fix the problem as best we can. (Brennan 2011, pp. 106–7)
This response is fallacious. Brennan correctly notes that enfranchisement of oppressed minority groups may not be sufficient to eliminate their oppression or to ameliorate the societal problems that made their oppression possible in the first place. But it does not follow from this lack of sufficiency that the enfranchisement of minority groups is not a functionally necessary condition or a useful tool for the elimination of their oppression, or for the amelioration of the societal problems that made their oppression possible. Without the robust participation of minority groups, the political process is unlikely to correctly identify the problems those groups face, nor to settle on effective solutions.
Minority participation is an important and necessary way to compel the privileged classes to fix what is wrong with society, for the privileged majority tends to benefit from what is wrong with society—indeed, the purpose of such unjust discrimination is often to serve the interests of the privileged class—and so tends to lack intrinsic motivation to set things right. Any motivation to set things right must be provided extrinsically, by the oppressed minority group itself. Brennan would have the Greens forfeit one of the most effective tools by which they might work on their own behalf to repair the injustices they are victims of, and be made simply hope that the Blues will fix it on their own initiative, and thereby benefit the Greens at their own expense despite their having no incentive to do so. As Martin Luther King, Jr. points out in Letter from Birmingham Jail, this is absurd: “we know through painful experience that freedom is never voluntarily given by the oppressor; it must be demanded by the oppressed.” Although enfranchisement is neither strictly necessary nor by itself sufficient for an oppressed minority group to end its oppression, an oppressed minority group that is enfranchised is in a better position to end its oppression than a disenfranchised one.
In Against Democracy, Brennan deals with this objection by claiming that it relies on questionable assumptions. First, he claims, it presupposes that disadvantaged voters will each vote for their own self-interest, or for the interest of the group they are a member of. But, he claims, most voters vote for what they take to be the common good. Second, this objection presupposes that disadvantaged voters know what is in their own interest, or that of the group they are a member of. Brennan claims that it is doubtful that disadvantaged voters know what outcomes would serve their interests, and such voters are also unlikely to understand the mechanisms that would bring those outcomes about. For example, it is unlikely that members of oppressed minority groups possess the background in macroeconomics necessary to understand which macroeconomic mechanisms and policies are likely to benefit their communities. If so, then disadvantaged voters are unlikely to vote for the right policies if they are permitted to vote. And third, this objection presupposes that disadvantaged voters will vote for representatives who will pursue policies that will benefit them. Brennan claims that this is false, too. Their representatives will give them what they want, not necessarily what they need (Brennan 2016, pp. 226–8).
Again, this pattern of reasoning is fallacious. It is true that the enfranchisement of poor and minority voters is not sufficient, by itself, to ensure positive outcomes for these groups, but that fact does not entail that such enfranchisement is not necessary for positive outcomes for them. It may be true that voters generally vote for what they see as the common good, but disadvantaged voters are less likely than the general population to overlook themselves and their problems when calculating what, in their view, the common good consists of. Restricting suffrage increases the probability that their needs will go accidentally unrecognized or be deliberately ignored. Furthermore, economics is a highly contentious discipline in which many disputes related to foundational and practical issues are fundamentally unresolved. It is therefore not clear that even experts in economics would be particularly effective predictors concerning the real-world implications and consequences of economic policies and practices.Footnote 14
Brennan then points out that “the fact that governments used to hide their racism beneath an epistocratic disguise does not show us that epistocratic exams are inherently objectionable.” (Brennan 2016, p. 224) But this fact does show that epistocratic restricted suffrage is in fact objectionable. For epistocratic restricted suffrage is extremely easy to use in objectionable ways, which is why its use is now prohibited by law. When one’s proposed political power structure is illegal because it has been used as a weapon against disadvantaged ethnic minority groups in the past, it is not enough to say that it is not one’s intention in proposing it that it be used as a weapon now. Rather, one must acknowledge this potential and make a sincere effort to design safety mechanisms and fail-safes that will ensure that the proposed political system cannot be weaponized. Brennan has given us no reason for thinking that epistocratic reduced suffrage would not be used as a tool of racist oppression and white supremacy if it were to be adopted in a contemporary political environment, nor has he thought to include in his proposal any kind of safeguard or countermeasure that would prevent such abuse. Brennan’s proposal instead goes out of its way to avoid taking these problems seriously.
Epistocratic Solutions are Ineffective
There is also reason to reject epistocratic restricted suffrage on the basis of its potential effectiveness. The point of epistocracy is that human individuals are significantly error-prone, and although there are various reasons why group decisions are potentially more reliable than those of individuals, whether group decisions are in fact reliable depends on the characteristics of the group. The obvious thought is that groups with higher average levels of education will make better decisions, and one can increase the average level of education in a group by expelling the least-educated members. Many people, for example, cite Condorcet’s Jury Theorem as a defense of democratic procedures: according to it, suitable groups are more likely to make correct decisions than individuals are, as long as the average person is at least slightly more likely than not to be correct. There is also an effect called the “miracle of aggregation,” which claims that if voters make random errors, then as long as a (potentially small) minority of voters is well-informed, democratic procedures will make decisions just as well as an electorate that was made up entirely of well-informed voters. This happens because the errors are random, which means that no single error is particularly common, and so no individual erroneous belief has much support. Finally, the Hong-Page theorem claims that a large and diverse group can make better decisions than a smaller, more expert group “because it can pool its diverse collective knowledge which, in the aggregate, is greater than that of the smaller group. But the participants must be genuinely diverse, agree about what the problem is and what would count as solving it, be trying to solve the problem, and be willing to learn from the others.” (Brennan 2016, p. 181ff; Somin 1998, p. 431).
However, critics such as Brennan point out that it is doubtful whether the electorate as we find it satisfies Condorcet’s applicability criteria. First, each voter must be independent—if their mistakes are systematic, then those systematic mistakes will infect the decision of the jury. The jury theorem also presumes that the voters are uniformly attempting to answer the same question. If the electorate is attempting to choose the policy that will best serve the public good, then the jury theorem predicts that a suitable group has a high probability of selecting the correct policy. But if only some members of the electorate see the election as a question of which policy will best serve the public as a whole, and various other sub-groups see it as a question of which policy will best benefit them, or their demographic group, or the demographic group for which they feel the most empathy, the jury theorem does not have this consequence. Third, the theorem doesn’t directly apply to cases where there are more than two available alternatives. Fourth, the jury theorem presumes that each voter has an independent probability of voting for the correct decision, but actual voters often fail to satisfy this independence condition. Instead, voters often vote in organized blocs, as when voters vote along party lines, or when large numbers of voters are all misled by the same unreliable media conglomerate.
Lastly, some political questions are such that it is genuinely controversial which position is correct. Maybe the question involves trade-offs between competing interests, or opportunity costs, moral values, or other similar conflicts in which there is no settled consensus as to how the competing values ought to be weighed. In such cases there is no way to determine whether the members of the electorate are sufficiently likely to choose correctly, because it is fundamentally unsettled which choice is correct. In such cases, the jury theorem is inapplicable.
Furthermore, Brennan argues that the flip-side of the jury theorem is that if most of the people are not mostly right, then juries make poorer decisions than individuals, and adding additional jurors increases the likelihood of an incorrect decision (Brennan 2016, p. 180; Estlund 2008, pp. 15–6). Similarly, he points out that the miracle of aggregation presumes that the mistakes of uninformed voters are random, and claims that they are not. Instead, uninformed voters make systematic errors and have systematic biases (Brennan 2016, pp. 177–9). Brennan argues that Hong-Page fails to apply as well; he denies that the electorate satisfies the necessary conditions of cognitive diversity, agreement about what the problems are, and willingness to learn from one another. Brennan’s argument, then, is that the actual electorate is error-prone in a non-random, systematic way that makes correct decisions less likely rather than more. He concludes that epistocracy is necessary to eliminate these systematic sources of error.
Brennan’s argument does not go through; epistocracy is not sufficient to eliminate the kinds of systematic error that Brennan mentions. For there are at least three kinds of errors that voters might make: (A) random, arbitrary falsehoods; (B) errors that result from latent cognitive biases and other systematic sources; and (C) failures to understand complicated or specialized subject matter. According to Brennan, errors of type (A) are not particularly important, and it turns out that epistocracy is not necessary to counteract these errors.Footnote 15 The miracle of aggregation tells us that democracy is sufficient to correct for type (A) errors because as long as the errors are random and unsystematic, groups make correct decisions much more reliably than individuals. For errors of type (A), epistocratic restricted suffrage and Millian plural voting are overkill. Even if less-educated people are more prone to these random, unsystematic errors than are those who are more educated, the miracle of aggregation predicts that as long as the errors are unsystematic, none of them will be popular enough to influence the outcome of an election. Instead, Brennan emphasizes that our mistakes are often systematic rather than random or arbitrary, owing to the systematic nature of our latent cognitive biases. Because many of these biases are hard-wired into us and are universally shared, they result in systematic patterns of error. That is, for Brennan, errors of type (B) are much more important.
But neither epistocracy nor Millian plural voting will correct for these type (B) errors, and Brennan’s argument to the contrary is unsound. Brennan correctly points out that these biases are systematic, and operate on us in a systematic fashion, and infers that that they result in the kind of non-random patterns of false beliefs that epistocracy can correct for. But this inference does not follow. Consider, for example, the systematic tendency to attend to evidence that serves to confirm one’s prior beliefs and to ignore evidence that serves to disconfirm them known as confirmation bias. The action of confirmation bias will result in a non-random pattern of false beliefs among a group of people only if that group’s prior false beliefs were antecedently subject to a non-random pattern, in which case confirmation bias will serve to preserve that previously-existing pattern, or the bias itself is selectively applied in a systematically selective way, in which case the pattern will be created by the action of the bias. Brennan provides no argument that either of these conditions is met. What’s more, there is no evidence that better-educated people are less subject to confirmation bias, or to other cognitive biases, than are less-educated people. Because many of these biases are hard-wired into us and are universally shared, they result in systematic patterns of error. Because these latent cognitive biases are not influenced by or correlated with one’s native intelligence or level of education, neither restricted suffrage nor plural voting schemes will correct for them.
The universal nature of latent cognitive biases compounds a major practical problem with administering epistocracy. The problem relates to the creation of fair and even-handed epistocratic exams that are politically and ideologically neutral. The authors of these exams will have to be selected, and their work monitored, with a great deal of care. But since well-educated people are no less subject to latent cognitive bias than anyone else, these biases will inevitably affect the work of those who establish the criteria for epistocratic voting restrictions.Footnote 16
More generally, cognitive biases represent a series of systematic tendencies to engage in fallacious reasoning and to draw invalid or mistaken inferences. But the specific outcomes of these tendencies depend not only on the nature of the fallacious pattern of reasoning or the form of the invalid inference, but also on the specific content of the premises used in the inference. If the public lacks a uniform set of antecedent background beliefs, then there is no reason to expect that their cognitive biases will yield a uniform set of (mistaken) output beliefs, even if everyone is uniformly subject to the same set of biases to the same degree. For example, the action of confirmation bias, however systematic, is compatible with the public possessing a diverse set of (potentially false) background assumptions that the action of confirmation bias then systematically preserves.
Similarly, the availability heuristic generates systematically similar beliefs in a group only if its members systematically favor the same recently acquired or highly memorable information in a systematically similar way. Framing effects generate systematically similar beliefs in a group only if its members are systematically subject to the same frames systematically operating on the same antecedent beliefs. There is no reason to suspect that these conditions are met. And if the public were to manifest these biases in a non-uniform way—if not everyone manifests the same bias in the same way at the same time on the same set of background beliefs—then there is even less reason to expect that the biases will yield a uniform set of output beliefs. There is therefore no reason to expect that our cognitive biases will yield a consistent or uniform set of mistaken beliefs, even if the biases themselves are universally shared and operate systematically. This undermines Brennan’s arguments that epistocratic procedures are necessary to counteract the shortcomings of democracy: the existence of a systematic set of latent cognitive biases does not entail a uniform or non-random pattern of output beliefs, and the epistocrats are no less subject to these cognitive biases than the general public. What’s more, once a certain threshold of knowledge has been crossed, additional knowledge is neither necessary nor sufficient for choosing effective, morally worthwhile policies. There is no shortage of well-educated racists, sexists, homophobes, and bigots, and any such person, however well-educated, is likely to vote poorly.Footnote 17
Additionally, Brennan’s criticisms of the jury theorem, the miracle of aggregation, and the Hong-Page theorem entail that democratic procedures, when applied to the United States’ electorate, are more-or-less incapable of yielding a correct decision. Brennan points out that the miracle of aggregation and the Hong-Page theorem both depend on the assumption that the group’s errors are random rather than systematic and that their background beliefs are diverse rather than uniform. If the errors are systematic and the background beliefs uniform, then the miracle of aggregation and Hong-Page predict that the mistakes will be compounded rather than canceled out. The jury theorem implies that if the members of the population are generally even slightly less likely to be correct than incorrect, then the probability that the group will choose correctly approaches zero, and that the probability is inversely proportional to the size of the group. But these results are far too strong—the electorate of the United States at least occasionally gets things right.Footnote 18 This suggests that Brennan’s diagnosis of the situation is incorrect, and that his conclusions about the ills of universal suffrage are not supported by the evidence he adduces.
In particular, Brennan’s contention that democratic forms of government are so ineffective that they violate the right to minimally competent government in a way that trumps the right to political participation is hard to accept. The United States in the 21st century is one of the most prosperous and most stable countries to have ever existed. Nearly all of the wealthiest countries on Earth have democratic forms of government with something that approaches universal suffrage. Over time, these governments have become more democratic; over time, they have become more respectful of the civil and political rights of their nonaristocratic classes and their ethnic and religious minorities; while at the same time becoming more economically prosperous. Perhaps these are a series of incredible coincidences, but I suspect not. But even if it is a coincidence, and these democratic nations have achieved their prosperity independently of their commitment to democratic ideals and universal suffrage, we are nevertheless left without a compelling argument for epistocracy. For Brennan’s argument is that epistocracy is necessary for the minimally competent government to which we are entitled, and the general prosperity of democratic nations indicates that, on the contrary, Brennan’s epistocratic reduced suffrage and Mill’s epistocratic plural voting are unnecessary to achieve competent governance.
This leaves type (C) errors: failures to understand complicated or specialized subject matter. There are two main ways to attempt to correct for such errors: the epistocratic way, which revokes the right to vote of the less well-educated or enhances the right to vote for the better-educated; and the democratic way, which attempts to increase the educational opportunities that are available to the type-(C) error-prone, in the hope of converting errors of type (C) into errors of type (A) as much as possible. If the citizenry, or certain segments of the citizenry, are poorly informed or uneducated in a way that interferes with (what Brennan regards as) the purpose and justifying principle of the government—its ability to “help us lead our lives together in peace and prosperity”—then the only morally defensible response is to work very hard to improve the citizenry’s access to information and educational resources. It is morally wrong to respond to this situation by revoking or diminishing the rights of citizenship from the poorly informed or poorly educated. For to respond in this way is to impose on them an unjust violation of their rights. If the situation is that uninformed voters are causing problems, epistocratic techniques are nevertheless not called-for. Epistocratic techniques are oppressive, and are therefore morally wrong.
When one makes a proposal about which political system to adopt, one should pause to ask whether the proposed system is already illegal. And if it is illegal, one should ask why it is illegal. And if it is illegal for a good reason, one should take this good reason seriously when designing the new proposal, in order to ensure that it is not the case that the new proposal should be illegal for that same good reason. Although there is no reason to suppose that Brennan intends for his proposed political system to be oppressive, he has taken no steps to ensure that his proposed system could not be used in the same way as the epistocratic restricted suffrage regime that characterized the Jim Crow-era South. Instead, he gestures toward the intrinsic/extrinsic distinction: the fact that restricted suffrage is highly susceptible to abuse does not show that it is bad in itself. Of course extrinsic badness is a kind of badness, and this response does nothing to show that epistocracy is not actually very bad. Although no system is immune to abuse, epistocratic restricted suffrage would be so conducive to such egregious abuse that implementing it fairly would be a practical impossibility.
My view is that democratic procedures with universal suffrage are the only way to create a government with legitimate authority to govern. Though I do not have space to argue for these claims here, my view is that undemocratic decision procedures are ceteris paribus less fair than democratic procedures, and that democratic procedures with less-universal suffrage are less fair than procedures in which suffrage is more universal. As I have argued above, disenfranchisement robs the disenfranchised of the only voice they might have in government, and robs them of any hand they might have in helping to decide how their local, state, and federal governments operate. And although these effects will be experienced even in the absence of malice and abuse, non-democratic procedures and non-universal suffrage systems practically invite malicious abuse. Though it is easy to react to the current political situation in the United States by losing faith in democratic procedures, my view is that to abandon democracy or universal suffrage would be a mistake that would make our government less legitimate, our procedures less fair, and would do real and lasting harm to vulnerable people.
Instead, my suggestion is that we should do whatever we can to remove any hindrances to citizens exercising their right to vote, and also to remove any hindrances to their becoming high-information voters. That means dramatically improving the quality of and access to educational opportunities, and making high-quality information about specific ballot items extremely accessible. The right way to respond to the current political moment is not to lose our faith in democracy, but to renew and deepen our commitment to it.Footnote 19
He makes a similar point in (Brennan 2011, pp. 7–8).
These include Condorcet's Jury Theorem, the miracle of aggregation, and the Hong-Page theorem. Brennan argues that the systematicity of voter error renders them inapplicable. We will discuss them in more detail in §4.
Brennan echoes this idea in (Brennan 2011, pp. 11–2), writing that “individual votes are of little instrumental value in influencing electoral outcomes or the quality of government”.
While this may be true in elections in which the electorate is large or the margin of victory wide, it is less true in close elections, or those in which the electorate is small, such as elections for positions on the local school board, city council, county commissioner, or even U.S. Congressional Representative. In such elections, the power of one’s vote is not diluted nearly to the extent that it is in, say, a U.S. Presidential election.
This is a reversal of his earlier position; Brennan advances a semiotic argument at (Schmidtz and Brennan 2010, p. 189).
It is not obvious that Anderson’s remark here is a genuine semiotic argument. There is of course a reading according to which equality demands that we use universal suffrage to symbolically recognize everyone else’s equal standing, but this is not the only reading. On another reading, universal suffrage is constitutive of, rather than symbolic of, this equal standing. The text Brennan cites does not decisively favor the semiotic reading, and elsewhere Anderson writes, “the democratic way of life realizes the universal and equal standing of the members of society, and is therefore justified as morally right.” Anderson’s claim here is not that democracy symbolizes or represents the universal and equal standing of the members of society; her claim is that democracy realizes this equality. This later text favors a non-semiotic reading of the argument.
It is also worth emphasizing that equality, dignity, and autonomy are all distinct moral values, and that recognizing someone’s equality is not identical with, nor necessary or sufficient for, recognizing their autonomy or dignity. I thank an anonymous reviewer for encouraging me to emphasize this point.
This demographic information was compiled by the United States Census, 2010. See Census.gov. For information about the Alabama Department of Licensing, see Alabama.gov.
Estlund (2008, pp. 215–9) acknowledges this.
Again, to be clear, Estlund ultimately rejects this proposal.
This way of framing the objection does not make it easy to believe that he is treating it with the seriousness it deserves.
I am grateful to an anonymous reviewer for encouraging me to emphasize this point.
Brennan’s anti-democratic argument focuses on type (B) errors.
I am grateful to an anonymous reviewer for encouraging me to stress this point.
I thank an anonymous reviewer for stressing this point.
Although not everyone will agree on precisely which decisions the of US electorate were correct, I suspect that there is a consensus that at least some of these decisions were correct (though perhaps not which ones were correct). Left-leaning liberals might point to the fact that the US elected Barack Obama to two terms as President; right-leaning conservatives might point to the fact that the US elected George W. Bush to two terms as President. Additionally, the electorate of the US has elected federal governments that enacted such pieces of legislation as the Clean Water Act, the Americans With Disabilities Act, Medicare, and the COBRA program, each of which is popular and effective.
I am grateful to Caroline Klocksiem, Seth Bordner, Hud Hudson, Marcy Lascano, Christian Lee, Rekha Nath, Jason Raibley, Neal Tognazzini, Mark Walker, Chase Wrenn, two anonymous referees for this journal, and audiences at California State University at Long Beach, the University of Alabama, and Western Washington University for helpful discussion of earlier versions of this paper.
Anderson, Elizabeth. 2009. Democracy: instrumental vs. non-instrumental value. In Contemporary debates in political philosophy, ed. Thomas Christiano and John Christman, 27–213. Malden: Wiley-Blackwell.
Astor, Maggie. 2018. Seven ways Alabama has made it harder to vote. New York Times. June 24, 2018. New York edition p. A12.
Barnes, Robert, and Marimow, Ann E. 2016. Appeals court strikes down North Carolina’s voter-ID law. Washington Post. July 29, 2016. Washingtonpost.com.
Brennan, Jason. 2011. The ethics of voting. Princeton: Princeton University Press.
Brennan, Jason. 2016. Against democracy. Princeton: Princeton University Press.
Caplan, Bryan. 2007. The myth of the rational voter. Princeton: Princeton University Press.
Editorial Board. 2017. In Alabama, federal intervention protects minority rights. Washington Post. January 7, 2017. Washingtonpost.com.
Estlund, David. 2008. Democratic authority. Princeton: Princeton University Press.
Gilabert, Pablo. 2012. Is there a human right to democracy? A response to Joshua Cohen. Revista Latinoamericana de Filosofia Politica/Latin American Journal of Political Philosophy 1: 1–37.
Griffin, Christopher. 2003. Democracy as a non-instrumentally just procedure. The Journal of Political Philosophy 11: 21–111.
Hajnal, Zoltan L., Lajevardi, Nazita, and Nielson, Lindsay. 2017. Do voter identification laws suppress minority voting? Yes. We did the research. Washington Post. February 15, 2017. Washingtonpost.com.
Mill, John Stuart. 2015. Considerations on representative Government. In On liberty, utilitarianism, and other essays, ed. Mark Philp and Frederick Rosen, 178–405. Oxford: Oxford University Press.
Nozick, Robert. 1989. The examined life. New York: Simon and Schuster.
Schmidtz, David, and Jason Brennan. 2010. A brief history of liberty. Malden: Wiley-Blackwell.
Somin, Ilya. 1998. Voter ignorance and the democratic ideal. Critical Review 12: 58–413.
U.S. Census Bureau. 2017. Median income in the past 12 Months. <https://factfinder.census.gov>.
Weber, Paul J. 2017. Texas voter ID law was designed to discriminate against minorities, judge rules. Washington Post.
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
About this article
Cite this article
Klocksiem, J. Epistocracy is a Wolf in Wolf’s Clothing. J Ethics 23, 19–36 (2019). https://doi.org/10.1007/s10892-019-09279-1
- Social philosophy