What follows is adapted from my paper, "Does the Demographic Objection to Epistocracy Succeed?"
The Demographic Objection to Epistocracy goes follows:
Most epistocratic systems give more knowledgeable citizens greater political power than less knowledgeable citizens. However, surveys of basic political knowledge routinely find that political knowledge is not shared equally by members of all demographic groups. Some demographic groups, particularly advantaged groups, tend to have more political knowledge (at least of the kind ANES, etc., study) than other, disadvantaged groups. In particular, whites tend to know more than blacks, the rich more than the poor, the employed more then the unemployed, and men more than women. Thus, the epistocratic electorate will tend to be whiter, richer, better employed, and more male than a democratic electorate. For this reason, epistocracy is unjust.
As stated, the objection leaves open precisely what makes epistocracy unjust. One version might hold that there mere fact that an epistocracy might have a voting electorate that is not demographically identical to the population as a whole is inherently unjust, regardless of what consequences follow. A second version holds that the system is likely to be unjust because disproportionate representation in the voting electorate will likely lead to outcomes that favor some groups over others.
The funny thing about this objection, though, is that is applies to real-life democracies. Yet democrats almost never realize that their arguments against other possible systems apply to their own favored system.
In modern democracies, including even in democracies with compulsory voting, the voting electorate—the people who actually vote—are not demographically identical to the eligible electorate or the citizenry of the whole. Instead, advantaged voters—such as ethnic majorities, higher income voters, men, the employed, the middle aged, etc.—generally vote at higher rates than disadvantaged voters than disadvantaged voters, even in compulsory voting regimes. De jure, a rich white person in the UK or Australia and the poor black person have the same voting power, but de facto, rich white people have more influence and power than poor blacks.
Thus, if we were genuinely bothered by the Demographic Objection, we would not just dismiss epistocracy tout court and simply assume that democracy does not also suffer from it. Instead, we would carefully compare different systems. If epistocracies are inherently unfair on the grounds that the voting electorate does not perfectly match the demographics of the population as a whole, then so are all democracies. Indeed, this is one reason why some political theorists advocate sortition—decisions via lottery—rather than democratic voting.
Two forms epistocracy—the enfranchisement lottery and enlightened preference voting—seem avoid both the Demographic Objection altogether. Or, at least, the objection is only very mild.
Consider: In enlightened-preference voting, everyone votes, and the government calculates what the voting public would want if it were fully informed, while measuring and correcting for the influence of demographic factors. (The raw voting data is made public, so that any calculations can be checked by independent newspapers, political scientists, and so on.) While this system may not perfectly correct all pernicious demographic influences, it certainly does so to a greater extent than any modern democracy, let alone any other form of epistocracy. After all, enlightened preference voting is simply a form of government based on the widely-used enlightened preferences method, which statistically determines how knowledge affects policy preferences while controlling for the effect of demographics. Political scientists already routinely use this method to assess how demographics and knowledge independently affect policy preferences. Indeed, if we wanted to know whether and how much an actual democratic voting is influenced by demographic factors, we must use the enlightened preference method to check.
In the enfranchisement lottery, the epistocratic electorate is selected at random, and incentivized (i.e., paid) to become competent. Voters are permitted to vote only if they undergo competence-building exercises.
Now, neither system is flawless from the perspective of the Demographic Objection. However, whatever the benefits or flaws of these two systems may be, they both are less flawed by the lights of the Demographic Objection than voting-based democracy is. Both government-by-simulated-oracle and the enfranchisement lottery de facto and de jure give advantaged voters less additional influence than democracies do. Ceteris paribus, as measured by the Demographic Objection, the following ranking roughly goes from better to worse:
1. Enlightened Preference Voting (epistocracy)
2. Enfranchisement lottery with competence-building exercise (epistocracy)
3. Universal suffrage with compulsory voting (democracy)
4. Universal suffrage with voluntary voting (democracy)
5. Values-only voting (arguably epistocratic or democratic)
6. Epistocratic veto (arguably epistocratic or democratic)
7. Plural voting (epistocratic)
8. Restricted suffrage (epistocratic)
That is, as we move from 1 to 8, advantaged voters get progressively more de facto (and in some case de jure) political power, while disadvantaged voters get progressively less. Once we examine the facts about voter turnout in democracy, the Demographic Objection does not on its face give us any special reason to favor democracy over epistocracy, but instead places common forms of democracy in the middle of the pack.