Tuesday, October 1, 2019

Meritocracy as the Best Form of Government

Jason Brennan (2016) argues against democracy in significant part by relying on many empirical studies demonstrating that the public largely is politically ignorant and has various cognitive biases that make a rule by the people unreliable.  For example, most voters don’t know the platforms of the two major candidates, carry a host of false social/political facts, aren’t swayed by public deliberation, and are deficient in instrumental rationality.  Instead, Brennan posits an epistocracy, which is rule of the knowledgeable.  This view has roots in Mill.  Voting power is distributed according to competence, where for example, those who have greater epistemic virtues can have votes that carry greater weight.  An epistocracy may also be set up such that only the knowledgeable can vote.  An epistocracy can better address the political psychology findings as compared to a democracy.
As the data shows that socially advantaged groups in the U.S., such as wealthy male republican Caucasians, generally have more political knowledge than many minority groups like women and African-Americans, Brennan anticipates the Demographic Objection (DO).  DO states that for an epistocracy, advantaged groups will have more voting power over others. Thus, we will have unfair policies favoring the advantaged.  Brennan responds by writing that experiments show that minority groups are unlikely to know how to promote their own interests and that advantaged groups mostly vote for their perceived national good, which will be to the benefit of minorities.  It’s better for minorities if we restrict voting power largely to, for example, rich white republican males.  
The problem with Brennan’s epistocracy is that the perceived national good from most of the advantaged group – viz., rich white republican males – can harm relevant minority groups.  It looks like most republican voters’ perceived national good is to keep Trump in office despite discriminatory rhetoric and policies.  For example, Gallup shows that a nearly unanimous majority (91%, Sept 3-15) of Republicans approve of Trump.  This isn’t beneficial to relevant minorities but can lead to a tyranny of the “knowledgeable.”
Also, the studies showing that minorities largely have deficiencies in knowledge of politics and in promoting their own self-interests fail to measure for deficiencies regarding knowing when their own respective group is being discriminated against.  This is a key omission from such studies as discrimination can be an issue that carries overriding importance for minority groups.  Contrarily, there is data suggesting that minorities groups are aware when they are facing discrimination and know how to vote against it.  For example, a Fox News poll shows that 83% of African-Americans disapprove of Trump on race relations.  Hence, I conclude that Brennan is unable to account for DO, and his epistocracy does worse than a democracy in light of DO. Minorities should be allowed to vote and have their vote carry substantial weight.
Brennan then may claim that an epistocracy can adjust to take into account knowledge of discrimination.  However, even so, the many other aspects of political knowledge, such as economics, political science, U.S. history, and sociology, will favor wealthy Caucasians which will give them a more powerful vote.  Given their superior overall knowledge, they likely should be able to drown out the minorities.  If not, then it’s questionable whether Brennan’s view is a true epistocracy or rule of the knowledgeable.  The DO objection remains.
While Brennan’s discussed studies against democracy are formidable, I advocate a specified version of a meritocracy, or rule by the merited, as it can better handle the worries stemming from the empirical data than a democracy.  Moreover, unlike an epistocracy, it can better account for DO.  Meritocracy has its roots in Confucius and Plato.  It stands opposed to the unadulterated rule by the largely ignorant masses that should be feared with a democracy.  
With a meritocracy, in order to run for office at the national level, political candidates must 1) take relevant classes, such as economics, informal logic, political science, history, environmental science, and political philosophy, with high marks, 2) pass non-ideological tests, and 3) have experience leading in local government while scoring high on various indices such as on decreasing crime, helping underprivileged groups, and maintaining a healthy economy.  

Medical doctors must take relevant classes, such as biology, physics, and chemistry, as an undergraduate student, pass tests on these courses with an extremely high GPA, attend medical school, and must take medical entrance and board exams.  They also must take periodic tests as a medical doctor to make sure that one is current on medical advancements.  They must be in residency to gain experience.  By analogy, politicians, with the fate of many more lives on their hands, must also acquire an education and experience in political leadership and must demonstrate their virtue and merit.  If we have such requirements for a medical doctor, then how much the more we should have such requirements for those future politicians who will make decisions on a nation’s healthcare, economy, education, warfare, environment, laws, etc.  There is good reason for having requirements to be a doctor given the gravity of the job and the technical skill required.  All the reason more to have stringent criteria for being a political leader given the gravity of the job and the more diverse technical knowledge required to perform the job well.  

Those many candidates who jump through these hoops then must be elected from a democratic vote.  Of course, public fully funded universities is a prerequisite for my system so that all may have an opportunity to run for office in terms of acquiring the requisite education.  A meritocracy diverges from an epistocracy in that everyone has an equal vote.  A meritocracy focuses immediately on the virtues of officials rather than voters.  
I contend that this meritocracy does better than a democracy in that, although not foolproof, it is more likely to have virtuous officials than a democracy alone in light of the political psychology data.  It also is able to address DO better than an epistocracy in that everyone carries an equal vote.  Moreover, officials will have taken classes in ethics and must demonstrate prior successful experience working with minorities.  Hence, my meritocracy appears better than an epistocracy and democracy. Winston Churchill once said that “democracy is the worst form of government except for all those other forms…” However, this is false as my meritocracy can at least make it more likely that our leaders have merit.

John Park
Philosophy Department
Sacramento State


  1. Hi John, thank you for this interesting post. I have two questions, both of them around the idea of how do we decide what knowledge is best for a meritocracy, and how do we measure it. First, in the first paragraph you use "knowledge" and "epistemic virtues" as if they are the same. Are they? I'd say that epistemic virtues include being sensitive to things we don’t know (something José Medina has explored in his book The Epistemology of Resistance), and this doesn’t necessarily translate into having any specific knowledge. This question then turns into this other one: what is more valuable for a meritocracy: knowing (certain) things (e.g. historical data, how the congress works) or having a certain approach to knowledge and ignorance (e.g. being sensitive to things you are unfamiliar with, like other people’s realities or ways of knowing).

    The second main question is: your second paragraph says that there is empirical data from the US showing that wealthy white males know more political science than women and African-Americans. Your argument tries to accommodate this data, or rather, the worries steaming from them, as you put it. I wonder how we should evaluate those data, and whether they are indeed worrying as to justify your proposal. My reasons for doubt have to do with my first questions above: how should we interpret that knowledge of political science? Do those people also have the epistemic virtues that we should expect of epistemically responsible people? Maybe knowing a lot of political science is not such an epistemic merit, maybe being sensitive to realities other than your own should be a much more valuable epistemic virtue. I wonder what kind of knowledge those data are measuring, and whether in defining that as the paradigm of what a knowledgeable person is, we are not already setting the stage in a very specific and biased way.

    1. Saray, thanks for your post. It's been a busy week, and I'm now just getting to all the replies. I do like your distinction between knowledge and epistemic virtues. I think there's a distinction there to be made, and I'll likely incorporate it as I proceed further on this project. So, in response to your first question, I'd like to say that they're not necessarily the same thing. For example, a child may have the epistemic virtues but not yet know a lot about politics. However, one who does know a lot about politics will likely at least have a few virtues, such as inquisitiveness, but may be lacking in others, such as open-mindedness. As far as which one is more important, I think both of them are important. It's important to have political epistemic virtues such as being open-minded, inquisitive, diligent, wise, creative, etc. These will be beneficial not only in regular day affairs but also when novel political situations present themselves. However, knowing certain things are also important, such as knowledge of global warming, discriminatory US immigration policies in the 1920's, the economics of trade wars, etc.

      Regarding your second main question of interpreting all the data: I think they largely are worrisome. That most voters don't even know the platform positions of the two main party candidates is alarming. Furthermore, the general tribalism and dogmatism discovered in voters does not leave me very optimistic. There are many other weaknesses found in the general voting public that I didn't have the space to include. I think they're strong enough to provide significant prima facie support to a meritocracy over a democracy.

      As mentioned above, I think having good information does imply that they have some political epistemic virtues, but not necessarily all. It's also important to have both knowledge and the virtues. My meritocracy attempts to account for this by having non-ideological tests for knowledge, but also taking classes like ethics to acquire political epistemic virtues. Moreover, candidates at the local level to advance to the national level must show that they have worked with minority groups to better their situation. Such experience is in part there cultivate virtues, like open-mindedness, in political officials. My meritocracy attempts to increase the likelihood that candidates have political knowledge and the relevant epistemic virtues.

  2. There are two big problems with your presentation of Brennan’s epistocracy. First, you seem to present Trump as the epistocratic candidate, *rather than the 180 degree opposite*. But Brennan has used Trump as evidence of how crappy democracy is (like here https://www.chronicle.com/article/Pox-Populi/236817).

    You seem to think that the largely white, republican, male endorsement of Trump reflects the endorsement of the knowledgeable (which is hilarious on its face). But it’s important to remember that high-information voters don’t toe a party line. Studies indicate the following mix of preferences among high-information voters: freer trade, more open immigration, less war, a less punitive criminal justice system, pro-gay rights, concerns about deficit spending and about climate change...

    It’s also useful to keep in mind that Trump won the election *in our current electoral system*. There’s no reason to think that Trump would have even been a candidate in some alternative system, where parties would face different incentives about which platforms and candidates to present to the electorate. But, since voters don’t have much incentive to be informed, parties don’t have much incentive to offer us a better choice than, e.g., between Trump and Clinton.

    The second big problem is that you seem to conflate epistocracy with the restricted franchise model (and one where only white republican males get to vote). But restricted franchise is a model of epistocracy that Brennan explicitly rejects. He doesn’t recommend the rule of experts, and certainly not the rule of white republican males. He suggests a different model of epistocracy he calls “government by simulated oracle” where everyone votes. Everyone also takes a basic knowledge quiz and provides demographic information. Then, stat monkeys estimate what demographically identical but informed versions of them would have done. It’s supposed to be a way to avoid giving power to some small elite group. Think of it as “enlightened preferences voting.”

    1. Kyle, thanks for this. Yes, I think your initial comments bring out the point that I should have made clear that my critique of Brennan is strictly to his 2016 book, although my critique of him will be more encompassing when I finish the project. I think what’s written in Brennan’s 2016 work has problems with the DO objection and thus, creates too substantial a likelihood that there can be leaders who will bring seeming injustice for certain minorities groups, even though Brennan might dislike such leaders.

      Yes, there’s some data showing high-information voters have some similarities in policy preferences on a variety of things as you mention. E.g., a majority of them don’t want to increase spending on welfare for the unemployed despite the U.S. having a rather minimal welfare state especially compared to, e.g., Scandinavian countries (Gilens 2001). They generally want less government intervention on healthcare, which can involve medicaid and poor people with pre-existing conditions (Althaus 2003). If we’re going to accept such findings, notice these already have adverse effects on those in lower socio-economic classes and for significant percentages of African-Americans and Hispanics. This is right under Brennan’s nose. An epistocracy will limit their political power when they need to be at least heard out on these issues. The DO objection remains.

      Moreover, I’m reluctant to rely completely on the Althaus 2003 study I’m assuming you’re referencing that tested participants on political information before asking them about their policy preferences to find out what high-information voters think. They didn’t also run any tests on analytic reasoning skills. This is a significant aspect of political decision-making, so their results are strictly speaking invalid as results from those who pass the information and analytic tests might lead to different results.

      Brennan at times uses studies (Gilens 2012) on groups that correlate with having high-information, such as having a high income, in order to draw a more moderate inference to try and get a picture of what high-information people believe. But if we’re going to do that, then once again most Republicans approve of Trump, 54% of male college graduates voted for Trump, and 63% of white men voted for Trump (The Guardian). Those making $100,000+ were generally split between Clinton and Trump (CNN).

      Regarding the 2nd issue, there are many forms of epistocracy that Brennan discusses in his 2016 book, some of which he dislikes and several he seems ok with. He’s not particularly clear on this matter in the book. I didn’t have space to discuss all of them, so I listed two general ones. One doesn’t allow everyone to vote, but the second does with different weights depending on competence. I address both options in light of the DO objection. The first one is most important to address in light of his 2016 work since he writes on the DO objection, “My view is that rather than insist everyone vote, we should fix those underlying injustices…excluding the bottom 80% of white voters from voting might be just what poor blacks need (p. 228).” He doesn’t depend on the simulated oracle model here, which regardless, relies on an invalid inference from the Althaus study and could lead to predictions of minorities’ supposed ideal preferences that from an eyeball or philosophical test could be non-ideal for such agents, as discussed above (Also, psychological laws are ceteris paribus laws (Fodor), and it will be very difficult to make general psychological predictions of ideal preferences given differences in disparate life events for each person. Handling all of such possible variable likely is too unwieldy at least with the technology we have available today).

    2. I still don’t know why you think votes for Trump in our current system says anything about anything except our current system. I mean, you can’t just imagine everything happening as it did, except at the end change the voting rule from our democracy to some version of epistocracy, suggest that Trump would win, and say “See? Epistocracy is bad and unfair.” That’s obviously silly. Which parties exist, how many parties there are, which candidates they run, the way the parties and candidates campaign, how voters vote and how many turn out in different places, etc., are all highly dependent on the voting system in place.

      Compare: it would silly for someone to say that the Pacers should have been crowned the NBA champs because they gave up the fewest points per game. Maybe there’s a good argument for thinking this is the way the NBA champs should be decided, but if it were the way, teams would draft, trade and play differently.

      Moreover, the fact that our democracy *actually produced* a Trump presidency shows that *it* is vulnerable to a particularly acute DO. Again, it was *our democracy* that “created too substantial a likelihood that there can be leaders who will bring seeming injustice for certain minorities groups.” That’s actually the point of the passage you quoted: “You think epistocracy would give disproportionate power to the advantaged and produce unjust results? Look around (in our current system)!” Brennan doesn’t need to rely on the simulated oracle model to make that point.

      You might want to look at his recent paper that directly examines DO.

    3. It's important to keep in mind that my most immediate target of using the data from the last presidential election is to challenge the apparent unanimity in policy preferences high-information voters have. With the correlation to high information voters, the data I present at minimum puts on a burden for Brennan to have the scant studies he relies upon regarding high-info voters to be replicated in different ways. It at least puts them into question (This modifies my initial position to a degree).

      You could say that it’s possible to change the results of how groups that correlate with high information voters vote if you change various aspects of the political system, but notice this also then applies to the studies Brennan cites. Changes to how the political system is can change high-information voters’ preferences. Both sets of results are embedded within our current political system in which there’s seeming conflict between both sets. Empirical studies then are warranted to see what they really believe in an alternate desired political setting and if there’s a consensus in preferences as it is an empirical question. If this is the case, one shouldn’t endorse an epistocracy since one doesn’t know what politically “smart” people will believe in in the alternate setting.

      Some further things about the Trump data: They bring to light the possibility that what people say they believe in a study is one thing, but who they vote for could be a different thing. When people actually vote, there’s more going on in the real world than in a laboratory, such as how attractive a candidate looks or how alpha their personality is. Voting results are more ecologically valid than results in a lab and should be trusted more. What ultimately matters is how they vote. Future studies should attempt to take this into account.

      Yes, I too think democracy does not fare well in light of minorities and some of their concerns. However, I think a meritocracy, although not perfect, can do better.

    4. “Changes to how the political system is can change high-information voters’ preferences.”

      I think this is just obviously not correct, John (unless it’s their preferences for candidates). The reason why Brennan thinks that epistocratic voting rules would keep Trump-like characters out of the presidency (indeed, not even within a mile of the race) is that high-information voters have policy preferences that would make Trump a bad candidate. After all, the party wants to win, and so wants to appeal to voters who will cause a win. There’s no similar argument for thinking that different voting rules would have anything like as straightforward effect in the content of anyone’s policy preferences. The big question is “what’s an effective way to make electoral outcomes reflective of knowledgeable preferences? (because, ack, see current administration).” The thing you worry about in your third paragraph is already something political scientists know how to control for.

      Given how elections work, I don’t see how your alternative proposal helps here, either. The problem (at least before the Cheeto in Chief) isn’t that candidates are dumb and uninformed. It’s that they’re picked by a dumb and uninformed electorate. Is there anything in your proposed merit training that will make elected officials ignore the preferences of the masses who elected them? Your response to Randy makes it sound like your answer is “virtue” which is cute.

    5. I think “high-info” voters could change their minds if you change the system. For example, having lots of prominent parties could lead to previously suppressed and largely unheard ideas and ways of thinking becoming more prominent and changing peoples’ minds for good or for ill. However, to reiterate what I said before, ultimately it’s an empirical question what people will believe in such an alternate system. Data needs to be collected. It’s common for psychological findings to be very surprising, so I’ll avoid spending more time on speculation. Hence, given that it’s an empirical question, at this time, one shouldn’t endorse an epistocracy.

      One factor mentioned in my response to Randy that can motivate officials to stay on course is continuing performance competency measurements of various indices every so number of years. Also, I take it that virtues have emerged on top from the attack on their globality in the situational psychology literature. Those who have the virtues tend to manifest them in life. Also, I'm tinkering with the idea of having longer terms in office for the senate and possibly elsewhere. A weakness of having short terms is that it provides incentives to look out for the short term interests rather than the long term interests or greater good of the country. Longer terms, where elections are not frequently just around the corner, can also help motivate politicians to think of the long term interests of the state.

  3. John, I'm inclined to agree with you that some of our problems stem from the fact that many of our politicians are genuinely scientifically illiterate. You don't mention science specifically, but I hope that is a big part of what you meant.

    But I actually don't know how to bet with respect to your implicit empirical claim, viz., that by drawing our politicians from a pool of scientifically educated people we will get more intelligent legislation. An argument against this is that education of this sort does not in any way prevent politicians from saying and doing whatever is required to get elected and to remain popular with their constituents. Highly trained doctors constantly do things to patients that are known to be of no medical value because that is how they make their money or that is what the patients want. Similarly, educated politicians should be expected to support whatever policies and legislation is popular with their constituency.

    It should also be noted that the data we have so far suggest that education is positively correlated with skill at rationalization, not at actively open-minded inquiry. Training in critical thinking has never been shown to have any value at all.

    I haven't read Brennan, but I'm not attracted to implementing anything like what he proposes and I think it's fairly obvious that nothing like that is going to happen here.

    The reality is that we have a more highly educated citizenry than we ever have and this will just continue to improve as educational standards continue to rise. (Though possibly not fast enough.) The real structural problem we are facing right now is the tyranny of the minority which results from both gerrymandering and our absurd practice of treating individual states with very small populations as individuals. The amount of power that the citizens of Kentucky wield in this country right now is as staggering as it is stupid.

    1. The structural problem is mainly a result of federal law requiring single-member House districts + first-past-the-post elections. According to Duverger’s law, this is also why we get our crappy two-party system. A proportional system of representation would lead to a lot more seats for many more political parties — which is why this is never going to happen here, either.

    2. Yeah, I'm inclined to say that many, if not most of the major problems with democracy as it exists in the U.S. today could be fixed by fairly simple reforms, such as doing away with first-past-the-post voting. Making that happen politically is a nightmare, of course, but they'd be much easier than switching either to epistocracy or meritocracy.

    3. Randy, yes, I do want to include scientific knowledge, such as environmental science. Having to pass such classes with really high marks and pass tests on them can increase the likelihood of better political decision-making in relevant areas.

      I think your worry of what people will do when they get in office is a good one. They could just throw all their training out the window. While I don't claim my meritocracy is perfect and that I can fully respond to your objection, I do want to claim that such occurrences are more unlikely to happen than with a democracy alone. Having prior experience in local gov't demonstrating competence on various indices and receiving the relevant background education can make it more likely that politicians will try to do the right thing and that they already have cultivated a number of good virtues. E.g., having prior experience working successfully with minority groups can help to make certain politicians more open-minded.

      I've been wrestling with whether to put in place continuing tests and competence measurements of various indices every 10 years for politicians on the national stage. This potentially could lessen your worry as well.

      Also, my meritocracy is generally a virtue political philosophy, and it will emphasize cultivating the virtues in voters too, such as by having classes on ethics, scientific issues relevant to politics, and critical thinking early on in a child's development. Children should also have experience working on group projects with other children of different backgrounds. All of this is open to empirical scrutiny, of course. I'll look further into the issue of critical thinking and its efficacy as you mention. I wonder if there are any longitudinal studies of its possible efficacy when people learn logic at an early age and continually get it pounded into them. What if people knew the fallacies, how to draw good inferences, and recognize bad ones in basic to medium level arguments as simple as they know basic arithmetic?

    4. Yes, I agree there are other measures to shore up our political system such as eliminating gerrymandering (Randy) and first-past-the-post elections (Kyle & Garret). However, even so, I believe that a meritocracy with such changes is even better than a democracy alone.

    5. Hi John, thank you for the post. I am arriving a little late to the conversation, but I had a few thoughts on your most recent reply to the comments:

      1. Is there a reason you call your view "meritocracy" rather than "aristocracy"? In fact, both your view and Brennan's view could be reasonably described as aristocracy, i.e. rule by the best, where you identify the best according to merit and Brennan identifies them according to expertise/knowledge. Brennan seems to be trying to ground Plato's classic case against democracy in empirical data. Because of this, it seems to me, that the burden on the defender of aristocracy is to empirically show that people really can acquire the knowledge necessary for ruling. I'm not sure that case is any easier than the positive case for democracy. Maybe Aristotle was right after all that there is such thing as the wisdom of crowds.

      2. Speaking of Aristotle, and since you mention virtues, how is it that the same people who are capable of acquiring all the specialized knowledge you need for ruling are also going to acquire the virtues of character necessary to put this knowledge to good use? As Randy hinted at, we are all familiar with very intelligent persons with the intellectual virtues, but none of the moral virtues. Now, as you mention, perhaps the moral virtues can (and should) also be encouraged in the citizenry. I would agree, but traditional educational classes (e.g. in ethics) seem like the wrong prescription. These are in the realm of acquiring knowledge, not necessarily in developing the habits of good character. I would be much more concerned here with maintaining a healthy culture, social institutions, family structures, religion, and so on. No amount of education can undo the influence of vice in these areas.

      3. On critical thinking, I believe there are studies showing that even experts in logic and critical thinking, or experts in psychology, make elementary errors in reasoning and are equally prone to psychological biases. This seems to be a larger problem for meritocracy/aristocracy, which is that nobody really does have the skills required for the best form of political rule. In that case, maybe democracy is the best we can do, given the limits of human nature and human understanding. In any case, I think you and Brennan make an interesting case for shifting the burden of proof back on to the defenders of democracy.

  4. Somewhat related: here's Brennan's lecture, "Why Most Americans Shouldn't Vote," from the Ethics Center's 2014 Fall Symposium, Virtue in Politics -- https://www.youtube.com/watch?v=ZFfZAhkTDZE

  5. Thanks for the article! Totally agree. But, as with so many good philosophical ideas, this is not very practical. People would be screaming Hitler and human rights about it.