The use of analogies is a standard analytical and rhetorical procedure in philosophical and scientific (where models are often presented as analogies) reasoning. The analytical power of analogies lies in the fact that they can convey information about the cause for or the nature of a not-well-understood/known phenomenon by drawing on our more disseminated and accessible knowledge about a different but still under some aspects similar one. If you grant that A and B are two phenomena sharing (for instance) a causal structure C, then if you understand how C leads to A, then you understand how C leads to B. Rhetorically, analogies tend to make a strong impression because if the postulated similarity between A and B is plausible at the surface, then you will be inclined to believe that whatever explains A should also explain B. In general, it requires a cognitive effort to assess the degree of similarity between two phenomena, and analogies are therefore a good heuristic shortcut to explain the world and to convince others.
However, analogies tend to lose their strength once you start to scratch the surface. This is because while two phenomena may appear to be similar, differences will become more salient as the examination proceeds more thoroughly. This is obviously so for historical events that are by definition unique (which explains why analogies are rarely used in this context) but also for “nomothetic” accounts of classes of phenomena, where upon examination it is realized that the underlying laws, causal structures, or properties differ. In this sense, analogies are always limited and never fully convincing. Below, I briefly consider two instructive cases of such analogies that have been recently discussed in the literature on the relative merits of democracy and epistocracy as political regimes.
The Jury Analogy
The first example is the jury analogy. This analogy is used by Jason Brennan to claim that citizens have a right to a competent electorate.[1] The idea is relatively straightforward. Most of us would agree that we are entitled to demand that a jury, asked to make a decision that can potentially greatly interfere with our freedom, is “competent” in the relevant sense. More specifically, we would like that (i) the jury uses all the evidence that is presented; (ii) it processes this evidence and all the relevant information in an unbiased way; (iii) jurors have the mental capacity to understand the case at stake and to properly use the information available; (iv) the jury does not based its decision on prejudices and other morally problematic considerations; and (v) that its verdict is due to the fact that jurors have been bribed. In a nutshell, we have a right to have a jury that is not ignorant, irrational, impaired, immoral, or corrupt. Most of us would feel entitled to reject the legitimacy of the decision of any jury that does not fulfill these demands.[2]
The second part of the argument consists of claiming that there is a similarity (which here is ontological and moral) between the case of the jury and the case of the electorate in a democratic regime. What is true of a jury is true of an electorate. We are prima facie entitled to demand that the electorate whose decisions affect us is not ignorant, irrational, impaired, immoral, and corrupt. The analogy proceeds on the assumption that the relationship between a jury and the defendant and the relationship between an electorate and a citizen is of the same nature. Therefore, defenders and citizens have the same right to demand that decisions that affect them satisfy the same requirements. The last part of the argument is the claim that empirically it is established that democratic electorates do not meet these requirements. Therefore, citizens have a right to contest the legitimacy of democratic decisions and to demand the restriction of suffrage by disenfranchising citizens who are ignorant, irrational, impaired, immoral, and corrupt.
The argument is strong at first sight because the jury case is essentially noncontroversial and because the claimed similarity with the electorate case is plausible. In a recent article in Inquiry, Brian Kogelmann and Jeffrey Carroll convincingly show however that the analytical and rhetorical strength of the analogy is undermined once we acknowledge a key difference between the two cases.[3] In the jury case, under all plausible moral conceptions, there is a strong reason to acknowledge that the potential harm caused by an incompetent jury to the defender is significant. In criminal trials, decisions tend to be binding, except for the eventual possibility of an appeal. Mistakes entail a formidably high cost for the defendant. Moreover, the deliberation process that precedes the decision-making can accentuate rather than lessen the biases and immorality of jurors. Moreover, the cost of enforcing the right to a competent jury is relatively low. After all, few citizens are asked to participate in a jury each year, meaning that is not too costly to evaluate their competence. Excluding an individual from a jury is also not too costly for this individual. For many indeed, participating in a jury is more a burden than anything else.
Things are different with voters. The incompetence of the electorate can surely have adverse consequences. In a representative democracy, the harm caused by voters’ incompetence is however fairly indirect and largely diluted by the fact that most of the decisions are taken by representatives who indeed are competent (or more competent). Moreover, in a democracy, an electorate will have many opportunities to overturn bad decisions because elections happen on a regular basis. Finally, there is a significant moral cost for the people who are disenfranchised, at least in a society with a strong democratic culture where the right to vote is viewed as a necessary condition for political equality.
The bottom line is therefore that the jury and electorate cases are actually fairly dissimilar. This is important if you agree with many philosophers that the existence of a right depends on the costs of its enforcement and the importance of the harm its enforcement prevents. There surely is a right for a competent jury, but probably not for a competent electorate.[4]
The Market Analogy
In a very interesting podcast, Nick Cowen discusses his recent paper co-written with Aris Trantidis “Include the Ignorant.” The discussion again addresses Brennan’s and others’ criticism of democracy based on a claim of the electorate’s incompetence but provides an argument in support of democratic institutions relying on a different analogy between markets and democracy.[5] The analogy is not new and has been made by many economists, starting with Kenneth Arrow in his book Social Choice and Individual Values.[6]
Cowen and Trantidis appeal to the Hayekian account of the virtue of market competition in terms of knowledge discovery. As Hayek famously argued in his paper “The Use of Knowledge in Society,”[7] the main problem that confronts any planning system in an economy is that knowledge and information are decentralized and privately held by economic agents. To set (administrative) prices and thus determine the best allocation, the planner would need to know, inter alia, what people want, how much it costs for firms to produce goods, or what technologies are available or likely to emerge. This information is not publicly available and is difficult to access. Moreover, preferences, costs, and technologies change in the course of time and knowledge must thus be revised constantly. The task of the planner is in this sense impossible, at least on a large scale. Hayek argues that the market process is a process of knowledge discovery where the relevant information and knowledge are under an ongoing process of creation, discovery, and dissemination. Market prices play an important role as part of this process as they synthesize the information about preferences and costs and reflect marginal changes. A way to put it is that prices inform agents by incentivizing them (prices force agents to make tradeoffs and these tradeoffs are reflected in price changes) and incentivize agents by informing them (prices give information about how others value goods and this information is used to make tradeoffs).
Now, a case can be made that political choices, especially voting choices, are similar to market choices, i.e., consumption and production choices. The act of vote reveals information and knowledge about what citizens believe is good and fair. To use Arrow’s distinction between tastes and values, while market prices give information about people’s tastes, voting patterns reflect people’s values. Political programs of candidates also reveal values and beliefs about what is feasible, in the same way that prices give an indication about costs of production and technological possibilities. Without elections, it is basically impossible to obtain this information. Political knowledge is in this sense dispersed and decentralized as “economic” knowledge.
To deepen the analogy and understand its implications for the democracy/epistocracy debate, Cowen mentions the example of the market for smartphones. Consumers in this market are typically ignorant of most of the technical details of the products they buy. That does not seem however to be a significant issue for the satisfaction of consumers’ preferences. This is because firms are in competition and are incentivized to discover what people want, even if consumers themselves don’t have this knowledge yet. Firms are more generally incentivized to acquire knowledge about what could appeal to consumers and how to produce and supply products meeting wants and expectations. This work even if all actors have a very partial understanding of the technology, of the characteristics of the products, and so on. This version of the market/democracy then contends that something similar happens in a democratic regime. Voters’ political knowledge is drastically limited but politicians are incentivized to discover what voters really want and to find the best means to achieve this. In a democracy, knowledge is dynamically aggregated and made public as people (voters and politicians) constantly revise their values and expectations in light of the information they receive.
I find this argument very compelling, and any case far more convincing than the reliance on mathematical theorems (e.g., Condorcet’s jury theorem) purported to show that a democratic regime is more likely to choose the “correct” options than any other regime. We understand with Cowen and Trantidis’s analogy why disenfranchising citizens may have adverse effects by undermining the process of knowledge discovery. The problem however is that there are significant differences between the market and the democratic processes that render the analogy relatively weak in terms of analytic power. I see at least two main dissimilarities that seriously affect the value of the analogy.
First, a major difference is that consumers and firms are directly and privately affected by their choices in the market. Consumers who are not able to identify products that correspond to what they want will not be satisfied and waste money. Firms that are not able to identify the best way to produce what consumers want will be eliminated. As Anthony Downs and many others have argued, the “demand side” in democracy behaves differently because voters’ choices have virtually no marginal effect on the outcome of an election. There is therefore no relation between one’s political choice and the effects that follow the result of the election. Incentives to acquire information are weak and most voters are ignorant. The analogy is more convincing on the “supply side” because, at least under normal democratic circumstances, parties are incentivized to discover the preferences of the median voter. Overall, the process of knowledge discovery is nonetheless far less effective in a democracy than on the market. If we want to keep with the analogy, what we have is more like what happens in a market process with respect to public goods.
A second major difference is that the market process essentially solves a coordination problem. From a Hayekian perspective, the market works such as agents’ plans are made compatible, thus limiting the waste of resources. Sure, in a democracy, we want policies to be responsive to what people want. But we also want policies to be good. Unless you identify goodness with what citizens want, there is no reason to consider that the two considerations will correspond. In other words, democratic choices are partly about finding the best or the correct solution to a given problem, and the latter does not reduce to satisfying people’s preferences. That’s why epistemic democrats tend to appeal to the aforementioned mathematical theorems to demonstrate that democracy is more likely to find the correct solution than epistocracy or any other political regime.
Finally, in a sense, Cowen and Trantidis analogy is incomplete. If you really want to push for it, then you should acknowledge that the democratic setting is more like a market with serious informational asymmetries that may hinder the discovery process. Writing in 1945, Hayek did not have the benefit of relying on the results of decades of information economics and mechanism design. Market relations are essentially principal-agent relationships. In this case, it is just not true that it is impossible to improve market outcomes by well-designed “constructivist” interventions. Hayek’s main insight remains valid: economics coordination is mostly about information processing and knowledge discovery. As it happens, however, we may have the ability in specific cases to design mechanisms of preference revelation that justify not fully relying on the market’s spontaneous outcomes. The same may be true with democracy. That militates for further reflection on how to improve the institutional setup to encourage individuals to engage in deliberative and knowledge-acquisition activities. Hybrid regimes combining elements of democracy and epistocracy may be relevant in this perspective.
[1] Jason Brennan, Against Democracy (Princeton: Princeton University Press, 2016).
[2] Ibid., p. 151-4.
[3] Brian Kogelmann and Jeffrey Carroll, “There Is No Right to a Competent Electorate,” Inquiry 0, no. 0 (2024): 1–23.
[4] That there is no right for a competent electorate does not mean however that citizens do not owe each other within the practice of democratic voting, thus justifying the establishment of epistemic constraints on voting. But the argument would not proceed based on any kind of analogy. See for instance Michele Giavazzi, “Political Equality and Epistemic Constraints on Voting,” Philosophy & Public Affairs n/a, no. n/a, 2024.
[5] The paper is not yet available though Nick Cowen has kindly sent it to me. The remarks that follow are however mostly based on the talk.
[6] My colleague and co-author provides an interesting historical perspective on the market/democracy analogy (in French): Alexandre Chirat, “Démocratie de la demande versus démocratie de l’offre : reconstruction et interprétation des analogies démocratie-marché,” Œconomia. History, Methodology, Philosophy, no. 12–1 (March 1, 2022): 55–91, https://doi.org/10.4000/oeconomia.12394.
[7] F. A. Hayek, “The Use of Knowledge in Society,” The American Economic Review 35, no. 4 (1945): 519–30.
Not strictly relevant, but I had a dig at Brennan here, pointing out that he was wrong about expert opinion on the Trans Pacific Partnership. Key quote
" Brennan has chosen to illustrate his argument for requiring voters to be better informed about the issues by picking an issue on which he himself is clearly not well-informed. This is, I think, a case where ad hominem argument is entirely appropriate."
https://crookedtimber.org/2017/07/06/against-epistocracy/