Rational Irrationality or Radical Ignorance?

The Myth of the Rational Voter)Caplan)We believe that Caplan has mistaken simple, inadvertent public ignorance—in this case, ignorance of economic theory—with deliberate (rational) avoidance of the truth (irrationality). In Caplan’s view, the public consciously decides to be ignorant of truths it does not want to know. “Don’t confuse me with the facts”—i.e., with the truth—is his summary of the public’s posture on economics (Caplan 2007, 102). Thus, the public’s so-called economic errors are actually perverse choices, not inadvertent mistakes.

— Stephen Earl Bennett and Jeffrey Friedman, “The Irrelevance of Economic Theory to Understanding Economic Ignorance,” Critical Review 20, 3 (2008), pp. 195–258.

In The Myth of the Rational Voter, Bryan Caplan argues that the best theory that accounts for the shortcomings of democracy — the enactment of bad policy — is his own “rational irrationality.” Rationality is defined as the search for truth, irrationality the opposite. Caplan further differentiates between different motives, where irrationality is caused by emotion and ideology, implying that truth-seeking stems from reason in the narrow sense of the term (e.g. logical deduction). In this world, there exists a tradeoff between reason and irrationality, where the latter offers good feelings at the expense of objectivity. Where the costs to a lack of objectivity are small, such as the case of voting, it becomes much more enticing to opt for a “feel good” action.1 The good feelings, I’d imagine, stem from a concern for social justice and helping fellow man — if, for instance, a voter supported raising the minimum wage out of solidarity.

Bennett and Friedman criticize Caplan’s “rational irrationality” as a contradiction in logic. They argue that Caplan’s theory must assume that the agent is knowingly embracing falsity. After all, the tradeoff is between feeling good and objectivity, implying that one knows that she is sacrificing truth. But, if the person knows that X is false, then why would she feel good about voting for it? It doesn’t make sense for the tradeoff to exist it all. I get the sense that Caplan goes back and forth between whether the voter is knowingly taking the false position or whether she genuinely believes hers is the correct one. This being said, if there’s one defense of his theory it’s that Caplan does suggest at one point that the tradeoff can be made subconsciously, where the mind literally papers over the objective truth to convince one that hers is the right belief. But, Caplan doesn’t make a detailed defense of this conjecture.

An important criticism of rational irrationality is that the person needs to actually know the truth in order to avoid it. I understand the point being made, but I’m not entirely convinced. An individual could attach an uncertain probability to her being right and a certain probability to each other alternative. She doesn’t even need to know the details of the alternative, only that there are concrete alternatives out there. A decision has to be made whether it’s worth exploring these alternatives (a decision also has to be made as to the value of exploring whether there are discrete alternatives). It’s possible that a person convince herself of being right to the point of purposefully avoiding having to discover contradictory ideas. For example, an economist might hold a strong position on business cycle theory. This economist knows that disagreement exists. But, the economist can also selectively choose which disagreements to familiarize herself with, allowing her to avoid the most challenging ones (at least, he ones she feels will be the most challenging).2 In a sense, this is still an ignorance based theory, but I think it’s close to Caplan’s argument as well.3

Truth be told, I’m more inclined towards Friedman’s alternative theory: radical ignorance. According to this theory, voters err because humans are cognitively limited. But, it’s not a rational ignorance — there’s no omniscient cost–benefit analysis to educating oneself. Rather, radical ignorance posits that there is knowledge that some agents don’t know about, making it impossible for that person to weigh the costs and benefits (how can you judge something you don’t even know exists?). It could be that economic theory tells us that protectionism harms society, but that voters aren’t aware of this theory. It could also be that voters don’t understand this theory very well — parts of it are unknown to them. Likewise, a person may not be aware of a particular application of a theory, so someone who doesn’t oppose the profit motive when it comes to her own shop may do so when it comes to big business (who, according to her, may enjoy “excess” profits).

I’ve been grappling with radical ignorance for a couple of years now. I’m not too convinced of the extreme case: complete ignorance of the unknown unknown. Instead, I think there are shades of radical ignorance. For example, it’s different to say that someone is completely ignorant of the unknown unknown and to claim that someone at least knows that there’s things she hasn’t learned yet (even if she doesn’t know what these things are). Similarly, a voter knows there are alternative positions and she has to decide to get to know them. The existence of choice doesn’t require omniscience. To me, the real value of radical ignorance theory is to force rational ignorance into the realm of realism. The individual simply does not have enough information to draw a perfect tradeoff curve, and that same individual may decide to interpret a data point in such a way that downplays the costs and benefits attached to it. The result is “non-optimal” decision making. There’s an important difference between being completely ignorant about something and just not knowing everything about everything, because the latter still allows room for a theory of choice.

Coming full circle, if there’s one thing in The Myth of the Rational Voter that can help radical ignorance theory it’s Caplan’s point on costs and incentives that impact decision-making. He argues that since in the private sphere the costs of a bad decision are internalized, it makes sense that voters don’t hold the same beliefs in everyday life (e.g. the average person who shops at the grocery store is unlikely to think that they’re being exploited for having to pay for food, yet may still be hostile towards the profit motive). He uses this to justify rational irrationality, but it can be taken in a different direction. In the market, loss forces the person to find and choose amongst alternative actions. In a democracy, this incentive may not be as strong (since the probability of costs is so low, and most of the costs are borne by others). This has to do directly with my point in the preceding paragraph: you can make a choice to explore the unknown unknown, and people do make cost–benefit analysis towards this end. Some institutional frameworks are better at forcing piecemeal changes than others.

I do think that Bennett’s and Friedman’s critique is damning. But, I think their theory of radical ignorance is stronger than it should be. If you make the requirements for radical ignorance less strict, there’s still room for Caplan’s book to provide a lot of valuable insight. The costs and benefits to voting impact the decision-making of the agent — whether ex ante or ex post (loss belonging to the latter) —; the democratic institutional framework simply doesn’t have a strong built-in process by which to punish poor decision makers. It has a weak system of forcing people to explore what’s out there and come up with alternative plans and ideas.4 Looking at it this way also helps us analyze the merits of democracy based on a dynamic model, rather than a static one that just looks at voters’ incentives ex ante.


1. Bennett, Friedman, and Caplan all accept the empirical evidence showing that the average voter votes with the social good in mind, rather than out of selfishness (narrowly defined).

2. I think that people avoid having their worldview shaken is an empirically valid possibility. This doesn’t necessarily imply the existence of rational irrationality (or they are actively choosing to believe something that they know is false), but it does suggest that people purposefully choose to discard information for the purpose of avoiding the displeasure of having your most important beliefs falsified. It’s extremely close to choosing to be ignorant.

3. It’s hard to tell whether Caplan fully unshackled himself from ignorance-based theories. He rejects rational ignorance for a variety of reasons, but he still holds that a better educated public would vote for better policy. Further, he argues that restricting voting to the best educated leads to better policy results. This contradicts his rational irrationality thesis, because these voters aren’t voting irrationality despite the fact that their vote matters as little as everybody else’s.

4. If I remember his piece, “Popper, Weber, and Hayek,” correctly, Friedman says as much, For example, p. xxiii,

In the absence of laboratory experimentation, piecemeal social engineering would, to be an error-correction process, require rigorous, informed counterfactual reasoning about social problems and the effects of putative political solutions to them.

  • http://www.facebook.com/critical.review.1 Critical Review

    Agreed, pretty much. But remember that radical ignorance doesn’t mean total ignorance, it means inadvertent ignorance.

    However, the main takeaway point of my critique of Caplan is, I think, that incentives cannot affect you if you don’t know that they exist. There is no incentive to read economic critiques of protectionism if you are totally ignorant that the critiques exist. And I do think that applies to 99% of the population, including a good proportion of non-economist intellectuals. However, there is also no incentive to read them if you’ve heard of them but have also heard that they’re based on unrealistic economic models, or are written by apologists for class exploitation, or if you’ve actually read one of them and disagreed with it (for whatever reason). These are still cases of radical ignorance because, assuming that there is a sound critique of protectionism, you remain ignorant of it or of its logic, so you have no incentive to revisit the topic or delve deeper. When you are inadvertently ignorant of the truth of X, you are necessarily ignorant of the value of knowing X, and thus of the value of learning X.

    If one reflects on that, I think one can see that *everything we don’t know* is either a case of radical ignorance or a case of resources that are inadequate to search or discovery. Now ask yourself how often you err because you just didn’t have the time or money to research something that you knew would be valuable to research, vs. how often you err because you just plain didn’t realize that something was true, because you thought it was false or had never thought about it at all or had thought about it incorrectly.

    To repeat, in the latter cases incentives can have no effect. So the upshot is either a call to drastically roll back economic imperialism (economic analysis should pretty much be banished from the study of politics, because the cases of misaligned incentives have been obvious for thousands of years and don’t require economic theory, while most of the problems are due to radical ignorance) or a call to drastically revise economics, so the obsession with explaining economic behavior solely by reference to incentives ends. For this reason I reject attempts to synthesize radical ignorance with rational ignorance or public choice, because while politics contains cases of all three, as does life in general, they are conceptually and practically distinct.

    For example, your notion that “weak incentives to explore what’s out there” lead to error assumes that voters know their votes don’t matter, such that they have a weak incentive to search. But a voter who knows his vote doesn’t matter is virtually a logical contradiction (one Caplan never even mentions: “the paradox of voting”). Awareness of the odds against a vote counting may account for the ignorance of nonvoters, but not that of voters.

    When we look at (talk to, read about in newspapers, read their political commentaries online) voters, what we actually see are people with various levels and types of “objective” information (those who skim a local paper, those who read a lawn sign, those who watch Fox, those who swear by Rush, those who devour the New York Times), *all* of whom know, in an abstract and meaningless sense, that they could, and in some Platonic sense “should,” learn more about politics and government (and economics, and medicine, and astrophysics, and everything else, literally); but all of whom, by virtue of being able to cast a non-random vote, must necessarily think that the amount and type of information they have is adequate. Only if they realized that their vote didn’t matter, and thus deliberately chose to be “ill informed” (in their own estimation), would economic analysis apply (rational ignorance theory, which your example expresses). But in that case, they would not vote.

    Am I saying that ignorant voters don’t know they’re *culpably* ignorant (i.e., not just ignorant in the abstract sense, but ignorant in a way that is likely to mean that their opinions are wrong)? Absolutely! Just talk to them!

    By the same token, I deny that anyone chooses not to read things that they think likely would change their opinion. Instead, they don’t read what they (think they) know is so biased or unreliable or extraneous or redundant that it probably would *not* change their minds, so there is no point reading it.

    So yes, choices are being made, probabilities calculated–to vote or not? to read this or that or not? But while costs and benefits are therefore being weighed, an epistemological analysis of the resulting decisions is in order, not an economic analysis, which would presuppose an “incentive to be right,” because the cost-benefit calculations *all* lead to the conclusion that, after reading this or that or not reading it, one’s opinions are correct. So everyone who votes or has an opinion about anything already thinks they are right, and therefore that they have responded to the incentive to be right.

    Lessons: 1. Opinions about the truth cannot be subjected to more than trivial incentives analysis. 2. To the extent that political action is undertaken by people who think they are right, economic analysis is irrelevant. 3. You’ve gotta choose: econ or pol sci.

    –Jeffrey Friedman

    • http://economicthought.net/blog JCatalan


      Just to clarify, I think it’s important to distinguish between incentives and their sources. The knowledge that my vote doesn’t matter may provide an ex ante incentive not to inform myself, or to not vote at all. But, even if I don’t hold this knowledge, and therefore vote, it’s conceivable to imagine a situation where an ex post loss forces me to reevaluate my position. I mean that democratic, and many other, institutional frameworks lack a process of forced error correction. If we restrict economics to the pure logic of choice I can see how it has little to say about things outside of the individual’s control. But, if we broaden economic analysis so that it can look at how humans develop institutions to better coordinate means and ends — especially considering spontaneous order — I think it has a lot to say.

      Like I wrote in response to Wladimir, I can see how my angle has little to say about how voters come to the decisions they embrace. You write that this is mostly the responsibility of psychology and cultural history, and I see how I’d be wrong to argue otherwise. But, at the same time, there’s at least two ways we can approach voting and democracy. I think I approach it as a study of democracy’s faults and the types of things that need to come into existence in order to build better political institutions. I think Caplan approaches it in a similar way, especially since he spends one of the final chapters talking about “democratic fundamentalism” — I don’t remember him being explicit, but it seems that he’s justifying his own preference for anarchism. But, of course, rather than talk just about what incentives are missing, Caplan then tries to make up (or discover) alternative incentives. Your point that he’s trying to rescue rational ignorance is on the mark. But, this doesn’t entirely preclude economic analysis from the study of politics (it just means that economists should be more careful about the tools they apply). Maybe I’m just trying to (erroneously) rescue economics, by offering it an alternative role.

      But just so that I don’t come off as overly obtuse, I accept the point that that economic theory has little to say about why people err. I also agree that The Myth of the Rational Voter tries to address an epistemological question with the wrong tools. Without having to go back and re-write this comment, maybe I’m (and Caplan, I suspect) confusing two separate realms of study: one is how voters come to hold erroneous beliefs and the other is how institutions can help force individuals correct erroneous beliefs (through signals that imply error).

  • http://www.facebook.com/wladimir.kraus.3 Wladimir Kraus

    Jon, I’m curious about this claim:

    “democratic institutional framework simply doesn’t have a strong built-in process by which to punish poor decision makers.”

    While the claim is very intuitive and hints at “tragedy of commons”-type of incentive analysis, I am wondering whether the private vs. public distinction is really the decisive one. Isn’t it rather an issue of simple vs. complex cases?

    For example, I do care about getting the best deal on soda and do spend time researching best prices, spend money trying different brands for their taste, read news about why diet soda might be harmful etc. But why I do invest time and resources? It is because their prices and my preferences for each of them are readily available and I kind of know where to look.

    On the other hand, consider some new drugs whose long-run effects are totally unknowable. Even individual patients with advanced degrees in medicine are as likely to make poor decisions as people completely ignorant of medicine.

    Similarly, people spend enormous amount of time and money learning about causes and issues that would never benefit them directly, e.g. global warming, recycling. To them, for whatever reason, the issue is *very simple* and obvious: we have to save the planet for our children’s children, make the world a better place etc.

    Maybe the key to understanding issues like voting behavior lies in something along the lines Jeff’s proposing: epistemological analysis in the context of a complex and confusing world? But whatever the alternative, the incentives-based analysis is clearly not up to the task.

    • http://economicthought.net/blog JCatalan


      I agree with you that as a subject becomes more complex the relevant person is going to suffer a higher probability of error. You make a good point that when the subject is simple it’s easier for us to relatively accurately weigh our preferences what different products, or courses of action, have to offer. But, it seems to me that this is more relevant for ex ante decision-making, or expectations formation. In the market, though, there’s the additional ex post correction, brought about by loss. Even where the subject is at its most complicated, the ex post punishment/reward mechanism, in a sense, forces the individual to keep looking for better, alternative theories on which to form expectations — loss can’t tell us the truth, but it can give us a strong signal that our current theoretical base is in some way erroneous. This is something that really draws me to radical ignorance, because it acts as a foil to show how the institutions of the market help maintain some degree of order despite the fact that 99.99… percent of the time our expectations (and therefore our knowledge) are at least partially erroneous. It forces us to keep searching for better preforming knowledge (trial and error).

      In the study of voting and democracy I can see how incentive-based analyses may not be entirely suitable. If these incentives don’t exist then they don’t offer an explanation of why agents vote the way they do. But at the same time by pointing out where incentives don’t exist it not only draws our attention to shortcomings, but it might offer answers as to how we can improve institutions of governance. In the context of Caplan’s book, his answer is not a very good one since he’s looking to explain how the public attains the ideological position it does, but thinking about incentives can help as understand why the public doesn’t attain the beliefs we think it ought to (or, more generally, why implemented policy isn’t to some standard).

  • Blue Aurora

    Jonathan Finegold Catalan, I have shown you before Dr. Michael Emmett Brady’s review of Friedman and Kraus’s 2011 book, Engineering the Financial Crisis.

    However, have you seen Dr. Michael Emmett Brady’s review of the 2008 paperback edition of Bryan Caplan’s The Myth of the Rational Voter?

    If not, here it is.