On a superficial reading, my critics and I have found little or no common ground. Public debates between economists have created the false impression that economists don’t agree about anything. In a similar vein, I believe that my debate with David Estlund, Loren Lomasky, and Jeffrey Friedman may create the false impression that the four of us don’t agree about anything either. Before I take issue with any of my commentators’ criticisms, then, I would like to take the time to point out important issues where we appear to be on the same page.
All of us agree that voters in modern democracies have a lot of false beliefs. More importantly, no one has taken refuge in the claim that these errors harmlessly cancel each other out. By implication, everyone on the panel seems to accept my claim—controversial with most economists and many political scientists—that the electorate suffers from systematically biased beliefs, and not merely random error.
Furthermore, with the possible exception of Estlund, no one on the panel disputes my substantive claims about the particular biases that the public has. I claimed that the public seriously underestimates the social benefits of the market mechanism, especially in international and labor markets, and is overly pessimistic about the past, present, and future of the economy. These are strong, controversial claims, so if any of the panelists dissented, I would have expected them to say so. (Of course, their next posts may do just that!)
If we can agree to this extent, I think we have made more progress than 95% of academic exchanges. Do I overestimate the extent of our consensus? If so, I’d like to hear about it during the next week. Let’s figure out how much we already agree before we try to show each other how wrong we are.
With that out of the way…
Reply to David Estlund
Estlund’s most important objection to my approach is that the right answers to political questions depend on a mixture of factual and moral beliefs. Economists and the well-educated may know more about economic policy, but their moral vision is narrow and unrepresentative. As Estlund explains:
It’s not that their education or economic expertise damages their competence. Rather, the demographics of expertise do not necessarily mirror the population in all other respects while simply adding on the benefits of factual and technical expertise. Those kinds of expertise often travel with other differences, such as race, class, gender, and religion…
It is precisely because of such concerns that—with the exception of religion— my statistical work controls for all of these factors. I do not naively compare the beliefs of the average economist to those of the average member of the general public. Instead, I compare average beliefs controlling for race, gender, income, job security, income growth, and more. It turns out that the raw differences are robust. Economists do not just disagree with the average American; they are almost as likely to disagree with rich, white male Americans. Scott Althaus similarly finds that political awareness generally pushes all demographic groups in the same direction. For example, he reports that better-informed individuals in all income brackets are more supportive of free-market policies.
Estlund wonders what the effect of a more restricted franchise would have been for racial justice in the 1950′s, but his doubts are ill-placed. It has long been clear to public opinion researchers that the well-educated are more tolerant than the rest of the population. Seymour Martin Lipset’s ground-breaking work on “working-class authoritarianism,” documented this by 1960. The finding that the educated are more tolerant has been amply confirmed by later research. If it were up to educated citizens, equal rights for non-whites would probably have come years earlier.
Given the time budget of this discussion, I hardly think it possible to completely change Estlund’s mind about democracy. But I am still shooting for a marginal effect. After all, Estlund tells us that “I agree with Caplan (against many) that the quality of political decisions matters, not just the process.” What additional evidence would it take to convince Estlund that he has been overestimating the quality of democratic political decisions, and underestimating the quality of the alternatives?
Reply to Loren Lomasky
Lomasky is quite right to point out the parallels between my analysis and Plato’s. But he understates how much modern social science has built on the shoulders of this—and many other—giants. To the best of my knowledge, Plato did not anticipate the most compelling objection to his whole story: the Miracle of Aggregation. No matter how low the quality of the electorate, unrestricted democracy will choose the same policies as a land governed by Philosopher-Kings as long as voter errors are random. In my work, I am not just repeating Plato’s argument. I am responding to modern social science’s more sophisticated defense of the democratic process. My focus is not on evidence that voters make mistakes, but on evidence that the Miracle of Aggregation fails because voters’ mistakes are systematic.
Of course, I do not show that all of voters’ mistakes are systematic. There are plenty of areas where the Miracle of Aggregation probably works as advertised. I point this out because Lomasky is too quick to infer that voter ignorance seriously relaxes politicians’ incentive to conform to public opinion:
A significant percentage of Americans cannot name their congressional representative. What is the likelihood, then, that they can tell you what this anonymous individual’s positions are on immigration, ethanol subsidies, Swiss cheese importation, aircraft carrier construction, or 1001 other matters on which legislatures pronounce? If voters are as intellectually maladroit as Caplan suggests, then they are incapable of mastery of their elected representatives.
It doesn’t have to work this way. In fact, it probably doesn’t. Put yourself in the shoes of a politician weighing whether to support further restrictions on immigration. You know most of the electorate has never heard of you, much less your position on immigration. But you still have a strong incentive to take the popular view, because it will make the minority of people who are paying attention more likely to vote for you. Yes, it’s just one issue. But winners don’t get to the top by saying “It’s just one issue.” They get to the top by saying “I need every vote I can get.” Unless the public is so “intellectually maladroit” that taking popular positions actually costs votes, we should expect politicians to pander.
To those who are not convinced, I present the following challenge: What existing policies are unpopular? If you test your guesses against a canonical source like the General Social Survey, it is striking how regularly the median respondent favors the policy status quo. Even in the rare cases where a majority seems to favor a reform, its preferred alternative often turns out to be simply impossible. Since no politician is able to deliver lower taxes, more spending, and a smaller deficit, for example, the system’s failure to deliver this contradictory package hardly shows that democracy is unresponsive.
In conclusion, I am puzzled that Lomasky wants “an airtight guarantee that those advertised as the Best and the Brightest are the genuine article and that, in addition, they are indelibly committed to serving the public good” before he’ll “sign up with the Plato/Caplan antidemocrats.” In real life, there are no airtight guarantees, just options that are more likely to be better than others. If economists ran trade policy, for example, there is no guarantee that we would become less protectionist. But given abundant evidence that economists are less protectionist than the general public, a less protectionist outcome is very likely. The same holds for a host of other foolish economic policies.
You can ask for more, but you won’t get it.
Reply to Jeffrey Friedman
Friedman faults me for mistaking mere voter ignorance for outright voter irrationality. It might surprise him to learn, then, that when I explain my work, most of my fellow economists initially offer his very objection. Like Friedman, the typical economist believes that ignorance is the root of all error, and resists my claim that irrationality is an important part of the story.
Fair question: Why do I make such a big deal out of this seemingly semantic issue?
To make my answer as clear as possible, let me start with the extreme case of religion.
You could say that people around the world accept the religion they were raised in simply because they are ignorant of the alternatives. You might be a Christian because you’ve never heard that there’s anything else to be.
But this is far too charitable. People know that other religions exist. But they generally dismiss the competition without consideration. Even though they make little or no effort to “comparison shop,” people tend to be highly certain that their religion is true. Most strikingly of all, if you present evidence that a person’s religion is false, he usually gets angry.
These are all symptoms of irrationality, not ignorance. If people were merely ignorant about religion, they would be agnostic, keeping an open mind until more information came in. They would not hastily conclude that the first story they heard was certainly true. And if someone gave them free information inconsistent with their current beliefs, they would be grateful, not angry.
Of course, religion isn’t the only subject that people are “religious” about. And in my experience as a teacher of economics, economics is another fine example. Most people haven’t studied economics, but that does not stop them from dogmatically embracing extreme views, and getting upset at people who question them. People who know nothing about international economics are rarely agnostic about trade policy. Instead, they confidently leap to the protectionist conclusion, and lash out at those who suggest that they might be mistaken.
Sometimes, of course, Friedman is right. Sometimes, when you explain an economic argument to a student, he says “I hadn’t thought of it that way before,” and changes his mind. But all too often, teaching economics to non-economists is like teaching evolution to Christian fundamentalists. The problem isn’t what they don’t know, but that what they know that ain’t so.
To turn Friedman’s argument around, I think that he’s the one with an unrealistic, stilted psychology that’s “vulnerable to caricature and dismissal.” Friedman seems to think that everyone wants to get the right answer—and given enough information, everyone would get the right answer. In contrast, I maintain that cognitive motives are more complex. People want more than just the truth. They also want beliefs that give meaning to their lives, and cement relationships with friends and family.
It’s not surprising, then, that people practice cognitive nepotism to shield their emotionally-charged beliefs from the facts. Of course, just as a businessman may fire his favorite nephew if he is sufficiently incompetent, people tend to become more reasonable if error is sufficiently costly.
Here is a final question for Friedman. He emphasizes that he’d “rather be ruled by open-minded ignoramuses than by doctrinaire ideologues.” If I were him, I’d like more information about how the beliefs of the ignoramuses and the ideologues differ. If we’re comparing rule by the median Russian voter in 1917 to a Bolshevik dictatorship, of course I’d prefer the former. But what if we’re comparing rule by the median American voter in 2006 to rule by a body that is markedly friendlier towards Friedman’s (and my) view that “we should rely a lot less on politics, and a lot more on markets”? The empirical evidence strongly indicates that educated Americans are one such group, and economists are another. I think that more influence for either would be an improvement. Why doesn’t Friedman agree?
 The main data set my research builds upon – the Survey of Americans and Economists on the Economy – does not contain a measure of religion. However, my colleague Lawrence Iannaccone has studied the connection between theological conservatism and support for the free market in detail, and found little evidence of a connection. [pdf] Thus, the omission of religion as a control variable is probably unimportant.
 Althaus, Scott. 2003. Collective Preferences in Democratic Politics: Opinion Surveys and the Will of the People. Cambridge: Cambridge University Press.
 Althaus (2003: 111).
 Lipset, Seymour Martin. 1960. Political Man. Garden City, NY: Doubleday.
 Nie, Norman, Jane Junn, and Kenneth Stehlik-Barry. 1996. Education and Democratic Citizenship in America. Chicago: University of Chicago Press.
 This phrase was coined in Converse, Philip. 1990. “Popular Representation and the Distribution of Information.” In Ferejohn, John, and James Kuklinski, eds. Information and Democratic Processes. Urbana and Chicago: University of Illinois Press: 369-88. See Hoffman, Tom. 1998. “Rationality Reconceived: The Mass Electorate and Democratic Theory.” Critical Review 12(4): 459-80, for an excellent discussion of the Miracle of Aggregation, Surowiecki, James. 2004. The Wisdom of Crowds. NY: Doubleday, for a high-quality popularization, and here for my class notes on the topic.
 Here’s how one of my students responded to this challenge.
 See e.g. Caplan, Bryan, and Edward Stringham. 2005. “Mises, Bastiat, Public Opinion, and Public Choice: What’s Wrong With Democracy.” Review of Political Economy 17(1): 79-105.
 I share Friedman’s view that evolutionary psychology can help advance our understanding of the patterns of economic errors. (See Rubin, Paul. 2003. “Folk Economics.” Southern Economic Journal 70(1): 157-71 for a first step in this direction). At the same time, I see no reason why an evolutionary psychologist would object to my claim that people are more likely to be rational as the private cost of error goes up.