Portal de Periódicos da CAPES

THE IRRELEVANCE OF ECONOMIC THEORY TO UNDERSTANDING ECONOMIC IGNORANCE

2008; Taylor & Francis; Volume: 20; Issue: 3 Linguagem: Inglês

10.1080/08913810802503418

1933-8007

Stephen Earl Bennett, Jeffrey Friedman,

Fiscal Policies and Political Economy

ABSTRACT ABSTRACT Bryan Caplan’s The Myth of the Rational Voter treats several immensely important and understudied topics—public ignorance of economics, political ideology, and their connection to policy error—from an orthodox economic perspective whose applicability to these topics is overwhelmingly disproven by the available evidence. Moreover, Caplan adds to the traditional and largely irrelevant orthodox economic notion of rational public ignorance the claim that when voters favor counterproductive economic policies, they do so deliberately, i.e., knowingly. This leads him to assume (without any evidence) that “emotion or ideology” explain mass economic error. Straightforward, unchosen mass ignorance of economic principles—neither “rational” nor “irrational,” but simply mistaken—is a more coherent explanation for economic error, and it is backed up by the vast body of public‐opinion research. Notes 1. For another example of such complacency, see Graber 2006 Graber, Doris A. 2006. “Government by the People, for the People—Twenty‐first Century Style.”. Critical Review, 18(1–3): 167–78. [Taylor & Francis Online], [Web of Science ®] , [Google Scholar]. 2. Caplan takes Page and Shapiro’s The Rational Public (1992 Page, Benjamin I. and Shapiro, Robert Y. 1992. The Rational Public: Fifty Years of Trends in Americans’ Policy Preferences, Chicago: University of Chicago Press. [Crossref] , [Google Scholar]) to be central to his story because he vastly exaggerates the importance of the “miracle of aggregation” in that book. Page and Shapiro were merely responding to a part of the public‐opinion literature that should, strictly speaking, be segregated from the literature on public error with which Caplan is concerned. The part of the literature to which the “miracle of aggregation” was a response originated in Section VII of Converse’s “The Nature of Belief Systems in Mass Publics” ([1964 Converse, Philip E. [1964] 2006. “The Nature of Belief Systems in Mass Publics.”. Critical Review, 18(1–3): 1–74. [Google Scholar]] 2006, 44–52). In Secs. III‐VI of the paper, Converse had shown that most people are (or were, in the 1950s) so inattentive to politics that they are ignorant of the tools for organizing political information typically used by sophisticated political observers: conservative and liberal “belief systems,” i.e., ideologies. But it occurred to Converse that, however ignorant of conventional ideologies most people were, they might be using their own “idiosyncratic” ideologies, rather than conservatism or liberalism (ibid., 44), to organize political beliefs. So he devoted Sec. VII of his paper to testing this possibility. Converse reasoned that if people’s issue positions were being “constrained” by ideologies of their own devising, then even though, perforce, he could not know which specific issue positions these ideologies would dictate, any given person should hold any given issue position consistently over time. However, Converse found that this was not the case for most people. Many of their issue positions fluctuated over time, as if they were so inattentive to politics that on many issues, they did not really have political “attitudes” at all. Thus, the attitudes that people appeared to have in a one‐shot survey sometimes turned out, in panel surveys over time, to be merely random nonattitudes. The “nonattitudes thesis” became Converse’s most famous finding, and commanded a good deal of attention in the subsequent literature—but for reasons that were almost completely orthogonal to any concern one might have about public incompetence or error. As Christopher Achen (1975 Achen, Christopher H. 1975. “Mass Political Attitudes and the Survey Response.”. American Political Science Review, 69(4): 1218–31. [Crossref], [Web of Science ®] , [Google Scholar], 1227) put it, the worry raised by the nonattitudes thesis was that a certain conception of normative democratic theory would lose “its starting point” without intertemporally stable public attitudes. This conception of normative democratic theory is democratic voluntarism (Friedman 1998 Friedman, Jeffrey. 1998. “Public Ignorance and Democratic Theory.”. Critical Review, 12(4): 397–411. [Taylor & Francis Online], [Web of Science ®] , [Google Scholar]), according to which the will of the people is sovereign. Without stable public attitudes, there would be no meaningful will of the people. But in the voluntarist conception of democratic theory, there is no such thing as public error or incompetence—except in executing the public’s (self‐legitimating) will. That is, a sovereign may err by appointing deputies who fail to carry out its will. But in the voluntarist view, the sovereign’s will itself, or, in the present context, the public’s “policy preferences” themselves, cannot be mistaken: that would contradict the notion of popular sovereignty. Vox populi, vox dei. Page and Shapiro had data showing intertemporally stable aggregate opinion. Thus, it seemed that individual fluctuations such as those limned by Converse in Section VII were random noise—and that at the aggregate level, one could hear, through the static, a public will. The “miracle of aggregation” answered the nonattitudes thesis by showing that, over time and in the aggregate, the “public will” did exist. But Page and Shapiro had bigger fish to fry: They wanted to know whether the public will, however existent and minimally “rational,” was also wise, i.e., “sensible.” And they certainly did not assume that they could infer this from the “miracle of aggregation.” The “miracle” is a deus ex machina brought in to answer a question—the non‐“existence” of a public will—that had been raised by a section of Converse’s paper that was interesting only to the extent that one assumes that the public will can do no wrong. To ask whether the public is sensible is to contradict that voluntarist premise. Caplan commendably rejects voluntarism, as Page and Shapiro did. He, too, has bigger fish to fry than whether the public will merely “exists.” But Caplan is so caught up in the economists’ theory of rational ignorance that he neglects the bulk of Page and Shapiro’s book, and the bulk of the public‐opinion literature. Rational‐ignorance theory makes a point prediction of zero public “information” about politics (just as the underlying theory of voting predicts that nobody will vote). Therefore, the “miracle of aggregation” seems, to Caplan, supremely important, since it shows that the random opinions that would result from zero information could be offset by a few well‐informed voters. Therefore, if Caplan can show that public opinion is nonrandom, he has restored at least the logical possibility of systemic error to rational‐choice analysis of public opinion—by disproving a “miracle” that is so remote from the realities of politics that political scientists pay it little heed. Scholars such as Page and Shapiro, who were intimately familiar with the realities of nonrandom public opinion, considered the “miracle” relatively unimportant, except as a response to an orthogonal branch of the literature, and went about arguing that despite the nonrandomness that Caplan takes to be so crucial, accurate media signals could render public opinion sensible. Caplan, failing to notice the context of the “miracle,” does not even attempt to rebut Page and Shapiro’s other grounds for confidence in the wisdom of public opinion. Had he done so, however, he might have had to consider whether reality is so transparent to the media as Page and Shapiro tended to think. And in that case, he might have had to reconsider his own epistemology, in which knowledge of the truth falls directly into voters’ laps, such that “lack of mental discipline, not lack of information” (Caplan 2007 Caplan, Bryan. 2007. The Myth of the Rational Voter: Why Democracies Choose Bad Policies, Princeton: Princeton University Press. [Google Scholar], 99) must explain voters’ errors. 3. Surveys over the years have indicated that the public consistently overestimates businesses’ profit margins (Lipset and Schneider 1987 Lipset, Seymour Martin and Schneider, William. 1987. The Confidence Gap: Business, Labor, and Government in the Public Mind , rev. ed., Baltimore: Johns Hopkins University Press. [Google Scholar], 176–78). Data from merged polls conducted for the Times Mirror/Pew Research Center for the People and the Press between 1987 and 2007 (Pew Research Center 2007b Pew Research Center. 2007b. Political Landscape More Favorable to Democrats: Trends in Political Values and Core Attitudes: 1987–2007, Washington, D.C: Pew Research Center for the People and the Press. 22 March [Google Scholar]) reveal that over the twenty‐year span, the percentage of the public believing that businesses make too much profit has exceeded that demurring by a thirty‐point margin (62 percent agreeing vs. 32 percent disagreeing). On the other hand, the mix of attitudes about whether businesses strike a fair balance between profit and serving “the public interest” has been closer (42 percent agreeing vs. 54 percent disagreeing). If these results seem inconsistent, perhaps it is because, as Seymour Martin Lipset and William Schneider (1987 Lipset, Seymour Martin and Schneider, William. 1987. The Confidence Gap: Business, Labor, and Government in the Public Mind , rev. ed., Baltimore: Johns Hopkins University Press. [Google Scholar], 178) noted, “most people do not have a very precise understanding of the term ‘profit.’” 4. Caplan rebuts two weak counterarguments: that economists are biased because they are rich and because they are politically conservative. He does not address the real problem, which is that economists are human beings. Whatever processes of “emotion or ideology,” i.e., psychology or culture, might bias an ordinary member of the public might also bias an economics teacher, hence an economics student, hence an economics Ph.D. Caplan seems to view “education” as the transference of truth from teacher to student—begging the question of how the teacher acquired this “truth.” Thus, the fact that “educated Americans in general” tend to agree more with economists than uneducated Americans do (Caplan 2006, 83) does not impress us, any more than we are impressed by the fact that even Soviet engineers who studied no economics formally would have undoubtedly tended to agree more with the labor theory of value than American engineers would have. “Perhaps education just increases exposure to brainwashing,” Caplan writes (ibid.). Yes: that is precisely the point about formal and cultural education. So we are at a loss to understand how he can conclude, without argument, that the “more likely” alternative is that educated people “think clearer [sic] and know more” (ibid.). (Knowing more is not at all inconsistent with having been “brainwashed.”) 5. Philip E. Tetlock’s Expert Political Judgment (2005 Tetlock, Philip E. 2005. Expert Political Judgment: How Good Is It? How Can We Know?, Princeton: Princeton University Press. [Google Scholar]) demonstrates that academic social‐science experts are no better at making accurate predictions than “dart‐throwing chimps.” 6. This was the point that, eventually, the first distinctively “Austrian” economists, Ludwig von Mises and F. A. Hayek, realized that they were making in the course of a scholarly debate over central economic planning that raged in the 1920s and 1930s: How would the central planners know what they needed to know? Mises and Hayek had considered themselves to be standard neoclassical economists, but their interlocutors were standard neoclassical economists who simply assumed that the central planners would know what they needed to know (Lavoie 1985 Lavoie, Don. 1985. Rivalry and Central Planning: The Socialist Calculation Debate Reconsidered, Cambridge: Cambridge University Press. [Google Scholar]). This gave Mises and Hayek the first hint that their own version of neoclassical economics was no longer “mainstream” (Boettke 1997 Boettke, Peter J. 1997. “Where Did Economics Go Wrong? Equilibrium as a Flight from Reality.”. Critical Review, 11(1): 11–64. [Taylor & Francis Online], [Web of Science ®] , [Google Scholar])—because it took epistemology too seriously. However, Mises and Hayek gave a hostage to fortune by making “market prices” their answer to the question of how participants in market economies would “know” what central planners, bereft of market prices, could not know (i.e., accurate supply and demand curves). This inadvertently fed into the illusion among epistemically naïve mainstream economists that costs and benefits are objectively knowable, both to economists and to the market participants they are studying—as if prices need not be interpreted, such that people with faulty interpretations will make economic mistakes and, perhaps, suffer economic losses. As subsequent Austrian critics never tired of pointing out, there is no real conceptual space in the mainstream neoclassical view for ex ante economic errors (Kirzner 1997 Kirzner, Israel M. 1997. “Entrepreneurial Discovery and the Competitive Market Process: An Austrian Approach.”. Journal of Economic Literature, 35(1): 60–85. [Web of Science ®] , [Google Scholar]). 7. This passage is quoted, in Caplan’s Reply, as the only evidence for his incorrect claim that we challenge the rationality of the rational ignorance hypothesis—i.e., that we claim that rational ignorance is self‐contradictory. In fact, we do no such thing. We do reject the rational‐ignorance hypothesis as empirically irrelevant: it is an implausible explanation for public ignorance, given the paradox of voting, but Caplan does not seem to understand the difference between implausibility and impossibility. The rational‐ignorance explanation is, as we say quite clearly (p. 213), a logical possiblity; ipso facto, it is not logically incoherent. “But as usual with such theories,” we write, “the question is whether the logic describes actual behavior. Many logically possible hypotheses are not actualized in reality.” Contrary to rational‐ignorance theory, millions of people vote, indicating that they probably don’t think that their votes, hence their political views, are inconsequential, as the theory maintains. This is an objection to the empirical relevance of the theory, not to its logical internal consistency. Similarly, we reject the empirical relevance of economic theory tout court to understanding political ignorance and ideology (as the title of our paper suggests)—but we have no problem with the logical coherence of economic theory. And we do not reject information “search theory” as a logically impossible account of what actually happens in either the economy or the polity: in fact, information searches happen all the time. We claim only that, to the extent that people start out genuinely ignorant (e.g., when they are new to email and have never heard of spam or tricks), such that they face important unknown unknowns, their informational searches are not necessarily likely to be accurate; the economics of information is logically consistent, but in overlooking the empirical possibility of unknown unknowns, it fails (as we say in the sentence immediately prior to the one Caplan chooses to quote) to take ignorance seriously. The only aspect of economic theory that we reject as logically impossible is Caplan’s theory of rational irrationality—not economic theory in general, not the economics of information, not rational‐choice theory, and not rational‐ignorance theory. 8. Caplan (2007 Caplan, Bryan. 2007. The Myth of the Rational Voter: Why Democracies Choose Bad Policies, Princeton: Princeton University Press. [Google Scholar], 137–40) makes a similar point in objecting to a predecessor to his approach, Geoffrey Brennan and Loren E. Lomasky’s hypereconomistic theory of “expressive voting” (1993). Caplan (2007 Caplan, Bryan. 2007. The Myth of the Rational Voter: Why Democracies Choose Bad Policies, Princeton: Princeton University Press. [Google Scholar], 137) begins his critique of Brennan and Lomasky’s theory by acknowledging his great debt to it. But then he backs away from it, because according to Brennan and Lomasky, “you can safely vote for ‘feel good’ policies even if you know they will be disastrous in practice” (ibid., 138). While Caplan does not see the underlying logical problem with this notion, he does think that it is “psychologically” untenable to maintain, as Brennan and Lomasky do, that people would support policies that they knew were disastrous—especially when, according to the theory of expressive voting, the purpose of such support is to make the supporters feel good. Caplan writes: Few protectionists see their policies as economically harmful. If they realistically assessed the effect of this “feel‐good” policy, supporting the policy would no longer make its friends feel good. (Ibid., 139) Thus, Caplan suggests, “support for counterproductive policies and mistaken beliefs about how the world works normally come as a package” (ibid., 140). We could not have said it better ourselves. However, we dispute Caplan’s subsequent words: “Rational irrationality emphasizes this link.” For, on the very next page, in explicating rational irrationality, he writes that “when voters talk about solving social problems, their primary aim is to boost their self‐worth by casting off the shackles of objectivity” (ibid., 141). We do not dispute that people tend to do things that make them feel good. But recall Caplan’s criticism of Brennan and Lomasky: It would not make people feel good to support what they knew was disastrous. By the same token, it would not boost their self‐worth if they knew they were casting off the shackles of “objectivity,” because this would amount to knowingly opening the door to supporting disastrous policies. Caplan has simply recast Brennan and Lomasky’s implausible voter knowledge that the voter is wrong into voter knowledge that the voter is being “subjective” rather than objective, which we think is a distinction that does not make the needed difference. For the decision not to “put on my thinking cap” to be irrational, I must think that it is at least more likely than not that if I “shackled” myself to the rigors of “objectivity,” I would change my mind about the truth of my subjective opinion—but that, since I prefer to “indulge” the subjective belief, I choose not to think about it. But now we are back in Brennan‐Lomasky territory (deliberate mistakenness), for to believe that X is more likely to be wrong than right would deprive X of its “feel‐good” power. (Or, as we would put it: To believe that X is more likely wrong than right makes it impossible to believe in the truth of X in any coherent sense.) Like Brennan and Lomasky, who portray the truth as obvious to all, Caplan portrays voters as hiding it from themselves. In order to do that, they still have to know that “the truth is out there,” objectively, and that it probably falsifies what they subjectively believe. Otherwise, they would not have to actively avoid it, nor would they know what to actively avoid. So the truth really is obvious to them, as it is to Brennan and Lomasky’s voter, even though Caplan bifurcates people’s psyches into a level that “subjectively” believes X to be true, and a “level” that (somehow) knows that X is, “objectively,” false. 9. The Pew Research Center’s data were released directly to Stephen Bennett, who is responsible for all analyses and interpretations. The authors thank Andrew Kohut and Scott Keeter at Pew for releasing the information. 10. Caplan (2007 Caplan, Bryan. 2007. The Myth of the Rational Voter: Why Democracies Choose Bad Policies, Princeton: Princeton University Press. [Google Scholar], 132) does point out that poll taxes tend to reduce turnout, which shows that (some) people must be weighing the costs and benefits of voting. However, he infers from this, illogically, that voters who are not confronted with a poll tax—or, more crucially, voters who are confronted with one, but pay it—know that their votes are inconsequential, albeit “subconsciously” (ibid., 132). Even if we concede for the sake of argument, however, that all voters and all nonvoters are engaging in the same type of calculus (rather than some of them not thinking about it at all, and others voting out of perceived duty), it doesn’t mean that all voters and nonvoters know the same thing (i.e., the odds against their votes mattering), and that the thing they know is true, and that, therefore, they assess the costs and benefits of voting accurately. If they did, nobody would vote. The paradox of voting remains untouched by Caplan’s invocation of the subconscious. 11. The American National Election Studies are conducted by the University of Michigan’s Center for Political Studies, and the data are made available by the Inter‐University Consortium for Political and Social Research. Stephen Bennett is responsible for all analyses and interpretations. 12. Indeed, Caplan, who went to Princeton, might have been taught by Blinder in graduate school. 13. As we have noted, public‐opinion polls indicate that the public consistently overestimates businesses’ profit margins (Lipset and Schneider 1987 Lipset, Seymour Martin and Schneider, William. 1987. The Confidence Gap: Business, Labor, and Government in the Public Mind , rev. ed., Baltimore: Johns Hopkins University Press. [Google Scholar], 176–78). But Lipset and Schneider note that views of profits, and support for business, vary considerably over time (ibid., 181–83). Data from merged polls conducted for the Times Mirror/Pew Research Center for the People and the Press between 1987 and 2007 (Pew Research Center 2007b Pew Research Center. 2007b. Political Landscape More Favorable to Democrats: Trends in Political Values and Core Attitudes: 1987–2007, Washington, D.C: Pew Research Center for the People and the Press. 22 March [Google Scholar]) reveal similar variations. 14. Adam Smith (or, rather, Bernard Mandeville) did not come up with his counterintuitive thesis because he had the “incentive” to think about greed more objectively than had centuries of moralists before him. He simply made a conceptual breakthrough—the reasons for which may be found in the untidy currents of intellectual history, or in unique aspects of his biography, but not in the elegant thesis that the supply curve for knowledge slopes upward.