CONSTITUTIONAL POLITICAL ECONOMY, VOL. 5, NO. 2, 1994
REVIEWS
GEOFFREYBRENNAN and LOREN LOMASKY (1993) Democracy and Decision: the Pure Theory of Electoral Preference. Cambridge: Cambridge University Press. x + 237 pp. Do voters behave in the voting booth the same way they do in the market? Geoffrey Brennan and Loren Lomasky's answer is an uncategorical " n o " . What motivates people is the same whether one is participating privately in the market or collectively in politics. What people do in the two choice settings, however, is quite different. Voters wear two hats: an instrumental hat which focuses on voting outcomes, and an expressive hat which focuses on the act of voting. Consider a loyal alumnus. Donating money to the athletic program at one's alma mater is an instrumental act as the donation is expected to lead to greater athletic success. Cheering for your alma mater while sitting in front of the television set is an expressive act. The cheering has no effect on the outcome but expresses the fan's preference for his team. Brennan and Lomasky's thesis is that voting, like cheering, is an expressive act. Orthodox public choice theory treats voting like donating money--as an instrumental act, a means to bring about an end. The public choice view, Brennan and Lomasky argue, is theoretically untenable. The logic of their argument is straightforward. In any large number collective choice setting, where decisions are made by majority-rule, the probability that an individual's vote will swing the outcome renders one' s vote meaningless with respect to the outcome. Unlike market choices where actions translate into predictable outcomes, the link between action and outcome in voting does not, for all intents and purposes, exist. What, then, does a voter communicate when she casts her ballot? To answer this, Brennan and Lomasky first suggest that all choices whether made as consumers or voters contain an expressive element which is different from one's preferences over the bundles of goods one is choosing between. In the market, instrumental and expressive preferences combine into a kind of revealed " n e t " preference. For example, an investor interested in maximizing the performance of his portfolio considers buying stock in a South African company, but at the same time does not want to lend legitimacy to apartheid. Here, whichever preference is stronger will be revealed in the market. Purchasing the stock implies the instrumental return is greater than the foregone expressive return. Expressing one's distaste for apartheid by not buying the stock entails an opportunity cost that exceeds the cost of violating one's moral principles. In voting, the insignificance of one's vote on the collective outcome makes the opportunity cost, in terms of foregone instrumental returns, of exercising one's expressive preferences insignificant. Consider our hypothetical investor now as a member of a collective decision-making body, say 247
CONSTITUTIONALPOLITICALECONOMY
shareholders voting to determine a company's pension portfolio policy. With a sufficiently large number of shareholder-voters, our investor would now vote against adding the South African stock even if the instrumental returns were several times greater than the expressive returns. Voting to buy the stock would entail the personal cost of violating the investor's ethical code with no associated benefit. The link between vote and collective outcome vanishes, but the link between vote and personal outcome remains. What is revealed by the individual vote is the voter's expressive preference and the collective outcome reflects the majority's expressive preference, and nothing more. Thus, while individuals are motivated as voters the same as they are in the market, their behavior differs. Brennan and Lomasky's basic argument is clearly correct. The mathematical realities of any individual voter having an impact in a large number election are inescapable. Yet, while the fundamental critique is on the mark, some difficulties remain. Foremost is the question: does expressive voting explain too much? Can empirical hypotheses be constructed and examined? The public choice literature is replete with empirical results that relate voting and political behavior to explanatory variables which proxy instrumental interests. Now, one could argue that these variables are also proxies for expressive interests. But if expressive preferences are positively correlated with instrumental preferences, one could continue to use instrumental theory without much concern. Brennan and Lomasky explicitly deny such a positive correlation and offer a number of reasons (whimsy, social approbation, self-reflection, to name a few) why instrumental and expressive preferences might diverge. If that's the case, any empirical result appears supportable on expressive grounds. Suppose the per capita monetary gain in legislators' districts from the passage of a certain bill exhibits a positive and significant relationship to a legislator's probability of voting in favor of the bill. One could then say the instrumental interests implied by the bill mirrored voters' expressive preferences. If the same explanatory variables turned out to be statistically insignificant, one could say that the voters' expressive preferences countered their instrumental preferences. Can expressive and instrumental models of voting be reconciled? I think so. A large fraction, if not the majority, of empirical results on voting behavior deal with voting by legislatures. At the legislative level, the electoral arithmetic is not so devastating. Individual representatives may have a significant impact on the outcome. Further, the existence of coalitions, lobbying, vote trading, party politics, and logrolling, mean that voting outcomes can be affected by a few votes, and hence an individual vote could be significant. It is generally accepted that on sensitive issues, opposing political parties woo undecided voters to gather the necessary majority of votes. If votes have value to others, then it appears that instrumental concerns are relevant determinants of voting behavior. Thus, while not denying that there is an expressive content in voting, the expressive aspect is likely to be less important i n ' 'small number" decisionmaking bodies because the instrumental aspect takes on a bigger significance. It is at this level that public choice has had the most to say. In elections 248
REVIEWS
involving large numbers of participants, the electoral arithmetic favors the expressive. Indeed, that voter turnout is low and voters are less than adequately informed is not surprising. What is surprising, from a purely instrumental view is that voter turnout and political acumen is as high as it is. Expressive voting theory helps explain that. In the latter part of their book, Brennan and Lomasky deal with the implications of expressive voting on constitutional design. Of particular interest is their critique of unanimity as a welfare norm. In both markets and voting, unanimity is viewed as a test of efficiency. Standard welfare theory says that a move from one allocation to another is Pareto-optimal if there is unanimous agreement over the move. The analog in collective choice is Wicksell's unanimity test. A collective choice is efficient (and therefore legitimate) if it secures unanimous consent. However, in a large number setting the instrumental dimension over which the vote is taken, budget size for example, disappears from the individual voter's calculus. All that remains revealed is the voter's expressive preferences. A mandate for change, or for the status quo, becomes no mandate at all with respect to the issue at hand. The Wicksellian test requires that individuals as voters wear both hats. Operationally, however, only the expressive hat shows up if the collectivity is large. If one were to summarize what this book is about it would be that there are no returns to instrumental voting, at least in theorizing about large number elections. Explanations of political behavior based on outcome differentials must either account for the expressive dimension or assert that probabilities of decisiveness are large enough to be relevant. And as corollary, social norms drawn from orthodox views of voting behavior must make the same assertions. Any student of public choice would benefit from reading this book. A second subtler message of the book is that institutions matter. Behavior is not neutral under the alternative institutions of market and collective choice. What happens under one set of institutions (the market) need not occur under another set of institutions (collective choice). That institutions matter is what makes studying constitutional political economy and public choice interesting. If politics were simply just another market, analyzable with the same set of tools and implications as private markets, public choice and constitutional political economy, would have ceased being interesting long ago. If economists have anything to contribute to the study of political economy, and I think they do, it is precisely because there are some inherent differences between market and public choice. Roger Faith University of Arizona YOUNG BACK CHOI (1993) Paradigms and Conventions: Uncertainty, Decision Making, and Entrepreneurship. Ann Arbor: The University o f Michigan Press. ix + 184 pp. In recent years, many economists have become interested in the evolution of social conventions, and a great deal has been written on the subject. Young 249
CONSTITUTIONALPOLITICALECONOMY
Back Choi's book is another contribution to this literature, but a strikingly original one. The usual way to look at conventions is in terms of game theory. A convention is interpreted as an equilibrium in a game which has more than one equilibrium solution. Evolutionary arguments are adduced to explain how we get to one equilibrium rather than another. Despite its history-dependence, this kind of analysis remains rooted in a model of the individual as an instrumentally rational chooser. Choi takes a very different approach, which has echoes of Joseph Schumpeter, Friedrich Hayek, and Israel Kirzner. The fundamental concept in Choi's system is what he calls a paradigm. He uses this term quite elastically, but the basic idea is that of a model of, or way of making sense of, the world. Before a person can choose an action, he or she-must accept some paradigm which defines the choice problem and distinguishes what is relevant from what is not. The conventional theory of rational choice assumes that choice problems have unique descriptions: in effect, the theory invokes a particular paradigm, which it then assumes is accepted by all agents. Choi takes the view that we choose among paradigms, and that this choice is underdetermined by o.ur experience and our reason; thus there is a place for creativity, imagination, and commitment in accepting one paradigm rather than another. A theory of human behavior should recognize that different individuals might use different paradigms; if, within a society, individuals converge on a common paradigm, this is a phenomenon to be explained, not something to be taken for granted. For Choi, a convention is a collective or social paradigm--a model of the world that is common to the members of a society. Much of the book is devoted to showing why conventions, so understood, tend to come into existence and to be stable. One source of stability amounts to the familiar game-theoretic argument about multiple equilibria. Each individual is trying to make sense of a world which includes other individuals. There is a state of equilibrium if each person, interpreting other people's actions from the viewpoint of her own paradigm, is led to act in a way that confirms those other people's confidence in their own paradigms. In many cases, such an equilibrium requires that interacting individuals accept a common paradigm. Choi's originality comes in the emphasis that he gives to other forces that favor stability. A paradigm imposes 'selective vision': things which do not make sense within the paradigm may not be seen. More importantly for Choi, conventions tend to become norms. Choi's argument here is intricate. The basic idea is that each of us is uncertain whether we are using an appropriate paradigm, and so we have a keen interest in studying how other people behave in situations like our own. We feel more secure if we find that other people behave as we do, since this tends to confirm the appropriateness of our own paradigms. If I read him correctly, Choi sees this desire for confirmation-which he calls a search for approval--as the ultimate source of our moral sentiments. Thus, our moral sentiments tend to favor conformism. Deviants suffer disapproval and ostracism, a process which tends to protect paradigms from exposure to conflicting ideas and experiences. 250
REVIEWS
Choi hopes to explain our moral sentiments as the consequence of fundamental uncertainty about the world and of our need for paradigms; he seems to want to avoid having to make assumptions about human nature. I am not sure that he succeeds in this. He begins by invoking individual interest to explain why we tend to imitate each other; to this extent, he is offering a rationalistic form of explanation for norms. But the explanation of moral conformism seems to depend on our being subject to some form of cognitive dissonance. It is one thing for us to approve our own actions because they correspond with what others would do in similar circumstances: if we are uncertain how best to make sense of the world, imitation may be prudent. But it is quite another thing to ostracize other people because their actions fail to correspond with what we would do: in Choi's framework, this amounts to censoring our experience so as to avoid confronting evidence which might cast doubt on the wisdom of our actions. Cognitive dissonance would seem to be a distinctive feature of human psychology, not a consequence of our need for paradigms. For Choi, the normal state of affairs or 'presumptive truth' is a social world governed by convention. But just as in Thomas Kuhn's account of science, 'normal science' is occasionally disrupted by paradigm shifts, so Choi's conventions sometimes collapse. A society governed by convention is likely to offer unexploited opportunities for mutual gain; over time, as underlying economic realities change but conventions remain stable, the extent of these opportunities grows. Entrepreneurship is the force which exploits these opportunities, and in the process, disturbs conventions--a Schumpeterian process of creative destruction. Choi gives a convincing account of how such entrepreneurship is met with disapproval, suspicion, and envy. He suggests that some of the dynamism of market-based societies stems from their ability to shield entrepreneurs from the hostility of the envious. As will be clear by now, this is an extremely ambitious book. The mode of argument is a priori: Choi hopes to deduce his conclusions from a few simple propositions. (At the end of the book, he summarizes his arguments, and exclaims: 'All this from uncertainty, inference, and paradigms?') Although Choi comments on the work of other scholars, he develops his argument almost from scratch, and is quite unabashed about treating his own theory as a rival to neoclassical economics. I do not think he can seriously expect to generate a paradigm shift of this magnitude. (The theory itself explains why not.) But one might realistically hope that this book will have some influence on economic thinking. Robert Sugden University of East Anglia ALEXANDER ROSENBERG (1992) Economics--Mathematical Politics or Science o f Diminishing Returns ? Chicago: University o f Chicago Press. xvii + 266 pp. Since the publication of his Microeconomic Laws in 1976, Alexander Rosenberg has occupied a prominent position in the debate on the methodology of 251
CONSTITUTIONALPOLITICALECONOMY
economics. The focus of that debate, and Rosenberg's views, have shifted somewhat over time, and this book acts in part as a retrospective of the debate over that period, with Rosenberg responding to critics and charting the development of his own thoughts. But this aspect of the book is secondary to its main purpose, which is to argue in favor of an answer to the question in the title which denies that economies is a science and claims that it is more appropriate to view economics simultaneously as a branch of constitutional politics and as a branch of mathematics. Rosenberg's argument against economics as science is broadly familiar from his 1983 paper "If Economics isn't Science, What is it?" In broad outline he argues first that science is characterized by improving predictive precision, and then that economics has shown no such improvement in predictive precision--indeed that many economists seem not to be interested in such improvement. Finally he argues that economics is incapable of improved predictive precision because of the intentional nature of economic explanations and the unobservable nature of the key intentional factors. Rosenberg's argument on this matter is controversial on a number of grounds--the claim that economics has not made progress in terms of predictive precision is contested by some; the claim that the intentional nature of economic (or, more broadly, social scientific) explanation makes predictive precision and, therefore, scientific status an unattainable goal is contested by others. But Rosenberg puts his arguments and rejoinders forcefully and cogently. In particular, the argument that intentionality is a irreducible feature of economic explanation and that this in itself cuts economics off from the mainstream scientific tradition is powerfully put. But if Rosenberg argues against economics as science, he argues at least as strongly against the post-modem critique of economics and science associated in particular with Donald McClosky's The Rhetoric of Economics. Indeed, Rosenberg is at his best when engaged in argument with an identified opponent, and his critique of McClosky in chapter 2 of this book is a fine example of his style of argument. The major part of this book is devoted to the critical debate on the cognitive status of economics, and it is only in the last thirty pages or so that Rosenberg focuses on the positive aspect of his position. Here he sketches out an argument in favor of viewing economics (and particuarly general competitive equilibrium analysis) as a branch of political philosophy, and then returns to the theme of the 1983 article to further defend the view of economics as a branch of mathematics. The view of economics as political philosophy is essentially that associated with constitutional political economy and James Buchanan in particular. Thus the inquiry into the behavior of markets is motivated not by a scientific concern to predict market outcomes, but rather by a concern with the choice of institutional structures. The key question for Rosenberg is then, how can the economic analysis of markets inform this institutional and constitutional debate if we believe (as Rosenberg does) that the economic analysis cannot provide predictions of market outcomes. In short, how much predictive power 252
REVIEWS
does a theory need in order to have normative bearing on institutional design? The constitutional concern is with the general operating characteristics of alternative institutional structures, and unless we can deduce (predict) these operating characteristics from economic models, how can economic models be put to the constitutionalist task? Rosenberg takes the point that in choosing institutions we may wish to guard against undesirable outcomes as well as promote desirable ones, so that the relatively coarse-grained generic predictions available from economic analysis based on assumptions about intentionality which focus on self-interest and treat all individuals as if they are 'knaves' will have some value in the constitutional calculus. Nevertheless, there is some doubt surrounding the ability of the constitutionalist interpretation of economics as political philosophy to progress beyond very general statements without requiring the sort of predictive power of economic models that Rosenberg denies. Readers of this journal will be disappointed that this aspect of Rosenberg's discussion is rather brief and sketchy--covering fifteen pages at most. Readers will also be disappointed if they are looking for a balanced overview of the methodological debate, of the type provided recently by Daniel Hausman. But the reader who is familiar with Rosenberg's earlier work will find the argument developed, defended and extended, while the reader coming to the debate for the first time will find a clear and powerful argument deployed with great skill. Alan Hamlin University of Southampton REFERENCES
Hausman, D. (1992) The Inexact and Separate Science of Economics. Cambridge:Cambridge University Press. McClosky, D. (1985) The Rhetoric of Economics. Madison: Universityof Wisconsin Press. Rosenberg, A. (1976) Microeconomic Laws. Pittsburgh: University of Pittsburgh Press. Rosenberg, A. (1983) "If Economics isn't Science, What is it?" Philosophical Forum 14: 296-314.
253