Public Choice 64: 299-300, 1990. ~ 1990 Kluwer Academic Publishers. Printed in the Netherlands.

Book r e v i e w

Larry M. Bartels, Presidential primaries and the dynamics of public choice. Princeton, N J: Princeton University Press, 1988. xxi + 374 pages. $45.00 hardbound, $14.95 paper. This book is a major contribution to the study o f presidential nominations, and it may be the best single book on the topic. As a book whose data and analysis are those of behavioral political science, it is not the intellectual style normally found in this journal. Yet the process of selecting presidential nominees in the United States is one of the most public of choices. Moreover, the book concludes with a serious challenge to public choice as we usually practice it. Bartels identifies "the classical theory o f public choice," drawing on C.L. Dodgson, Kenneth Arrow, and Charles Plott regarding the problems of aggregating preferences. Bartels sees this theory as an extension of a liberal tradition in politics with origins in the seventeenth century. The distinctive element in this theory is "the assumption that individual preferences over a set of relevant alternatives are taken as the unchanging, unanalyzed, predetermined basis for public choice" (p. 300). Although Bartels overstates the degree to which this assumption characterizes the liberal tradition, (see for example The Federalist Papers on momentary passions, and irregular and violent propensities), much of modern economics and public choice is well described by such a premise. The idea that individual preferences are exogenous and fixed is a very useful heuristic premise for purposes of analyzing how they may be aggregated by alternative voting procedures. This assumption has been a basis for some of the greatest contributions of public choice. From Bartels' perspective, however, the idea that individual preferences are the unanalyzed basis for rational economic behavior provides an inadequate basis for dealing with some important aspects of political life. He shows that the process o f presidential nomination is in many respects at odds with the classical theory of public choice. Instead of a fixed set of alternatives, the set of candidates changes from week to week. Instead of a fixed electorate, decisionmaking is fragmented among different self-selected electorates at different points in the process. Clearly the world of presidential nominations is one in which preferences are dynamic and endogenous rather than fixed and given. The major achievement of this book is an imaginative and insightful modelling of the process of preference formation within an institutional structure that affects the content of preferences as well as their aggregation. The key institutional feature is the series of individual state primaries in which the outcomes of any given primary may affect both the identity and the popular support of candidates in subsequent primaries. Media coverage of the campaign as a horse race provides the public with a basis for judgments of the victory chances of the competitors. Bartels models these expectations, rendering them predictable. More importantly, he models three distinct ways in which expectations of victory lead to increased support

300 for a candidate. Not surprisingly, voters with low information are much more susceptible to having their choices influenced by expectations and m o m e n t u m than are high information voters. Although a desire to support a winner is clearly relevant in this process, Bartels shows how underlying political predispositions are also an important part of the preferences formation process. Bartels builds on the microfoundations provided by this model of individual preference formation to provide an analysis of primary elections, whose outcomes are a function of underlying political predispositions of the voters, current candidate standing in the horse race, and the level of information at that stage of the campaign. This is built into a typology of recent campaigns, and an illustration of the ways in which m o m e n t u m can affect or fail to change the outcomes of different contests. In some ways Bartels sees the nominating process as sidestepping the difficulties of preference aggregation so well explained by public choice. He sees the process as one of " m u t u a l shaping and adjustment of individual preferences in a truly social process" that m a y result in "coherent majority choices" (pp. 304-305). Although he discusses this process as an escape or vitiation of A r r o w ' s formal dilemma, he acknowledges the arbitrariness of the process, and the absence of a basis for judging the desirability of the outcome. Presidential Primaries addresses another central public choice concern: the choice of alternative institutional arrangements. The b o o k goes beyond an analysis of preference formation and aggregation in the context of the institutional features described above. It also analyzes the likely impact of major rules changes such as regional primaries, a single national primary or approval voting. In my view the analysis of preferences is more penetrating and authoritative than the analysis of institutional alternatives, but the former is a m a j o r achievement by itself, and it is an essential basis for intelligent institutional engineering. Bartels has done quite enough in this volume to merit a place as one of the leading books on presidential nominations. While his b o o k might be read as an attack on public choice, I suggest that we view it as an expansion of our scope and a contribution to our field. William R. Keech Department of Political Science CB#3265 The University of North Carolina at Chapel Hill Chapel Hill, NC 27599-3265