Keith T. Poole, Howard Rosenthal. Congress: A Political-Economic History of Roll-Call Voting. New York: Oxford University Press, 1997. x + 297 pp. $85.00 (cloth), ISBN 978-0-19-505577-1.
Reviewed by Douglas Dion (Purdue University)
Published on H-Pol (December, 1998)
Over a number of years, Keith Poole and Howard Rosenthal have pursued the thesis that voting in Congress can be adequately captured by a single underlying ideological dimension (and, at most, by two dimensions). This book represents the culmination of that research effort and is required reading for anyone interested in spatial models of legislatures. What I would like to do in this review is to set the book within its theoretical context and then to set out some of the strengths and weaknesses of the work.
Poole and Rosenthal's work sits at a critical juncture in the literature on legislative politics. In many ways this juncture represents as much of a generational gap as an intellectual one. Political analysts, including prominently the late William Riker (to whom this book is dedicated), argued that politics was rife with disequilibrium. This disequilibrium was generated by what was taken as a given--that legislatures deal with a number of issues. So, for example, the House of Representatives must determine how much money to allocate to a set of military construction projects. Thinking of each spending project as an issue, it turns out that there is no outcome that is preferred by a majority to every other. If, for example, I am cut out of a particular distribution, then I have an incentive to form a new coalition by cutting some of the coalition members out and reallocating their money to form a new coalition (something that requires giving some of the old coalition members a bit more than they had before). But this new coalition is itself open to the same strategy, and so we proceed. It turns out, however, that if issues are confined to a single dimension, then this kind of disequilibrium does not occur.
While multiple issues looks like a realistic assumption, the prediction of disequilibrium does not appear to be correct. A key problem for disequilibrium theorists has been finding actual cases of disequilibrium. As a result, more recent theoretical work on legislatures has tended to assume that legislators have preferences over a single dimension, although the link between particular policies and their eventual outcomes is a matter of uncertainty. These informational theories have done an excellent job of predicting legislative behavior, but at the cost of making what appears (at least from the disequilibrium perspective) to be a fairly unrealistic assumption.
Poole and Rosenthal's work appeals to both ends of this theoretical divide. By taking an empirical approach to determining the number of dimensions, Poole and Rosenthal avoid presuming the case in favor of either single or multiple dimensions. Furthermore, their results neatly blend the important strains of legislative politics theory together. Disequilibrium theorists are right--there are often multiple issues--but in terms of actual voting, Congress can be fairly well understood using a unidimensional model. This places informational theories on a stronger empirical basis, as well as providing disequilibrium theorists with the (continuing) challenge of determining why multiple issues should boil down to single-dimensional votes.
A further strength of the Poole and Rosenthal work is the empirical demonstration of the influence of race on congressional voting. At a time when rational choice theories face strong theoretical challenges from constructivist and interpretivist camps, the inability of rational choice theories to deal adequately with the special nature of race in American politics should rightly be viewed as a source of embarrassment. In doing so, they help bring a set of classic anecdotes about sophisticated voting--stories about minimum wage, the Powell amendment and so on--into a unified theoretical framework. Sophisticated voting, the argument goes, occurs when issues of race cut across ideological cleavages. The absence of sophisticated voting therefore is a testament not to the weakness of rational choice theories, but to the hegemonic power of white politicians to suppress potentially divisive racial issues throughout much of American history.
Students of American political development will also, I think, find a great deal of interest in the discussion of realignment. Contrary to both the Burnham critical elections periodization as well as Silbey's partisan-based breakdowns, Poole and Rosenthal find no evidence in Congressional roll call voting for treating either 1896 or 1932 as realigning elections. This has larger implications than Poole and Rosenthal bring out--it suggests, for example, that the links between realigning elections and public policy set out by David Brady may need some rethinking. (Indeed, I would have preferred to have seen the authors discuss Brady's work at greater length. Brady's book is not cited, nor is a discussion of his article on the topic especially prominent.)
More problematic, at least to those who are more interested in the ramifications of the work for understanding American political history, is the absence of a historical narrative. The presentation of the results places a premium on exposing the unidimensional nature of roll call voting. As a result, the implications of this view for transforming our understanding of Congressional history are left relatively underdeveloped. What, for example, should we infer about theories of the New Deal based on the continuities in the roll call data? I could imagine a host of theories that could be spun out here. Or consider the 1850 realignment. Poole and Rosenthal rightly find evidence of the breakdown of the Whigs prior to 1860, but in many ways this backs up what we already know from scholars of the 1850s like Holt, Foner, and Gienapp.
Even outside of the question of macro-changes, the location of individuals in voting spaces could lead to a greater appreciation regarding the role of individuals, particularly party leaders. One of the primary results of Poole and Rosenthal's work is the claim that individual members of Congress are fairly fixed ideologically--in their recurring language, members "die with their ideological boots on." True enough, but some individuals do change more than other individuals--and investigating those changes might help illuminate questions about the more crafty members of Congress (for example, was Henry Clay the opportunist he was often painted to be?). Furthermore, while individuals may stay fixed, the relationship of party leaders to the bulk of their party could change (imagine graphs showing where Thomas Brackett Reed stood in relation to his party placed next to similar graphs showing the relative location of John Carlisle and his party).
There is a great deal to learn from this book. Readers going into it, however, should be forewarned that they should be prepared to wade through a bit more technical armature than they might expect from a history of congressional voting. The authors start on the right track in Chapter Two with an excellent discussion of spatial location, but with the introduction of "cutting lines" things get a bit murkier. In part this is a function of the way the authors view policies--since politics boils down to a single dimension or at most two, policies can be thought of as lines passing through these two dimensions. The discussion up to that point, however, has thought of issues as in essence replications of the underlying dimension. Things get even hairier in Chapter Three, where the important Proportional Reduction Error (PRE) criterion is introduced. Add in a bit of combinatorial logic and a lot of exploding parameter estimates, and even a relatively informed reader (which I flatter myself to be) will find a need to slow down.
A final criticism. Although the book's main aim is to present a unidimensional model of Congress, every figure representing the spatial locations of members of Congress shows locations in two dimensions. This strikes me as quite a missed opportunity. It would be of great interest to see the changes that have taken place in the relative locations of House committees compared to floor medians. Linking this to actual policy changes, for example, would help to illuminate the theoretical issues involved in preference outliers more than anything else could (especially more than the attempt to retrofit a theory of policy change onto the convex hull of member ideal points). Exploring such diagrams could help show the utility of the unidimensional model much more than the dry multi-page listing of median positions.
All of which suggests that the true intellectual legacy of this book may be less in the arguments that have been set out--as compelling as they are--as in the formidable collection of historical data that Poole and Rosenthal have made available to the scholarly community. Scholars will be mining that information for quite a while and in doing so extending our understanding of the historical changes that have taken in place in Congress over the past two hundred years.
This review was commissioned for H-Pol by Lex Renda <email@example.com>.
Copyright (c) 1998 by H-Net, all rights reserved. This work may be copied for non-profit educational use if proper credit is given to the author and the list. For other permission, please contact H-Net@h-net.msu.edu.
If there is additional discussion of this review, you may access it through the network, at: https://networks.h-net.org/h-pol.
Douglas Dion. Review of Poole, Keith T.; Rosenthal, Howard, Congress: A Political-Economic History of Roll-Call Voting.
H-Pol, H-Net Reviews.
Copyright © 1998 by H-Net, all rights reserved. H-Net permits the redistribution and reprinting of this work for nonprofit, educational purposes, with full and accurate attribution to the author, web location, date of publication, originating list, and H-Net: Humanities & Social Sciences Online. For any other proposed use, contact the Reviews editorial staff at firstname.lastname@example.org.