Volume XXXIII, Number 4
November 2008




        The Quarterly is devoting special sections in this and the next several issues to a set of articles that deal with one of the central problems of legislative research: how to best measure the policy preferences of individual legislators and of legislative parties. The answer to that question is much more problematical in comparative legislative research in which an ever-growing number of legislatures must be taken into account than it once was in studies of single legislatures where roll-call vote data were readily available. Most of the articles on this subject were first presented at a research conference at Washington University, St. Louis. The conference was organized by Brian F. Crisp and Matthew Gabel and was sponsored by the Quarterly, by the Center for New Institutional Social Sciences and the Center for Political Economy of Washington University, and by the Claus M. Halle Institute for Global Learning and the Center for the Study of Law, Politics, and Economics of Emory University. The special section in this issue opens with an essay that introduces both this subject generally and the first set of articles that deal with it.


        The remaining two articles in this issue consider perennial questions in U.S. congressional research: how can we explain variance in the success of legislators in getting their bills passed, and how can we explain voting behavior in elections in which voters choose both their candidate for a seat in the House of Representatives and their candidate for president. On the first of these subjects, Gary W. Cox and William C. Terry reexamine previous work showing that members of the majority party have greater success in getting their bills passed than members of the minority but that holding leadership positions in committees does not clearly improve success. The authors use a negative binomial regression model to predict legislative success and test that model with a unique set of data. It includes all bills passed by the U.S. House from 1973 to 1999 and for the first time permits multiple observations of the status of members between the parties and between leadership positions. Contrary to previous findings, the authors demonstrate that holding leadership positions in committees does boost legislative productivity, especially in the majority party. They also show that the larger the majority party’s share of seats, the larger the increase in their members’ legislative productivity. Combined with previous findings by Cox and McCubbins (1995) that the majority party determines the left/right direction of the policy content of enacted bills, this article concludes that the party that holds the majority affects not only the volume of legislation that is passed but its policy direction as well.


        Michael Peress reexamines the accepted explanation for split-ticket voting in elections, where voters choose their presidential candidate from one party and their congressional candidate from the other. This behavior presumably reflects the "policy balancing" strategy of voters with moderate policy positions who split their vote. It is the dominant explanation for divided government and gridlock. But this theory presupposes a strong institutional role for the majority party in Congress which is at variance with current theory that emphasizes the majority’s negative agenda control. Peress develops a model of strategic voting in multi-office elections consistent with current theory that presupposes that the majority party faces an executive with veto power and has only limited institutional powers, principally a negative control over the agenda or "gate-keeping" power. In such a setting his model predicts that the moderate voter would strategically vote for a congressional candidate and a presidential candidate of the same party to avoid gridlock. This behavior would achieve a moderate outcome because policy is proposed by the median legislator. Peress is able to test the predictions derived from his model by using data on the policy preferences of voters from the thermometer scores of the American National Election Studies and by using survey data on the voting behavior of respondents. He finds that very few voters with straight policy preferences actually split their ballots and that approximately one third of voters with split preferences also cast straight party votes. These voters engage in what Peress calls "policy stacking" rather than "policy balancing." He concludes that strategic voting actually reduces the likelihood of divided government. Split-ticket voting, Peress suggests, reflects split preferences, not strategic voting.

                                                                                                          —Gerhard Loewenberg
Comparative Legislative Research Center


 Return to November 2008 Titles

 Return to LSQ home page