A Reassessment of Party Voting in the US House

A Reassessment of Party Voting in the US House* Richard Fleisher Fordham University Jon R. Bond Texas A&M University John Owens The University of Westminster This paper examines variability in party voting in the US House for the period 1953-2004. The paper makes a case for a new operational indicator of party voting that explicitly takes into account the number of non-conflictual votes that occur in each Congress. Redefining the percentage of party votes as a percentage of conflictual roll-calls held during the two year period provides a better measure of the degree to which party structures conflict in the US Congress. We use the refined measure of party voting as a dependent variable in an analysis exploring the effect of several variables including those previously identified by earlier studies as generating party voting. Consistent with previous research, our analysis of the contemporary time period (1950s to the present) with better data and more appropriate time series methods provides evidence that party voting on the House floor is a function of both external and internal forces. However, while past studies consistently report that external forces have the greater effect, we find that the short term effects of internal and external forces are similar in magnitude. * The Departments of Political Science at Fordham University, Texas A&M University, and the Centre for the Study of Democracy at the University of Westminster have been most helpful and supportive, and we are grateful to these departments for the support provided for this study. We are also grateful to Jeff Cohen, Pat Hurley, Paul Kellstedt, Dave Peterson, and Dan Wood for help with data, analysis, and interpretation. The Nuffield Foundation granted financial support to John Owens under grant SGS/LB/99/0248. We are grateful to Stan Bach for his technical advice in coding rules. The usual disclaimers apply. A Reassessment of Party Voting in the US House It is common to observe that political parties in the US Congress are less disciplined than parties in most parliamentary democracies (Cheibub 2006; Owens 2003). Yet, partisanship in Congress varies over time. For example, from relatively high levels of party voting in the House following the 1932 realignment, we see a long-term secular decline. After reaching a low point in the early 1970s, party voting began to rise and continued to rise through the remainder of the 20 century. During the 1990s, party voting in the House reached levels nearly as high as those observed during the New Deal. High levels of party voting continued into the 21 century, though there seems to have been some attenuation in recent congresses. Past research has established the general causes of these trends: party voting in Congress is a function of social and political forces external to the institution as well as the power and behavior of party leaders in Congress. Of these two sets of causes, the evidence indicates that external forces have the stronger effect (Aldrich and Rohde 2000; Brady, Cooper, and Hurley 1979; Cooper and Brady 1981; Patterson and Caldeira 1988; Rohde 1991). Although this research has contributed much to our understanding of party voting in Congress, several limitations point to the need for additional research. First, the primary focus of this research has been on explaining long secular trends. Understanding the rise and fall of these trends is clearly important, but party voting also varies considerably from one Congress to the next. Second, measures of key external and internal variables used to analyze trends spanning long historical periods were crude, because more appropriate data such as surveys of public opinion were not available. Third, the statistical models in these studies accounted for some, but not all, of the potential challenges to estimation inherent in time series data. Finally, several important studies analyzing these trends were based on data that contain few if any observations after the secular decline in party voting turned around in latter third of the 20 century.