The Marginal and Time-Varying Effect of Public Approval on Presidential Success in Congress

Abstract

We analyze the relationship between public approval and presidential success in Congress using time-varying parameter regression methods. Cues from constituency, ideology, and party dominate congressional vote choice, so the effect of public approval of the president is typically marginal. Because the strength of these primary cues varies through time, the effect of public approval on presidential success should also be time varying. Analysis of conflictual roll-call votes from 1953 through 2000 using the time-varying Kalman filter reveals that the effect of public approval on presidential success is marginal and changing through time. These models assume that the time variation is a stochastic process, and finding time-varying relationships may indicate model misspecification. Our theory, however, suggests that this time variation depends on a systematic factor—partisanship. A better specified model that allows systematic parameter variation confirms that the level of partisanship conditions the relationship between public approval and presidential success in Congress.