Solving a Puzzle: Aggregate Analysis and Economic Voting Revisited

This paper proposes an alternative explanation for the lack of congruence between aggregate and individual level studies of economic voting through a reexamination and extension of the aggregate time series. While aggregate level studies find a strong relationship between the state of the economy and the congressional vote, individual level studies performed in the latter half of the postwar era find little or no relationship between personal economic conditions and vote choice. It is argued that this lack of congruence exists because the impact of short-term economic fluctuations on voting in congressional elections has decreased dramatically over time. Using aggregate data, the relationship between economic changes and voting patterns is found to be time dependent; economic conditions are no longer major determinants of House elections. An examination of Senate elections indicates that this trend is limited to the House, in that economic fluctuations continue to influence senatorial contests. This House-Senate difference, it is proposed, stems from the different electoral dynamics of the two chambers.