Partisan Differences in Economic Outcomes and Corresponding Voting Behavior: Evidence from the U.S.

The present empirical work examines the differences in economic outcomes delivered by partisan governments, and the way in which voters take this into account. Autoregressive models of output growth, unemployment and inflation, augmented with political variables; and probit binary choice models of voting decisions, incorporating expectations about inflation and unemployment, are estimated for U.S. post-war data. The analysis confirms that partisan differences in economic outcomes are actually observed in the data. U.S. unemployment rate exhibits a distinct partisan cycle, behavior of output growth and inflation rate partly supports the partisan differences hypothesis. Thus suggesting that each party can be ``instrumental'' in solving particular economic problems. In line with this logic, U.S. voters seem to believe in the asymmetric abilities of parties to fight inflation and unemployment. Most interesting empirical findings include evidence that U.S. citizens tend to vote for the left party (Democrats) when high unemployment is expected, and for the right party (Republicans) when high inflation is expected. This relation is especially robust for Presidential elections. There is also evidence pointing to the presence of electoral inertia and absence of ``midterm'' electoral cycle in the U.S.