Theories of the political business cycle are by now common. They range from the elegant theoretical work of Nordhaus (1975) to the essentially empirical work of Tufte (1978). They all assume that politicians manipulate the economy to aid their reelection efforts; in particular, politicians manipulate macroeconomic policy to create a boom just before election day. Golden and Poterba (1980) provide an excellent review of this literature. Both Tufte and Nordhaus provide casual evidence for the existence of a political business cycle. Using inflation and unemployment rates as indicators of macroeconomic outcomes, Tufte and Nordhaus predict at a minimum, that unemployment should be decreasing before election day. If presidents are really manipulating the economy very well, then unemployment should be at a minimum in the October before the election, and hence should be rising after the election. In the Nordhaus version, the economy pays for the pre-election boom with a post-election surge in the inflation rate. Nordhaus also suggests that the president should engineer a recession early in his term to reduce inflationary expectations. Tufte (1978: 20) presents graphs of the unemployment series and claims that visual examination shows them to be generally consistent with the hypothesis. However, he provides no statistical tests of the hypothesis, nor does he estimate the quantitative impact of elections on unemployment. Tufte does not examine the inflation series. Nordhaus restricts himself to comparing unemployment rates in the two years preceeding and following an election, with a simple nonparametric test supporting the hypothesis. The literature provides several good empirical tests of the political business cycle hypothesis. Both Paldam (1979) and McCallum (1978) test for the existence of four year cycles, and find none. However they do not test for discontinuous changes in the series around election day. Golden and Poterba examine whether the use of policy instruments changes near election day (and