Political models of the business cycle have typically been ignored because they appear inconsistent with rational behavior and because empirical evidence is inconclusive. This paper addresses the second issue, demonstrating for U.S. real GNP, unemployment, and inflation that electoral cycles (persistent patterns across electoral terms) are significant, but apparently only for Republican incumbents, and that partisan cycles (persistent differences between parties) are also significant. These findings are consistent with the conjecture that a minority party is more constrained by electoral concerns, whereas a majority party is freer to pursue partisan objectives. Copyright © 1990, Wiley Blackwell. All rights reserved