[Under the hypothesis of a myopic electorate, vote-loss-minimizing behavior by the party in power subject to a dynamic inflation-unemployment relation is shown to generate a stable electoral policy cycle. The pattern of unemployment and inflation in the United States during the four presidential election periods from 1957 through 1972 is then examined for evidence of whether or not the administration believes the electorate is myopic. The conclusion is that the myopic hypothesis does a superior job of explaining aggregate demand policy, as reflected by the unemployment rate, during the second and third election periods, while the hypothesis that the administration believes the electorate is rational does a better job in the first and fourth periods.]