The time horizon of economic voting is unsettled. Each aggregate time series study offers a different specification, virtually all retrospective. The survey research, which does look at individuals, mostly assumes that economic voters are myopic and retrospective. Unfortunately, the available data have permitted the testing of few rival hypotheses. However, new data from the Surveys of Consumer Attitudes allow further exploration. These rich surveys show that, at least in the 1984 elections, the American voter did respond to immediate past economic conditions. And more noteworthy, they responded strongly to future economic expectations, whether personal or collective, short-term or long-term. These sharp prospective results call into question standard explanations of economic voting, which have relied on some variant of retrospective theory.