Studies of OECD countries have generally failed to detect real economic expansions in the pre-election period, casting doubt on the existence of opportunistic political business cycles. We develop a theory that predicts a substantial portion of the economy experiences a real decline in the pre-election period. Specifically, the political uncertainty created by elections induces private actors to postpone investments with high costs of reversal. The resulting declines, referred to as reverse electoral business cycles, are larger the more competitive the electoral race and the greater the polarization between major parties. We test these predictions using quarterly data on private fixed investment in ten OECD countries between 1975 and 2006. The results suggest that reverse electoral business cycles exist, and as expected, depend on electoral competitiveness and partisan polarization. Moreover, simply by removing private fixed investment from gross domestic product (GDP), we uncover robust evidence of opportunistic cycles.