The hypothesis on the strategic use of debt argues that governments issue more debt when facing higher probabilities of electoral defeat. Testing this hypothesis has proven challenging, since measures of those probabilities are potentially endogenous. However, during presidential transitions, the probability of electoral defeat becomes one if the incumbent was defeated in the election and zero if the incumbent was reelected. I thus use ex post electoral outcomes as a proxy of the probability of electoral defeat to construct measures of the electoral surprise and to estimate their impact on the budget deficit. Monthly data from Latin American democracies in 1980–2005 reveal that higher magnitudes of surprise defeats (wins) produce higher (lower) deficits when the executive controls the legislature, but there is no effect when the legislature is controlled by the opposition. While previous studies find that unified governments facilitate the manipulation of fiscal resources for electoral gain, here I show that such manipulation is extended even beyond electoral defeats, during presidential transitions.