Temporal aggregation in political budget cycles

We build a panel with quarterly data from Latin American and OECD countries over the 1980-2005 period. Annual data strongly underestimate the effect of political budget cycles (PBCs) when a pre-electoral fiscal expansion is followed by a post-electoral contraction, since the effects cancel out in the calendar year of elections, but not in the four quarters up to elections. Quarterly data show there is a significant fiscal expansion in the four pre-electoral quarters that extends, perhaps, to the first post-electoral quarter. In the next three post-electoral quarters, there is a compensating fiscal contraction only in Latin America, so PBCs contributed to public debt build-up in the OECD. Our results contradict a widespread consensus on PBCs being only a developing country phenomena - a conclusion which might have been affected by temporal aggregation - and imply that studies of electoral cycles should be done from now on with quarterly, not annual, data.