Does opportunistic election calling pay?

This paper examines whether governments that call elections opportunistically do better at the polls than those which do not. The literature gives rise to conflicting expectations regarding this question. On the one hand, the political economy literature rests on the assumption that governments call elections opportunistically because it raises the probability of winning a new term in office. On the other hand, Smith’s work suggests that voters are sophisticated enough to infer from an opportunistically called election that the government expects its performance to decline if it holds out for longer. Hence voters are expected to discount observed performance (Smith 1996, 2003). Given the strategic importance of this question for the survival of governments, the complete absence of comparative work, which examines whether early election calling pays is deeply puzzling. This paper offers a first empirical investigation of it. We use the most comprehensive comparative dataset available on elections in West and Eastern Europe from 1945 (or democratization) to the present and address the difficult problems of causal inference presented by the comparison.