Political Careers, Government Stability, and Electoral Cycles

Essay 1: This essay investigates the impact of voter support on the representation of women in the political profession. The empirical analysis exploits two-stage elections in the United States and Italy to hold the selection of candidates constant. In two-stage elections, candidates are admitted to the second round of voting based on the outcome of the first round. I find that among candidates who marginally qualify for the final round, women are 20 percent less likely than men to be elected to the US House of Representatives and 40 percent less likely to be elected mayor in Italian municipalities. Using a difference-in-discontinuities design, I then show that the gender gap in the probability of being elected has long-lasting effects on career trajectories. Women are substantially less likely than men to win future elections and to climb the political hierarchy. My findings suggest that one of the reasons that few women reach the top in politics is that female candidates face hurdles at the beginning of their careers. Essay 2 (with Felipe Carozzi and Luca Repetto): This essay studies how political fragmentation affects government stability. Using a regression-discontinuity design, we show that each additional party with representation in the local parliament increases the probability that the incumbent government is unseated by 5 percentage points. The entry of an additional party affects stability by increasing both the probability of a single-party majority and the instability of governments when such a majority is not feasible. We interpret our results in light of a bargaining model of coalition formation featuring government instability. Essay 3 (with André Reslow): This essay introduces macroeconomic forecasters as political agents and suggests that they use their forecasts to influence voting outcomes. We develop a probabilistic voting model in which voters do not have complete information about future states of the economy and have to rely on macroeconomic forecasters. The model predicts that prior to a referendum, it is optimal for forecasters with an economic interest (stake) in the outcome and influence over the public to publish biased forecasts. We test our theory using high-frequency data at the forecaster level surrounding the Brexit referendum. The results show that forecasters with greater stakes and influence released much more pessimistic and incorrect estimates of GDP growth under the Leave outcome scenario than other forecasters. Essay 4 (with André Reslow): This paper documents the existence of Political Forecast Cycles. In a theoretical model of political selection, we show that governments release overly optimistic GDP growth forecasts ahead of elections to increase the reelection probability. The bias arises from lack of commitment if voters are rational and from manipulation of voters’ beliefs if they are naive and do not expect the incumbent to be biased. Using high-frequency forecaster-level data from the United States, the United Kingdom, and Sweden, we document that governments overestimate short-term GDP growth by 10 to 13 percent during campaign periods. Furthermore, we find that the bias is larger when the incumbent government is not termlimited or constrained by a parliament led by the opposition. Consistent with the model, we also find that the election timing and amount of available information determine the size of the bias at different forecast horizons.