The full paper is free to read for a limited time at: bit.ly/4n4Ca56. It is also available on my website: drive.google.com/file/d/1ko1o... 8/8

These results suggest that politicians follow their party line and do not flip-flop on policies. However, “high-quality” politicians with good rhetorical skills may strategically focus on non-policy issues when their party platform is not the most popular among voters. 7/8

Discourse moderation is not random. Candidates who enter the runoff with an electoral disadvantage are more likely to moderate than those who are very likely to win. Conditional on winning, MPs who moderated their discourse tend to be more productive in office. 6/8

Discourse moderation does not necessarily mean policy moderation. Instead, I find evidence that candidates switch from partisan policy issues to neutral non-policy issues (e.g., their personal attributes or ties to the local constituency). 5/8

Candidates moderate their discourse when they need to appeal to a broader set of voters. In the French two-round system, candidates who make it to the runoff issue a new manifesto before the second round. Those present in both rounds reduce their extremeness in the runoff. 4/8

Using text-as-data methods, I first show that candidates tend to follow the party line: their ideological tone is consistent with the ideological ranking of their parties, from left to right. 3/8

I study candidates' campaign strategies by exploiting a novel dataset of 30,000 campaign manifestos issued by French parliamentary candidates (1958-1993). These data provide a systematic record of the campaigns run by thousands of individual politicians at the local level. 2/8

Would politicians say anything to get elected? In practice, they face many constraints. For instance, adopting policy positions opposite to that of their party or contradicting their previous policy announcements may be costly. 1/8

I am on the search committee for an applied microeconomist and happy to answer any questions about the job!