This paper investigates whether social media access is associated with increased probability or intensity of ethnic conflict in Myanmar. In this context most people use mobile phones, and particularly the Facebook app, to access the internet. To distinguish the effects of social media from those of the broader internet, I exploit geographic variation in mobile phone coverage as a proxy for Facebook availability. Despite evidence of a hate-campaign utilizing Facebook to reach wide audiences, I do not find that social media access is associated with increased probability or intensity of conflict. The only exception to the null result is variation related to the Rohingya crisis: in this regional setting suggestive evidence points to Facebook availability being associated with slightly higher probability of conflict. 

Populist parties have gained significant power in European politics in the last decades, raising concerns over the potentially contagious effect of populism. I study how populist party representation in local councils affects other parties' ideological positions. I use variation created by close elections to identify ideological shifts resulting from a change in party representation, holding voter preferences constant. I use candidate level data from a voting advice application to derive ideological positions. I model candidates' responses using item response theory to obtain measures of ideology that are comparable across election years. The results show that higher populist representation causes ideological convergence among established parties'  political candidates. One additional seat to the populist party causes a 10 % decrease in the interquartile range of ideological positions. The convergence takes place only on the liberal-conservative dimension, while positions on the economic dimension are unaffected.

This paper investigates how political alignment affects the implementation of punitive welfare measures in the UK. In particular, we examine whether a legislator's party affiliation affects the rate of sanctions to unemployment benefits in the MP's constituency. We use a regression discontinuity design based on close elections to compare the sanction rates across constituencies that are marginally aligned or unaligned with the central government. We find that implementation of the sanction regime is significantly more lenient in constituencies won by the government parties. The RD estimate indicates a drop of .8 percentage points at the cut-off, implying on average 18 % lower sanction rates in Coalition controlled constituencies. Our findings suggest that legislators are able to influence national, rule-based policies, even within a highly centralized system. Such pork barrel politics that can undermine institutions that should be neutral to local partisan considerations.the government parties. Our findings suggest pork barrel politics can also influence the allocation of economic "bads", even within a highly centralised system, and can undermine institutions which should be neutral to local partisan considerations.

The Effects of Female Political Leaders: Evidence from an Exogenous Government Change in Germany (with Luisa Dörr, Klaus Gründler and Niklas Potrafke)