Research

Working Papers

Power Struggles After Autocracy - Evidence from Post-Uprising Tunisia
Current version: July 2023

What is the Ideal Number of Women in Politics? - The Gender Gap in the Perceived Deficit of Women in Politics
with Pablo Selaya (UCPH), and Sina Smid (CBS)
Current version: July 2023

The Making of a Ghetto - Residential Moving and Neighborhood Segregation
with Jack Melbourne (Bocconi)
Current version: July 2023

Work in Progress

Trauma and Refugees’ Employment – Evidence from People Displaced from Ukraine
with Mette Foged (UCPH, Economics), Karen-Inge Karstoft (UCPH, Psychology), Anne Agathe Pedersen (UCPH, Psychology)

Other

Local Election Predictions - A Machine Learning Approach
Project website (in German)
We explore the potential for machine learning algorithms to predict local election outcomes. In order to do so, we collect micro-level data on candidates from several sources. We use different prediction algorithms including neural networks to train estimators based on past election results. Then, we apply these estimator to the current candidate lists to predict results of upcoming elections. So far, we publicly covered three local elections in Bavaria (March 2020), North Rhine-Westphalia (September 2020), Hesse (March 2021), and one mayoral election in Stuttgart (November 2020).