The Economic Consequences of Political Turmoil

Abstract

This thesis investigates the long-run consequences of political turmoil, and how economies recover from it. Four empirical studies provide evidence from different contexts and different types of civil violence. The first empirical study in Chapter 2 develops a novel measure of local conflict exposure, tests its determinants, and provides evidence from causal Shift-Share estimations that conflict exposure decreases economic activity in the medium run. Most importantly, this chapter points to an important heterogeneity in conflict resilience related to locations’ geographic characteristics; locations with spatially clustered population patterns are more likely to suffer persistent consequences of civil violence. Chapter 3 analyzes the general equilibrium effects of civil conflict in international trade relationships. Leveraging a novel estimation approach derived from the structural gravity model of international trade, the chapter shows that importers turn to alternative export providers when a trade partner decreases shipments during conflict, and that trade flows remain relocated even after the end of the conflict. Hence, civil conflicts have persistent negative effects on countries’ standing in international markets. The fourth chapter investigates the local effects of German casualties during World War I. This chapter draws from a new dataset compiled from, among other things, geocoding 8.5 million war casualties and digitizing historical census data. Continuous Difference-in-Differences estimations show that locations with more war casualties specialized on manufacturing employment, as employees lost in the emerging tertiary sector could not be compensated. We trace this vulnerability of the tertiary sector to its reliance on White Collar workers, who distinguished themselves with specific skills and education. Finally, Chapter 5 traces the individual long-run consequences of Kenya’s struggle for independence. In the 1950s, the so-called “Mau Mau” rebellion challenged the British colonial power, and was countered by the indiscriminate incarceration of people into re-education camps that belonged to tribes associated with the rebellion. Using a newly compiled dataset based on colonial records and historical census data, among other things, Triple-Differences estimations provide causal evidence that people that were subject to incarceration are less trusting, wealthy, and educated still 50 years later. This lack of trust can further explain ethnic voting patterns which culminated in the large-scale violence conflict after Kenya’s 2007 elections.