Explaining Sub-national Variation in Pogrom Violence: The Case of Kristallnacht

• In Event: Time in a Bottle: Analyzing Historical Conflict Data

Abstract

Why do pogroms result in violence in some localities, but not in others? Studies of pogrom violence offer multiple explanations for variations in anti-civilian attacks. These studies have highlighted the particular influence of rumors and misinformation, patterns of interethnic integration, and territorial and electoral contestation on pogrom outcomes. However, the pogrom literature fails to explain variations in a crucial case: the Kristallnacht (“Night of Broken Glass”) pogrom against Germany’s Jewish population from 9 – 10 November 1938. The Kristallnacht pogrom resulted in the destruction of thousands of synagogues, the deaths of dozens of German Jews, and the first mass deportations of Jews to Nazi concentration camps. The events were a critical juncture in the Nazi genocide against European Jewry.

I argue that revealed pre-violence support for anti-minority groups explains spatial variations in anti-civilian attacks during a pogrom episode. A pogrom, a relatively brief episode of selective violence involving mass, collective participation by non-state forces, has high social and emotional costs for its participants. Would-be participants confront a basic information problem: to enact the intimate, selective violence that characterizes the pogrom, participants must know that the pogroms will achieve a sufficient level of collective participation in violence to outweigh the social and emotional costs. To succeed in mobilizing violence at scale, pogrom organizers must find ways to overcome these social barriers to collective action. Revealed political identities solve this problem by signaling a locality’s collective willingness to sustain the social costs of violence.

To evaluate this hypothesis, I introduce a new geocoded dataset of more than 2,000 Jewish houses of worship in pre-1938 Germany, based on archival information gathered by a Jewish memorialization project. The dataset demonstrates significant subnational variation in levels of violence during the Kristallnacht pogrom. These descriptive findings challenge most historical narratives of the November violence, which present the state-directed violence as all but totalizing in its direct effects on Germany’s Jewish population. I use a multiple regression analysis, clustered at the municipality (gemeinde) level, to test the effects of Weimar-era Nazi Party vote share on spatial variations in pogrom violence.

Author

• Daniel Solomon