Working Paper: CEPR ID: DP18209
Authors: Konstantin Sonin; David Van Dijcke; Austin L. Wright
Abstract: The massive violent protest at the US Capitol on January 6, 2021 was a rare event for a mature democracy. We investigate its drivers using granular location data from 40 million mobile devices. Leveraging a novel approach for estimating spatially dispersed protest participation, we show that political isolation amplified the effect of partisanship on participation. Mobilization also increased sharply in states with narrow Trump losses and in counties with a Trump-to-Biden swing in the election-night voter tally. The latter effect was driven by isolated communities, consistent with a model in which individuals in such communities are relatively more sensitive to information from their preferred sources. Our findings shed light on the broad factors and specific triggers that result in violent collective action.
Keywords: collective action; protest; elections; regression discontinuity; geospatial analysis
JEL Codes: P00; D74; D72; C31
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| political isolation (P33) | participation in the January 6 protests (K16) |
| partisanship (D72) | participation in the January 6 protests (K16) |
| political isolation + partisanship (D72) | participation in the January 6 protests (K16) |
| communities with more Biden-supporting neighbors (R23) | participation in the January 6 protests (K16) |
| perceptions of electoral unfairness (K16) | participation in the January 6 protests (K16) |
| electoral outcomes (K16) | participation in the January 6 protests (K16) |