Agglomeration over the Long Run: Evidence from County Seat Wars

Cory Smith, University of Maryland
Amrita Kulka, University of Warwick

We study the long-run effects of urban growth using a regression discontinuity (RD) design on close elections that selected "county seats" (capitals) in the frontier United States. An election victory signaled a community's prominence in the mobile frontier period which proved attractive to new migrants. By 2010, election victors are causally 310% larger and have 8.5% higher Zip-level income. The timing of these effects helps characterize when historical shocks matter for long-run outcomes. The effects are primarily driven by elections that occur soon after a county is created, indicating that patterns of economic geography become locked in after a certain period. Economic transitions occur relatively quickly, with most economic and demographic transitions occurring within two decades of an election victory. In contrast to other important historical shocks, county seat status did not involve significant public investments and only transferred a small number of jobs. Changes in government employment and per-person public goods spending are respectively small or indistinguishable from 0, indicating that these mechanisms do not explain the results.

See paper

 Presented in Session 209. Urban Displacement: Local Governance and Conflict