We study the stability of regional population structure on exogenous temporary shocks. The analysis exploits a quasi-experimental setting caused by the World War II, which led Finland to relocate 11%
of its population. We show that the population distribution was not stationary nor followed a random walk. Instead, the results suggest that a temporary shock set off a self-reinforcing process, where municipalities receiving the largest flows of displaced persons grew faster also in later periods. These findings are robust to using two distinct sources of exogenous variation and to controlling for previous growth trend, industry structure and geographical indicators.

Joint paper with Aki Kangasharju (Government Institute for Economic Research (VATT), Finland), January 2009