This article explore how the relation between economic shocks and local unemployment can be mitigated by labor market size. We exploit a quasi-natural experiment by studying the economic impact of 357 local shocks both negative and positive generated by the reform and the restructuring of the French army. Exploiting a geo-referenced dataset of unemployment spell over an extensive period of time (2005-2014), we are able to measure the impact of these local shock on the rate at which unemployed workers find a job. To construct a credible counterfactual for each zone which experienced a closure, we use an interactive fixed effects model. We show that contractions in military personnel reduce the local likelihood of finding a job. Moreover, our results reveal some heterogeneity in the local economy's resilience. In line with our theoretical model, we show that city size is a relevant explanation for the observed heterogeneity in resilience: the likelihood of finding a job is less affected in denser area by a relative equal-sized shift in labor demand.