Locating public facilities : theory and micro evidence from Paris

This paper investigates the problem of the optimal location of public facilities. I develop a quantifiable model in which the central planner decides on a location strategy, which includes the geographical location and the capacity of public facilities, while anticipating how individuals and firms will react. The central planner's objective is to maximize aggregate welfare. I calibrate the model to fit the economic and geographic characteristics of the Paris metropolitan area at a 1km x 1km geographic resolution and focus on secondary schools as an example of public facilities. The counterfactual analysis, which compares the optimal and observed location strategy between 2001 and 2015, suggests that adopting the optimal strategy in any year would have increased welfare growth by about 12%. Half of the effect is attributable to improvements in channels other than shorter commutes to the public facility, mostly via lower housing prices and shorter commutes to the workplace. The analysis also reveals that the observed location strategy disproportionately favored short commutes in central locations and led to a mis-allocation of residential and commercial activities between the center and the periphery.