Journal of Economic Geography Advance Access published online on February 11, 2009
Journal of Economic Geography, doi:10.1093/jeg/lbp002
| ||||||||||||||||||||||||||||||||||||||||||||||||||||
Does public housing occupancy increase unemployment?
*F.R.S.-FNRS and Université catholique de Louvain, CORE, Voie du Roman Pays, 34, B-1348 Louvain-la-Neuve, Belgium. email <claire.dujardin{at}uclouvain.be>
**Université de Lyon, Lyon F-69003; CNRS, GATE, UMR 5824, Ecully, F-69130, France. email <goffette-nagot{at}gate.cnrs.fr>
JEL classifications:: R2, J64
This article shows that living in public housing has no effect on the probability of being unemployed in France, once we account for the endogeneity of public housing. We estimate a simultaneous probit model of unemployment and public housing. On a first sample for Lyon, we instrument public housing with the gender composition of children. On a second national sample, the instrument is the city-level share of public housing. Both samples yield the same conclusion, which we justify by showing that a small amount of selection on unobservables is enough to eliminate the positive effect found in naive estimates.
Keywords: public housing, unemployment, simultaneous probit models, instrumental variables
Date submitted: 12 March 2007
Date accepted: 29 December 2008