GATE Working Papers
November 1, 2006
The objective of this paper is to provide microeconomic evidence for the so called “Oswald’s hypothesis”, which is whether homeownership results in negative outcomes in the labour market. To estimate this effect we use two data base, comparing results from British Household Permanent Survey and French part of European Household Panel Survey. In a first step, a multinomial logit model for the choice of tenure status is estimated. Estimated probabilities of being either homeowner, public or private renter are then used to explain the length of an individual unemployment spell. This flexible method of estimation accounts for both censoring and selection bias, without constraining the shape of the hazard rate of leaving unemployment. Results suggested strong differences between French and British household behaviour. Home-ownership has a positive effect on unemployment duration in France but no significant effect is detected in Britain. However we find a positive impact of public renters on unemployment duration in Britain. These stylised facts seems to confirm the existence of a real spillover effect between labour market and housing market.