Analysis of Neighbourhood Effects and Work Behaviour: Evidence from Paris

Abstract

This paper highlights the effects of being located in a deprived neighbourhood on unemployment. Interest is focused on the consequences of neighbourhood effects. The paper uses the 1999 Population Census for Paris and the three surrounding sub-regional administrative districts in order to estimate different models that take into account the potential endogeneity bias of the residential location choice. The study first runs a bivariate probit model that includes the residential location as an endogenous variable. A probit model is also run on a sub-sample of households living in public housing with the idea that for them the location choice is exogenous. Whatever the method used, it is shown that living within the most deprived neighbourhoods, in terms of local composition, decreases the probability of employment.

Keywords:

• Unemployment
• residential segregation
• neighbourhood effects
• endogeneity bias
• housing choice
• housing economics

Acknowledgements

The author would like to thank anonymous referees for their comments and suggestions that have greatly improved this paper. The author would also like to thank Samia Benallah, Oana Calavrezo, Yannick L'Horty, Mathieu Narcy, Sanja Pekovic, Patrick Sillard, participants of the internal seminar of CEE, participants of the internal seminar of Evry (EPEE), participants of the Spring Meeting of Young Economists 2009, participants of the 2nd Doctoral Meeting of Montpellier, participants of the Journées de Microéconomie Appliquée (JMA) 2009 and participants of the Association Française de Science Economique (AFSE) 2009 for useful remarks and comments that have greatly improved this paper. All remaining errors are those of the author. Census data as the Iris level have been provided by the Centre Maurice-Halbwachs

Notes

¹ Co-ordinates on the first axis vary from − 4 for the most favourable to +11 for the most deprived.

² The creation of such a variable is necessarily because in the case of bivariate probit and probit regressions explained variables have to be dummy variables.

³ For example, the Gautreaux Program or the Moving To Opportunity Program in the United States (see Katz et al., 2001; Kling et al., 2005 for more details).

⁴ The probit model is a specification adapted for a binary response model which employs a probit link function. This model is generally estimated using maximum likelihood procedure. The idea behind maximum likelihood parameter estimation is to determine the parameters that maximise the probability (likelihood) of the sample data. The advantage of the method of maximum likelihood is that it is considered to be more robust and it yields estimators with good statistical properties.

⁵ Nevertheless, the last two methods present very close results.

⁶ A marginal effect of an independent variable x is the partial derivative, with respect to x, of a prediction function f(x).

⁷ The methodology is as follows: first an equation is run to explain the fact of living in a deprived neighbourhood with the instrumental variables. Then, predicted values of this first-stage equation are simply squared and cubed and introduced in the second equation.