Thursday 6 August 2015

Better residential than ethnic discrimination! Reconciling audit andinterview findings in the Parisian housing market

François Bonnet (CNRS, UMR Pacte), Etienne Lalé (Department of Economics, University of Bristol), Mirna Safi (Department of Sociology and OSC, Sciences Po) and Etienne Wasmer (Department of Economics and LIEPP, Sciences Po)

Abstract can be found: 

We started out this research with two questions. First, is there a banlieue effect operating in the Parisian housing market, i.e. does living in a deprived neighborhood per se undermine the prospects for residential mobility? Second, is it possible to separate this effect from discrimination caused by another potential stigma, namely the ethnic origin (North-African background) of those living in the French banlieues? Answering these questions is of importance for a broad range of academics studying discrimination and the cumulative effects of residential and ethnic/racial inequalities. The French context makes this task all the more challenging and interesting. Indeed, the French Republican model embraces a color-blind, universalistic model of race relations, which notably implies rejecting ethnicity as the basis of categories for official statistics. Recent studies, partly in the aftermath of the 2005 urban riots in France, have challenged the myth of a color-blind society.[i] They provide evidence of discrimination in the labor market. However, at present, there is almost no audit study based evidence of discrimination in the housing market in France.

A distinctive feature of our research is to make complementary use of two methodological designs. We conducted: (i) an experimental paired-testing audit study involving (fictitious) housing applicants and (ii) a series of face-to-face interviews with real-estate agents in Paris and the Paris region. Thus, our research has potential of offering both statistical and discursive evidence on discrimination in the housing market. After juxtaposing the findings from the audit and the interviews, we do find an interesting paradox:

  • While the current (alleged) residence of housing applicants has a significant negative effect in the audit, real estate agents clearly deny its relevance as a discriminatory factor affecting access to housing;
  • Real estate agents overwhelmingly report that ethnic origin has a discriminatory impact, whereas ethnic origin has no significant effect in the audit when we control for the current residence of housing applicants.
We offer several hypotheses that may solve this apparent paradox. We especially favour the following two hypotheses which, in our view, should not be seen as mutually exclusive.

First, the findings could reflect statistical discrimination whereby real estate agents seek to proxy the risk of non-payment of the rent (insolvency). In particular, residential and ethnic origins could correlate with access to housing simply because real estate agents use these characteristics to infer the risk that truly matter to them. Real estate agents deny the relevance of residential origin as a discriminatory factor with good reasons, in that only insolvency is relevant. In the meantime, if residential origin is strongly correlated with the risk of insolvency, then controlling for this variable explains why ethnic origin does not play a statistically significant role in the audit study data.
The other hypothesis is that there is an ethnic stigma and that residential origin is used to proxy ethnicity. That is, it may be that overt information about ethnic origin (like, for instance, an African name) is not used to discriminate against housing applicants because such biased decisions appear highly undesirable. Strong correlates with ethnicity (like, for instance, residential origin) are instead used to select housing applicants. This would explain why real estate agents report that residential origin is not a discriminatory factor while ethnic origin is. This would also help understanding why, in the audit study, overtly signalling ethnicity through the housing applicant's name did not result in discriminatory behaviours.

The more important conclusion of our article is that the complementary use of different methodological designs helps overcome the shortcomings of each. In this respect, the paradox we find is anything but a weakness of this research: it proved very instrumental in developing rich hypotheses to understand discrimination.

[i]      To our knowledge, the most comprehensive study is the audit conducted jointly by the ministry of labour and the International Labour Organization in 2008. The results, presented in a report written by Eric Cediey and Fabrice Foroni (“Discrimination in access to employment on grounds of foreign origin in France: A national survey of discrimination based on the testing methodology of the International Labour Office”), reveal  that employers prefer candidates perceived as being of “national” origin (French) to strictly identical candidates of African immigrant background in 80% of all cases.

No comments:

Post a Comment