Discrimination in recruitment is a significant barrier to the labour market participation of people from migrant backgrounds – a form of participation that is all the more crucial as it is a key factor in integration. This study compares the success rates of identical CVs in which only the names have been changed. The results show that, for identical applications, a candidate from an immigrant background receives 40 per cent fewer invitations to job interviews than a candidate identified as ‘native-born’. For female applicants, this discrimination disappears when a skill demonstrating a good command of the French language is added. These findings pave the way for new tools to combat discrimination that improve the information available on a candidate’s skills, such as a certificate of proficiency in French.
Discrimination in recruitment in France: findings and courses of action
Presentation
Key Results
- For identical applications, a candidate from an immigrant background (from the Maghreb or of an origin not clearly identified by employers) receives 40 per cent fewer invitations to job interviews than a French candidate identified as ‘native’.
- Discrimination on the basis of origin is more pronounced against men than against women.
- For female applicants, discrimination on the basis of origin disappears when they demonstrate a good command of the French language.
Method and Data
A study by R. Aeberhardt, E. Coudin and R. Rathelot (2010), researchers at CREST affiliated with the IPP, sought to distinguish the explainable portion of differences in employment rates between ethnic groups from the unexplainable portion, which includes discrimination. The objective characteristics taken into account include both factors that are directly valued by employers (such as qualifications or professional experience) and factors linked to the decision to enter the labour market (family circumstances or a spouse’s salary, for example).
These differences in characteristics account for up to a third of the observed differences in labour force participation rates, whilst the residual differences correspond to unexplained variations.
To isolate the proportion strictly attributable to differential treatment of individuals based on the group to which they belong, it would be necessary to compile and measure a comprehensive list of all individual characteristics likely to affect the employment rate. Measuring the extent of actual discrimination is therefore by no means straightforward.
To overcome these difficulties, researchers have developed an original method based on the controlled submission of fictitious applications in response to job offers. This approach, known as ‘Testing’, involves responding to job offers using fictitious applications (CVs and cover letters) designed to present qualifications that match the profile sought. This database of applications is linked to a list of identities – which are also fictitious. Each identity therefore corresponds to a fictitious candidate whose first name and surname have been chosen in advance. These identities constitute the main variable of interest: they are designed to suggest as clearly as possible that the ‘candidates’ belong to the groups against which the level of discrimination is to be measured. As many applications as there are identities in the study are sent in response to each job advertisement. When these applications are submitted, the CV and cover letter differ from one candidate to another in order to minimise the risk of the study being detected.
The study focuses on the accountancy sector and, more specifically, on the roles of accounting assistant, accounting clerk and accountant. The results are based on responses to 504 job offers in the Île-de-France region, collected between September 2011 and February 2012.
