Discriminations fondées notamment sur l’origine dans l’accès aux biens et services sur plusieurs plateformes collaboratives

This project is being carried out in partnership with the Défenseur des droits, a department of the French Prime Minister.

Project holder:

• Morgane LAOUENAN (Université Paris 1, CNRS, LIEPP)

Research team:

• Guillaume Chapelle (CY Cergy Paris Université, LIEPP)
• Pierre Deschamps (SOFI, Université de Stockholm, LIEPP)
• Dylan Glover (INSEAD, LIEPP)
• Xavier Lambin (ESSEC)

Project description:

The aim of this study is to describe the phenomenon of discrimination on online platforms in France, and to propose a methodology that goes beyond simple descriptive analysis in order to gain a better understanding of the mechanisms behind any discrimination.

To carry out this assessment of discrimination based on origin in access to goods and services, we have chosen to focus on two leading French collaborative platforms: a car-sharing platform that connects drivers with free spaces in their cars and passengers (platform A) and a platform that allows users to sell and buy several types of goods (platform B).

Unlike Airbnb, few studies have assessed discrimination on these two platforms.

It may be interesting to assess discrimination on both sides of the market. These two platforms allow us to look at the behaviour of both drivers/sellers and passengers/buyers.

The characteristics of these two platforms allow us to identify certain new mechanisms (discrimination based on payment/interaction security for platform B and discrimination based on reputation for platform A).

The large volume of transactions enables us to study several underlying mechanisms of discrimination (according to the level of security, interaction, competition).

The preliminary results of our study can be summarised as follows. Firstly, concerning platform A, analysis of the scraping data suggests that drivers of North African and African origin receive lower incomes for similar journeys. This result is not confirmed by the testing, where we find no detectable difference between drivers of different origins, but an advantage for women who have a higher probability of receiving a reservation on their journeys. Can we therefore conclude that there is no discrimination against drivers of North African or African origin on platform A? The results of the experiment are not very precise and do not allow us to reject discrimination against these drivers on the same scale as that observed during the scraping. The imprecision of the results may be linked to greater variability in demand, linked to COVID, and to problems in setting up the experiment. The results of the experiment at this precise point do not therefore allow us to confirm or invalidate the results of the scraping.

The second part of the experiment on platform A looks at driver behaviour. Do drivers on this platform respond more often to requests for clarification from women or passengers of European origin? Gender seems to be particularly salient: female drivers respond less often to requests for clarification, while female passengers receive answers to their questions more often. The origin of the fictitious passengers, on the other hand, does not seem to have any effect on the response rate.

The results of the study on platform B seem more surprising. Analysis of the data collected shows that the adverts of people with a first name of African or North African origin last longer. This may suggest the existence of unfavourable behaviour towards members of this group. However, the results of the testing experiment reveal that minorities receive a significantly higher number of SMS and email contacts for their ads. At this stage, it is still difficult to explain this result and its interpretation remains delicate. In particular, it should be remembered that these contacts with potential buyers do not constitute formal sales and discrimination could persist when the sale is finalised. In addition, the trial focused on a limited number of properties. We are currently continuing the experiment to gain a better understanding of the results.

Project results:

Read a summary of the results:  Les discriminations en raison du genre et de l’origine supposée sur deux plateformes collaboratives, Défenseur des droits, mars 2023.

Consult the study report: Les discriminations en raison du genre et de l’origine supposée sur deux plateformes collaboratives, Sciences Po LIEPP Report, october 2022.

Poster of the project - June 2023