Many studies in Anglo-Saxon countries report lower wages for the homosexual population, with given observable production characteristics. This gap suggests the presence of discriminatory behaviour, although it is not sufficient to prove this formally. In France this issue has not yet been the subject of any statistical study, due to a lack of suitable data.This article attempts to fill this gapusingdata from the Labour Force Survey, by selecting individuals cohabiting with a person of the same sex which they declare to be a“friend”. This approach, the only one possible to date,is indirect and partial in nature: it excludes homosexuals who live alone and therefore only targets part of the population of interest. Conversely, it may include individuals who live together with a person of the same sex for reasons other than sexual orientation. Additional filters are provided to try to limit this bias; they consist of excluding individuals who are most likely to live together for mainly economic reasons (students, pensioners, people with low wages, etc.) or due to migration. An econometric approach is then used to compare the wages of this group with those of the rest of the population. It highlights a negative differential for men which cannot be confined to the production characteristics observed and some of which could therefore be interpreted as discrimination. It is around - 6 % to - 7 % in the private sector and- 5% to - 6% in the public sector. No such differential is observed however for women ; this asymmetry is consistent with results of studies in other countries. Qualifications are not sufficient to avoid this wage disadvantage:in the private sector, the gap is even wider among skilled workers than unskilled workers. It also increases with age.