Localised unemployment rates are constructed from three sources in metropolitan France:

• the Employment survey, which provides the number of unemployed and employed persons in the ILO sense;

• administrative data on employment from the social declarations of companies and the self-employed, which make it possible to exhaustively and finely localise employment at the level of the territories;

• data on jobseekers registered at the end of the month with Pôle emploi in category A (DEFM A), which makes it possible to pinpoint the location of unemployed people. This last source differs in level from the results of the Employment survey, since it does not follow the ILO concepts implemented in the survey. But, assuming that this difference is evenly distributed, its structure can be used to break down unemployment from the Employment survey.

The calculation of localized unemployment rates in the DOM excluding Mayotte is based on a specific methodology, leading to a statistical hazard in the short-term variations. In order to study the structural and long-term effects, the series of unemployment rates in annual average are to be preferred in the DOM excluding Mayotte.