This article aims at providing a test of the spatial mismatch hypothesis on the Bordeaux metropolitan area. Starting with a theoretical survey of the complex links between residential segregation and local labour markets, we then propose a framework allowing for a better understanding of the impact of physical distance on spatial mismatch. The results (at the quarter section level – French IRIS 2000-) confirm the existence of two different effects of spatial friction depending on the distance considered. Furthermore, the results provide some evidence of the existence of a spatial hysteresis effect affecting specifically long-term unemployment.residential segregation, spatial mismatch hypothesis, unemployment.