Many social scientists consider improvements in urban travel facilities during the last decades of the 20th century to have significantly weakened proximity constraints, and therefore think that home-workplace proximity is no longer a major factor in household location choice. They believe that better transport conditions give households more freedom in selecting their homes and workplaces, regardless of the distance separating these two locations. We argue that this point of view underestimates the costs of daily commuting (which remains an increasing function of trip length) and consequently overestimates households' tolerance to commuting. The aim of our research is to measure how much weight households give to home-workplace proximity in actual location choices. Using survey data from the Paris-Ile-de-France metropolitan area, we carry out analyses to estimate two values: (i) the share of home and workplace changes (observed between 1991-2002) which either bring home and workplace closer to each other or keep these two places close to each other, and (ii) the "life expectancy" of job-residence combinations imposing long-distance commutes.