This paper examines the effects of collectively agreed increases in real minimum wages on employment transitions and hours among manual workers in the Swedish retail sector over the period 200105. The findings indicate that increases in real minimum wages are associated with more separations, whereas hours are less affected because separated workers put in relatively fewer hours before being separated. Among the young, however, both employment and hours are negatively affected. Labour-labour substitution seems to be important, since increases in minimum wages promote employment among workers with higher wages than those directly affected by the increases. The assumptions of the econometric model were tested by imposing fictitious minimum wages on lower-level non-manuals in the same industry, with turnover characteristics similar to manuals but covered by a different collective agreement with non-binding actual minimum wages.