According to transport economics theory, the effect of flexible hours should be that workers spread out their starting times with the result that rush-hour travel conditions improve. But is this really what happens regardless of territorial contexts and transport systems? Paris and San Francisco region are facing serious congestion and are also the French and American cities with their countries' largest share of workers with flexible hours. In those contexts, do employees with the possibility of flexible working hours in fact avoid peak-time travel? Drawing on an international literature review and data from the national household travel surveys for the Paris region and the Metropolitan Area of San Francisco, we reveal new findings that may even appear paradoxical. Far from leading to staggered workplace arrival times, in Paris, flexible hours are associated with a significantly higher probability of arrival at work during the morning peak period. In San Francisco, flexibility has no significant impact on peak period. In an era of flexible schedules, our research tempers the idea of any systematic impact of flexibilisation oriented toward desynchronisation. It reveals that the impact of flexible hours depends on territorial dynamics, where the "social distribution of flexibility" plays an important role.