Using an inventory of local and/or non-statutory transfers in thirteen French towns and cities including Paris, Lyon and Marseille, we measure the gains from returning to work for recipients of means-tested benefits (RMI and API) by type of household. The reforms of national and statutory benefits carried out during the 2000s, especially those affecting the employment tax credit (PPE), failed to ensure that the recipients of meanstested benefits always stood to gain from returning to work. In most localities and for most types of household, taking a part-time job on the minimum wage represents a loss of income compared to staying on welfare, and full-time work does not always present any advantage. The effects of the reforms were offset by the effects of other measures such as the widespread provision of travel assistance by regional councils, the development of social tariffs for telephone and electricity, or exemption from paying the television licence fee. We then simulate the effects of introducing the RSA in place of the RMI. We take into account the way that local and/or non-statutory transfers ("droits connexes") are modified by the increases in national transfers. We observe that the RSA does indeed eliminate those areas where returning to work is unprofitable, for almost all localities and types of household. Lastly, we show that the marginal tax rate of 38 % chosen by the government in its bill is very close to the upper limit compatible with a back-to-work incentive.