We use a combination of ex-ante and ex-post evaluation methods to evaluate a major welfare policy implemented in France in 1989. The policy granted an allowance (the Revenu Minimum d'Insertion, RMI, of up to 45% of the French full time minimum wage) to every individual above age 25 and below a threshold household income. The ex-post evaluation relies on the specificity of the Eastern part of France. In Alsace-Moselle, since 1908 and during German occupancy, residents benefited from a very similar transfer system (called “Aide Sociale”). Our estimates, based on double and triple differences, show that the RMI policy was associated with: a 3% fall in employment (among unskilled workers 25-55 years old) , leading to an estimated loss of 328 000 jobs; a decline in the job-access rate; and a 5-month increase in the average duration of unemployment. We find considerably larger disincentive effects for single parents. In a second step, we build and calibrate a matching model with endogenous job search effort, using the difference-in-differences estimates. It predicts that, if a 38% implicit tax rate had been maintained as in the 2007 reform (RSA), instead of a 100% implicit tax rate due to the RMI, the increase in unemployment would have been approximately half of its actual value, and the increase in the duration of unemployment would have been limited to only 2.5 months.