We assess the impact of lowering the cost of learning to drive in France by randomly assigning candidates to either of two groups of 18 to 25 years olds. Young people in the “test group” were given a €1000 voucher to pay for their driving lessons and were suported by a welfare centre throughout the time they were learning. Young people in the “control group” retained all the other welfare benefits for the underprivileged. The vouchers were given to 10 000 young people most of whom were not in education, employment or training. We investigate three types of outcome covering driving, housing and employment status. We analyse the specific role of local support in passing the driving test and we specifically take into account the possibility of spillover effects between treated and untreated individuals.