In 2005, the French government introduced a plan called Aide Complémentaire Santé (ACS). Aimed at households whose living standards were just below the eligibility threshold for supplementary universal medical insurance (Couverture Médicale Universelle Complémentaire: CMU-C), ACS offers a “healthcare voucher” as an incentive to take out additional insurance coverage. Although the number of recipients has grown slowly since 2005, it remains modest. There are two possible explanations: (1) information on the existence of ACS, how it works, and how to apply for it is inadequate; (2) the subsidy is too low, making supplementary coverage too expensive even after deducting the voucher. To test these two hypotheses, the University of Paris-Dauphine conducted a controlled social experiment on a sample of 4,209 insureds potentially eligible for ACS. A randomly selected group was offered a larger subsidy and a different form of access to information on the plan. The results demonstrate robustly that an increase in the “health voucher” slightly improves ACS use and allows better targeting of effectively eligible persons. However, ACS remains a complicated system that has difficulty reaching its target. In all, just 17% of insureds applied for ACS. A mere 9% of the insureds invited to an information meeting showed up, and the invitation largely discouraged the other insureds, nullifying the effect of the voucher increase. Only 55% of insureds who applied were granted ACS. The others were turned down, in most cases because they exceeded the means-test limit. We therefore conclude that, in addition to the complexity of application procedures, access to the plan is inhibited by the difficulty in targeting the eligible population and the significant uncertainty over eligibility that this entails.