Redistribution program in developing countries often "leak" because local officials do not implement programs as the central government intends. We study one approach to reducing leakage. In an experiment in over 550 villages, we test whether mailing cards with program information to targeted beneficiaries increase the subsidy they receive from a subsidized rice program. On net, beneficiaries received 26 percent more subsidy in card villages. Ineligible household received no less, so this represents substantially lower leakage.