Balochistan Province of Pakistan initiated two pilot programs attempting to induce the creation of private schools for poor girls. Randomized assignment to treatment and control groups are used to measure program effectiveness. The pilot schools were successful in urban areas, but relative failures in rural areas. Urban schools benefited from larger supplies of children not served by government schools, better availability of teachers, and more educated parents with higher incomes. Use of experienced school operators in the urban pilot was another critical difference. All urban schools appear self-sustaining or else require a modest subsidy, whereas only one rural school may survive as a private school. These pilots show that private schools may offer a viable alternative supply of educational services to poor urban neighborhoods in developing countries. However, they are not likely to offer solutions to undersupply of educational services to rural areas.