In this paper, we use a unique two-stage experiment that randomized access to school vouchers across both markets and students in rural India to estimate the revealed preference value of school choice. In the first step of the research design, we develop an empirical model of school choice subject to liquidity and credit constraints that is estimated using data from only the control markets. Based on this exercise, we estimate that the voucher generated welfare gains exceeding four times the average private school's annual tuition on average to the students induced into private schooling. The second step of the research design will validate the estimated welfare impacts by comparing model predictions for a simulated voucher program in control markets with the data from the treatment group. The results in this paper are based on the first step (using only control data) and this draft serves as a pre-commitment to the model estimates and predictions before examining the experimental data.