This study intends to estimate the causal effects of human capital investment on economic development.We do so by adopting a novel natural experiment in the early 20th century China called the "Temple Destruction Movement," under which the local governments were required to close Buddha\Tau temples and take over temple assets to construct modern primary schools in the local areas.Our Difference-in-Differences estimates indicate that places with more temple assets ex ante constructed more schools during the movement.In addition, an Instrumental Variable approach provides evidence that regions constructed more schools during the movement period experienced faster economic development later on.