This paper analyzes the impact of voter-approved school bond issues on school district balance sheets, local housing prices, and student achievement. We draw on the unique characteristics of California's system of school finance to obtain clean identification of bonds' causal effects, comparing districts in which school bond referenda passed or failed by narrow margins. We extend the traditional regression discontinuity (RD) design to account for the dynamic nature of bond referenda, since the probability of future proposals depends on the outcomes of past elections. By law, bond revenues can be used only for school facilities projects. We find that bond funds indeed stick exclusively in the capital account, with no effect on current expenditures or other revenues. Our housing market estimates indicate that California school districts under-invest in school facilities: passing a referendum causes immediate, sizable increases in home prices, implying a willingness-to-pay on the part of marginal homebuyers of $1.50 or more for each $1 of facility spending. These effects do not appear to be driven by changes in the income or racial composition of homeowners, and the school bond impact on test scores cannot explain more than a small portion of the total housing price effect. Our estimates indicate that parents value improvements in other dimensions of school output (e.g., safety) that may be not captured by test scores.