For better or worse, fiscal decisions made through property tax referenda allow local political markets to work. Demand, supply, and voting process components of such markets are estimated for those Oregon K-12 school districts that held referenda between 1981 and 1986. Various attributes of the median voter were related to school spending, but supply decisions by school boards and administrators were also important. Large districts used state aid to substitute for local property tax revenues on nearly a one-for-one basis, while relying on reversion budgets (inadequate property tax bases and implicit threats of school closures) to extract greater-than-desired spending levels from the median voter.