In the relatively young sub-discipline of Public Choice, there are a number of theoretical models predicting the type of equilibrium that would be expected in the production of goods and services in the public sector. These models arrive at a number of different conclusions; but one major class of models predicts that public sector output will cater to the demands of the median voter, while another class of models predicts that the amount of public sector output will exceed that level most preferred by the median voter. More information can be generated about the nature of equilibrium in the production of public education than for most other goods and services produced in the public sector, because of the existence of millage issue elections that are held in many states. Two aspects of millage issue ·elections are emphasized in this analysis. First, millage issue elections . act as a constraint on the behavior of school boards, since the school board cannot enact a millage rate without first receiving the approval of the voters. Second, millage issue elections act as a revealed preference experiment, since voters are being asked to state their preferences on the millage issue presented in the referendum. These two aspects of millage issue elections generate information about the nature of equilibrium in the production of public education, and provide the basis for using millage issue elections as a vehicle for studying the nature of equilibrium in the production of public education. A model for interpreting the results of millage issue elections is developed, followed by a survey of institutional differences in public school finance from state to state. Some empirical propositions are forwarded, based on the hypothesis that there is a general tendency for school expenditures to exceed the level most preferred by the median voter. Two empirical analyses fail to support the propositions, but are in agreement with the results that would be expected if public education were being produced at close to the level most preferred by the median voter. / Ph. D.