The median voter hypothesis (MVH) represents a much-used as well as much-criticized tool in the hands of public finance economists. To evaluate the MVH, this paper applies the Cox specification test using data from general purpose municipal governments. The Cox test allows for possible simultaneous rejection of the MVH and all tested models, thus providing stricter criteria than applied to date in empirical MVH analyses. The test results reveal that the MVH is appropriate for explaining the aggregate behavior of municipal governments, but not specific services. Further, the single service and single tax base assumptions are not critical to the empirical MVH performance.