Erin Adele Scharff and Darien Shanske

Volume 74, Issue 3, 823-868

Traditional theoretical literature on fiscal federalism urges cities to finance themselves with taxes on immobile sources. Thus, the literature sees real property taxes as the best source of local revenue; real property, after all, cannot be easily moved. This same literature eschews local income taxes because it is easy for a taxpayer to move, thus allowing exit from local income tax obligations.

Practice here seems to follow theory: cities do not tend to levy income taxes. However, this general trend has caused scholars to overlook important exceptions. In fact, many cities impose income taxes and have for a long time. In this Article, we argue these exceptions are not a vestigial mistake. The persistence of local income taxes suggests that the traditional fiscal federalism view is too absolute. Rather, local income taxes can play a useful role in the municipal revenue toolkit, provided that such taxes are not too administratively burdensome and levied at a low rate that considers the city’s competitive position.

In this Article, we provide an account of local income taxes in practice. Because the existence of these taxes has been obscured by the theoretical dismissal of local income taxation, this account itself is a contribution to the existing local finance literature. We then argue that the traditional skepticism of local income taxes should be tempered. Evidence from agglomeration economics suggests some limits on the traditional mobility story; agglomerations make exit difficult for many industries and professions. Moreover, to the extent these agglomerations enable a segment of the population to earn more, it can be both fair and efficient to raise money through local income taxes.

We make this case for local income taxation well aware of increasing skepticism of local tax authority in an era of rising remote work. But the move to increased remote work does not doom local income taxes, either as a practical or legal matter. Remote work will not eliminate agglomerations so much as shift where they occur.