Property taxes continue to be the main source of local revenue in the United States, but the incidence of these taxes is difficult to determine. Because tax incidence is central to evaluating equity, fairness, and the accomplishment of policy objectives, understanding who bears the tax burden is of critical importance. Unfortunately, two characteristics about metropolitan rental markets – renter mobility and heterogeneity of public services – significantly hinder empirical efforts to identify the extent to which landlords are able to shift property tax burdens onto tenants.

Tax incidence in urban rental markets is difficult to pinpoint because the relative mobility of renters allows them to move when the combination of rental prices and services in one location become less attractive than in another. This limits the landlords’ ability to shift tax burdens and has driven most empirical efforts to examine incidence across jurisdictions. Because each community's bundle of public service levels, quality, and costs tax rate determine its property tax rate, however, identification of tax incidence is elusive.

This project takes advantage of a natural experiment to examine, for the first time, the impact of property tax changes on renters within the same jurisdiction. An unusual feature of the property tax in New York State – the Homestead Tax Option (HTO) – and longitudinal rental market data from the American Housing Survey (AHS) create an opportunity to estimate incidence on rental properties in which public services remain constant. Further, because the policy is most likely to affect the least mobile group of renters – those in affordable housing – the ability to escape higher rents by moving is restricted. 

The HTO creates a dual-rate property tax system for residential (homestead) and commercial (non-homestead) property. Mechanically, it ties post-policy property tax rates to a fixed, pre-policy distribution. This drives effective tax rates (ETR) for each property classification in opposite directions as the value and stock of homestead and non-homestead property change over time. This variation creates an exceptional opportunity to study tax incidence on rental housing, because buildings with fewer than four units are classified at the homestead rate, while those with four or more units are taxed at the non-homestead rate. 

The cities of Rochester and Buffalo have adopted the HTO, resulting in the ETR for non-homestead rental property (4+ units) being more than double that for homestead rental property (<4 units) in each city. Because the vast majority of affordable housing in each metropolitan area is in large buildings located within the central city, an opportunity for landlords to shift tax increases onto immobile renters exists.

The research thus employs a difference-in-difference design to test the hypothesis that landlords of non-homestead properties will shift some portion of the increased tax burden onto renters. This analysis utilizes AHS unit-level panel data spanning 15 years, with each unit linked to HTO classification. These data are ideal because they allow for unit-level fixed effects and controls for potentially time-variant unit (e.g., maintenance) and neighborhood characteristics (e.g., new parks) that may affect rental price.