Local employment, poverty, and property value effects of geographically-targeted tax incentives: An instrumental variables approach

a r t i c l e i n f o The federal Empowerment Zone (EZ) program is a set of tax incentives targeted to areas of select cities. I estimate the effect of the EZ program on employment, poverty, and property values by comparing areas that received an EZ to areas that applied (and qualified), but were rejected. Because of endogeneity concerns, I use political representation to instrument for EZ designation. OLS results show a positive and statistically significant effect of the program on employment and poverty. IV estimates suggest the program had no effect on employment and poverty, and instead had a large statistically significant effect on property values. Geographically-targeted tax incentives are an increasingly popular policy for economic redevelopment. The first federal tax incentive with a geographic targeting mechanism, the Empowerment Zone (EZ) program , began in 1995 as the start of a decade-long trend toward using the federal tax code to subsidize areas experiencing economic hardship. In addition to the nine original EZs, Congress established 29 new zones since 1997. The Office of Management and Budget estimates that in terms of foregone revenue, geographically-targeted tax incentives will cost 1 $1. 2 Despite the growing popularity of geographically-based tax incentives among policy makers, there is not a consensus about how they affect targeted populations. I use the federal EZ program to test the effect of geographically-based tax incentives on local employment, resident poverty rates, and median property values. Previous estimates of the federal EZ program, Busso and Kline (2006) and HUD (2001), find large positive effects on employment and large negative effects on poverty rates; however both of these studies treat EZ assignment as strictly exogenous. Krupka and Noonan (2009) estimate a substantial positive effect of the federal EZ program on local property values, a result that is robust to several specifications, some that account for endogeneity of zone assignment. I examine potential for zone designation to be endogen-ous to outcomes of interest and use instrumental variables to provide estimates that correct for the potential endogeneity bias. I estimate the effect of the EZ program with two different methods. The first, an OLS differencing design, compares the relative outcomes for EZ designated areas and their surrounding city with rejected applicants and their surrounding city before and after the program. The primary assumption in the first method is that EZ designation is not correlated with expected changes in …