Panel Paper: Evaluating the Effectiveness of Georgia’s Job Tax Credit

Thursday, November 8, 2018
8222 - Lobby Level (Marriott Wardman Park)

*Names in bold indicate Presenter

Laura A. Wheeler, David L. Sjoquist, Shiferaw Gurmu and Clay Crook, Georgia State University

Proposal Abstract: States regularly use income tax credits to incentivize businesses. However, there is little empirical research focused on whether the credit is successful in creating jobs that would not have otherwise existed. This research proposes to evaluate Georgia’s Job Tax Credit program by measuring the causal effect of the credit on Georgia employment and wages.

The credit subsidizes new jobs created in certain industries, with the value of the credits being based on the number of jobs created and the county in which the establishment is located. Employers located in counties ranked as most economically distressed receive a tax credit equal to $3,500 while employers in the least distressed counties receive a tax credit of $750 per new employee. Once a firm has met the qualifications for the tax credit, the credit is awarded each year for five years. Counties are ranked from one to 159 based on an equally weighted average of the unemployment rate, the per capita income, and the number of residents with income below the poverty level.

This research evaluates the credit using separate macro- and a micro-level analyses. The first level of analysis considers the macro-level implications of the credit by focusing on the county employment outcomes. We use changes in employment at the county level as a metric by which to measure success of the credit. This approach captures possible spillover effects between firms that take the credit and those that do not. The county-level approach uses a sharp regression discontinuity design to explore the degree to which counties with higher tax credits experience increases in employment or wage growth compared to their lower-tax credit neighbors. This model uses county-level employment and wage data and standard control variables for the 1998-2015 period.

For some economically distressed census tracts, the value of the tax credit is higher than those available in the remainder of the county. The presence of these census tracts creates an additional level of geographic diversity that can be exploited by our estimation technique. To take advantage of this diversity, we run a set of regressions on these census tracts and their neighbors which mirrors the county-level models.

Although the county and census tract analyses have some benefits, they also have several limitations. First, because they are based on aggregate data, they cannot directly model the effect of the tax credit. Second, an insignificant coefficient on the tax credit variable may be the result of offsetting employment gains and losses or it may be an indication that the incentive has no effect. To address these shortcomings, we employ a micro-level analysis as the second level of analysis. This firm-level analysis enables measurement of the elasticity of hiring with respect to the wage subsidy and provides clarity on the impact of the incentive. The firm-level approach uses a fuzzy regression discontinuity design to explore the degree to which firms which take the tax credit experience increases in employment or wage growth relative to similarly-situated firms that do not take the credit.