Projects

Measuring Racial Inequality in Tax Data

A major limitation faced by researchers working within the Internal Revenue Service (IRS) data infrastructure is the absence of any information on race and ethnicity. The same limitation also affects researchers who use tax data in other environments, such as the U.S. Government Accountability Office and the U.S. Department of the Treasury. Additionally, the lack of race and ethnicity measures is an important constraint for researchers using other administrative data—for instance, Social Security Administration information on earnings or state-level information on unemployment insurance income—to carry out research on economic inequalities. This project proposes the application, validation, and extension of a new method that addresses the problem of lack of self-reported race/ethnicity in tax and other administrative data. It will take advantage of a unique opportunity to access restricted data through a new collaboration with researchers at the U.S. Census Bureau. Through an internal Census project, the project will be able to produce gold-standard validation studies of the method’s ability to estimate racial and ethnic inequalities in economic outcomes with data lacking self-reported information on race/ethnicity.

Pablo A Mitnik