The Social Security Administration (SSA) is interested in giving researchers better access to its administrative data, however, much of this data is covered by confidentiality agreements and cannot be made publicly available. We have created a public use version of the Social Security Administration’s historic earnings records matched to the 1990 to 1993 SIPP panels for the Urban Institute DYNASIM microsimulation model. This paper describes the statistical matching method we used to generate this public use file. It then compares the imputed file with actual administrative data to show that the imputed file has a high degree of similarity to the actual data by cohort, gender, race, education, and marital status.