Using Social Security Data to Estimate Earnings Inequality

Abstract

We use detailed information on labor earnings and employment from Social Security records to document the evolution of earnings inequality in Spain in the last two decades. As is common in administrative records, our measure of labor earnings is top and bottom-coded. We compare the prediction performance of two censoring correction methods, using tax files that are available for the most recent years. According to our results, earnings inequality shows a marked hump-shaped pattern, inversely related to the business cycle. To assess the importance of variation in unemployment rates in the results, we use two different approaches to impute income values to the unemployed. We find that taking unemployment into account magnifies the changes in inequality over the period, although the qualitative pattern remains.