THE CONFIDENTIALITY AND ANALYTIC USEFULNESS OF MASKED BUSINESS MICRODATA

Abstract

In this paper, I discuss the problem of providing microdata on businesses to researchers. These microdata must satisfy two conditions: First, they must provide confidentiality to individual firms. That is, we should not be able to identify the XYZ corporation by looking at the microdata alone or in connection with other data. Second, the microdata must give reliable economic analyses. That is, if we calculate summary statistics or run regressions, we should get similar answers from the microdata as we would get using the entire data base. At last year's ASA meetings I talked about this problem (Spruill {i}). In this paper, I'll quickly review the issues involved, the confidentiality criteria developed, and early work using test data. Then I'll present the new work using actual tax data, discuss some recommendations and conclusions, and talk about where we go from here.