Estimating suppressed data in regional economic databases: A goal-programming approach

To avoid disclosure of individual establishment information, data records may have to be suppressed in regional economic databases, with values represented by flags. This paper investigates this suppression process and presents a goal-programming optimization approach to estimate these flagged data, using the 2000 County Business Patterns (CBP) database as a case study. The approach minimizes the sum of weighted deviations between the estimates and target values, subject to constraints related to county and sector total employment, as well as to flag and establishment size intervals. The model is tested using Ohio and Arizona data, for both sources of inconsistencies and parameter selection. A decision-theoretic analysis of the test results points to specific strategies that yield the best estimates of the suppressed data.