that UI wage records typically do not include the physical establishment for workers employed by multiunit businesses. For these workers, only the identifier for the multiunit business is recorded. The LEHD program staff developed a multiple imputation technique to assign a place of work to these individuals based on the size and hiring patterns of establishments within the multiunit business and the relationship between the place of residence of each worker and the location of each business. This is an example of using imputation to improve the quality of data. Imputation can therefore be useful for two purposes—when there are missing data and when confidentiality must be protected.

The potential benefits of more efficient data sharing among the U.S. statistical agencies have long been recognized. The National Research Council’s 1993 report, Private Lives and Public Policies: Confidentiality and Accessibility of Government Statistics, included a section on “Barriers to Data Sharing within Government” in which the authoring panel wrote (p. 6):

The panel concluded that barriers to sharing data on both persons and businesses for statistical purposes should be removed subject to strict controls to protect confidentiality. Recommendation 7.4 of that report articulates the panel’s position that interagency data coordination should include the sharing of lists of businesses by federal and state agencies for statistical purposes. Recommendation 7.5 asked for new legislation to expand confidentiality to records collected by all statistical agencies; some of this recommendation was subsequently accomplished by CIPSEA. Although this report cited several examples of reduced sharing as a result of the provisions