Download in PDF format         Get free software to read and download PDF files.

     WASHINGTON, D.C. – Today Vice Chairman Jim Saxton of the Joint Economic Committee (JEC) released the text of a letter he sent to Treasury Secretary Summers last month. Saxton pointed out that the release of Treasury tax distribution statistics does not comply with normal standards of disclosure. These standards are set forth in federal rules by OMB Circular No. A-130, which states in part: "Agencies should inform the public as to the limitations inherent in the information dissemination product (e.g., possibility of errors, degree of reliability, and validity) so that users are fully aware of the quality and integrity of the information."

Dear Secretary Summers:

As you know, over the years the Joint Economic Committee (JEC) has conducted research on a wide variety of economic topics including statistical and technical issues related to employment and unemployment, the Consumer Price Index, household and family income, monetary policy, the International Monetary Fund, and other issues. I am writing today in reference to conceptual and technical issues raised by the Treasury's tax distribution tables prepared by the Office of Tax Analysis (OTA).

A review of these Treasury distribution tables by the JEC was unable to find disclosure of common data limitations such as standard errors and measures of statistical reliability. Further inquiry suggested the possibility that the OTA itself may not have such information because it is not possible to derive it from statistics that include imputations and data tabulated from multiple sample sources. In other words, it appears that the mixing of data from different sources and the reliance on imputation may make it impossible to determine the statistical reliability of the data.

Would you please confirm whether it is correct to state that the Treasury's tax distribution tables do not include the disclosure of standard errors and measures of statistical reliability? Furthermore, if so, would you also please explain why these disclosures are not made? Is it possible to derive such measures, or does the nature of the data preclude this?

Given the central role the Treasury's tax distribution tables often play in highly charged policy debates, it is important that everyone dealing with them have a clear understanding of these statistical issues. Imputation of income in past discussions, for example, has often proven to be fairly controversial. While Henry Simons noted that imputation was necessary to construct measures based on the Haig-Simons income concept, he himself described the process of imputation as "highly conjectural."

Publication of data limitations would inform the public of the reliability of the statistics including these conjectural elements. The mixing of data from various samples also raises questions that would be clarified by publication of data limitations. Moreover, the publication of the data limitations of the distribution tables would also be consistent with the concept of transparency and with federal rules governing the dissemination of data from government agencies.

Thank you for your help in clearing up these questions.

                                                   Sincerely,
                                                    Jim Saxton
                                                    Vice Chairman
                                                    Joint Economic Committee

      For more information on taxation and related issues, please visit our website at www.house.gov/jec.

Press Release: #106-74

Return Home