Affiliations: Statistics of Income Division, Internal Revenue Service, Washington, DC 20224-0002, USA
Correspondence:
[*]
Corresponding author: Yan K. Liu, Statistics of Income Division, Internal Revenue Service, 1111 Constitution Avenue, NW, K-Room 4112, Washington, DC 20224-0002, USA. E-mail: [email protected]
Abstract: The Statistics of Income Division (SOI) of the IRS started a panel sample of individual income tax returns for Tax Year 2007 for longitudinal analyses. SOI edits Sales of Capital Assets (SOCA) transactions reported on individual tax returns for this panel. The panel sample, together with a small, yearly refreshment sample has also been used for cross-sectional SOCA estimations. This has allowed SOI to get annual SOCA estimates at a small expense. The cross-sectional weights are derived using weight calibration method. The budget allows SOI to select and edit a cross-sectional SOCA sample once every 5 years, most recently for Tax Year 2012. This paper compares the 2012 SOCA estimations based on the panel sample to those based on the 2012 cross-sectional SOCA sample and evaluates the current sample and weighting procedure for the purpose of cross-sectional SOCA estimations. It also provides simulation scenarios to see the balance between a larger, yearly refreshment sample and the associated efficiency gain in SOCA estimations.
Keywords: Calibration weighting, cross-sectional estimation, panel sample, R statistical software
