Using Auxiliary Information to Adjust for Non-Response in Weighting a Linked Sample of Administrative Records

• Activity Number: 37
• Type: Topic Contributed
• Date/Time: Sunday, August 11, 2002 : 4:00 PM to 5:50 PM
• Sponsor: Section on Government Statistics*
• Author(s): Barry Johnson*+ and Paul McMahon
• Affiliation(s): Internal Revenue Service and Internal Revenue Service
• Keywords: Bias ; Weight ; imputation ; Record Linkage

Abstract

Federal estate tax returns are a rich source of information on the assets and liabilities associated with a decedent, as well as data on beneficiaries of the estate. When linked with income tax data for the decedents and their beneficiaries, and with data on lifetime gifts reported on Federal gift tax returns, the resulting database provides a unique opportunity to study a variety of important economic issues relating to the transfer of wealth and the accumulation of capital. However, in creating such a complex, linked database, it is inevitable that, for a variety of reasons, a number of records would not be found.

In this paper we detail steps taken to weight these linked data files. We adjust the linked record weights in two stages. First, an adjustment factor is created to balance the original population totals, essentially treating unmatched records as non-respondents. Next, we employ auxiliary data to adjust the sampling weights, then compare these results to estimates from other administrative records sources.