Lost but Not Forgotten: An In-Depth Analysis of Attrition in the 1999--2011 Individual Income Tax Return Panel

• Activity Number: 75
• Type: Contributed
• Date/Time: Sunday, August 3, 2014 : 4:00 PM to 5:50 PM
• Sponsor: Government Statistics Section
• Author(s): Victoria Bryant*+ and Jessica Holland
• Companies: Internal Revenue Service and Internal Revenue Service
• Keywords: Attrition ; Panel data ; Administrative data ; imputation

Abstract

Policy research increasingly uses panel data because of their ability to examine causal relationships. However, missing data can undermine validity and misrepresent results. Using a panel of Individual Income Tax Returns for Tax years 1999-2011, this paper will explore the various forms of data loss and their implications. Preliminary results suggest that of the 177 million taxpayers represented in 1999, by 2010 only 140 million taxpayers remained. This paper will expound to investigate the composition and attributes for missing data to better understand the disappearance, persistence and potential for reappearance. Because this panel is drawn from administrative records the very nature of their absence calls into question the unbiasedness of analysis. We will thus assess the observed return drop rate over time to identify possible correlations to larger economic cycles. Additionally, we will explore the advantages and disadvantages of imputing missing information from various other sources of income tax data. We will then discuss the implications of this data imputation method on the panel and the potential for its application to other similar panels.