Using Matched Household and Administrative Data to Measure Response Bias in Earnings

• Activity Number: 120
• Type: Topic Contributed
• Date/Time: Monday, August 4, 2014 : 8:30 AM to 10:20 AM
• Sponsor: Business and Economic Statistics Section
• Author(s): Barry Hirsch*+ and Christopher Bollinger and Charles Hokayem and James Ziliak
• Companies: Georgia State University and University of Kentucky and U.S. Census Bureau and University of Kentucky
• Keywords: non-response bias ; measurement error ; earnings ; CPS ASEC ; hot deck imputation

Abstract

Earnings non-response in household surveys is widespread, yet there is limited evidence on how response bias affects measured earnings. This paper examines the patterns and consequences of non-response using internal Current Population Survey worker records matched to administrative data on earnings for 2005-2010. Non-response across the earnings distribution, conditional on covariates, is found to be U-shaped for men and women, with left-tail "strugglers" and right-tail "stars" least likely to report earnings. Household surveys report too few low earners and too few very high earners. Non-response is ignorable over much of the distribution, but there exists trouble in the tails.