Labor & Workplace Research
Data Snapshot
During U.S. household surveys, a substantial proportion of individuals refuse to report earnings. The Census Bureau imputes earnings for nonrespondents, but little is known about their accuracy or how nonresponse varies with the level of (unobserved) earnings. In Trouble in the Tails? What We Know about Earnings Nonresponse, researchers use CPS household surveys matched to administrative tax records makes it possible to address these questions. The U-shaped pattern of nonresponse implies that standard measures of earnings and income inequality are understated.
The chart at the right shows the rate at which male and female employees refuse to report their earnings to the US Census Bureau.
Each panel shows the nonresponse rate for a 3-pt moving percentile average across a common DER earnings distribution for men and women. The nonresponse rate is weighted using inverse probability weights for ASEC-DER linkage.
Sources: U.S. Census Bureau, Current Population Survey, 2006-2011 Annual Social and Economic Supplement. Social Security Administration, Detailed Earnings Record, 2005-2010.

Notable Research
Some Like it Hot: Assessing Longer-term Labor Market Benefits from a High-Pressure Economy
High unemployment during recessions is mitigated by recent prior exposure to a high-pressure economy, both in its intensity and duration. Although a hot economy helps workers in a subsequent recession, it does little to narrow gaps in well-being between advantaged and disadvantaged worker groups.
October 1, 2018: Julie Hotchkiss and R. Moore
Trouble in the Tails? What We Know about Earnings Nonresponse
In U.S. household surveys, a substantial proportion of individuals refuse to report earnings. The Census Bureau imputes earnings for nonrespondents, but little is known about their accuracy or how nonresponse varies with the level of (unobserved) earnings. Using CPS household surveys matched to administrative tax records makes it possible to address these questions. Imputed earnings are reliable on average, but not with respect to important non-matched attributes. Overall, nonresponse varies little with respect to actual earnings. The exceptions are very high nonresponse rates among low-wage workers and high nonresponse among the very highest earners. The U-shaped pattern of nonresponse implies that standard measures of earnings and income inequality are understated.
June 2, 2018: C. Bollinger, Barry Hirsch, C. Hokayem, and J. Ziliak