"...We examine the relation between administrative income data and survey reports for selfemployed and wage-earning respondents from 2000 - 2015. The self-employed report 40 percent more wages and self-employment income in the survey than in tax administrative records; this estimate nets out differences between these two sources that are also shared by wage-earners.

We provide evidence that differential reporting incentives are an important explanation of the larger self-employed gap by exploiting a well-known artifact – self-employed respondents exhibit substantial bunching at the first EITC kink in their administrative records.

We do not observe the same behavior in their survey responses even after accounting for survey measurement concerns.."