We detect sharp increases in the incidence of missing answers to the relevant question starting in 2007, when the U.S. Census Bureau introduced the Respondent Identification Policy. We show evidence of nonresponse selection by both observable and unobservable worker characteristics that correlate with EE mobility. We propose a selection model and a procedure to impute missing answers, thus EE transitions. Our imputed EE aggregate series restores a close congruence with the business cycle after 2007, including the COVID-19 recession, and exhibits no downward trend since 2000.