Misclassification Errors and the Underestimation of the US Unemployment Rate

摘要

This paper examines misclassification errors in labor force status using CPS data. Similar to previous studies, we show that there exist considerable misclassifica-tions from unemployed to not-in-labor-force and from unemployed to employed. The results at least partly reflect the intrinsic difficulties in classifying labor force statuses of certain groups of people, such as marginally attached workers (especially discouraged workers) and part-time workers for economic reasons, into three distinct categories. We correct for such errors and show that the official US unemployment rate significantly underestimates the true level of unemployment in the United States. For the period from January 1996 to August 2011, our corrected unemployment rates are higher than the reported ones by 2.1 percentage points on average, with differences ranging from 1 to 4.4 percentage points and always statistically significant. In addition, our estimates suggest that unemployment might be much more sensitive to business cycles than previously thought, as the degree of underestimation is larger in magnitude when unemployment rate is higher.