The Real U.S. Unemployment Rate is Twice the Official Rate, and the Phillips Curve

Abstract

The official unemployment rate has become an inadequate measure of labor market conditions. This poses a major challenge for basic research as well as for the formulation of adequate economic policy. We propose a new definition of the unemployment rate by weighing part-time workers with 62.5%, the proportion of the time they work relative to full-time workers. We provide new monthly estimates of the unemployment rate for the period 1994-2019 and find that their average during this 25-year period was 10.1% or 4.4 percentage points above the average of the official rate of 5.7%. The gap between the two rates fluctuated between 3.6 and 5.6 percentage points and rose in wake of the recession of 2008 reaching a peak in 2014 only to decrease slowly thereafter back to its pre-recession level of 4 percentage points. The Phillips curve is investigated with the new unemployment rate as well as with U3 and U6 in seven specifications for the period 2008-2019 confirming the very shallow slope found in other studies. However, in one of the specifications the slope is much steeper, mysteriously reminiscent of the coefficients estimated for the 1970s providing a conundrum for further study.