«Is the U.S. Unemployment Rate Today Already as High as It Was in 1982? John Schmitt and Dean Baker March 2009 Center for Economic and Policy Research ...»
The decline in the coverage rate of the CPS, a phenomenon that has affected almost all government and private household surveys, has had an important impact on the survey’s ability to estimate the national employment rate. In earlier work, we demonstrated that in 2000, the CPS overstated the employment rate by about 1.4 percentage points, based on a comparison of employment rates in the CPS with an adjusted measure for employment in the 2000 Census covering the same period.6 (The coverage rate for the CPS has fallen further since 2000.) These earlier estimates suggest that about one-third – 0.5 percentage points –of this overstatement of employment corresponds to an understatement of the unemployment rate relative to its true rate.7 We also showed that the deterioration in coverage rates between 1986 and 2005 likely added 0.6 percentage points to the overstatement in national employment rates (with the overstatement of employment increasing from
1.2 percentage points in 1986 to about 1.8 percentage points in 2005).8 Assuming, as suggested above, that about one-third of this rising overstatement of employment rates corresponds to a rising underestimate of the national unemployment rate, the long-term decline in CPS coverage rates means that recent unemployment rates understate unemployment by about 0.2 percentage points relative to 1982.
Combining the Two Effects
Figure 1 summarizes the impact of both the demographic and the survey effects on comparisons of unemployment rates between 1982 and February 2009. The first bar shows the official unemployment rate for 1982 – 9.7 percent. The second bar shows the official unemployment rate for 2008 – 8.1 percent, or 1.6 percentage points below the 1982 rate. The third bar gives our estimate of the unemployment rate in February 2009 if we had the same, much younger age structure now that we had in 1982 – about 9.3 percent, or just 0.4 percentage points below the average rate for 1982. The fourth column presents our estimated unemployment for 2009 assuming that the CPS had the same coverage rate in 2009 that it had in 1982 – about 8.3 percent. The final column combines our estimates of both effects. The resulting unemployment rate – which puts the 2009 rate on a comparable basis with the 1982 rate – suggests that the gap between the current unemployment rate and that of 1982 is now just 0.2 percentage points.
Conclusion The aging of the US population and the reduction in coverage rates of the main US labor-market survey mean that the official unemployment rate does not provide a completely consistent basis for comparing labor-market slack between the early 1980s recession and the current economic downturn. We calculate an unemployment rate for February 2009 that takes these demographic and statistical issues into account.
Putting the February 2009 unemployment rate on a basis that is more directly comparable to the unemployment rate in 1982, the February 2009 rate rises from 8.1 percentage points to 9.5 percentage points, just 0.2 percentage points below the 9.7 percent average for 1982. Adjusting for the aging population raises the February 2009 unemployment rate about 1.2 percentage points;
adjusting for the change in the coverage rate of the CPS raises the February 2009 unemployment rate an additional 0.2 percentage points. Together, both effects eliminate almost all (1.4 percentage points) of the 1.6 percentage-point gap between the 1982 and the February 2009 unemployment rates. The current unemployment rate, therefore, is nearly identical to the rate reached in the year with the highest annual unemployment rate in the postwar period.
Is the U.S. Unemployment Rate Today Already as High as It Was in 1982? 7 References Schmitt, John and Dean Baker. 2005. “Correcting Employment Rates in the 2000 Decennial Census using information from the CPS Census 2000 Match,” Center for Economic and Policy Research Briefing Paper.
Schmitt, John and Dean Baker. 2006a. “Missing Inaction: Evidence of Undercounting of NonWorkers in the Current Population Survey,” Center for Economic and Policy Research Briefing Paper.
Schmitt, John and Dean Baker. 2006b. “The Impact of Undercounting in the Current Population