Interpreting the Diverging Employment Statistics
September 28, 2003

The payroll survey says the number of jobs is still falling.  Meanwhile, the survey of households says the number of employed people is on the rise.  What's the best take on the U.S. labor market?  Probably that overall employment is rising slowly, but that jobs remain scarce.

The payroll survey is large and statistically robust, but it does not include self-employment, and it has difficulty picking up new jobs at startup firms, especially in an economic recovery, when those types of firms are most likely to be the ones hiring.  The payroll survey counts jobs -- if a worker held two jobs at the same time, that counts as two jobs in the payroll statistics.

The household survey is small, and is therefore subject to more random fluctuation.  The household survey counts employment -- it picks up the self-employed, and counts workers with two jobs as one employee. 

The following chart shows how the two surveys have started to diverge:

What is happening?

1.  Self-employment is rising.  Almost a half a million additional people reported being self-employed in the household survey so far in 2003.  Some of that is no doubt involuntary and at much-reduced compensation.  For example, the long-term unemployed may go into business for themselves when it becomes clear that re-employment is unlikely.

2.  New startup firms are hiring, but the data aren't showing up in the payroll survey yet.

3.  People are losing those second and third jobs.  The number of people with more than one job has fallen by about 1 million since its peak in the late 1990s, according to the economics website The Dismal Scientist (www.dismal.com).

On balance, the labor market is probably best characterized as "flat" for policy purposes.  It is not as bad as the declining payroll numbers indicate, but is not improving as much as the household numbers imply.

This interpretation the jobs and employment picture is validated by the number of new unemployment claims, which continues to hover around 400,000 a month, economists' rough benchmark for a flat labor market.

Likewise, the total growth of wages and salaries in the economy is hovering around 2 percent a years, a number that indicates a weak labor market.



Links:
The Dismal Scientist from Economy.com Behind the Employment Numbers by Sophia Koropeckyj (Sept 8, 2003).  (Subscription required -- trial subscription available at www.dismal.com)