Jobs, jobs, jobs

A couple of interesting posts that continue to assuage recession fears and household uncertainties. First, the WSJ says this (subscription required):

This year the great jobs undercount has continued. An analysis by economists at First Trust Advisors in Chicago has found that from April 2006 through February 2007 (the latest revised numbers available), the federal government originally reported almost one-half million fewer

Americans working than actually were working. On average, the first monthly forecast for this period was 113,000 jobs; the second report from a month later counted 138,000 new jobs; and the final revision was 157,000 jobs. Whoops. So this year, one of every four new jobs flew below the government's statistical radar -- which counts as improvement, we guess.

What's the problem? Well, the Labor Department seems to be using old statistical techniques to measure the dynamics of the new economy. The government publishes two different job estimates. The one that gets big play in the media is the establishment or "payroll" survey -- which asks some 300,000 large employers about their hiring patterns. The second is the smaller household survey, which randomly calls people on the phone and asks whether the adults in the household are working.

For at least the last two years the household survey has been far more accurate. No one knows the precise reason. Former Department of Labor economist Diana Furchtgott-Roth suspects that the survey of big and established employers is failing to account for the huge increase in the number of Americans who are self-employed, contract workers, or work for start-up business ventures that are invisible to Washington.

"We know when Ford lays off 20,000 workers. We don't always detect in our sample the small start-ups, like Joe's Auto Parts, and other entrepreneurial enterprises that define the new economy," says Ms. Furchtgott-Roth. Robert Stein, senior economist at First Trust Advisors, agrees that Labor estimators seem to think "most Americans work in factories and punch time clocks."

These mistakes can have political and policy consequences. An official jobless rate that is higher than the real rate can lead Congress to spend more on jobless benefits, for example, or contribute to protectionist pressure against foreign goods. It can also influence close elections, because the media tend to play up jobs news. In retrospect, we now know that John Kerry's claims of a "jobless recovery" in 2004 were fiction. With the latest revisions, job growth in this expansion has been similar to the pace of the 1990s expansion.

..., the failure to detect nearly 1.5 million jobs -- the equivalent of ignoring every worker in North and South Dakota and Montana -- suggests there may be a problem in the government's statistical methods. And that provides an opportunity for politicians to create mischief and make policy blunders. America is working, even if the DOL's statistics don't always say so.

And from the clear and brilliant Steve Conover:

This chart shows the big picture: how the job churn compares to the total number of jobs in the private sector. Notice the growth trend in total jobs since 2003. The “creative destruction” process is working just as Schumpeter described sixty years ago. Net growth is the result.