In a phone conversation yesterday, John Williams at Shadow Government Statistics warned me not to read too much good news from the better-than-expected jobs figure.  The government’s seasonal adjustments aren’t, well, adjusting properly.  They’re still keying off “typical” fluctuations in employment.  But of course today’s economic climate is anything but typical.  Yesterday the official unemployment rate ticked down a tenth of a percent to 9.4%, but according to Williams it should have ticked up a tenth of a percent to 9.6%.

There are big seasonal changes in employment that the Bureau of Labor Statistics corrects for in order to reduce the volatility of the unemployment rate.  For instance, each year employment spikes ahead of the holidays as companies add workers, and then drops as those workers are let go.

July usually sees a regular pattern of planned automobile production line shutdowns to accommodate retooling for the new model year, but recent disruptions to the auto industry have changed pattern this year. Without the usual pattern of shutdowns, the government’s computers nonetheless responded by creating the usual offsetting boost in jobs, not only in the auto industry, but in supporting industries as well. The auto industry itself was alone among durable goods manufacturing industries in showing a reported, seasonally-adjusted monthly gain in July, up by 28,000 jobs.

Besides bad seasonal adjustments, Williams has problems with the so-called “birth-death” model, which “adds a fairly consistent upside bias to payroll levels each year, currently averaging 76,000 jobs per month.”  The genesis of the birth-death model was after the early ’80s recession, when employment figures didn’t catch jobs being added by new small businesses.  However, when a company like Taylor Bean & Whitaker stops reporting its stats, say because all employees were fired en masse, BLS assumes the company is still in business.  (For how long, I’m not sure)  The bottom line is that, in recessions, you’re losing more jobs from failing businesses than you’re gaining from emerging ones.  Hence the upward bias of the model during recessions.

But according to Williams the biggest problem with the official unemployment rate—”U-3″ in BLS parlance—is that it excludes both the underemployed and workers who have become “discouraged” and stopped looking for work:

During the Clinton Administration, “discouraged workers” — those who had given up looking for a job because there were no jobs to be had — were redefined so as to be counted only if they had been “discouraged” for less than a year.  This time qualification defined away the long-term discouraged workers.

Add all the underemployed and the disappeared and you have Williams “alternate” measure, which pegs unemployment at 20.6%, not 9.4%.

For more of Williams work, I recommend a subscription to SGS.

And for more on the unemployment rate, check out this helpful post from EconomPic Data.