De-linting the fuzzy math

Tim Kane writes in the New York Times that the increase in US jobs is being underestimated. The cited figure of 308,000 jobs comes from a payroll survey. This approach has some problems. Here’s one:
The payroll survey’s biggest problem is that it systematically double counts workers when they change jobs. Since somewhere between 2 percent and 3 percent of the work force changes employers every month, payrolls tend to be noisy. The illusion of lost jobs in recent years occurred because job turnover declined after 2000, first with the recession, then even more sharply after 9/11. As a result, 1 million jobs have been artificially “lost” in the payroll survey since 2001.
Since November of 2001, the US lost 323,000 jobs – or gained 1.9 million jobs, as stated by a household survey. What gives?
An even bigger problem with the payroll survey is the evolution of what constitutes work. We can think of the payroll survey as counting all workers at traditional firms, plus some workers at start-up companies who have payroll records. But the payroll survey doesn’t count individuals who are self-employed – despite the fact that their ranks have surged by at least 650,000 in just two years.
Then there are limited liability companies, a new form of business the Joint Tax Committee says is growing at an annual rate of 34 percent. Consider, too, the rise of consultants, like a marketing executive who was once on the I.B.M. payroll but who has switched to a freelance consulting role with I.B.M. None of these employees are counted in the payroll survey.



  1. Steve

    Excellent and interesting post. I missed that link. This clearly presents an explanation of how bad data might be used by politicians. Surely the Democrats wouldn’t do that? πŸ˜‰ I’m familiar with “limited partnerships” in several states which are likely included in the “limited liability companies” description. These types of entitites for entrepeneurs have been seeing growth for years. They’re found in all areas of commerce and I can’t conceive of how they’d be included in any count.
    They appear much like the independent consults and others also mentioned on IRS rolls and are generally invisible to all but the state agencies, usually a sec. of state or trea., where they register and nothing more. Stock, value, and income are just as invisible to government, other than the IRS. With the incredibly miserable IT system the IRS has does anyone want the IRS keeping track of employment? (For that matter does anyone want the IRS involved at ANY time?) Thanks again for the post.

  2. BigFire

    Personally, I lost a job in November of 2001, and got a job in October 2002. By the fuzzy logic standard, only the fact that I lost my job in 2001 counts.