It is hard to imagine that economic statistics could ever be anything but boring (unless you happen to be a nerd geek like me), and I frequently recommend them to friends as a guaranteed cure for insomnia (my resume also works well). Once in a great while, however, one comes along that is worthy of mention outside the arcane world of the obsessed. At present, there is a modestly fascinating phenomenon occurring in—of all things—employment.
Much has been written (by me and others) over the past couple of years about our “jobless” recovery. Indeed, it is only in recent months that the labor market has shown signs of life. By one measure, however, the world looks very different.
When we measure employment each month in the US, we actually do it twice. In one survey, the Bureau of Labor Statistics (BLS) asks businesses “How many folks are working for you?” This one is called the “Establishment Survey.” When we talk about job growth (or lack thereof), those are the numbers we use. From the beginning of the current expansion in late 2001 through the end of last year, this measure shows a net loss of about 70,000 jobs (hence, the name “jobless recovery”).
In the other survey, the BLS asks people at home “How many people that live here are working?” That one is known as the “Household Survey,” and, since it can also tell us how many are not working, it is the basis for the widely watched unemployment rate. By this measure, there has actually been a gain of some two million jobs over the past couple of years.
It is not uncommon for these two measures to differ over time for understandable reasons. For example, moonlighters can be counted twice in the Establishment Survey. Farmers don’t get picked up at all. The Household Survey, however, always shows more people working and is generally regarded as the quirkier of the two. Some people don’t want to admit they are out of work; some are tinkering with something that really has no prospect of bringing in any income; and some domestic goddesses feel (quite justifiably, I might add) that they work full-time and then some even though they are technically not part of the workforce (you can tell that guys made up the system and did so in much different times than today).
In any case, they all generally move in the same basic direction over time despite some minor monthly gyrations. It is quite strange for one to show a healthy gain in jobs while the other shows a loss over an extended period of time. This phenomenon suggests to me that we may not be picking up some employment trends in the proper way. For example, the Internet and other gains in infrastructure and communications technology has made it much more practical for both self-employed persons and others to work at home. Thus, the Establishment Survey may be measuring some very legitimate concerns. Similarly, with the increasing trends toward employee leasing, outsourcing, use of temporary workers, and contract labor, the existing measurement instruments may be missing a lot of gainfully employed people. I could go on and on, but you get the picture.
Economists and others have historically used the Establishment Survey to measure “job growth.” At the moment, however, it may be missing more folks than usual and, in our dynamic and ever-changing world, it may be time to revamp the way we go about calculating employment. I am not suggesting that the jobless recovery was a myth; there was well-known and widely-chronicled sluggishness in the labor market while gross domestic product percolated along at a reasonable and at times frenetic pace. I am saying, however, that it may not have been as bad as it seemed. In any case, while I think that both surveys will be pointing upward in the future, we still need to figure out a way to get it right (or, at least, “righter”).