No one ever said that economics was an exact science. This is particularly true with many statistics related to economic performance such as employment estimates, income measures, and most of the other numbers thrown out in various media outlets. We do the best we can to describe what’s going on, but let me explain something here—it’s largely guesstimation.
On the first Friday following the end of every month, US unemployment numbers are released. That little percentage sets off a media frenzy as analysts compare the figures to expectations, history, other countries, and anything else they can think of. Wall Street takes notice, and the stock market may move hundreds of points on the news—whether good or bad. In short, it’s probably the most-watched economic data series bar none.
What many people don’t realize is that it’s far, far, far from precise. The process works like this. Each month, local agencies conduct surveys of their area to determine how many people are unemployed (and employed) and the size of the total available workforce. (You divide these two numbers to get the unemployment rate, with the unemployment as the numerator and total workforce as the denominator.) All of these local agency results are then compiled at the state level—then at the national level.
While every effort is made to ensure accuracy, there are several potential pitfalls with regard to the procedure. First, the information gathering process consists of surveys, meaning that only a small percentage of all households are contacted. Second, local agencies vary in terms of the quality of their staff and resources to conduct the surveys. Third, there could be issues with the compilation at the state and national levels. Fourth, this all happens in a matter of days. Naturally, there could be data issues, and this series is never revised.
But the problem goes beyond the quality of the information. A larger issue is whether the unemployment rate is even a good measure of economic health. There are at least two scenarios where such a statistic can be very misleading. Take, for example, the case of a rapidly growing economy. In such an instance, many people may be moving in to take advantage of the opportunities. Because of this increased pool of available workers, some of whom may still be in the transition phase and thus indicate that they are jobless, the unemployment rate may actually increase. So you could have an economy in great shape, but rising unemployment. (This situation happened in Texas a few times during the Oil Boom of the late 1970s and very early 1980s.)
On the other hand, if an area has seen persistent hard times, people may begin to drop out of the workforce. The way the process is set up, to be counted as unemployed, a person must be looking for a job. So here you could have an economy in bad shape, with people giving up on finding work and essentially dropping out of the available workforce, but the unemployment rate actually improving. This pattern has also occurred occasionally.
Another issue is the variance between this measure of employment and others, such as the one based on surveys of businesses. Recently, these two series have diverged, with the number of jobs indicated by the household survey going in a different direction than the business establishment series indicates. This could be an indication of several things, including rising numbers of self-employed people (which wouldn’t show up on the business survey). The opposite problem could arise in the case of people with two jobs, as they could be counted twice. Which is more accurate? Depends on your perspective.
A final problem worth noting is seasonality. As a simple example, changes in the weather can impact construction jobs in random and unpredictable ways. Such matters can’t be accounted for by traditional adjustment methods that must be implemented in advance.
I’m certainly not complaining about the quality of economic data. It’s frankly amazing that it can be compiled with as much accuracy as it is, given the inherently difficult nature of the process. The problem is that many people—from the media, to businesspeople, to Wall Street gurus—take the unemployment rate as absolute gospel. Billions of dollars change hands, political careers are made or broken, consumers and investors decide to spend or pull back, and article after article after article appears every time the unemployment rate is released.
Instead, we should all remember that this isn’t a precise accounting; rather, it’s a quick-and-dirty look at the state of things. Use it to get a feel for whether the economy is growing, shrinking, or stagnant, but don’t go much further than that. In fact, I focus more on the job growth (or decline) than the rate itself. After all, it’s guesstimation.