The CEO of one of the world’s most prominent polling and analysis firms, Jim Clifton of Gallup, wrote a bombshell column last week: The unemployment rate reported by the federal government and widely cited by the media is “a Big Lie.”
As AEI’s Michael Strain argued on the Corner, Clifton is wrong, or at least way overstating his case: A statistic is just a statistic, and the federal government and the media don’t offer the unemployment rate with a particular agenda. In fact, they don’t just offer the unemployment rate by itself, anyway, nowhere close to it.
Like all statistics, though, it can leave out important context. To the extent that people just hear about the unemployment rate dropping but nothing else, they are indeed missing important facts — maybe even being misled.
But it’s no well-kept secret that the unemployment rate can overstate health of the labor market. The current Federal Reserve chairman and her predecessor repeatedly say so. Financial markets follow their long statements and musings with rapt attention.
Their failure at this tough task is overstated. Month to month, in fact, if the media and the BLS do rely on one number to report the employment situation, it is not the unemployment rate — it’s the number of jobs created in a given month. That is the first number in the BLS report every month and it’s the one I cite most prominently when I write it up.
But the number of jobs created in a given month isn’t a great statistic either. For one, it’s a big, gross number, not a percentage. It means little to a layman, with the exception of laymen who remember about how many jobs are created each month in a good or bad economy, how many are needed to keep up with population growth, and more.
Really, there is no one statistic the media or the government can use to describe the unemployment situation. People will always prefer some numbers besides the one used, just like I’ll cite Tom Brady’s overall playoff record before his Super Bowl record, and Peyton Manning fans will do the opposite.
But let’s do our best: What is a decent single number to encapsulate the unemployment situation? How do we get to a “real unemployment rate”?
One number that’s sometimes referred to as the “true” or “real” unemployment rate is a statistic called U-6, one of the six unemployment measures produced by the federal government’s Bureau of Labor Statistics (the standard unemployment rate is the “U-3”). It’s the broadest measure the BLS produces: It counts those unemployed by the typical, U-3 definition, plus those who’ve stopped looking for work but still want a job, plus those who are working part-time because they can’t get a full-time job.
Defining “unemployment” more broadly makes the rate much higher, and scarier sounding, than the 5.6 percent U-3 rate: U-6 is currently 11.3 percent.
That isn’t necessarily a much more meaningful number: We shouldn’t count people stuck working part-time as entirely “unemployed” (they could be working 2 hours a week, as Gallup’s CEO pointed out, but they could also be working 25 hours a week), and how badly does someone who hasn’t looked for a job in twelve months really want one?
Most important, citing U-6 doesn’t actually add much to the discussion of our current labor-market woes. Using it instead would just bias media coverage marginally toward pessimism, no matter what party the president belongs to.
It sounds worryingly higher — 11.3 percent, instead of 5.6 percent — but it has dropped just as fast as the widely reported rate. Here’s U-3 in blue versus U-6 in red:
So if the government or the media decided to use the so-called real unemployment rate, they’d still have been reporting most months over the past four years that it dropped — and at a faster rate than the fake one.
The neatly, if not precisely, parallel image in the above chart, though, misses an important fact about the jobs market under President Obama: Many Americans have left the “labor force” — i.e., when polled they say they don’t have a job and don’t want one — which means they’re not counted even in the U-6 rate.
Can we take that into account? Again, many or even most media reports do — “labor-force-participation rate” has made an improbable entry into the common political lexicon — and this number is complicated, too, because people leave the labor force for reasons besides economic weakness — America is aging, for one, and more people are staying in school longer. We can try to fix some of these problems by just looking at the labor-force participation rate among people who are in their peak working years, often defined as 25 to 54 years old. In blue, the overall labor-force-participation rate, and in red, the working-age one (with its vertical axis adjusted):
As you can see, this measurement looks grim, though not as grim as the labor-force-participation rate by itself, the metric you usually hear, suggests.
What we’d ideally like to do is use the information provided both by the labor-force-participation rate and the unemployment rate while taking into account factors like aging.
We can try to do so without getting too fancy: Just look at the share of the whole working-age population that’s employed.
This number over the past few years depicts a picture most economists agree on: The jobs market has recovered in significant ways, but remains far from fully healed.
But there are plenty of ways this measure falls short: For one, it suggests the jobs market didn’t improve at all between 2009 and 2011, roughly, and then improved as much in 2014 as it did in 2012 and 2013 combined. And certainly, the jobs market isn’t obviously better now than it was in, say, 1978. (The reason it never dropped to late-1970s levels is because there are now more women in the labor force; the reason growth has been choppy is more complicated.)
So we haven’t found a “real unemployment rate” yet — one that can be used consistently to compare the unemployment picture month to month, year to year, and in larger historical context. Such a number, perhaps obviously, doesn’t exist.
And we haven’t even gotten into taking into account facts about jobs besides whether people have them or not — how much they pay and how many hours people are working at them. (There is, finally, the question of whether or not the BLS actually gather the numbers honestly and accurately: The Gallup CEO thinks they do, and he does know the business well.)
That’s why economists use all kinds of numbers to assess the quality of the labor market — they are aware, as we all are, that just one number can’t capture what’s going on, and that the best numbers to tell the story vary all the time.
It’s a bit like assessing quarterbacks in football or players in any other statistics-heavy position in any sport. Unlike QB rating or WHIP for pitchers, though, there isn’t even a universally agreed upon fancy, composite number for the labor market. And no one thinks the best QB rating or the best WHIP ever settles the question of the best pitcher or quarterback ever, because times change.
Meanwhile, no one’s accusing ESPN of failing to do their jobs when they use TD totals or ERA. This brings us back to Strain’s point: Implying the government and the media lie by using one number to describe the labor market attributes a crazy level of agency or bias to an innocuous and necessary choice.
America’s jobs situation is improving steadily but it’s still not great. I’m going to need a few numbers to explain any more.
— Patrick Brennan is opinion editor of National Review Online.