The Weekly Sift

Yes, some “real” unemployment rate is roughly double the official 5.1%. But there’s nothing sinister about that, and the job market really is gradually improving.

Some 19th-century wit — maybe British Prime Minister Benjamin Disraeli or American humorist Mark Twain or somebody else — once said that there are “lies, damned lies, and statistics“.

This week the Bureau of Labor Statistics issued its monthly jobs report, in which it asserted that the economy added a good-but-unspectacular 173,000 jobs in August, bringing the unemployment rate down to 5.1%, the lowest it’s been since early in the Great Recession of 2007-2009.

As happens every month, a number of pundits and politicians then blew a lot of hot air about how these numbers hide the “real” unemployment rate, which is much higher than 5.1%. Some even made it sound as if an evil government conspiracy is trying to fool the public into thinking things are getting better when they’re actually getting worse. Ben Carson expressed this idea in May:

What you have to know is that you can make the unemployment rate anything you want it to be, based on what numbers you include and what numbers you exclude.

Well, pretty much whatever you want to include or exclude, the BLS tracks that too, and publishes it for everybody to see. The wonks at the BLS refer to the “official” unemployment rate as U-3, but they also keep track of  U-4, U-5, and U-6, each of which defines unemployed more broadly than the previous U. U-4 includes people who would like a job, but are too discouraged to look for one; U-5 adds people who want a job, but haven’t looked recently for some other reason; and U-6 adds people who are working part-time when they would rather work full-time.

U-6 is what people (like the conservative Washington Examiner and liberal Bernie Sanders) are pointing to when they say the “real” unemployment rate is 10.3%. And that’s perfectly reasonable thing to say: 10.3% of potential workers wish they could work full-time, but haven’t found jobs that let them do so.

What isn’t reasonable is the conspiratorial that’s-what-they-want-you-to-think attitude that Carson and others are promoting. For example, The Daily Caller quoted Sanders’ comment accurately, and then inaccurately claimed: “That dose of reality is like a wet blanket on President Obama’s recent claims that the economy is improving.” It was nothing of the sort.

You see, the 10.3% U-6 number wasn’t smuggled out of the BLS by some whistle-blower; it was published in the same report as the 5.1% U-3 number. And while 10.3% unemployment sounds a whole lot worse than 5.1%, in the context of similar measurements taken over time, it tells the same story: The job market has been getting consistently better since very early in President Obama’s administration.*

This graph (constructed with tools at Macrotrends), shows U-3, U-5, and U-6 over the last ten years. All three measures of unemployment bottom out in March, 2007; climb sharply to a peak October, 2009 (nine months into the Obama administration), and then decline to a level that is still a bit above the March, 2007 low.**

So U-3 is 5.1%, down from a peak of 10.1% but still not at the 4.4% pre-recession low. And U-6 is 10.3%, down from a peak of 17.4% but still not at its 8.0% pre-recession low. The two stats tell the same story.

But if you really want to make the slow-but-steady economic uptrend sound like smoke and mirrors, you selectively quote another stat openly published by the BLS: the labor participation rate, the percentage of people over 16 who are either working or looking for work.

The LPR has been going down throughout the Obama administration and now stands at 62.6%. So Carson, Paul Ryan, and other Republicans like to point to it as proof that things have actually been getting worse.

The problem with using the LPR as a measure of economic health is that good news can drive it down too, as people who have some economic slack choose not to work: older people retire, younger ones stay in school, couples let one spouse focus on raising the children, and so on.

If you just show a graph of the plunging LPR during the Obama administration, it looks like something must be going horribly wrong. But you see a different pattern if you take a longer view.

This half-century graph makes it apparent that the major trends in LPR don’t have a lot to do with the ups and downs of the business cycle. Otherwise, you’d have to conclude that the 1960s were some economic hellscape, rather than the relative good times they actually were.

What you’re mainly seeing in that big hump-in-the-graph is the life cycle of the Baby Boom generation, added to the effect of middle-class women entering (and staying in) the job market through the 1970s and 1980s. That Boomer-retirement trend is affected on the margins by the economy (as 60-somethings decide whether to retire this year or next year), but barring some catastrophe that keeps 80-year-olds looking for work in large numbers, the LPR should continue its downward trend for years to come, independent of who is president or what policies they implement.

So yes, there are some damned lies going around in the guise of statistics. But the notion that the economy and job market have been slowly getting better for the last six years is not one of them.

* If you really want to get wonky about this, we’re talking about the difference between a quantity that varies with time and its first derivative. Slightly less wonky: the difference between the raw total and the trend.

As far as I’ve noticed, Sanders has been using these numbers responsibly, to claim that the economy still needs a lot of jobs, and so could use the massive infrastructure project he has proposed. I haven’t exhaustively searched his speeches, but I haven’t seen him question the reality of the upward trend in the job market.

** It’s arguable that the 2007 low isn’t a fair comparison, since a lot of those jobs depended on the soon-to-pop real estate bubble.