Everyone apparently agrees—the August jobs numbers were lousy. Despite the fact that the unemployment rate fell for the second month in a row, continuing a march downward that started in late 2009, journalists and analysts left, right, and center believe that the decline has been illusory. The Wall Street Journal’s Real Time Economics blog declared, “Unemployment Rate Drops for Wrong Reasons.” The Washington Post’s Wonkblog upped the ante, saying unemployment dropped “for the worst reason.” According to liberal economist and blogger Jared Bernstein, “of the two ways you can reduce unemployment, we got the bad one.” Conservative American Enterprise Institute blogger, James Pethokoukis concurred, as did Brookings economist and blogger Justin Wolfers, as did libertarian Reason writer Peter Suderman.
Who says the nation is hopelessly divided?
But in fact, while changes in the official unemployment rate have sometimes given too rosy a picture of the state of the recovery, joblessness really has declined and continues to do so (just not as quickly as we’d like). What do analysts mean when they say unemployment has fallen for the wrong reason? The unemployment rate compares the number of jobless who are looking for work to the combined pool of people either working or looking. People who are not seeking a job do not enter into the calculation. The conventional wisdom is that unemployment has fallen not so much because the economy has produced more jobs for the people who want them but because fewer and fewer people believe looking for work is worth their time. The unemployment rate will decline even if there are fewer jobs, so long as they are divided among an even more-diminished pool of Americans either working them or trying to get them.
There is a kernel of truth to this claim, which comes in shorter-run and longer-run variants. The short-run version looks at an employment report like last week’s, notes that unemployment fell, but then insists that it did so because so many people “dropped out” of the labor force. Last month, employment fell by 115,000 according to the Bureau of Labor Statistics survey of households from which the unemployment rate is determined. But because the size of the labor force—people working or looking for work—fell by over 300,000, the unemployment rate declined nonetheless. This short-run assertion tends to be offered by commentators who advocate more monetary stimulus from the Federal Reserve Board.
The longer-run version of the argument most often comes from those who believe that President Obama’s fiscal stimulus was ineffective or counterproductive. They argue that if labor force participation was as high as it was when Obama took office, the unemployment rate would be 10.8 percent rather than 7.3 percent. Their concern is only secondarily with what the latest jobs report says; they worry more about the years-long decline in labor force participation. Of course, many people make both the longer- and shorter-run argument.
Consider the longer-run first. From December 2007 to October 2009, extending from the peak of the previous business cycle to the high point for unemployment, labor force participation declined from 66 percent to 65 percent. Since then it has fallen to 63, while the unemployment rate has dropped from 10.0 to 7.3 percent. The conventional wisdom says that the improvement in unemployment should be discounted because of the drop in labor force participation.
While this is a reasonable point, decline in the size of the labor force does not necessarily indicate a weak economy. People leave the labor force—or decide not to enter it—for a variety of reasons. Older workers retire. Younger people stay or enroll in school. Women give birth, and new moms (and some dads) take time off to raise their children. Midcareer adults may care for a sick relative. Some working-age adults are themselves sick or disabled. Others are fulltime homemakers.
One reason that labor force participation has declined—and would have declined at least somewhat absent a recession—is that the population is aging. The retirement of the baby boomer generation would produce declining labor force participation even in a strong economy. Barclays analysts last year estimated that nearly 40 percent of the decline in labor force participation between the end of 2007 and the end of 2011 was explained by reduced participation among workers older than 54 who indicated to the Bureau of Labor Statistics that they did not want to work.
In addition, school enrollment has been rising among 18- to 24-year-olds for a couple of decades now. Figures from the National Center for Education Statistics show that among 22- to 24-year-olds, for instance, school enrollment in October increased from 16 percent in 1980 to 21 percent in 1990, then to 25 percent in 2000 and to 29 percent in 2010. Since the trend began before the recession, rising enrollment does not simply reflect younger people fleeing the weakened labor market.
The monthly household employment survey asks people who say they did not work or look for work the previous four weeks whether they wanted a job. Among those who did want to work, some have mitigating circumstances that prevented them from being available for work the previous four weeks (such as family responsibilities). Some have not looked for work in over a year, for various reasons. Others looked for work sometime in the previous year but stopped for reasons having nothing to do with the job market. For assessing the strength of the recovery, the most relevant subset of people outside the labor force who want to work is the group known as “discouraged workers”. Discouraged workers have looked for work in the past year but did not look for work in the previous four weeks specifically because of their negative impression of the labor market.
The following chart shows discouraged workers as a share of people outside the labor force and how their ranks have changed over time. The group constitutes a small fraction of those not in the labor force, highlighting the point that changes in labor force participation can be driven by many factors other than the state of the economy. Discouraged workers became a bigger share of non-participants through 2010. However, since then, the share has fallen. Comparing this trend to that for the unemployment rate suggests that for a few months, the initial decline in unemployment from its late-2009 peak might have been an artifact of the still-falling labor force participation of would-be workers. But falling unemployment since 2010 does not appear to have been driven by declines in labor force participation because workers frustrated about jobs have not become a bigger share of those outside the labor force.
Another way of making the same point is to consider an alternative rate of joblessness published by the Bureau of Labor Statistics that takes into account discouraged workers (the so-called “U4” rate). The next figure plots this rate against the official unemployment rate.
Historically, the two rates have tracked one another very closely, but during 2009 and early 2010, the U4 trend did look worse than the official trend in unemployment, rising more through late 2009 and then initially failing to decline. As a consequence, the level of joblessness is higher today using the U4 rate than using the official one. Note, however, that the trends have paralleled since early to mid-2010—both indicating a strengthening labor market. One reaches the same conclusion comparing the unemployment rate to a jobless rate that incorporates all people outside the labor force who want to work.
One objection to these analyses is that they assume that the people outside the labor force who do not want a job have no bearing on how we should interpret unemployment trends. Adults saying they are out of the labor force due to a disability generally do not wish to work. Since the prevalence of work-limiting physical ailments has not risen over time, one might think that neglecting the disabled in assessing unemployment trends is unproblematic. However, receipt of federal disability benefits has soared in recent decades. Such benefits have become a shadow unemployment program for many Americans lacking a college degree. If enough people dropped out of the labor force to live off of disability payments rather than experiencing unemployment, that would produce an overly-rosy picture of the recovery.
The household survey data indicate that the share of adults out of the labor force reporting that they are disabled rose from 14 percent in 2007 to 15 percent in 2011 (the most recent year available in the data I obtained). The share of adults—in or out of the labor force—that was disabled rose from 4.9 percent to 5.5 percent. To gauge the possible impact of the increase in disability rolls on the official unemployment rate, I used the survey data to look at how the 2007 to 2011 rise in unemployment changes when the disabled are added as “jobless”.
Unemployment rose 4.3 points over these four years. Adding in disabled adults bumps the increase up to 4.9 points. For comparison, adding in discouraged workers as jobless instead of disabled adults bumps the 2007-to-2011 increase in unemployment up to 4.7 points. While the change in discouraged workers ceased having an effect on the unemployment trend after 2010, the increase in the ranks of the disabled continued to impact the trend in 2011.
Clearly, the rise in in discouraged workers and workers receiving disability payments did affect the trend in the unemployment rate. Had there been no change in the share of discouraged workers or in the share of disabled between 2007 and 2011, I estimate that the unemployment rate would have risen from 4.6 percent of the labor force to 9.7 percent instead of rising to 8.9 percent. However, their combined impact did not reverse the decline in unemployment that occurred after 2009. Unemployment (averaged across months) first declined from 2010 to 2011, dropping from 9.6 to 8.9 percent. Had discouraged workers and the disabled stayed at their 2010 shares, unemployment would have fallen by the same amount. Discouraged workers ceased affecting the unemployment trend three years ago, while receipt of disability benefits appears to have peaked in 2010 or early 2011. The decline in unemployment since 2010 has not been overstated by the official rates.
The conventional approach to assessing whether unemployment declined for the right reasons would be to add back the entire decline in the labor force between 2010 and 2011 as jobless workers. That method produces an estimated “true” 2011 unemployment rate of 9.8 percent. This divergence points up the problem with calculations suggesting unemployment would be above 10 percent if labor force participation had not declined after early 2009. Much of the decline in the labor force would have occurred even absent the recession, as discussed above. It is inappropriate to add students and retirees back to the labor force as jobless workers if they would have stayed out of it even in the absence of recession. We don’t know what the counterfactual unemployment rate today would be if labor force participation declines that were going to happen anyway had occurred but additional declines since the beginning of 2009 had not.
Like the longer-term arguments discounting unemployment declines, the shorter-term ones also routinely fail to distinguish between bad and benign drops in labor force participation. In last week’s report, for instance, the unemployment rate fell from 7.4 to 7.3 percent while the labor force participation rate fell from 63.4 to 63.2 percent. Several dozen stories were filed about frustrated workers dropping out of the labor force—and a thousand tweets tweeted—with scarcely any recognition that the U4 rate also declined, and by as much as the official unemployment rate did (8.0 to 7.8 percent).
This blind spot matters a lot. Forty-six months have passed since unemployment hit its peak in October 2009. Seasonally adjusted, unemployment declined in 23 of those months. In 15 of those 23 months, labor force participation also fell. In the conventional view, that means that during the recovery unemployment has fallen “for the wrong reason” two-thirds of the time. However, in 11 of those 15 months where unemployment fell “for the wrong reason,” the U4 jobless rate fell by at least as much as the official unemployment rate. Over all 23 months with an unemployment decline, the U4 rate fell at least as much in 16 of them.
Again, it’s not that declining labor force participation has had no impact on official unemployment figures, or that the official rate is a perfect indicator of the health of labor markets. But the decline in labor force participation should not affect the conclusion that the economy is strengthening, albeit more slowly than anyone would like. By continually interpreting good news as bad and depressing economic confidence, commentators themselves exert a drag on the recovery. There are aspects of the recovery worth worrying about. But commentators who point to declining labor force participation are worrying for the wrong reasons.