By Aspen Gorry, Aparna Mathur, and Michael R. Strain
There was considerable confusion recently around the July jobs report. The payroll survey reported an addition of 163,000 jobs, but at the same time the unemployment rate ticked up a tenth of a percentage point? We added jobs, but at the same time the number of employed persons shrank by 195,000? How could these things be true at the same time? More generally, how should you read and interpret a jobs report? Here's our advice: read it as a qualitative story representing a coherent narrative, not as a list of facts.
The most important key is not to latch on to any of the specifics. The unemployment rate increased to 8.3 percent, as was widely reported in the press? Yes ... but barely. In fact, the unemployment rate was 8.217 percent in June and 8.254 percent in July. So it didn't increase by 0.1 — it increased by 0.037.
The economy added 163,000 jobs, as was splashed all over the headlines? Yes ... kind of. The economy actually lost over 1.2 million jobs moving from June to July. How do you get from a loss of 1.2 million to a gain of 163 thousand? The 163,000 number comes from a technical adjustment done to smooth employment patterns across the year. This is called "seasonal adjustment" and it accounts for the fact that in the summer months many layoffs occur that are temporary in nature. For example, auto companies shut down to retool plants and teachers leave school for the summer. Since these auto workers and teachers aren't receiving a paycheck at the time of the survey, their jobs "disappear." But the jobs aren't permanently gone — they're just on summer vacation.
In order to avoid classifying these "missing jobs" as job losses, the government estimates how many of these purely seasonal missing jobs exist and adds them back. In the month of July, seasonal adjustment "added" a total of 377,000 jobs, and relative to June, the economy added 163,000 jobs. As a result, instead of showing an employment decline, the jobs report showed a significant uptick in employment. (It works both ways, of course. In May, the seasonal adjustment subtracted jobs.)
This makes July's 163,000 number an estimate, not a fact.
Taking the Qualitative Approach
It should be obvious that the seasonal adjustment could be slightly off. Changes in the way that auto companies handle temporary layoffs, the budget shortfalls at the state and local level which have resulted in teachers being permanently let go, unusually warm winters, massive uncertainty over government policy, strong economic headwinds from Europe, structural changes to the labor market caused by the Great Recession — all these, and more, could render the seasonal adjustment factors slightly, or seriously, inaccurate.
So controlling for seasonal factors, did the economy add 163,000 jobs in July? Maybe. But maybe it added 180,000. Or maybe it added 145,000. You get the picture.
In addition, the 163,000 number is preliminary in the sense that it will be revised in the following two months' reports. In many years, the average monthly revision has been substantial — 73,000 jobs in 2008 and 37,000 jobs in 2011, for example. You can bet that the August and September reports will not show that the economy added 163,000 jobs in July. The July number may be higher, it may be lower, but it won't be the same.
If you can't trust the specifics, then what good is the report? Our advice is to approach the report as if it were telling a qualitative story, not a quantitative story.
What's the Story From July?
We had a good month. The unemployment rate was basically unchanged. The labor force shrank a bit, but not dramatically — same with the employment-to-population ratio. The economy added significantly more jobs than were expected based on the state of the economy and recent data releases. Moreover, the gains were widespread across industries. All in all, the story from July is that the labor market did much better than expected.
Of course, overall the labor market is in terrible shape. Employment is still extremely low. We have a jobs gap which is measured in millions. We have had forty-two consecutive months with over eight percent unemployment. We have millions of workers who have been unemployed for over six months.
This highlights another reality: the qualitative story of one month's report can be different than the qualitative story of the overall labor market. To fix the labor market, we need more reports that tell a story like July's. Actually, we need to beat July. Often.
Aspen Gorry, Aparna Mathur, and Michael R. Strain are economists at the American Enterprise Institute