How Do Jobs Numbers Work?
Tuesday, August 9, 2011
Here’s a rough guide to the numbers everyone is talking about.
Americans are intensely interested in the state of the job market but trying to get a good sense of it recalls the story of the blind men describing an elephant by feeling it. The description depends on which part is felt. Without attempting a tusk-to-tail description of job statistics, here is a rough guide to the much-discussed numbers.
1. How is the job market doing, anyway?
There are two big headline numbers used to answer this question: the number of jobs in the economy (non-farm payroll employment) and the unemployment rate. Each comes from the federal government’s Bureau of Labor Statistics (BLS). The latest figure for the first number shows 117,000 net new jobs in July; it comes from a monthly “Current Employment Statistics” survey of approximately 140,000 employers and their hiring. The second number shows unemployment decrease to 9.1 percent; it comes from the monthly Current Population Survey, a sample of 60,000 households.
If life were simple, these two numbers would move in lockstep. It’s not and they don’t. The biggest reason is that the unemployment rate is defined as the number of people who are looking for jobs and cannot find them. Not only does this beg the question of what it means to look for a job (Visit the unemployment office? Update your LinkedIn listing? Read help wanted ads with morning coffee?), but it means that when workers stop looking for jobs, the unemployment rate can fall even with a declining number of jobs. A growing labor force also means the unemployment rate can rise when job growth is positive but slow.
Some of these other numbers paint a distinctly darker picture: broader measures of unemployment show 16.1 percent of the work force underemployed.
Because these are two different surveys with limited samples, it is also possible for them to tell different stories, and economists will argue about which survey is more reliable. A more comprehensive picture of the job market would resolve some of that uncertainty, but would be difficult to field monthly.
2. Why do some people say the unemployment rate paints too rosy a picture?
The two big numbers are just crude summaries of a more complicated situation. For example, suppose someone would like to be working 40 hours per week, but can only find part-time work (say 20). Does he have a job? Yes. Is he unemployed? No. But he’s underemployed and likely dissatisfied. Or what about a woman who looked in vain for a job for so long that she gave up? She would gladly take work if offered, but she is too discouraged to actively hunt for it.
Each of these scenarios is captured in more detailed statistics. The monthly report from the BLS includes information about the average hours worked in a week, average wages, and unemployment rates using several different definitions of underemployed and discouraged workers. It also looks at the number of people who have been unemployed for a long time—a rut from which it can be difficult to break free. Some of these other numbers paint a distinctly darker picture: almost 45 percent of the unemployed (over 6 million people) have been out of work for at least 27 weeks, and broader measures of unemployment show 16.1 percent of the work force is underemployed.
3. Is it really one big labor market, from data engineers to short-order cooks?
When workers stop looking for jobs, the unemployment rate can fall even with a declining number of jobs.
No. Those workers are not interchangeable. Nor does a job in Oregon necessarily do much for a high-school graduate in North Carolina.
The job numbers capture some of this, but not all. BLS surveys break jobs down into broad categories (manufacturing, construction, services) and then into narrower subcategories (wood products, residential building, motor vehicle dealers). The surveys do not report a geographic breakdown of jobs (likely because of sample limitations) and they are unable to capture skill mismatches. Both can be very important for job market health. One sees surprising newspaper stories, even in the midst of sharp recessions, about how manufacturers cannot fill jobs. The explanation, as one gets past the headline, is that manufacturers are not looking for high-school grads but rather for skilled workers who are good at math and can handle sophisticated computerized machinery. Those workers can be scarce, even when the high-school grads are lined up out the door.
4. What if you had a really big data set of employers and job seekers? Could you get a better picture?
Probably. There would be some big pitfalls to watch out for, though. That celebrated gain of 117,000 jobs in July? If you go to Table B-1 of the report and look, you’ll find that there were 130,920,000 jobs in July 2011 and 132,151,000 jobs in June 2011. That looks like a loss of 1,231,000 jobs, not a gain of 117,000. What gives?
Those big numbers are not seasonally adjusted, while the headline payroll employment number of 117,000 is. It really depends what question you want to ask. Seasonal adjustment helps answer whether the job market is doing better or worse than we would expect. Traditionally, lots of jobs disappear in the month of July. Thus, if 117,000 fewer jobs than we expect disappear, we say the labor market is showing some positive signs.
How do we know what we expect to see? That comes from estimation. It’s also part of the reason that each monthly announcement also includes revisions to the previous month’s numbers, which can sometimes be substantial. For example, in the most recent report, nonfarm payroll employment for May was revised from an increase of 25,000 to 53,000, while June was revised from 18,000 additional jobs to 46,000—a revised total of 56,000 more jobs than first reported, in addition to the 117,000 in July.
The two big numbers are just crude summaries of a more complicated situation.
5. Are these government surveys the only measure of the labor market?
No. There are a number of other measures and reports out there: mass layoffs; first-time claims for unemployment benefits; Challenger, Gray & Christmas’s job cuts. Each of these provides a different look at how workers are leaving their jobs. They are all different impressions of the elephant.
These first-time glimpses and second-look revisions at the state of the labor market all dribble out on a regular schedule. Leaders use them to assess just how much suffering there is in difficult times. Markets treat them as an important gauge of how the economy is doing and compare them both to their ex ante guesses about the numbers and to benchmarks for what might indicate a recovery. Friday’s numbers beat expectations but lagged behind targets for a healthy recovery. The composite picture is still one of a U.S. labor market that is faltering badly.
Philip I. Levy is a resident scholar of the American Enterprise Institute.
FURTHER READING: Levy writes on the U.S. economy in "Will Exports Save the U.S. Economy?" "The Feds as Stock Speculators," and "Trade Fireworks." Rea Hederman provides more context in "What is Unemployment?"
Image by Rob Green/Bergman Group.