How reliable is the unemployment rate?

John W. Schoen Senior Producer • Profile • E-mail

Oct. 8, 2004 - Q: Is it true that unemployment numbers nationwide are false?  I have heard that once a person “exhausts” their benefits, then they are no longer counted as “unemployed”.  They just fall into a big dark hole!  … No wonder the numbers keep dropping: all the people out of work for more than 26 weeks are not being included in the numbers!  So the economy is NOT turning around! — Cliff D., Kansas City

A: Actually, no. The national jobs data that are reported every month aren’t based on state unemployment insurance claims. And no one would suggest that the monthly numbers are “false” in the sense that unemployed workers are somehow being deliberately undercounted for political purposes. If they were, you’d expect to see much bigger increases in new jobs reported every month.

That hasn't happened. In fact, President Bush will almost certainly be the first President since the 1930’s to end his four-year term with fewer Americans on the payroll than when he got started -- based on the latest Bureau of Labor Statistics payroll numbers.  (Before you flood our inbox with mail accusing the Answer Desk of liberal media bias, please check the numbers for yourself.) True, there are a few months left in Bush's full term, but it would take more than 600,000 new jobs to erase the Bush "jobs deficit." That's not very likely.

But those payroll numbers tell only part of the story. The “unemployment report” is, in fact, two separate reports. And those reports often tell very different stories -– which accounts for much of the “spin” in the presidential campaign. Let’s take a closer look:

The government’s monthly “household” survey interviews people at home (some 50,000 households in 792 areas) during the week that includes the 12th of the month. The basic question the BLS asks is: do you have a job or, if not, are you available for a job and looking for work? Even if your unemployment benefits have run out, if you’re still looking for work, you’re counted as unemployed.

This survey also includes a category called "discouraged workers” for people who say they’ve given up looking for a job. This group is sometimes referred to as the “hidden unemployed” because they don’t show up in the official unemployment rate. And their numbers have been growing this year, even as job growth has picked up.

A second, separate monthly BLS report is based on a “payroll” survey. This one tries to track how many jobs have been created or lost in the latest month by calling or mailing a questionnaire to a sampling of more than 390,000 businesses. This survey also asks questions about how many hours employees worked each week and how much they were paid. The results also help track which industries are hiring or laying off workers.

FREE VIDEO Behind the numbers Joe Lavorgna, an economist at Deutsche Bank in New York, tells CNBC that the September employment report doesn't accurately reflect other signs of strength in the job market. CNBC

Some economists have pointed out that the payroll survey doesn’t do a very good job of counting self-employed workers. For example, if a telecom middle manager is “downsized” by her company and then sets up a full-time consulting business in her garage, she’ll be counted as employed under the household survey. But when the BLS calls her former employer, the payroll survey will still count her job as “lost” until the company begins hiring again. Since most jobs are coming from small businesses these days, the “official” payroll statistics almost certainly miss some of these self-employed workers.

And there are factors like weather -- including the series of hurricanes that smacked Florida last month -- that can skew the numbers for a given month.

Then there’s the question of whether the “new” jobs being created are as good as the “old" jobs being lost. The BLS doesn’t measure that directly, and there’s really no reliable way to answer the question based on the government data. You can, though, track hours worked and wages earned. The average wage is up on Bush’s watch, but the number of hours worked has dropped.

So the jobs data are “accurate” only in the sense that the BLS conducts its surveys the job market the same way every month. That consistency is one the report’s greatest strengths. But no method has yet been devised to track the employment status of each and every American worker. Until such a method is devised, the BLS numbers are the best we have.

And now you can spin them like a pro.

MORE FROM ANSWER DESK View the archive Check out answers to earlier reader questions. Or click on a topic below for more specific questions and answers: Accounting 101 Credit and debt Economic explanations Energy enigmas Home sweet home Miscellaneous mysteries On the job Saving and investing Small business Taxes

MORE FROM TECHNOLOGY & MONEY

Technology & Money Section Front

Your shopping type: Black Friday? Cyber Monday? Can your name affect your career? Is Black Friday really worth the effort?   Navigating Black Friday   Is Black Friday the best day for deals? 10 tips to find seasonal work this year   Experts share financial advice   Stop holiday overspending before it starts   Is the recovery real on Wall Street?   Study: Wall St. execs cashed in before collapse Technology & Money Section Front   Add Technology & Money headlines to your news reader: