Larry DeBoer
Professor of
Agricultural Economics
Purdue University

Visit Larry DeBoer's Indiana Local Government Information Web site

Download the audio files or subscribe to our podcast.




Download the audio of Capital Comments: MP3, WMV

Job Numbers for the D's and R's

The presidential campaign is heating up, and one big issue is the economy. The Republicans say that it's not so bad, and it's getting better. The Democrats say that it's been stalled for years and has barely begun to turn around. And, ever ready to please, the U.S. Bureau of Labor Statistics (BLS) has numbers for them both.

The BLS takes two surveys each month to try to get a handle on what's happening to working people. One is the Current Population Survey. The BLS contacts about 60,000 people every month, asking whether they are employed or unemployed or neither. Call that the people survey. The other is the Current Employment Statistics survey. The BLS contacts 400,000 companies and government agencies, asking how many employees they have. Call that the business survey.

During the past three-plus years, since President Bush took office, these two surveys have produced different results. One says things are really bad. The other says things are pretty good. 

The people survey says that employment topped out at about 138 million in January 2001. It says that employment bottomed out one year later at 136 million, just about two million less. Since January 2002, the economy has added two-and-a-half million jobs, so employment is up half a million since the beginning of 2001.The R's point to this survey.

The business survey says that employment topped out at about 133 million in March 2001. It then fell for most of the next two-and-a-half years, bottoming out in August 2003 with 2.7 million jobs lost. Since then, employment has increased by 750 thousand, so it's still almost two million less than it was at its peak. The D's point to this survey.

You can see all these numbers on the BLS website, at http://stats.bls.gov (no "www"). Click on "employment and unemployment," and you'll see links for the payroll survey of businesses and the population survey of people.

It's very unusual for the two surveys to give such different results for so long a period of time. There are several ideas about what's going on -- which means we don't know for sure.

The people survey includes more kinds of jobs. In particular, the people survey includes self-employed people, while the business survey doesn't. Could the increase in the people survey's job numbers be an increase in self-employed people -- folks setting up shop on their own instead of working for someone else?

Lots of people have more than one job. A moonlighter is counted once in the people survey as one employed person. A moonlighter is counted twice in the business survey, as an employee for two businesses. If people are losing their second jobs, they'll still be employed, as far as the people survey is concerned. Only the business survey would show an employment decline.

New start-up companies can't be included in the business survey right away, but the people survey counts employees when they get jobs at new companies. If there are a lot of new start-up businesses, the people survey would show a jobs increase. For a while, the business survey wouldn't.

Here's a picky technical reason. The people survey estimates are made by scaling the size of the sample up to the size of the working-age population. Population has to be estimated each month, and immigration especially requires a lot of guesswork. If immigration has been over-estimated (maybe because it's been restricted since 9-11), then the people survey numbers are being scaled up too much.

So, don't be too confused. (A little confusion is appropriate.) When the R's say that there are more jobs than ever, they're right, by the people survey. When the D's say that jobs are way down since the president took over, they're right, by the business survey. We probably won't know which survey is right by election day.

I doubt that very many people vote based on job numbers from the BLS, however. People vote, I'll bet, based on their own experience in the job market the experiences of their relatives and friends, and on what's happening to businesses in their communities. When you add them all together, the voters know which survey is right.




Writer: Larry DeBoer
Editor: Olivia Maddox