- The Washington Times - Tuesday, January 10, 2006

Reviewing the U.S. employment picture for 2005, the Bureau of Labor Statistics (BLS) reported in its Jan. 6 press release total employment in December “was 2.6 million higher than a year earlier.” A later page of the same release reported that in the same period “payroll employment grew by 2 million.” That’s a big difference.

The media widely reported the 2.0 million figure, derived from a survey of businesses and government agencies. However, only a small handful reported the 2.6 million, which comes from a nationwide household survey. As the BLS has said repeatedly, “Both the payroll and household surveys are needed for a complete picture of the labor market.”

There are definitional differences between the two surveys. Total employment is the more comprehensive measure and includes agriculture, the self employed, unpaid family workers, private household workers, and workers on unpaid leave, which the payroll data do not. Total employment is a people count while the payroll numbers are a job count.

Last month total employment was 142.8 million compared with a payroll job count of 134.5 million, a 6 percent difference.

One reason the media consistently ignore total employment is the smaller sample of the household survey compared to the payroll survey. A smaller sample means month-to-month changes in the total employment data have a wider band of error than monthly changes in the payroll numbers.

But the media don’t seem to realize that the smaller sample size of the household survey, while relevant in evaluating monthly changes, poses no problem in interpreting large year-over-year total employment changes. By ignoring the longer-term message of the household numbers, the media are throwing out one of the twins with the bath water.

To help explain trend differences between the two employment surveys, the BLS regularly adjusts the household survey employment totals to be more consistent with the definition of the payroll data.

However, the reconciled household employment total for December 2005 still remained 2.6 million above a year earlier, the same as prior to the definitional adjustment. Therefore, the reasons for the discrepancy between the two yearly employment changes lie elsewhere, a problem the BLS is studying. But uncertainty on differences between the two employment counts is no reason to ignore either’s message.

One possible explanation is the payroll data may not fully pick up growth of new firms, brought into the survey only with a lag. Also, the payroll survey excludes off-the-books employment, some of which may be captured in the household survey. Greater illegal immigrant labor could be a factor. Such reasons suggest the 2005 measured increase in the payroll count may be understated, though there could be other unknown influences affecting the employment growth gap.

From early 2003 to early 2005, the trends in reconciled total employment and payroll employment were similar. Total employment started moving ahead faster than payroll employment last spring, and rose considerably faster in the second half of 2005.

When the BLS reported the 2.6 million year-over-year rise in December total employment in its press release, it did not state any limitations or reservations. The Bureau doesn’t knowingly hand out bad numbers to the media.

There’s no good reason to ignore the good news.

Alfred Tella is former Georgetown University research professor of economics.

Sign up for Daily Newsletters

Manage Newsletters

Copyright © 2019 The Washington Times, LLC. Click here for reprint permission.

Please read our comment policy before commenting.


Click to Read More and View Comments

Click to Hide