NCPA - National Center for Policy Analysis

Two Surveys -- Two Different Job Growth Rates

April 12, 2004

Depending on which survey the Labor Department reports, the employment numbers announced by the Labor Department each month may be higher or lower, according to Tim Kane of the Heritage Foundation.

The Labor Department uses two sets of research data -- payroll surveys and household surveys. Payroll surveys have been used since 1939 to record jobless rates each month, but they can present a misleading picture in comparison with household surveys:

  • The payroll survey records the number of jobs, not workers, but it counts workers twice when they change jobs.
  • Therefore, when job turnover is low, as it has been since 2001, the survey shows that 1 million jobs have been lost.

However, the Labor Department's household survey shows a higher employment rate.

  • Since November 2001, the payroll survey has shown 323,000 fewer jobs, but the household survey indicates 1.9 million more overall jobs.
  • The payroll survey does not include the self-employed, who have increased by about 650,000 over the last two years; freelance consultants and limited liability companies are not included in the payroll survey.
  • Jobless claims are actually 10 percent below the national average, and the unemployment rate is currently near what economists refer to as the "natural rate" of unemployment for the United States.

Economists often prefer using the payroll survey due to its larger sample size, but the survey underestimates job growth when compared to the household survey.

Source: Tim Kane, "Labor's Lost Jobs," New York Times, April 7, 2004; see also Tim Kane, "Will the Real Unemployment Rate Please Stand Up?" Heritage Foundation, March 24, 2004; and "Household Data," Bureau of Labor Statistics, U.S. Department of Labor, February 6, 2004.

For Bureau of Labor Statistics


Browse more articles on Economic Issues