Solving A Labor Market Enigma

Are we running out of workers?

No one denies that U.S. labor markets today are tight. But just how tight are they? Those who worry about inflation point to the government's survey of payroll employment, which has registered more than 1.5 million additional jobs since January. With a 4.2% jobless rate and scant labor-force growth, that suggests the economy could soon be running out of workers.

As L. Douglas Lee of Economics From Washington observes, however, the government's other employment measure, its survey of households, tells quite a different story. It shows a mere 154,000 people added to job rolls since January--suggesting that slowing employment growth may well keep wage pressures from erupting.

Which picture is correct? Most experts favor the payroll survey because it is based on business records and uses a larger sample. Moreover, the two estimates of job growth often diverge because of differences in coverage. The household survey includes such groups as farm workers and the self-employed, for example, whereas the payroll survey counts jobs rather than people employed--so its numbers are inflated by those holding multiple jobs.

Still, in terms of growth rates and the numbers added to job rolls, the two measures have tended to track together over time--until the mid-1990s. According to a new study by Mark Schweitzer and Jennifer Ransom of the Federal Reserve Bank of Cleveland, even after adjusting for the differences in coverage, a widening gap between the employment growth rates--amounting to several million workers--has surfaced in recent years.

The two researchers can't explain this growing disparity, but they note that the payroll survey numbers imply a far lower jobless rate than today's 4.2%. By contrast, the household survey's more moderate estimate of job gains implies greater productivity growth and less labor-market tightness.

Another strong possibility, observes Lee, is that the household survey has been understating both job growth and labor-force growth in recent years because it is keyed to available population data and thus missed increases in the employment of illegal aliens. After the 1990 census, he notes, some 1 million people were added to the household survey's job count.

The notion that population estimates may be too low also intrigues some government experts. That explanation would close the puzzling gap between the household and payroll surveys' job counts and would probably leave the jobless rate close to its present level. But it would allay fears that unemployment is a lot lower than reported--and would help explain why labor costs have stayed relatively restrained.

Before it's here, it's on the Bloomberg Terminal.