Why the unemployment rate is really higher than it looks.
"People are finding work," President Bush proclaimed yesterday in New Hampshire. "There's an excitement in our economy." Evidently, President Bush failed to read the first paragraph of the most recent Employment Situation Summary, which showed that the mammoth U.S. economy added a paltry 1,000 payroll jobs in December. He probably skipped right to the second paragraph, which showed that the unemployment rate fell in December to 5.7 percent from 5.9 percent in November.
The Bureau of Labor Statistics measures employment in two ways. The Establishment Survey gathers data directly from 400,000 companies and then estimates how many Americans have payroll jobs. The Household Survey, based on surveys of 60,000 households, determines how many people are working and produces the unemployment rate. Occasionally, the two surveys show divergent trends in job growth—especially when an economy is coming out of recession. According to the payroll survey, the number of jobs fell 232,000 over the course of 2003 on a seasonally adjusted basis. But according to the Household Survey, which includes farm workers, the self-employed, and people who may work off the books, the number of Americans working rose by 1.03 million in 2003 on a seasonally adjusted basis.
Last October, I dubbed the debate over the two surveys, and the emerging campaign to ignore the payroll numbers and focus on the household numbers, "antidisestablishmentarianism." Last week I described how it has become central to the Republican defense of President Bush's economic stewardship. The comparatively strong Household Survey figures also bolster the Republican case for refusing to extend the federal Temporary Extended Unemployment Compensation program. (Yesterday, the Center on Budget and Policy Priorities charged that about 375,000 people will see their benefits expire in January amid a lame labor market.)
What accounts for the gap between the two figures? The payroll survey is less likely to capture the self-employed, newly formed businesses, or domestic employees. So it could be that the millions of Americans who have been laid off are busy starting companies, or working full-time as self-employed consultants. All of this entrepreneurial energy would show up in the Household Survey and be good news for the economy.
Alternatively, the millions of Americans who are self-employed could simply be frustrated in their efforts to find full-time, salary-and-benefits-paying work at established companies. In other words, as Barry Ritholtz, chief market strategist at Maxim Group and an emerging blogger, has suggested, they're self-employed because they're unemployed. That would be bad news for the economy, and it probably wouldn't show up in the Household Survey.
Or would it? The Household Survey yields other data, including alternative measures of labor underutilization. One of the measures gauges the "total unemployed, plus all marginally attached workers plus total employed part time for economic reasons." Marginally attached workers are "people who are neither working nor currently looking for work but indicate that they want and are available for a job and have looked for work at some time in the recent past," according to Steve Haugen, an economist at the Bureau of Labor Statistics. Those working "part-time for economic reasons" are people who have had their hours reduced, or who work part-time but would prefer a full-time job.
This chart shows the paths of this alternative measure—let's call it the adjusted unemployment rate—and the unemployment rate over the past several years. In December 2003, the adjusted unemployment rate was 9.9 percent, compared with 5.7 percent for the unemployment rate. In other words, on top of the 5.7 percent of the labor force who said they didn't have a job, a low figure by recent historical standards, 4.2 percent of the labor force was either marginally attached or wanted to work full-time but couldn't. That's a high figure by recent historical standards.
The chart shows a persistent and relatively stable gap between the two measures. But in 1999 and in most of 2000, when the economy was adding payroll jobs, when people who wanted full-time work at companies large and small could find them with comparative ease, those falling into this "alternative measures" category represented a smaller percentage of the labor force—about 3 percent.
The persistence of large numbers of frustrated full-time job seekers doesn't explain away the difference in job figures in the Household and Payroll surveys. But it does mean the Household Survey—which is supposed to be signaling robust job creation—contains hints that there may be high levels of slack in the labor market. That doesn't bode well for the creation of payroll jobs or for increases in benefits and wages for those who have them. In fact, the Labor Department reported yesterday that benefits and wages grew a meager 0.7 percent in the last quarter of 2003.
For several months now, economists of all stripes have been forecasting that payroll jobs—a notorious lagging indicator—would begin to rise smartly. But with the economy having turned in what appears to be a disappointing fourth quarter, we may have to wait a while longer.
Daniel Gross is the Moneybox columnist for Slate and the business columnist for Newsweek. You can e-mail him at firstname.lastname@example.org and follow him on Twitter. His latest book, Dumb Money: How Our Greatest Financial Minds Bankrupted the Nation, has just been published in paperback.