--------------------------------------------------------------------------------
October 26, 2003
ECONOMIC VIEW
A Statistic That Tells Only Half the Story
By DANIEL GROSS
VERY Thursday morning, investors, economic forecasters and, increasingly, political pundits pounce on the Labor Department's numbers for new jobless claims. The report provides a sort of weekly thumbs up or thumbs down on the economy's progress and, hence, on the Bush administration's economic policy.
Last week, the department reported that first-time claims for unemployment benefits fell by 4,000, to 386,000, in the week ended Oct. 18. The four-week average - more useful because it smoothes weekly fluctuations - held steady at 392,250. Labor Secretary Elaine L. Chao hailed the data as another tile placed in an increasingly shiny mosaic. "With the president's jobs and growth plan now having a real, positive and lasting impact on the economy, we expect to see increased job creation in the coming months," she said.
Should we really be encouraged? After all, aren't 386,000 more people - roughly the population of Sacramento, Calif. - now standing on unemployment lines?
Yet the four-week average is below what economists and forecasters commonly regard as the magic number: 400,000. Jennifer Kelleher, who is an economist at Economy.com, which provides economic, financial and industry research, said that at least since 1990, "when the first-time claims fall below 400,000, the overall trend in employment will be flat or up."
In fact, over the last 13 years, a sub-400,000 figure has been a pretty good coincident indicator for job growth. From October 1990 to October 1992, a period of job loss, the first-time claims figure remained stubbornly above 400,000. From mid-October 1992 through May 2001, a period in which the economy added a stunning 23 million jobs, the four-week average never topped 400,000.
There are other reasons to pay heed. The figure is a rare weekly gauge of the economy's health. And the claims data is hard and comprehensive: state labor offices tally the claims filed and forward them to Washington. By contrast, the Bureau of Labor Statistics arrives at the monthly figures on household employment and business payrolls by sampling a small slice of the economic pie.
Some economists suggest that the unemployment claims numbers are worth watching not for what they say but for who else is watching them. Alan Greenspan, the chairman of the Federal Reserve, "pays a lot of attention to that figure," said Lawrence F. Katz, an economics professor at Harvard. "The market is always trying to guess what the Fed's likely reaction to the data will be." Even when the government shut down in the mid-1990's in a budget battle, Mr. Greenspan arranged to get the weekly data.
But it is not as if the economy starts to lose jobs the week the metric exceeds 400,000 and the job-production machinery creaks into gear the week it falls below. More recently, the 400,000 figure has been less a tipping point than a fuzzy guidepost.
In June 2000, when first-time claims were only about 280,000, the economy lost jobs. From March 2001 through July 2002, when the number of payroll jobs declined each month, the four-week average ranged from 365,000 to 496,000. It did not top 400,000 until June 2001 - three months into the job-letting. In the fall of 2002, the number of payroll jobs grew while the claims figures were above 400,000.
WHAT has been maddening about this recovery is that claims have not fallen much below 400,000. After all, the difference between 390,000 new claims and 401,000 new claims - on a job base of 129 million-is not significant. "All these numbers are so tentative and so close to borderline that nobody is jumping out and saying that jobs are ready to recover right now," Ms. Kelleher said.
Ultimately, the claims tell only half the story. A rapidly changing economy with flexible labor markets, like that of the United States, is constantly putting people out of work. For much of the last decade, jobs were so plentiful that many displaced employees found new work quickly, without submitting claims for benefits. An increase in claims is not a symptom that companies are suddenly shedding jobs. It is a sign that displaced workers are not finding new positions quickly.
Based on the recent trends in first-time claims - they have been below 400,000 for three consecutive weeks - the Labor Department anticipates an increase in payroll jobs for October. But here, too, there is another nice, round figure to watch. If payrolls add 150,000 jobs in a month - a figure not recorded since May 2000 - economists say that will be a sign that the economy is in serious job-creation mode. |
