The September employment report due Friday is expected to bring more bleak news about the jobless recovery, with most forecasters predicting the figures will show an eighth straight month of job losses. But the report also will include a tantalizing bit of data that could support economists who say the labor market is not nearly as bad as it looks.
The monthly payroll survey, generally considered the most accurate barometer of the employment situation, shows that the economy has lost 2.8 million jobs since January 2001, including 1.1 million since the recession ended in November 2001. The latest losing streak, which is contributing mightily to President Bush’s growing political vulnerability, is entirely without precedent for an economy this far into an expansion.
Meanwhile other economic indicators show an economy that gained strength in the second quarter and then expanded rapidly in the third quarter, possibly at its fastest rate since the end of the technology boom in 1999. Although consumer confidence unexpectedly dropped in September, the manufacturing sector expanded for a third straight month, consumer spending has been strong and housing sales have risen to record levels.
All that good economic news has made some analysts suspicious of the persistently weak payroll figures.
Gerald Cohen, senior economist at Merrill Lynch, pointed out that the payroll figures are subject to significant revisions, and he said that historically the numbers have been revised upward in the early stage of a recovery. On Friday, the Bureau of Labor Statistics will offer its preliminary estimated revision of the March 2003 payroll employment figures, which will become the reference month for a comprehensive annual revision to be published in February 2004.
A sharp upward revision could mean President Bush will get some statistical help on the economy when the full revision is published next year.
“I don’t think we’ll ever look back on this period and say we have created 800,000 jobs,” Cohen said. “The question is, will we see that we didn’t lose 1.1 million? The labor market may not be as weak as the data suggest.”
A revision in the payroll figures could help close a puzzling gap that has emerged between the so-called establishment survey and the separate household survey, which is used to compute the politically potent unemployment rate. While the total number of payroll jobs has declined steadily, the unemployment rate peaked at 6.4 percent in June and fell to 6.1 percent as of August.
Analysts generally expect the September figure to inch higher, to 6.2 percent, but the household survey has indicated the economy has been adding jobs lately rather than losing them. Some analysts believe the household survey does a better job of capturing employment at small firms, although others say the figures lack credibility. For example the household survey includes nearly 1 million workers who describe themselves as “self-employed,” many of whom may in fact be “between jobs.”
Mark Zandi, chief economist at Economy.com, agrees the payroll figures are likely to be revised upward, but he said the household survey is unreliable. And he points out that a raft of evidence, including persistently high unemployment claims, supports the notion that the labor market is weak.
“The job market is bad. It’s just a question of how bad,” he said.
A recent paper by John Kitchen, chief economist for the Republican-controlled House Budget Committee, suggests that payroll figures are distorted at a time of economic transition because government estimators survey only firms that have survived the recession, presumably missing many startup firms. He points out that in 1992, during the last “jobless” recovery, the Labor Department initially reported the economy lost a net 32,000 jobs during the six months leading up to the presidential election. Later revisions showed that the economy actually added 477,000 jobs in that period.
Kitchen said he believes “it is impossible to know” whether the inaccuracies in the economic data had an effect on the 1992 presidential election, in which the first President Bush was ousted in a vote that largely hinged on economic issues.
The direction of the economy also looks likely to play a huge role in the upcoming 2004 election. In a survey done late last month by the Pew Research Center, only 45 percent of registered voters said they would favor President George W. Bush, compared with 43 percent who said they would favor a Democratic candidate — any Democratic candidate. That represented a statistical dead heat, compared with the 10-point advantage President Bush enjoyed as recently as July.
The survey made clear that the economy and jobs were by far the most important issues to voters questioned in the survey, mentioned as the No. 1 concern by 49 percent of those polled. And only 26 percent said they thought Bush was doing “as much as he can” to boost growth.
While economists debate the accuracy of the jobs data, there are many who believe the employment situation is grim indeed.
Don Straszheim of Straszheim Global Advisers said that investors who ignore the labor numbers to hop on the stock market bandwagon are “whistling past the graveyard.”
“I think we’re going to get another fairly grim employment report,” he said. “I still think it’s going to be not a jobless, but more accurately a low-job-creation recovery.”
Straszheim estimates the economy will add only 67,000 jobs a month over the next six months, compared with the 115,000 needed just to absorb new entrants into the labor force.
As a result, he expects economic growth to slow to the 3 to 4 percent range over the next several quarters from the estimated 5 to 6 percent rate of the just-ended third quarter.
Straszheim points out that 21 months into the economic recovery, the number of jobs in the economy is down 2.1 percent from its pre-recession peak. At the comparable time after the 1990-91 recession, the economy was already above its previous peak. And 21 months after the deep recession of 1981-82, the economy had surpassed its earlier peak by 3.4 million jobs, or nearly 4 percent.
The magnitude of those differences suggests they are unlikely to be revised away.