Wait a Few Months Before You Believe the Numbers
morr...@mailinator.com
http://www.nytimes.com/2007/05/26/business/26charts.html
May 26, 2007
Off the Charts
Wait a Few Months Before You Believe the Numbers
By FLOYD NORRIS
IT is the newest economic statistics that usually get all the
attention as investors and analysts try to gauge the health of the
economy.
But sometimes the statistics that take the longest to arrive can
provide the most important information, particularly when they point
to inflection points in the economy.
So it may be with jobs data that the Bureau of Labor Statistics
released this month for the third quarter of 2006. The new data calls
into question the previous conclusion that employment grew at a strong
rate in late 2006.
And it indicates that many small businesses, which had been leading
the way in job creation, are now suffering. As is shown in the
accompanying graphic, companies with fewer than 50 employees lost
workers in the quarter, while larger ones kept hiring, albeit at a
reduced pace.
It also appears that 8,000 more businesses closed than opened in that
quarter, making it the worst quarter by that measure since the third
quarter of 2001, when an economy already in recession was jolted by
the Sept. 11 attacks.
The data is included in a quarterly report, titled "Business
Employment Dynamics," that comes from reviewing employment at every
company in the United States that is subject to state unemployment
compensation laws. By that measure, private-sector employment rose by
just 19,000 jobs in the quarter.
The widely reported data from the bureau's monthly survey of employers
concluded that the quarter had a net gain in private-sector jobs of
498,000. That led economists to conclude that employment growth was
holding up well even though the overall economy had slowed, growing at
just a 2 percent annual rate.
A big difference was in construction employment, which the quarterly
study found contracted by 77,000 jobs in the quarter, in contrast to
the increase of 34,000 jobs shown by the monthly surveys.
"The data show we had two consecutive quarters of job losses in
construction," said David Talan, an economist at the bureau, noting
the small decrease shown in the second quarter of last year.
The figures do not cover exactly the same things, as a small
proportion of employers - notably railroads and religious
organizations - are not covered by unemployment insurance. And Kirk
Mueller, a branch chief in the section of the bureau that deals with
current employment statistics, said differing seasonal adjustment
factors could affect the results.
Eventually, the monthly numbers will be revised to reflect the results
of the quarterly survey, but that will not be done until data is
available for the fourth quarter of last year and the first quarter of
this one, and Mr. Mueller said it was too early to conclude that major
adjustments would have to be made.
Robert Barbera, the chief economist of ITG, an investment advisory
firm, said that he thought the third quarter had been an inflection
point in the economy and that the end result would be a substantial
reduction in earlier employment estimates.
One reason the monthly survey can be inaccurate is that it has to
estimate the number of jobs created by new businesses, which are by
definition not included in the survey. It may be that those estimates
were too high.
It will not be until next February that the monthly numbers are
revised in the next benchmark revision. We may learn then that the job
market was not as strong as it seemed to be in late 2006 and early
2007.