, one way or another, is going to be blamed on the rain, to paraphrase a couple of lip-synching
Hurricane Floyd resulted in the mass evacuation of residents from several coastal states the same week the
conducted its jobs survey. Though most components of the most important monthly indicator shouldn't be too adversely affected, the hurricane shortened the
average weekly workweek
of millions of residents on the Eastern seaboard, rendering that component useless.
The average forecast for new nonfarm payrolls is 218,000, according to
forecasts. The unemployment rate is expected to remain unchanged at 4.2%, while average hourly earnings are expected to rise 0.3%. The workweek, the component expected to be the most heavily impacted by the hurricane, is forecast to fall by 0.2 hours, to 34.4 hours per week.
"If you assume, say, a couple million didn't work for two days, that's going to impact average hours worked," says David Resler, chief economist at
, who's forecasting a 230,000 uptick in nonfarm payrolls.
So, like the January 1996 report, which was thrown off kilter because of the blizzard that dropped three feet of snow on the Northeast, tomorrow's figures are likely to look a little funny. (The blizzard lopped off 0.4 hours from the average workweek, dropping it to 33.9 hours per week, but the affect on other components was minimal.)
The possible market effect of these one-time catastrophes suggests that the Labor Department should estimate, or make an adjustment, to offset what is obviously a unique distortion that the entire market shouldn't take at face value. But economists say that's just not possible.
"The problem they have is, and the reason these one-off events show up in the data, is seasonal adjustment programs don't handle one-off events," says Ethan Harris, senior economist at
. "They don't like to make ad-hoc adjustments."
Economists don't expect Hurricane Floyd to have an effect on payrolls. Employees are counted as long as they work at least one day a week, and the hurricane didn't cause too many people to miss an entire week of work. (Although in January 1996, job growth, which was cruising at a 12-month average of 181,000, fell 7,000 and gained it all back with a 454,000 gain in February 1996.)
Economists are predicting a softer September payrolls number (218,000 falls short of the 228,000 12-month moving average), because they're looking at history. September used to be a bigger month for jobs as teachers were added back to the government payrolls after summer recess. But two economists cited the recent trend of leaving teachers on payrolls during the summer as why actual figures have fallen short of monthly forecasts in September. "In 10 of the last 11 Septembers, payrolls have been weaker than expected," says Harris, who nonetheless is forecasting a gain of 235,000.
Countering this, the manufacturing sector is expected to rebound after shedding 63,000 jobs in August.
Henry Willmore pointed out in a comment yesterday that this was the largest decline in a nonstrike month since the 1991-1992 recession, calling it a "statistical fluke." (Barclays is forecasting a 220,000-payroll increase.)
The market's reaction to this report, barring a further decline in the household unemployment rate, probably hinges on the increase in average hourly earnings. The consensus forecast, a rise of 0.3%, would bump up the year-over-year average to 3.6% from last month's three-year low of 3.5%. A larger increase could spook the bond market -- already near its highest yields of the year.
"The market is at its very low right now. It will probably get some relief barring an outlier
in payrolls or earnings on the high side," says Bill Quan, economist at
Aubrey G. Lanston
, who is calling for a 235,000 increase in payrolls and a 0.4% increase in hourly earnings. "More important is next month. There are two sets of data coming out between now and the next
Federal Open Market Committee