WASHINGTON — WASHINGTON -- In reports raising fresh doubts about the durability of the economic recovery, the government said yesterday that the number of payroll jobs fell in August for the first time in more than a year and that its chief forecasting index turned down in July.
These and other sobering results more than offset improvements elsewhere among figures released yesterday, including a slight decline in the August unemployment rate, to 6.7 percent from 6.8 percent in July.
The index of leading indicators, intended to predict the course of the economy six to nine months in advance, edged down a tenth of a percent in July. It was the seventh month in a row in which the Commerce Department index moved in the opposite direction from the preceding month; it had gained a tenth of a percent in June.
The monthly employment report is widely thought to be the best indicator of the economy's performance, as well as a window on the financial well-being of the nation's households. But yesterday's report appeared to be at odds, at least superficially, with other recent data, including upward revisions in output for the spring quarter and for 1990, 1991 and 1992.
"It's kind of confusing," said Robert T. McGee, an economist at Tokai Bank in New York, noting various crosscurrents in yesterday's labor-market figures.
As is often the case, jitters about the economy prompted a scramble to buy bonds, sending long-term interest rates skidding below 6 percent, the lowest level in 25 years.
"The economy has slowed here in the summer," said Paul W. Boltz, an economist at T. Rowe Price Associates in Baltimore. "These payroll figures really look worrisome."
The one-tenth of a point decline in the unemployment rate was considered too small by the Bureau of Labor Statistics to be statistically significant.
Although the shrinkage in payroll jobs was relatively modest at 39,000, including losses from flooding in the Midwest, a gain of about 150,000 had been expected.
And not only was there a further 42,000 drop in manufacturing -- bringing the total factory job loss since February to 239,000 -- a gain in service jobs was only about a third as big as the average monthly increase since early 1992.
In Maryland, the number of new jobs outpaced the number of new job seekers in July, as the state's unemployment rate dropped to 6.6 percent. Maryland's statistics lag behind the nation's by a month.
The middle of the summer, typically the peak of the year's employment, saw the number of employed Marylanders jump by 44,849, to 2,638,880.
Baltimore, which has suffered from high unemployment, saw the state's biggest job market improvement, as the city's unemployment rate dropped nearly half a percentage point, to 11.1 percent.
In St. Mary's County, however, the rate jumped sharply -- to 10.3 percent, from 7 percent in June -- far worse than last July's level of 7.8 percent.
Labor Secretary Robert B. Reich said the administration was "disappointed" by the drop in factory and other goods-producing sectors and by too many people being forced to work in temporary or part-time jobs. After a sharp rise in July, the "part-time for economic reasons" category swelled 42,000, one-fourth as much, in August.
But the department's survey of households showed a jump of 409,000 in the number of people with jobs, enough to more than absorb an increase in the work force and thin the ranks of the unemployed by 108,000.
Rather than hiring new workers, employers are asking their current employees to work longer hours. Employees were called on to work an average of 12 minutes a week longer in August, driving hourly and weekly earnings sharply higher.
Although longer workweeks are traditionally a precursor of hiring, many analysts suspect that uncertainty over factors like higher health care costs may be inhibiting hiring.
While the July decline in the leading indicators was smaller than many economists had expected, it was viewed direly by those familiar with the work of Gerald Anderson and John Erceg of the Federal Reserve Bank of Cleveland.
The often-cited rule of thumb is that a shift in direction by the indicators, confirmed immediately by two similar readings, points an impending turn in the economy. But the researchers found that in calling downturns, at least, four out of seven negative readings are more reliable than three in a row.
Now that the four-of-seven condition has been met, Professor James F. Smith of the University of North Carolina said yesterday, the chances are 2 in 3 that there will be a business peak -- that is, the beginning of a recession -- within 14 months.