The department said this lack of hiring translated into a net job loss on business payrolls of 356,000 in 2002, the first full year of recovery from the 2001 recession.
The job movements tracked by the study are similar to those in the department's monthly survey of non-farm business payrolls. However, the new version is based on the actual unemployment insurance reports that 6.4 million private-sector businesses are required to file on a quarterly basis.
The new study did not shed any light on one of the biggest mysteries currently facing economists — the divergence between the jobs picture being reported by the business payroll data and a separate survey of households.
The monthly payroll data show that since the recession ended in November 2001, the economy has lost another 1.14 million jobs through August of this year. According to the household survey, however, the economy has gained 1.41 million jobs during that same time period.
Labor Department officials said a small part of the wide discrepancy may be dealt with Friday when it releases an update of its monthly survey, using more complete tax records covering all private-sector businesses.
The business payroll survey does a good job of tracking current businesses and those that shut down operations, but it can miss, for a time, new businesses being formed. The annual benchmarking on Friday seeks to correct that problem.
However, many analysts believe major changes in the nation's workforce may be involved, which may explain much more of the divergence in readings of the labor market.
Some suggest that companies, facing tough competition from overseas and unable to raise prices of their products, are trying to hold down costs by turning more and more to contract employees. These employees would not show up on the business payroll survey but would be picked up by the survey of households, which would list them as self-employed.
Analysts said there were several reasons for this trend, including a reluctance of companies to hire fulltime workers until they have more assurances that the recovery is sustainable. Also, companies often save money by not offering benefits, such as health insurance, to contract employees.
"We are just not getting the job creation by companies. Businesses are doing more outsourcing of the work," said David Wyss, chief economist at Standard & Poor's in New York.
The new Labor Department report, which provided quarterly data from mid-1992 through 2002, showed how much churning occurs in the labor market.
For the final three months of last year, there was a net job loss of 70,000. That figure was the difference from the 7.82 million jobs that businesses cut during that period and the 7.75 million jobs that were added during the final three months of last year.
The new report said that, of the job gains, 6.11 million occurred at businesses that were expanding payrolls and 1.63 million occurred at new establishments.
Of the job losses, 6.19 million occurred at businesses that were laying off workers and 1.63 million occurred at establishments that were shutting down.
The Labor Department did not provide any details of what kinds of businesses made up each of those categories, but said it hopes to provide expanded information as it further develops the new quarterly report, "Business Employment Dynamics."
The next report, on the January-March quarter of 2003, is scheduled to be released in November. Officials said that, because of the amount of time it takes to assemble information from the unemployment insurance reports, the data will be reported with a lag of about eight months.