U.S. Payroll Slump

U.S. Payroll Slump That's Probably Deeper Than Estimated Hinders Recovery

By Courtney Schlisserman -

The worst U.S. employment slump since the end of World War II was probably even deeper than currently estimated, giving the economy a larger hurdle to overcome.

About 366,000 more jobs may be subtracted from the payroll count for the 12 months through March when the figures are officially updated in early 2011, the Labor Department reported today in Washington. Following last year’s 902,000 adjustment, the revision would cap the biggest back-to-back reduction since at least 1979, the first year of available data.

“The hole we’re digging out of to make up for lost jobs is probably even greater than we had thought,” Michael Feroli, chief U.S. economist at JPMorgan Chase & Co. in New York, said in an interview.

The revision may mean the worst recession since the 1930s resulted in the loss of 8.73 million jobs, up from the 8.36 million currently on the books. The Labor Department’s report today showed employment has grown by 613,000 so far this year after employers cut payrolls by a larger-than-forecast 95,000 workers in September.

Once a year, the government revises its payroll figures after combing through tax records from the unemployment insurance program that covers practically all businesses. Those records are only available after a lag, explaining why it takes more than a year to make the tabulations.

The government’s initial monthly estimates are based on a sample of about 140,000 businesses and government agencies encompassing approximately 410,000 worksites.

Government’s Model

Since not all of those employers respond to the government’s survey, statisticians also include employment projections based on a formula, known as the birth/death model, to determine the influence from the formation and demise of businesses.

The Labor Department estimates the number of companies that may have folded using data collected over the previous five years. By the same token, it plugs in an estimate for the formation of new businesses to account for their hiring.

Annual revisions tend to be largest around turning points in the economy because by using the five previous years of data as a benchmark, the government ends up assuming too many businesses were created at the beginning of recessions and too few started at the beginning of expansions. The Labor Department last year said the model didn’t work as well as it usually does and is studying improvements to refine it.

Improving the Data

“We’ve been researching it for years, and we’re looking at different ways to possibly improve it,” Chris Manning, the national benchmark branch chief at the Bureau of Labor Statistics, said today in an interview.

Last year’s miss was mainly concentrated in the first quarter of 2009 when the world’s largest economy was in the midst of the recession. This year, the estimated reductions were spread out through the 12 months, Manning said.

It’s not yet clear how much of this year’s revision is a result of a breakdown in the birth/death model, Manning said, and he’ll have a better idea when the final first-quarter figures on the quarterly census of unemployment and wages are released, which should be in the next couple of months.

The Labor Department is considering ways to improve their estimates, Manning said. One would be adjusting the birth/death model more frequently, based on the quarterly census of unemployment and wages, he said.

The projected revisions would put the government’s payroll figures more in line with the ADP Employer Services report. Figures from ADP, a provider of business outsourcing services, are currently underestimating the government’s private payroll count by 562,000 workers.