The unemployment survey flap

Bottom line is there's not much change in jobs since Jan 2001 but 8 million more adults in the most jobless recovery since they started keepin stats.

"Household numbers not that much stronger than payrolls

Democrats have been hammering Bush for months, saying he has the worst economic record since Herbert Hoover presided over the Great Depression.

Republicans counter that the economy is expanding at the fastest pace in 20 years and that hiring isn't nearly as rare as Democrats say.

Each side points to a government survey to bolster their arguments. Democrats say payrolls are down 2.2 million since Bush took office. Republicans say employment is actually up about 500,000.

The election could turn on which arcane government statistic is most trusted: The payroll survey of 400,000 business establishments or the employment survey of 60,000 households. Read about the payroll data and the household survey.

New research from the government suggests the two alternative views of the U.S. labor market aren't as far apart as they appear.

Both surveys show a weak job market since the recession began in March 2001, just two months after Bush took office, according to a paper written by economists at the Labor Department's Bureau of Labor Statistics. Read the paper.

Over the past eight years, the two surveys, when properly compared, show nearly identical job growth, the economists say. During the late 1990s' boom, the payroll survey showed stronger growth, but since the recession hit in 2001, the household survey has caught up.

The two surveys are designed to look at different aspects of the labor market. The payroll survey tracks employment from the employers' point of view, looking at the number of workers in each industry, their hours and their pay.

The household survey looks at jobs from the workers' point of view, measuring the employment rate for various demographic groups and tracking the number of multiple job holders or the number working part-time.

The two surveys have been almost impossible to compare directly, until now. The payroll survey showed businesses employed 130.2 million workers in February, while the household survey showed that 138.3 million Americans said they had a job.

Most of the 8.1 million gap is due to differences in each survey's coverage. The household survey includes agricultural workers, self-employed workers and unpaid household workers. The payroll survey counts multiple jobholders twice.

A second complication is that the household survey is adjusted for new population estimates only infrequently, making direct comparisons with past data unreliable, the government says.

Labor Department economists now have created an adjusted household survey more comparable to the payroll survey.

As of February, the adjusted household survey showed employment of 130.4 million, not significantly different from the 130.2 million reported in the payroll survey.

Since Bush took office in January 2001, the adjusted household survey shows total employment growth of 214,000, an average of less than 6,000 new jobs a month. In the meantime, the adult population grew by 8.5 million.

While 6,000 jobs a month is nothing to get excited over, it's better than the payroll survey's average monthly job loss of 59,000. At least the sign is right: It allows the Bush administration to claim that employment has risen during his term.

Whichever survey you trust, it's been the weakest job market of any recovery since the government began tracking the data in the 1930s.

Most economists, including staunch Republicans such as Federal Reserve Chairman Alan Greenspan and nonpartisan economists who work for the Bureau of Labor Statistics, say the payroll survey is the more accurate reflection of the job market.

Economist John Irons says the fastest way to lose credibility in his profession is to claim the household survey is better or that economists disagree about which one is superior.

"The payroll survey is better, and it is what everyone uses -- there is as much disagreement on this as there is on the question of whether or not the earth is round," Irons wrote.

Those reassurances haven't stopped some critics from arguing that "something is different" in today's economy that the payroll survey is missing; namely large number of entrepreneurs who are starting their own businesses or consultants who have a strange "neither-here-nor-there" relationship with their "employer."

Sen. Robert Bennett, R-Utah, has held hearings into the "mystery" of the two surveys.

Commentator Larry Kudlow wrote in the National Review that the "softer payroll survey" is not believable.

Economist Tim Kane of the Heritage Foundation said the payroll figures are just "an illusion." Kane argues that the new adjusted and smoothed data prove that the payroll survey should not be trusted.

"There is simply no way that the economy could grow this rapidly and not experience significant job growth," said Brian Wesbury, chief economist for GKST. Wesbury says the payroll survey is missing the "eBay economy" that's driven by entrepreneurs and consultants.

However, Wesbury's thesis is shot down by Robert Barro, a Harvard economics professor writing in the Wall Street Journal, who notes that self-employment as a percentage of employment is no higher today than it was in the mid-1990s. Self-employment can only explain a tiny fraction of the gap between the two surveys.

Other critics of the payroll survey say it misses start-up firms, pointing to the embarrassing revisions after the fact from the early 1990s that showed payrolls were much stronger coming out of the recession than the contemporary data showed.

But the BLS learned from that mistake and it now finds and tracks new firms much faster than it used to. The recent benchmark revision to the payrolls showed the government had actually overestimated the number of new firms being formed.

Acknowledging that the payroll survey has its flaws, Fed Gov. Ben Bernanke said, "the truth probably lies in between the two series," adding that "somewhat greater reliance should probably be placed on the payroll survey."

Barro agreed. "It seems best to weigh both (surveys) when evaluating the recent performance of the labor market." Barro wrote in the Journal.

If Bennett, Kane, Kudlow and Wesbury are right that the household survey is the accurate gauge of jobs, one of the Democrats' major talking points against Bush loses its bite. President Bush would no longer be lumped in with Hoover as the only presidents to have net job losses over a full four-year term.

Ultimately, it won't be economists or statistics that decide the matter. Voters will make a judgment in November based on their own experiences and those of their friends and families.

Call it the voting booth survey."

http://cbs.marketwatch.com/news/story.asp?siteid=mktw&dist=nwtpf&guid=%7BA451F958%2D98CF%2D4845%2D8E44%2D55B7BC1BAB39%7D