ASSUMING THEY have time for anything besides responding to each other's attacks, President Bush and Senator John Kerry will be spending a lot of time talking about the economy and jobs. Kerry claims that Bush has overseen a recovery without job creation, while the president claims that strong growth in gross domestic product and productivity indicates that new jobs are just over the horizon. They're both wrong.

Since the end of the 2001 recession, the nation has been adding employment. But Bush and Kerry are failing to get America's job story right in a fundamental way. And unless they begin to get it right, America's workers -- and the families and economy that depend upon them -- will be the real losers.

It's understandable how the jobs story can be so confusing. Suppose that you wanted to know how many Americans found work from the end of the recession in 2001 through the first two months of this year. According to one of the two leading sources for such data -- the current employment statistics, also known as the payroll survey -- the number of wage and salary workers on employer payrolls fell by more than 620,000. But according to the other leading source -- the current population survey of households -- employment in the nation increased by nearly 2.3 million over the same period. Not only do these numbers -- both of which are drawn from monthly surveys -- move in opposite directions, they represent a staggering gap of 3 million jobs, a gap between the two surveys that is 10 times greater than that observed after the previous five recessions.

We think we've resolved these seemingly dueling data. We have found what may be the beginnings of an undermining of America's long-established employer-employee relationship that workers should have employment stability, fair wages, and a chance for advancement. The change carries enormous implications for our next president.

While the two surveys appear at odds, both are correct. Their findings are different because they measure different kinds of employment that change at different rates. While the payroll survey counts only jobs on the formal payrolls of firms and government agencies, the household survey includes self-employed workers, independent contractors, farm workers, private household workers, and people working under the table or otherwise off the books. Fueled partly by legal and illegal immigration, nonpayroll employment has risen sharply and largely explains the gap between the two employment surveys.

So when Susie Smith was a full-time programmer for Acme Computers, she was counted by the payroll survey. But after Susie was laid off and then brought back by Acme as a consultant, she was no longer counted by the payroll survey, though she was still counted as employed in the household survey. Similarly, Joe was in the payroll survey when he was a machinist with full-time wages and benefits, but as a self-employed landscaper, he now shows up only on the household survey.

This isn't just some quibble over data points. If the payroll survey is right and the household survey is just flat out wrong, as Alan Greenspan and many argue, and if employment fell by more than 600,000 rather than growing by the household survey's 2.3 million, it would mean that the last month's national unemployment rate would have been 7.6 percent, a full two percentage points higher than the official jobless rate of 5.6 percent. In Massachusetts, which has led the nation in payroll job losses, that would imply a state unemployment rate of more than 10 percent.

That's not the case. When properly interpreted, the two surveys together reveal the real emerging story line: Unable to find regular payroll employment, many workers are accepting second choice self-employment, contract labor, or off-the-books work arrangements. In other words, the growth in nonformal payroll employment over the past two years has acted as a labor market safety valve. American workers are finding that for now, their best and maybe only alternative lies somewhere between a regular wage and salary job and unemployment.

Maybe this is their choice. Or maybe state and federal wage and hour law enforcement has become so lax that employers flout payroll requirements. Or maybe the reason is that firms are able to take advantage of short-term excess labor supply conditions.

We don't really know, which is why we need a more informed understanding of what is taking place in US labor markets. Rather than superficial analyses of employment data that are often based more on ideology than methodology, we need to get into companies to observe actual hiring policies, into the homes of workers to understand their employment relationships, and into the operations of labor market intermediaries to observe what is really going on.

It is not enough to just say employment is expanding. We need to know how it is expanding. And we need to know what it means for the average American worker.

Paul E. Harrington is associate director and Andrew M. Sum is director of the Center for Labor Market Studies at Northeastern University.

http://www.boston.com/news/globe/editorial_opinion/oped/articles/2004/03/23/what_the_jobless_statistics_dont_reveal/