(Reuters) – It’s a mystery that has puzzled even Federal Reserve Chairman Ben Bernanke: if the U.S. economy is growing rapidly, why isn’t it creating jobs?

Friday’s hotly anticipated employment report for March may muddle matters even more. Economists polled by Reuters had widely divergent views, with one looking for an increase of 400,000 jobs — which would be the strongest in a decade — while others thought it may show another small net decline.

The consensus expects a gain of 190,000 jobs, which would mark only the second month of job growth since the recession started in December 2007, and the largest increase since March of that year.

Government jobs are expected to account for the bulk of the growth, thanks to the once-a-decade Census, which requires taking on hundreds of thousands of temporary workers. While the jobs pay well ($22.00 an hour in San Francisco; $11.75 in Ames, Iowa,) they last only a few months.

Bernanke and his central bank colleagues are well aware that Census hiring will skew readings, and have cautioned that unemployment will likely remain near 10 percent all year.

The Fed and private economists are trying to answer the bigger question of why the labor market shed 8.4 million jobs during this recession. Although the downturn was the deepest since the Great Depression, the job losses were even more severe than most forecasters had predicted based on models that compare economic growth and employment.

Bernanke offered two possible explanations.

“One is that maybe the recession was deeper than we thought,” he said in response to a question from a member of Congress last week. “The other is that the productivity gains were greater than we thought they would be when firms were able to cut their work forces and still maintain output.”

CAUSE FOR OPTIMISM

The first theory gained support when one of Bernanke’s staff economists wrote a research paper suggesting that the most commonly used measure of U.S. economic growth, gross domestic product, had understated the depth of the recession and overstated the recent recovery.

The economist, Jeremy Nalewaik, argued that a lesser known measure called gross domestic income may give a more accurate assessment of the business cycle. GDP looks at spending to measure the size of the economy, while GDI focuses on income.

Based on GDI, the economy began contracting in 2007, not 2008 as GDP data indicates. It also shows growth did not resume until the final quarter of 2009, while GDP showed the economy had expanded in the third quarter as well.

If GDI is indeed a more accurate gauge, there is reason to think employment will soon rise. Data released last Friday showed GDI jumped at a 6.2 percent annual rate in the fourth quarter, even faster than GDP’s 5.6 percent pace.

That would also help explain why payrolls were still contracting eight months after GDP indicated economic growth resumed. Employment gains normally lags economic growth by a few months, so if the cycle turn came in October rather than June, it would make more sense to see job growth now.

If the U.S. economy does indeed show large job gains for March, it would pull ahead of the euro zone, which is expected to report on Wednesday that the jobless rate ticked up to 10 percent in February from 9.9 percent the prior month.

The U.S. unemployment rate is at 9.7 percent, and the consensus view is that it will hold there in March. What happens after that is open for debate.

Some economists think it will inch up again later this year, for a somewhat counter-intuitive reason. As the labor market improves, discouraged workers may decide to start looking for work again. Those who give up the search are not officially counted in the unemployment rate, but when they jump back into the labor pool they are.

David Rosenberg, chief economist at money manager Gluskin Sheff in Toronto, is more concerned that the economy will weaken just as the Census jobs are disappearing. Government stimulus spending will be fading later this year, and the Fed may be cutting back on its extraordinary economic support.

“I would be looking for a second-half growth relapse that sees the unemployment rate climb back to a new cycle high once the Census hiring effect subsides,” Rosenberg said.

(Editing by Leslie Adler)