WASHINGTON (Reuters) - The sharp winter decline in the unemployment rate has put a spring in President Barack Obama's step ahead of his re-election bid. But some economists think quirks in the jobless data mean progress could stall out soon.
Several Wall Street economists believe the government is mismeasuring seasonal shifts in the labor market, and suggest the jobless rate's sharp winter drop was partly an illusion.
If their research is on the mark, the unemployment rate could change little in the coming months as pay back, robbing the Obama campaign of what otherwise might have been steady progress in the lead up to the election.
That could deny Obama the crucial support of swing voters in battleground states who might be swayed by a fall in unemployment in the months before the election.
"We think that the improvement over the last few months dramatically overstates the underlying improvement," said Andrew Tilton, an economist at Goldman Sachs in New York.
"You will not see that rate of improvement going forward."
Goldman Sachs expects the jobless rate to end the year at 8.2 percent, barely below January's reading of 8.3 percent.
The unemployment rate has fallen by six-tenths of a percentage point from October's level of 8.9 percent. That is an unusually rapid decline and a growing band of optimists expect it to fall below 8 percent by year end.
But the recent steep drops have puzzled analysts who question whether the economy was growing fast enough to bring unemployment down so quickly.
Searching for an explanation, several Wall Street economists have crunched the unemployment numbers and now believe the darkest days of the 2007-2009 recession left a lasting impact, distorting the outcome of the government's adjustments for normal winter lulls in employment.
The government uses computer programs to filter out seasonal changes like the drop in construction jobs every winter. Without these adjustments, it would be harder to gauge the health of the labor market, as the jobless rate might rise every winter when it gets too cold to work outside. The raw unadjusted rates rose in both December and January, for example.
But because the computer programs make adjustments based on what happened in the recent past, the millions of jobs lost during the winter of 2008-2009 may have tricked the machines into expecting the winters that followed would be similarly bad.
And when the raw data for jobless rates did not rise as much as expected this winter, the computer programs appear to have over-adjusted the data downward, resulting in big drops in unemployment, JPMorgan economist Michael Feroli wrote in a note on Tuesday. In hindsight, some researchers on Wall Street argue the same effect was present last winter.
Even the Federal Reserve is puzzled by the sharp declines in the jobless rate. Fed Chairman Ben Bernanke has said officials do not expect further substantial drops this year, and economists at the central bank are studying what lies behind the recent apparent improvements in the labor market.
February jobs numbers are due on Friday and the unemployment rate is seen holding steady.
Potentially worrisome for Obama, who stands for re-election on November 6, is that the Wall Street research suggests the seasonal adjustments make it less likely the unemployment rate will continue declining through the summer and fall.
According to Nomura's research, this effect could be pronounced in the October employment report due on November 2, four days before the election. Even if the underlying jobs market is slowly improving, the seasonal adjustment could make a mediocre October jobless rate climb upward.
"You could see a rise in unemployment a few days before the election," said Jeff Greenberg, an economist at Nomura in New York.
The government analysts who produce the data defend their methods for adjusting the numbers. Some real shifts in seasonal employment may have taken hold during the recession, meaning the winter of 2008-2009 may not have been an anomaly.
They argue that excluding the grim winter of 2008-2009, as the Wall Street economists have in their research, is a judgment call that might eventually prove to be wrong
"We don't want to change our seasonal adjustment methodologies on the fly," said Tom Nardone, who helps oversee the Labor Department's efforts to measure employment.
Still, the department has made changes to other procedures after getting stung by the recession. Last year, it tweaked the way it estimates how many companies are starting up or closing down. Problems with a prior model led the government to underestimate job losses in 2008 and 2009.
Other organizations are also modifying their reports to take into account seasonal adjustment issues.
The private Institute for Supply Management in January revised seven years of data on manufacturing and the service sector, tweaking the way it seasonally adjusts the readings to treat parts of the winter of 2008-2009 as anomalous.
The Fed made similar adjustments last year to its data on industrial output. And the U.S. Census Bureau, which collects much of the economic data released by the government, is also looking into the recession's impact.
"The Census Bureau is fully aware of the particular challenges the Great Recession posed in these procedures," Mark Doms, the agency's chief economist, said in an e-mail.
Seasonal adjustment issues would not be enough to change the unemployment rate's long-term trend. The math involved takes gains from one part of the year and moves them to another. But they could take away some of the recent wind gathering in Obama's sails.
No U.S. president since World War Two has won re-election with an unemployment rate above 7.2 percent. Some political analysts, however, say it is the direction of the unemployment rate, rather than its level, that matters most to voters.
If the seasonal adjustment bites into labor data in the fall, it might affect the decisions of voters who don't already lean toward Republicans or Democrats, said James Thurber, a political scientist at American University in Washington.
Those independent voters, who are particularly important in states with close races, can wait longer than others to make up their minds.
"If the unemployment rate stagnates, it will be harder for Obama to win in the battleground states," said Thurber.