Sharp drop in new claims for unemployment

New applications for U.S. unemployment benefits fell sharply last week, putting them at their lowest level in almost four years, government data showed.

Yet at least part of the plunge stems from seasonal quirks that might prove temporary, making it difficult to gauge the underlying trend in claims and determine whether layoffs will continue to decline.

Jobless claims sank by 50,000 to a seasonally adjusted 352,000 in the week ended Jan. 14, the Labor Department said Thursday. It's the lowest level since April 2008.

Economists surveyed by MarketWatch had estimated claims would decline to a seasonally adjusted 375,000 from 402,000 in the prior week. Claims from two weeks ago were revised up by 3,000.

The number of claims usually gyrates in January owing to the end of seasonal jobs filled during the holiday period running from Thanksgiving to Christmas.

In the first or second week of January, claims tend to rise to the highest level of the year on an unadjusted basis. That's usually followed by a sharp decline.

Labor economists try to adjust for seasonal variations, but their efforts sometimes fall short. The gap between the government's estimate for last week's decline and the actual decrease was unusually large, resulting in the biggest one-week drop in claims since 2005, a Labor official said.

As a result, economists say, it's hard to figure out the direction of claims until late January or February.

"While the massive drop in claims is good news for the labor market, it probably needs to be taken with a grain of salt," Daniel Silver of JPMorgan said in a note.

Economists say the four-week moving average is especially important at this time of year because it smooths out seasonal oddities. The average of new claims over the past four weeks fell by a much smaller 3,500 to 379,000.

Still, the four-week average puts claims at a level historically associated with a modest rate of hiring and a recovering labor market, though the data is not directly linked to employment since it reflects the number of people who have lost jobs.