Last week we pushed the case for thinking that the recent decline in new jobless claims was more than a statistical glitch. That was a slightly harder sell after looking at the numbers through November 27, a week that suffered a hefty gain. But the broader trend still pointed down, we opined, a view that returned to the good graces of the statistical goods with this morning’s update. Yes, there’s encouraging news in the latest tally of initial jobless claims, which fell a solid 17,000 last week. But while the widely cited seasonally adjusted number offers fresh encouragement today, a new front in the war on job growth may be opening up once we consider last week’s seasonally unadjusted jobless claims total.

But first the good news. Seasonally adjusted jobless claims dipped to 421,000 last week. As the chart below shows, that’s near the lowest level in more than two years. Indeed, other than the week through this past November 20, the latest reading is the lowest level in more than two years. It’s been a long time coming. After nearly a year of treading water in this series, the trend seems to have broken through the glass floor. That suggests the labor market will continue to recover, if only mildly.

Surprising? Not really. The economy was either headed into a new recession or the bias was set to tip toward growth on the margins. As we’ve been discussing in recent months, the evidence has been building in favor of the growth argument on multiple economic fronts. It’s no shock, then, to see initial jobless claims trend lower…eventually. Eventually seems to have arrived. Better late than never.
So far, so good. But then we stumbled across a rather frightening surge in the unadjusted claims numbers for last week. This raw measure of new filings for jobless claims soared by more than 169,000 for the week through December 4—the biggest weekly gain by far since January 2010. Yes, the unadjusted trend in jobless claims is a wild statistical pony and therefore it’s subject to volatility levels that can make even battle-scarred statisticians dizzy. In that case, the usual caveat about ignoring any one number goes double here. By definition, the unadjusted numbers are subject to a wide range of seasonal quirks that can and do mislead us in deciphering the true trend. Nonetheless, no one needs an excuse to wonder about what could go wrong these days, and so the reversal of fortunes in unadjusted initial claims speaks for itself, as the second chart below reminds.

It’ll take another week or two to figure out if the unadjusted claims are just statistical noise on steroids. For the moment, we’re assuming just that. There’s too much contrary evidence swirling about to give in to fears tied to any one negative number, albeit a rather large negative number. But forgive us for being a little paranoid. We’ll sleep better after seeing the unadjusted trend fall back to earth. But while we’re waiting, the guessing game just got a little more challenging.