Yearly decline in average jobless claims of nearly 30,000 still not pointing to recession

There has been a bit of talk about jobless claims since the abnormally large fall in claims two weeks ago. Those numbers were skewed by reporting anomalies that were adjusted in the subsequent week. Looking beyond the week to week data, however, one sees a discernible trend down in jobless claims. And this is not consistent with recession.

When I started Credit Writedowns during the recession in 2008, I used to look at the jobless claims data a lot. Previously, I had noted that a rise or fall in the change in jobless claims was a meaningful number. Put simply, recessions occur when jobless claims go up a lot and recoveries occur when they go down a lot. I used to look at both continuing claims as well as initial jobless claims. But because this recovery has been so weak and so many people are either on jobless claims for a long time or have been forced out of the labor market, I now only look at initial claims as meaningful.

The number that is most important here is the year-on-year change in unadjusted 4-week average initial jobless claims. That is quite a mouthful. But the reason  I use this number as the primary one is three fold. First, I don’t like seasonal adjustments. Especially at economic turns, seasonal adjustments tend to be wrong, making it difficult to compare data across economic cycles. So I use unadjusted numbers. Second, if I use unadjusted numbers, I am forced to make strict year-to-year comparisons because the seasonal adjustments, while flawed, are good at taking the natural seasonality out of the data. It is impossible to compare April data to December data without seasonal adjustments. That’s the reality. So I only use unadjusted year-over-year data. Third, I use 4-week average data, the standard number the US Department of Labor uses, to smooth out week to week fluctuations.

The year-on-year change in unadjusted 4-week average initial jobless claims is -27,000. That means that 27,000 fewer people filed for jobless claims each week these last four weeks than did a year ago. We would need to see this number at +50,000 for it to be consistent with recession.

Bottom line: The U.S. is not in a recession. It is in a tepid recovery.

Also see the following posts for how I applied the data to predict recession and recovery in 2008 and 2009 and continued recovery in 2010:

• Jun 2008: Another Perfect Recession Indicator
• Sep 2008: US unemployment claims way up: consistent with recession
• Apr 2009: Jobless claims may signal the end is near
• May 2009:  Both initial claims and continuing claims now pointing to recovery
• Aug 2009: Revisiting employment indicators for signs of recovery
• Sep 2010: Jobless claims still not pointing to imminent double dip recession

Note: Many U.S. companies are international and have significant exposure to economies in Europe, Asia and Latin America, where economic growth rates are declining or negative. This will hurt the U.S. stock market even if the U.S. economy is doing relatively well. Moreover, the fiscal cliff is still something to fear as far as economic growth and earnings go.

Source: US Department of Labor