We don't like it when economists bounce back and forth between using seasonally-adjusted and non-seasonally-adjusted economic data just to make a point. Seasonal adjustment factors are normally quite good. Moreover, using them is justified. However, every once in awhile, looking at the underlying data (before it is seasonally-adjusted) can tell us things that the adjusted data does not. And we believe this is true for this past week's report on initial unemployment claims from the Labor Department.
The headline (seasonally-adjusted) number showed 473,000 new claims for unemployment payments. This was down 31,000 from the week before, the largest weekly drop since February. While this drop was a relief from recent increases in claims, the seasonally-adjusted level of claims is still relatively high.
Digging deeper into the report we noticed that the non-seasonally adjusted data tells an even more optimistic story. The actual number of claims – the raw data – fell to 380,935, the lowest level since Septmeber 2008, when the financial panic was just beginning. This is a very encouraging sign. It suggests that the underlying labor market is healthier than seasonally-adjusted data make it appear. Moreover, given the pattern of seasonal adjustments, we expect a decline in "seasonally-adjusted" claims to under 400,000 by early October.
|