Thursday, July 28, 2011
For most of this year, the weekly unemployment claims have been jerked around by seasonal adjustment factors that didn't quite match the typical seasonal variations that occur. As a result, claims were under-reported in late February and March, and over-reported in April. I've been blogging about this for a long time. Now I think this little sideshow is finally coming to an end.
The top chart shows the seasonally-adjusted data, while the bottom chart shows the actual data. Note that in actual data claims have been relatively low and flat since February, with a modest spike in July (which is typical in July because auto manufacturers usually lay off workers as they retool their assembly lines). In the adjusted data, there is a huge spike in April which does not show up at all in the raw data. This was largely the result of earlier-than-expected auto layoffs. The seasonal factors expected the actual data to be weak in April, but instead it was higher than usual, so that turned into a large rise in the adjusted data.
In any event, the predictable seasonal events are now in the past, and we won't have another until later this year when workers hired to prepare for the Christmas season start getting laid off. What's happened so far this year is that all the confusion created as a result of volatile claims numbers has been much ado about almost nothing. Despite the headlines, claims have been fairly stable for most of the year; the economy is neither accelerating nor decelerating. It's steady (and slow) as she goes, although I think there is reason to expect activity to pick up somewhat in the months to come.
Posted by Scott Grannis at 7:57 AM