One of the more annoying aspects of business press reporting is its participants' singular focus on seasonally adjusted data to the exclusion of the underlying figures.
Many reports on the economy at least tag the figures reported as seasonally adjusted; but there seems to be a trend away from doing even that. For example, the Associated Press has routinely labeled weekly initial jobless claims as seasonally adjusted (examples from about a year ago are here, here, and here), but Thursday's adjusted claims figure of 331,000 and the 348,000 from a week earlier went unlabeled (as seen here and here, respectively). Additionally, none of the three main wire services (AP, Bloomberg, Reuters) described yesterday's reported increase in employment as "seasonally adjusted" (though the AP's Christopher Rugaber did report that the unemployment rate of 6.6 percent was seasonally adjusted). In failing to do so, they all were in essence telling readers that the economy really added 113,000 jobs in January. The truth is that it lost over 2.8 million of them:
As seen above, January is historically a month during which far more employees lose their jobs than get new ones. Among the reasons: Those who were hired for the Christmas shopping season are no longer needed.
The real question in looking at January's results is whether its job losses were greater or less than previous years. The answer is that losses (pending revisions) were worse than the previous two Januarys. They were the same as January 2010, which absolutely no one remembers as a good period for the economy. You would have to go back to 2008 and 2009 to find worse results.
January 2014's seasonally adjusted result of 113,000 really doesn't tell us much. One could argue that it should have been higher, because the raw number of jobs lost is barely more than a year ago, when the seasonally adjusted figure somehow came in at 197,000. But the argument that it should have been lower is almost as strong, because the identical 2.87 million jobs lost three years ago generated a seasonally adjusted figure of only 18,000.
Seasonal adjustment calculations attempt to look at the previous five years in an attempt to discern trends and present current data in context. In a normal, consistently growing economy, these calculations can be useful. But the fact of the matter is that the job market has ranged from miserable to up-and-down for over five years, and the seasonal relationships, to the extent they can be discerned at all, are no longer what they once were.
Those who treat the seasonally adjusted numbers as the be-all, end-all — and this would unfortunately include economists and analysts trying to serve their corporate and individual employers and clients — should be digging deeper to find out what's really going on.
Journalists who won't even apply the seasonally adjusted label are clearly misleading their readers, listeners, and viewers. It's as if they believe that their audience is either too stupid to understand the truth, or doesn't need to know it.
Cross-posted at BizzyBlog.com.
