Either an axiom for unemployment statistics is being rendered obsolete or something fishy is going on at the Department of Labor. The Dallas Federal Reserve mentions an old axiom on initial unemployment claims in an economic assessment. The Dallas Fed explains that initial jobless claims have trended only slightly lower so far this year, remaining above the approximate 400,000 claim level commonly associated with positive job growth.
The axiom goes as follows. If the initial weekly unemployment claims hovered above 400,000, then the labor market is likely contracting. If the figure remained consistently below 400,000, the labor market is likely expanding.
Historical data appears to back up this general rule of thumb for employment data. Yet in 2010, there is a noticeable divergence from this trend. In fact, the Department of Labor’s own statistics appear to contradict one another so far this year.
The non-farm payroll data clearly diverges from the initial weekly unemployment claims data. For the red bars in the chart below, we simply subtracted the average initial claims from 400,000 so that a positive figure would indicate net job growth and a negative figure would indicate a net job loss. Based on initial claims, it would appear our country should be shedding jobs. Instead, we are averaging higher job growth this year than in 2007 according to monthly non-farm payroll data. Does that make any sense? Yes, we know about the Census jobs inflating the data this year…but that doesn’t explain everything.
click to enlarge
Despite the fact that initial claims have not yet fallen below the key 400,000 mark all year, the Department of Labor reports that the economy has added jobs every month of this year. Economists expect the economy to add 540,000 jobs in the latest report to be released this Friday. This, despite initial weekly unemployment claims still averaging well above 400,000 thus far for the month of May. If indeed the economy does add more than a half a million jobs in May, our average monthly job creation so far for 2010 will exceed the rate at the height of the housing bubble.
So is the labor market improving at such a historic rate? Not according to initial claims. And not according to Gallup–which still shows companies are hiring at a slower rate than in 2008.
What exactly is going on here? Some in other blogs suggest that it is has something to do with the birth/death rate model.
Yet the divergence seems much more pronounced given the historical data. As we can see in the first chart, the largest gap between weekly initial claims and non-farm payroll statistics took place in 2008. Much of this can be explained by rather weak growth early in the year followed by severe hemorrhaging of the employment market near the end of the year. There were simply more data points for weekly claims that prevented a spike as severe as the monthly data.
No such explanation can explain the 2010 data. Perhaps there is some sort of politically-driven data manipulation. At first glance, this theory doesn’t make sense as both statistics are reported by the Department of Labor. But the initial claims are actually reported by the states and compiled by the DOL–so they are less prone to manipulation at the DOL. So perhaps everything is accurate and this is just a temporary divergence in the data. We simply don’t know.
We just think the Department of Labor should take the time to explain why the Dallas Fed and many other economists are suddenly wrong when using the rule of thumb on initial claims. This could all be just one big misunderstanding. Maybe the rule of thumb has been done in by some fat fingers at the Department of Labor. At least that explanation would make some sense.
For investors, the only thing worse than bad employment statistics are employment statistics you can’t trust.
Disclosure: I do not have a position in any stocks mentioned in the article.
