The slow but inexorable decline of jobless claims

Initial jobless claims for the past week were 505,000, tied with last week for the lowest since January. This brings the 4-week average down to 514,000, the lowest since November of last year and the 11th consecutive week of declines. 

Clearly, fewer people are losing their jobs.  But, 505,000 is still a large number – one consistent with a net loss of 150,000-200,000 jobs per month. And there are still 5.6 million people with continuing unemployment insurance claims, 1.8 million more than at this time last year.

I see this as reflective of an inexorable but slow decline in layoffs complicated by a lack of new jobs, which will keep unemployment elevated as far as the eye can see. So, yes, the employment market in the United States is improving. But it is improving slowly, leaving many long-time unemployed workers bereft. This is what is called structurally high unemployment – and its happening at a level much higher than even I thought likely.

Here are two nuances in the data flow.

Dichotomy in seasonal and non-seasonal data.

All of the numbers above are seasonally-adjusted data. But, I tend to think they overstate the number of job losses (this particular point puts me somewhat at odds with David Rosenberg).  My reason is simple: we are in the recovery part of the business cycle, which means that any seasonal adjustment bias will tend to elevate the numbers.

For example, when you look at year-on-year numbers, the seasonally adjusted continuing claims data show an average 1.82 million more workers filing claims. But, the non-seasonally adjusted data show 1.56 million.  That’s a pretty sizable difference. This leads me to conclude that the seasonal adjustments may be overstating continuing claims.

Year-on-year data much worse for continuing claims

On the other hand, the initial claims numbers are flat with year-ago levels. While, as I indicated above, continuing claims are greatly elevated. To me, this points to the likelihood that the unemployment rate will rise much higher. Last year, we were seeing 500,000 layoffs as well. Yet non-farm payrolls were declining at a greater rate (380K in October, almost 600K in November and almost 700K in December). I see only two ways this is possible: 1. more people are getting hired today so the net job loss is less; or 2. more likely, non-farm payrolls understate the number of job losses because of losses not tracked in small businesses.

I have not been able to reconcile these two points. I hope to see a trend which explains both points develop in the coming weeks.