Friday, May 8, 2009

Net Claims is What Matters

Ahead of today's unemployment report...

Ed, from Credit Writedowns, had a post last week which sparked a back and forth between the two of us about a decline in initial claims marking the end of the recession. I wasn't / still am not so sure this time around. Ed stated:

So, jobless claims are definitely a number to watch as we head into the spring and summer. Absent claims numbers averaging 700,000 by mid-to-late summer, it will be safe to say, we are on the road to recovery. What kind of a recovery we get is another entirely different question.
A summary of my response is as follows:

I understand that a bottom in the 4-week moving average for initial claims (i.e. firings) has historically been a leading indicator for an economic bottom, but that needs to correspond with the hiring of those that had previously been laid off in areas of the economy that are rebounding. In other words, the total number of unemployed people is what is relevant and while that has historically correlated with initial claims, it has diverged this go around.

Those individuals already unemployed, have remained unemployed FAR longer this cycle than past cycles, hence the reason why continuing claims continues to grow, even while initial claims reverses course. Thus, even with a decrease in initial claims, continuing claims will continue to rise (if the number of people laid off is greater than the number of people rehired into the workforce, the overall unemployment rate will continue to rise as well).

An example... if the number of people laid off drops from 600,000 to 500,000 (a SUBSTANTIAL drop), the unemployment rate can still grow if either of the following two things happen (all else equal):
  1. The workforce grows (we need to hire people just to keep the unemployment rate level)
  2. The number of people rehired + those that fall out of the workforce is less than the number fired (i.e. 500,000)

Initial claims are ONLY new jobless claims filed by individuals seeking to receive state jobless benefits, thus only those people that have been fired.

Continuing claims take into account those people that have already been fired for at least a week and continue to receive state jobless benefits.

Thus, if firms are firing less, but hiring by a smaller amount than that... continuing claims rise even though initial claims fall.

Here are charts of this exact phenomenon. It shows the one-year and 4-week changes in the number of those fired (i.e. the initial claims) less than number of people that have been hired (or fallen out of the workforce). To do this I took the previous week’s initial claims and removed the change in the continuing claims.

Rolling 12-Month

Rolling 4-Week

To conclude… even though initial claims is slowing / declining, the above figure keeps rising (to be a drag all it needs to do is stay above zero). In most recessions, firms that are strong will begin to hire more people when they sense a rebound than are fired by struggling companies, thus when initial claims flatten it has marked a bottom. I personally question whether this will happen this time. From what I’ve heard in the job market, many firms still have a hiring freeze through the year and/or are still shedding capacity.


  1. I'd be really interested in seeing a charts for Fired less Underemployed. How many part time workers really want more work?

    For what it's worth...

    I work as a W-2 employee on a project by project basis. One of the things I've noticed - and this seems to be a relatively new phenom - is that I am being contracted for fewer days. So what used to be a typical 5 week project (25 days) is now shaved down to 4.5 weeks (22 days).

    Yeah, it's just 3 days. But multiply that times 100 people and it's a year of employment in man days that's gone poof.

  2. thanks for the comment. i'll see what i can put together...