Unemployment Claims: Glass Half Empty or Half Full?
Early-'80s comparisons get you only so far.
As the initial unemployment claims data roll in week after week, there have been many breathless and dire comparisons to 1981 numbers. It's not without reason--the four-week moving average of jobless claims (see below) paints a scary picture. Just part way through the current recession, which began in December 2007, initial claims statistics in the recent era are worse than all recessions except the back-to-back recessions of the early 1980s.
However, an analysis like this misses the fact that the economy and the number of people employed have grown dramatically over the last 40 years, rendering the raw number of claims virtually worthless.
Instead, we think it is more appropriate to look at initial claims as a percentage of workers covered by unemployment insurance. In other words we are dividing the four-week moving average of initial unemployment claims by the number of people covered by unemployment insurance. Using that data, the graph below shows a worsening employment situation but one that is not at epic proportions.
At the moment, we have just barely passed the percentage claims level of the remarkably mild 2000 recession and are still below the somewhat worse 1990 recession. There are two ways to interpret this data. The glass-half-empty interpretation suggests an awful lot of room to get worse. Clearly the disturbance in the financial markets has been much worse than any of our previous recessions and could serve to amplify normal cyclical pressures. If initial claims moved to the average of the last six recessions (0.64%), and we apply that number to an employment base of about 133 million workers, the moving average of initial unemployment claims could approach 850,000 (versus 582,500 as of Jan. 31). If we approached the worst levels of the '70s and '80s, we could conceivably go over 1 million initial claims.
|Peak Initial Claims in Past Recessions|
|Peak Initial Claims/|
Covered Employment (%)
|* Through 02-05-09|
The glass-half-full story is that seasonally adjusted claims have been in a relatively narrow range since mid-November, and they appear to have stabilized and not spiked higher. We also note that layoffs are highly concentrated in the December/January period that is rapidly drawing to a close (initial claims are seasonally adjusted, though). The ratio has also been moving up for more than a year, which could suggest the number may begin leveling out, based on past recessions. However, the claims percentage could also move down slightly, then bounce back up and perhaps repeat the cycle again, as in the 1970 recession. If the recession ends up looking something like the 1990s, we think the initial claims as a percentage of employment could get as bad as 0.5% or roughly 665,000. This would correlate relatively closely with our unemployment forecast of just over 8%.
We tend to favor the half-full story. In general, we think that analysts estimating more than 10% unemployment are a bit high, and aren't fully considering the large changes that have occurred in our economy. Today a much higher percentage of our economy is in the services sector, which is less volatile than manufacturing. In fact, many of our manufacturing jobs have moved overseas (or have been eliminated by massive increases in productivity). Therefore, we believe a disproportionate number of the job losses resulting from falling consumer spending will fall not on us but on our trading partners. Also, a higher proportion of our economy is working in government, health care, and education, which also tend to be less affected by the economy. Finally, employers seem to have a few more tools at their disposal to trim costs than the very blunt instrument of wholesale job cuts. These include cutting 401(k) matches, freezing salaries, and reducing hours. Whether these actions are effective in eliminating the need for more extensive job cuts or are merely delaying the inevitable job actions remains to be seen.