Changing Nature of Unemployment

Highlights include Wells Fargo & Co. (WFC) and Goldman Sachs Group, Inc. (GS).

To the average citizen, the unemployment rate is one of the most important measures of how well the economy is doing. However, for many on Wall Street they tend to dismiss it because it is (rightly) seen as a lagging indicator.

That does not, however, mean that employment data cannot be useful in looking at how the economy will perform in the future. Certainly unemployment is a good measure of the amount of distress in the economy. People without jobs do not have the income needed to buy products and services. They are not adding to the economy.

However, not all unemployment is created equal, and it would not be healthy for the economy to have an unemployment rate approaching zero.

The nature of unemployment is different during a recession than it is during an expansion. The most noticeable and important difference is that during an expansion, people who are out of work tend to find new jobs relatively quickly. This sort of unemployment does relatively little damage to a person's long-term financial condition.

It is when unemployment lingers that it starts creating real hardships. In the first graph below, we chart both the average or mean (blue line) and the median (red line) duration of unemployment. As one would expect, they tend to track each other closely, and the mean is always higher than the median (since you can't be unemployed for less than zero weeks).

The green line on the graph tracks the percentage of the population that is employed. Effectively those people have to support those who are not working, either due to age (kids and the elderly) or through safety net programs. After all, one way to look at the overall economy is the total number of hours worked times the output per hour. Fewer people working means fewer hours worked.

Notice the precipitous decline in the green line over the last year. Some of that is due to demographics, but much more is due to the economic downturn. Discouraged workers, those that would like to work but have given up hope and not looked for a job in the past month get excluded from the unemployment numbers, but are included when one looks at the employment-to-population ratio.

Some of the decline in the ratio -- and its utter failure to get anywhere close to the previous expansion peak in the last expansion -- is probably tied to long term joblessness. After all, most people don't get discouraged after being out of work for three or four weeks. However, when you have been sending out resumes for months and months with no effect, then one starts to get discouraged.

A few things are noteworthy, though. The first is that while the length of unemployment rises during a recession, it will continue to do so for a long time after the economy begins to expand again. Note that the peak duration of unemployment followed much further behind the turn in the economy in the last two downturns than it had in previous recessions. The average length of unemployment is now the longest it has ever been during a recession and is rapidly closing in on the peak that followed the early 1980's recession.

The second thing to note is that even during expansions, there has been a pattern where the low point of unemployment duration is much higher than in the previous recession (especially true if you count '80-'83 as one long recession rather than two separate downturns). This was particularly noticeable during the last expansion, which even at its best point could not drive the average length of unemployment down below the longest average period of unemployment following the mid-1970's recession.



Keep in mind that under normal circumstances, unemployment insurance runs out after 27 weeks. While for much of the country there are now extended unemployment benefits in place (except where certain states have disallowed this), if the average is now 20 weeks, it means that there are lots and lots of people that are over that 27 week mark.

The second (green) graph shows the total number of unemployed, and then two key sub-groups -- the long term and the short term unemployed. There should be a bit of an upward bias to all three of these numbers due to population growth, but the number of short term unemployed (orange line) is really notable for its stability.

One would think that when unemployment is rising quickly, as it is now, that the number of short-term unemployed would increase rapidly, but that is not the case. The increase in long-term joblessness (red line) is much more dramatic.

We are now at an all-time record in both the total number of unemployed and the total number of long-term unemployed. Given the tendency of this to peak long after the economy has turned, it is not unreasonable to expect the total number of long-term unemployed to top the 4 million mark before this is all over.

There are generally two stages that people's budgets go through when one loses a job. The first comes immediately after losing the job, and a second round of belt tightening comes when the unemployment benefits run out.

These are the people who are most at risk of defaulting on their mortgages and credit cards. When housing prices were going up, they had the option of trying to sell the houses. Now if they are under water, they have no choice but to walk away. Thus even though we have had some good recent earnings news from financial stocks like Wells Fargo (WFC) and Goldman Sachs (GS), I still think that there are lots of problems still ahead for the sector.

Keep in mind that there is no indication that the overall rate of unemployment has stopped rising, which means that these problems are going to continue to get worse for some time to come. Indeed the slope of the rise of the unemployment rate (green line) is as sharp as anything we have seen in the last 40 years.

Even if the "green shoots" that people are seeing in the economy turn out to be crops and not weeds, it is likely that unemployment will still continue to rise through at least the middle of 2010. Even if the economy were to stop losing jobs this month, population growth would tend to keep the unemployment rate rising. As a rough rule of thumb, zero job growth would equate to about 0.5% rise in unemployment over a year. We are a long way away from getting to zero job growth.



So given the way that unemployment tends to be a problem that persists well after the economy has turned, are there any employment related measures we can look at to tell if the economy is turning around? Yes -- the four-week average of weekly new claims for unemployment is an excellent indicator of the economy turning.

In the third graph (from http://www.calculatedriskblog.com/) notice that new claims peak just before the recession ended in each of the last five downturns. Occasionally there are false signals, so the decline needs to be pretty sharp to be confident that happy times are here again.

From the current level of 657,000, I would want to see new claims fall to about 620,000 (a decline of 35-40K) or so before I declared victory on the economy. Even that level would still mean that unemployment would be rising, and it is not inconceivable that we could set a post-war record for unemployment -- but if that happens, it will be likely that we are at the valley floor for the economy.

That is a good first step, but will not mean that we are climbing the mountains on the other side of the valley. For the legions of long-term unemployed, that valley might seem like Death Valley, but it will still be a good sign when we get to that point.

Read the full analyst report on WFC