US Unemployment Rate

(Click here for the Slide-show version)

Here are two sample headlines:

• "Half a million more people employed"
• "Half a million more people unemployed"

Can both are true? Yes, they can. Read on to understand how.

To compute an unemployment rate, we ask:

1. how many people want to work (how big is the "labor force"?)
2. how many of them cannot find work (how many, of the labor force, are unemployed)

This graph shows the total population of the US, broken down into four groups:

• first (between the top-most blue line and the red line), we have children (under 16 years) and people in the military or in prison, etc.
  
  
• next (between the red and green lines), we have civilian adults (16+) who do not want to work today. They could be going to school full-time, or retired, or stay at home parents, etc.
  
  
• finally, we're down to the "civilian workforce" (the green line) or civilians (16+) who want a job and are either employed or unemployed. Of these, about 90% are employed today (the bottom-most light-blue line)

Reconsider the same data in percentage terms.

See the baby-boom (higher % of kids) in the late 40's and the 1950's.

Also, note the recent divergence, where the proportion of "people who want work" has reduced slightly. Many of these are "discouraged workers", who have gone back to school, or have retired earlier than planned, or are sitting home for a while waiting for things to turn around. (Aside: In part, this is balanced by others who are working now, but who will drop back into "not looking for work". e.g. A spouse loses a job and the other spouse goes to work because of skills that are in more demand, even though that was "not the plan".)

Next, let's ignore the children and the military. Let's take "civilians 16+" as 100%, to better understand its components. We get the chart on the right.

Notice that more people joined the "workforce" from the 1960s to the 1980s. This reflected an increasing number of working women.

Indeed, while women who wanted to be part of the workforce grew from 32% to 68% (of civilian women over 16), the corresponding percentages for men fell from 86% to 71%. The net result is seen on the graph.

Go back to the headlines at the top of this post. "Half a million more people employed, compared to last year" can be consistent with a report that says "Half a million more people unemployed, compared to last year". This is because the "labor force" can fluctuate by over a million from year to year. if the workforce grows by 1 million people, but only half of them get jobs, and everyone else keeps theirs, both headlines will be true. Fodder for spin-masters.

Current situation: Unemployment is higher than it has ever been since WW-II. In all post war recessions, the rate at which the unemployment ("headline") rate returns to "normal" has more or less mirrored the rate at which it deteriorated. This chart from the excellent Calculated Risk blog illustrates the pattern.

The pattern may not repeat: this time might be different. Nevertheless, I took that chart and simply drew a "reflected" upward leg (orange), assuming we have hit bottom. Then, I drew a second slower upward leg (yellow).

These two give some idea of the range of possibilities for near-term levels of unemployment, if this recession follows a pattern like the other post-war ones.

It is a major assumption, and I don't particularly support it. Still, it is something like a default hypothesis, against which to test other theories, and against which to track actual near-term progress. I hope to update this about once a quarter.