On Friday, we learned that the U.S. unemployment rate had dropped to 9.4%. An initial reaction to the news might have been celebration: it is the lowest rate the U.S. has seen since May 2009. But if you dug into the numbers, you quickly saw that the big decline didn't occur because of hiring. Instead, more workers quit looking for jobs, taking them out of the workforce and no longer counting them as unemployed. This implies two things about the nature of the unemployment problem: many Americans are having trouble finding a job for an extended period, and they're temporarily giving up.

In fact, the data demonstrates these trends quite clearly. The following chart shows how unemployment duration has changed since 2004. It provides the percentages of unemployed Americans who have been out-of-work for 27 weeks or more and for 26 weeks or fewer:

unemp duration full 2012-12.png

This is an extremely disturbing chart. The long-term unemployment percentage continues to hover around its high of around 45.6% hit last spring. It began to decline a little in the fall, but ticked back up to 44.3% in December.

This data looks even scarier when you consider the current long-term jobless numbers in a historical context. Since 1948, the highest portion of unemployed who were jobless for more than 26 weeks was 26.0%, hit back in 1983. Compare that to the current level of 44.3%. And back then, the group of long-term unemployed consisted of just 2.9 million Americans. Now it's 6.4 million.

And that doesn't even capture the entire picture, because some people who have been unemployed for a long time have stopped looking, at least temporarily. You can get an idea of this trend by looking at the labor participation rate. It's a measure of what portion of the population is considered in the labor force, meaning they're either employed or unemployed and looking for work. Here's another chart:

labor participation 2010-12.png

First, note that the chart is crafted to zoom in on how this percentage has changed since before the recession. If it were graphed from zero to 100, then the change would look negligible.

I chose these bounds for the vertical axis, because the change is anything but negligible. At current population estimates, if the labor participation rate hadn't fallen since it peaked in late 2006, then there would be another 4.3 million Americans in the labor force. These people do not have jobs, but are not considered unemployed. If they were, it would raise the number of unemployed people from 14.5 million to 18.8 million. That would increase the national unemployment rate from 9.4% to 11.9%.

This is, however, probably a slight exaggeration. Some of those who have left the workforce may be gone for good. Considering that some baby boomers are beginning to retire, their relatively large numbers could be a factor in labor force shrinkage. But that probably isn't the story for most of these Americans. Some are technically discouraged (about 1.3 million), and some have other reasons for temporarily not looking for work. According to the Bureau of Labor Statistics, 6.2 million Americans who weren't considered part of the labor force wanted a job in December. So this the 4.3 million estimate above is probably a conservative guess.

That means we'll likely see many of these Americans rejoin the workforce in coming months or years. So when you think about the number of jobs that must be created to get unemployment down to a reasonable level of around 5%, there are likely a several million additional jobs needed that official statistics don't even account for.

The nature of long-term unemployment and the declining workforce aren't statistics you generally hear tossed around in superficial coverage of the unemployment problem. But they're both very significant. One shows the serious harm that the problem is causing a huge chunk of unemployed Americans, while the other shows that even the relatively ugly headline numbers are probably too optimistic.

We want to hear what you think about this article. Submit a letter to the editor or write to letters@theatlantic.com.