At 8:30 on the morning of the first Friday in December, the Bureau of Labor Statistics (BLS) reported that the unemployment rate had fallen to 10% in November from 10.2% the month before. Hooray! Headlines heralded the unexpected drop. Stock prices surged. Enthused White House press secretary Robert Gibbs: "We're moving in the right direction."
By late morning, though, stocks were slumping. Commentators began to focus on concerns with the numbers. By the following Monday, Federal Reserve Chairman Ben Bernanke was warning that "we still have some way to go before we can be assured that the recovery will be self-sustaining."
So much for that fall in unemployment, huh? It was a telling reaction, indicative of the still gloomy national mood, the perceived fickleness of monthly economic indicators — and the diminished status of the unemployment rate as a statistic. Once the indispensable, largely unquestioned measure of the state of the job market, it is now treated with suspicion and disdain. With good reason, because the unemployment rate fails to accurately reflect just how bad things are out there.
Each month, interviewers contact 60,000 households — most by phone, some in person — and ask about the employment status of household members age 16 and over. Those who don't have jobs but have looked in the past four weeks are classified as unemployed. After some statistical adjustments to extrapolate the data from those 60,000 households to the total U.S. population, the number of unemployed is divided by the size of the labor force (employed plus unemployed), and there's your rate. Measured that way, unemployment still isn't as bad as it was at the lowest point of the 1981-82 recession, when it hit 10.8%. And it's nowhere near what it was in 1933, when the rate peaked somewhere around 25%.
This method of calculating unemployment was pioneered by the head of the Massachusetts Bureau of Statistics of Labor in 1878, and it has its merits. It's simple. It's straightforward. And it provides a pretty accurate count of those who really, really want jobs. But it also misses millions of people who may not be actively looking for a job but would happily take one if offered. Those ranks surely swell in a deep recession or during a time of economic turmoil that destroys entire job categories (like autoworker). The government's statisticians are aware of this, and since the 1970s the BLS has published broader measures of unemployment that include at least some of these people. In 1994 the broadest measure — which counts as unemployed those who have looked for work in the past year but not the past four weeks, plus part-time workers who would rather be working full time — was dubbed U-6 unemployment. During this recession, it has gotten far more attention than ever before. U-6 unemployment was at 17.2% in November, down from 17.5% the month before and up from 8.4% two years ago. These figures aren't strictly comparable with those from before 1994, but the New York Times has taken a stab at recalculating the earlier numbers — with help from the BLS — and estimates that U-6 unemployment peaked in December 1982 at 17.1%. Meaning this recession is worse.
Even these figures leave out people who say they want a job but haven't looked in the past year. Economist and gadfly John Williams, whose online newsletter Shadow Government Statistics has gained a big following lately, adds them in, makes a few tweaks and gets to 21.8% unemployment in November, down from 22.1% in October.
Such measures still rely on people's own assessment of whether they want to work. A BLS study a decade ago found that these self-assessments aren't all that reliable. So how about the simplest possible job-market measure, the employment-to-population ratio? Among Americans ages 25 to 54, it was at 75.1% in November, down from 80.3% in early 2007 and — with the exception of October's 75% — the lowest it's been since 1984. Because of the entry of women into the workforce, the ratio trended upward from the 1960s through the 1990s. If you look just at men ages 25 to 54, the picture is much more dire. Their employment-to-population ratio of 80.6% in November is the lowest since the BLS began keeping track in 1948. It's 4 percentage points lower than it was in the depths of the early-1980s downturn.
There are certainly other factors at play here besides just a tough job market — more stay-at-home dads, more rich loafers, more prison inmates. But it also may be a sign that these are in fact the worst times for American workers since the 1930s. Which helps explain why there was so little excitement about that drop in the unemployment rate to 10%.
Monday, December 21, 2009
The Real Jobless Rate
In the December 21, 2009 TIME magazine article "The Real Jobless Rate," Justin Fox explains why the unemployment rate may not be the best measure of labor market conditions.