Unemployment for the year ending March 2010 was worse than previously stated, the Bureau of Labor Statistics announced Friday as part of its monthly jobs report. There were 366,000 more Americans who lost their jobs than previously counted. But those depressing numbers, or something close to them, won't be officially added to the final official count until the January 2011 monthly jobs report is released in February. Those aren't the only numbers that demonstrate the economy to be worse off than it appears. But let's go one step at a time.
The chart below includes the BLS's revision.
Click here for larger version of this Calculated Risk chart.
Last year, in February 2009, another BLS revision showed that a shocking 902,000 more jobs than previously estimated had been lost in the previous year. That was the biggest revision ever, both in absolute and percentage terms, clocking in at a 0.7 percent overall change for the period. In the two previous years, a total 382,000 more jobs than previously estimated had to be added to the numbers. Add Friday's announcement of 366,000 lost jobs that went previously uncounted and, all told, 1.65 million more jobs were lost in the past four years than the BLS's original counts have included.
Such BLS revisions - the benchmark revision - have been standard procedure since 1979. Throughout the year, the bureau relies on a sampling technique that allows it to present monthly estimates of the nation's job situation. The sampling covers some 140,000 business and government establishments responsible for about one-third all of nonfarm payroll employees. This, however, is an inexact approach. So, every year, the bureau undertakes a major revision of unemployment data it has gathered to get a clearer and better calculation of joblessness. The revision is announced in October and finalized in the jobs report released in February. You can read about the methodology behind this revision here.
This revision, call it an update, covers 98 percent of the nation's non-farm employment in about 9 million business and government establishments.
Past annual revisions, either adding or subtracting from job loss estimates, have also been substantial. In 1984, for instance, BLS reported that 353,000 fewer jobs had been lost in the previous year than estimated. In 1991, the revision found 640,000 more jobs had been lost than previously estimated. Typically, it's after volatile transition periods between contraction and growth that the largest revisions occur. Such transition periods also present problems for the birth-death adjustment, a BLS formula that has generated considerable controversy. A few analysts consider the adjustment to be a cooking of the numbers. That's not the standard view, however. They think the formula works fine when the economy is stable.
Whatever one's perspective on the methodology, however, the benchmark revision announced Friday, together with the earlier revision, showed what had long ago become obvious – the recession that began in 2007 was significantly worse for rank-and-file Americans than had been shown by the job loss that was estimated while the downturn was in its acceleration phase.
That the recession and its aftermath were, and are, worse than previously reported is something we get reminded of nearly every day.
For one thing, a fair number of the private-sector jobs being added to the economy since January aren't equal to the relatively high-wage positions that were lost. Of the 64,000 jobs added in September, for instance, 34,000 were in low-wage sectors such as restaurants and leisure activities.
For another thing, 3.1 million more Americans would be in the labor force right now if the participation rate were the same as it was when the recession began. These millions have no jobs, but statistically, they've vanished. They just aren't counted anymore. In the real world, some have retired, some have enrolled or re-enrolled in college, and some have just plain given up in despair that they'll find anything. That ought to affect the unemployment rate, but it doesn't. Even the alternative U6 rate - which leapt to 17.1 percent in the latest report - fails to take into account people who have left the labor force unwillingly. It's not fair to blame the statisticians. They have their marching orders from the policy-setters. But those numbers are giving us a much rosier picture of what's happening in the economy than what's actually happening.
As the folks at the National Employment Law Project pointed out Friday, for those without jobs who are still counted, the average duration of joblessness is now at 33.3 weeks, more than eight months, with some 9 million Americans collecting unemployment insurance, including 5 million on extended benefits.
But those benefits will be cut off at the end of November unless Congress renews them—putting millions of jobless workers on edge once again about how they will afford to pay their mortgages and take care of their kids. …
“When Congress returns in mid-November, it will have barely two weeks to act in order to avoid cutting off unemployment benefits that serve as a lifeline for millions of hardworking Americans. We’re talking about people who have worked all their lives and never could have imagined being at risk of plummeting into poverty and homelessness, but are now at the mercy of the worst job market in generations.
That Congress, I'm sure I don't have to remind anyone, will contain more obstructionist Republicans – perhaps many more – than it does now. We know full well what to expect from them on the jobs front.
Even under the best political circumstances, there are no silver bullets for the job crisis. And, yes, it is a crisis. For instance, America needs a new trade policy and an industrial policy. But there are some things that can be done that do not depend on reordering our country's interface with the rest of the planet. One of those is repairing old infrastructure and creating new, as davej recently hammered home. Getting Washington to do these things, however, is another story.
• • • • •
[gjohnsit has a more acerbic take on these numbers, and their market-moving ability, here.]