Follow-Up on Employment: Conflicting Signals

U.S. unemployment: are we at the end, or the end of the beginning, or what? Different sources give different answers, or rather, accommodate different interpretations. I think we still have a long way to go. This post comments further on ideas I considered yesterday in "In The Long Run, We're All Temps."
A personal note: Although I discuss workers below by the hundreds of thousands, I realize that every job counts. Friends of mine are unemployed today, and in the past I, too, have been unemployed for long periods. It's miserable, and aside from the little bits of hope you get now and then from calls and e-mails, feels like house arrest. (In 2001, when I was out of work for almost a year, I was even grateful to be called for grand jury duty for a month.) Complicating matters further, it's now holiday time, when we're all supposed to be out shopping and nominally cheery.

OK, let's look at some of the numbers. First, consider the government's report on jobs losses in November, compared to losses in May, six months earlier. (I choose six months for no particular reason, but May was one of the big months for job losses.) Selected details appear below.

[Click on the table for a larger image]

The change in job losses from May to November, that is, comparing just the two months, was an improvement of about 300 thousand, at the top of the rightmost column. The change was roughly even between goods-producing and services business. But 100 thousand, or about a third of the total favorable change, is in temporary employment. In total, 1.9 million people of the 131 million U.S. work force are employed by temporary help services.

It would be preferable that companies had enough confidence in the outlook to hire people outright rather than as temps, but apparently they do not. Overall, employers have not even brought back their own part-timers to full-time in great numbers: the BLS reports that in May, 5.8 percent of the workforce was working part-time against their wishes (for you statistics fans, the difference between series U-5 and U-6, on page 19 of the Current Employment Situation linked to above), and in November the comparable figure was 5.9 percent. Maybe next month.

There are two alternative measures to the official reports. One is the ADP Employment Report, from the people who calculate payrolls for 23 million Americans in all walks of life. Here's what ADP says about November. Be attentive to the last paragraph -- small businesses have been hurt the worst:

Nonfarm private employment decreased 169,000 from October to November 2009 on a seasonally adjusted basis, according to the ADP National Employment Report®. The estimated change of employment from September to October was revised by 8,000, from a decline of 203,000 to a decline of 195,000.

...

Large businesses, defined as those with 500 or more workers, saw employment decline by 44,000 while medium-size businesses with between 50 and 499 workers declined 57,000. Employment among small-size businesses, defined as those with fewer than 50 workers, declined 68,000. Employment losses among small-size businesses were the smallest since July of 2008.

A part of that survey asks whether companies are adding or subtracting workers. It showed more companies hiring than firing in both October and November.

In the past, two such months of gains in the I.S.M. employment component always came after the recession was later determined to have ended, and usually after the unemployment rate had begun to decline. The only exception was in the recession that ended in February 1961. Then, the unemployment rate peaked in May 1961, which was the also the second month that the I.S.M. employment indicator showed growth in manufacturing employment.

Before this cycle, the I.S.M. indicator has never showed consecutive gains before the unemployment rate hit its cyclical high.

Even if the unemployment rate has peaked, long-term unemployment remains a major problem.

...[T]he number of workers who have been without a job for more than 26 weeks rose to 5.9 million in November, the highest ever and more than double the number in January.

The median duration of unemployment is up to 20.2 weeks, and the average figure is now 28.5 weeks. Both are records, indicating that those who are finding jobs are not the people who have been out of work for a long time.