Discerning the Truth In Employment Data

Just as it’s been the principal benchmark that financial markets have used to evaluate the economic recovery for the past three years, job creation has become the central economic focus of the U.S. presidential campaign.

Benjamin Disraeli’s statement on the three types of lies is appropriate because the way in which the employment statistics are presented is varying drastically depending on whether it’s the campaign of Senators John Kerry and John Edwards or that of President George W. Bush and Vice President Dick Cheney using the data.

Two quick examples of the selective use of the employment data illustrate this:

-The Kerry-Edwards website baldly states that “since January 2001, Bush has presided over the loss of 2.9 million private sector jobs.” In fact, private sector jobs did decline by 3,355,000 from when the recession began in March 2001 until bottoming out last August, but in the 10 months since, 1,553,000 jobs have been created, so the net decline is a smaller, albeit still large, 1.8 million.

-The Bush-Cheney site spotlights only the job gains since last August as if the losses before that never happened and only the last 10 months matter. In addition, the Bush-Cheney site highlights a conclusion that more of the job creation has been high-paying than low-paying, although the definition of a high-paying jobs cited would seem surprisingly low to most people.

The misuse of the jobs data by both sides is only likely to increase in the months to the election — which will include the release of three more jobs reports — at a time when financial markets remain keenly sensitive and reactive to any signal of labor market health.

By Any Measure, A Weaker-Than-Usual Jobs Cycle

The most accurate way to put the employment data in context is to compare its cyclical behavior this time around with past recessions. Looking at four of the recessions of the past 35 years (the short 1980 downturn is excluded because the recovery only lasted a year), the 1,636,000 jobs decline from the March 2001 peak to the November 2001 trough was the second largest, only exceeded by a 2,824,000 decline between July 1981 and November 1982.

However, when expressed in percentage terms (which corrects for the long-term growth in payrolls between recessions), the latest cycle showed a 1.2% decline in jobs from the peak in activity until the recession hit bottom. That’s a fairly middling experience: there was a 1.2% decline in 1969-70, a 1.6% decline in 1973-75, 3.1% in 1981-82 and 1.1% in 1990-91.

The big change is in the 30 months since the recession ended. From November 2001 through June 2004, payrolls increased by just 430,000, or 0.3%. By contrast, 5.6 million jobs were added (up 8.0%) in the 30 months after the 1969-70 recession ended, 5.9 million jobs were added (up 7.6%) after the 1973-75 recession, 8.5 million jobs were created (up 9.6%) in the 30 months after the 1981-82 recession, and 2.1 million jobs were added (up 2.0%) after the 1990-91 downturn.

While the job losses during the latest recession were fairly average, the rebound afterwards is clearly the weakest of the past generation, largely because job losses continued well after the recession itself ended.

Federal Reserve Chairman Alan Greenspan’s monetary policy testimony Tuesday and Wednesday presented an opportunity for legislators to try to put a partisan spin on the data. Some asked leading questions, hoping to trick the Chairman into making a reply that would favor one party’s position over the other. Greenspan, for the most part, was too wily for that.

For instance, Sen Paul Sarbanes (D, Md.), the ranking Democrat on the Senate Banking Committee, asserted that this is the first time since the depression that the U.S. economy hadn’t regained all the jobs lost 39 months after the recession began. He then went on to assert that “If we (had) just matched the average (of post war recoveries), we’d have about 6 million more jobs currently.”

Greenspan’s reply cited “two basic forces at play,” the much stronger productivity at work this time around which substituted for hiring, and the fact that “we’ve had the shallowest recession in the post World War II period, so that the normal rebound that we experienced in a lot of recoveries…was not possible.”

Are High- Or Low-wage Jobs Being Created?

The Bush-Cheney campaign cites a BusinessWeek analysis that takes issue with the charge that most of the jobs being created in the U.S. are low-paying. The magazine cited Bureau of Labor Statistics research that said “48% of American workers belong to occupation/industry groups where the median pay is $559 a week or more,” adding that employment growth in those “higher-paying groups” accounted for well over half of total job growth during the past year.

The Bush campaign touts this as evidence that the “economy is creating more high-paying jobs than low-paying jobs.”

In fact, a $559 weekly pay equals $29,058 per year, which compares poorly with the $29,548 level of per capita after-tax income. (That’s after-tax income to every man, woman and child in the country, well below income per employee of $61,353, an estimate based on the total disposable personal income in May divided by the number of employed in June.)

Even so, the implied income looks a bit on the high side. The BusinessWeek analysis was done using unpublished data compiled by the BLS (obtained by Dow Jones Newswires) from the household survey of employment.

According to Randy Ilg, an economist at BLS, the detailed data are compiled for 11 occupation groups (such as management, professionals, sales, and production) in each of 14 industries. There are 154 separate occupation/industry groups.

Dow Jones looked at these occupation/industry by whether they had gained jobs or not between August 2003 and June of this year. For the 78 groups (50.6% of the total 154), the median weekly income was $504 ($26,208 per year), while the median weekly earnings for all 154 groups was $541 ($28,132 per year).

That implies job gains occurred mainly in lower-paid jobs, but that the difference was small and the conclusion depends heavily on the dates chosen.

It’s clearly possible to use the employment statistics to support whatever story one wishes to tell. Whether it’s a lie or not depends on the listener’s willingness to believe.

— John McAuley, who writes about the economy for Dow Jones Newswires, is a former Wall Street economist. He holds a Ph.D. in international trade and macroeconomics and has been teaching economics at Fordham University since 1974.

 

By John McAuley

Business, Information, job