Income gapology 101

Posted: Aug 26, 2004 12:00 AM

 A widely circulated Associated Press story says, "The wealthiest 20 percent of households in 1973 accounted for 44 percent of total U.S. income, according to the U.S. Census Bureau. Their share jumped to 50 percent in 2002."

 Author Leigh Strope missed no time getting to the real point -- political missionary work. "President Bush touts a strong economy that is growing, but polls find ... John Kerry is trusted more on the economy, with Democrats talking regularly of ?two Americas' divided between the rich and everyone else. The argument has merit, some private economists say."

 In what sense could this allegedly "growing gap between the haves and the have-nots" be called "news"? The cited Census report is a year old. And it deals with 2002 -- which was not a time when President Bush was touting "a strong economy."

 By using the word "jumped," the AP story clearly implies the income share of the top 20 percent "jumped to 50 percent in 2002." In fact, the share fell to 49.7 percent in 2002 from 50.1 percent in 2001. To describe a falling number as a "jump" seems remarkably shameless, even in an election year.

 Cautious readers might notice that, regardless of the nonexistent jump, the AP story claims to be talking about what happened "over two decades," which is creatively defined as starting with 1973. And although the story claims the share at the top "has steadily increased," it actually rose from 46.9 percent in 1992 to 49.8 percent in 2000 (when Clinton was president), but has not risen at all since then.

 To reach all the way back to 1973 in search of something to blame on President Bush raises a statistical problem I discussed in "Musky Income Myths" this February.  "It makes no sense," I wrote, "to compare long-term growth of average income in any top income group with growth below. Only the top group has no income ceiling, and the lower threshold defining membership in that top group rises whenever incomes in general are rising."

 In 2001, you needed an income above $83,500 to qualify for what the AP story mislabels "the wealthiest 20 percent" (income is not wealth). In 1973, you only needed an income above $62,069 (in 2001 dollars). Neither figure seems unduly rich, but the larger point is that this way of defining "the rich" changes over time. When the Census Bureau averaged all the income above $83,500 in 2001, nobody should have been surprised that they came up with a larger figure than in 1973, when they were including incomes $20,000 lower. The reason this bar keeps moving higher has a name. It's called progress.

 After telling us that a falling share going to the top fifth constitutes a jump, the AP story next tries to tell us that rising incomes are down. "The U.S. jobs market is soft, sending wages down. ... More than a million jobs have been added back to the 2.6 million lost since Bush took office, but they pay less and offer fewer benefits. ... Three in five pay below the national median hourly wage."

 It is undeniably true that precisely 2.5 out of five jobs always pay less than the median, because that is what median means. With a little careless rounding, anyone could always claim three out of five jobs pay either more or less than the median. But nobody really knows whether "new jobs" pay more or less than old jobs, because the data are not collected that way. Strope's claim that "new jobs are concentrated in ... lower-paying industries" is dubious and irrelevant. Would she conclude that heart surgeons are poorly paid because they work in health care or that gourmet chefs are poorly paid because they work in food services?
To defend the claim that wages are "down," the AP story says "on a weekly basis, the average wage of $525.84 is at the lowest level since October 2001." Unfortunately, that figure (which is up from $491.64 in October 2001) is only for "production and nonsupervisory workers" and is diluted by including part-time jobs. Median weekly earnings of full-time wage and salary workers were $639 in the second quarter -- up 3.7 percent from a year before. When we include benefits, total compensation in private industry was up 4 percent between the second quarters of 2003 and 2004 -- a larger gain than in any year from 1994 to 1999.

 Hundreds of newspapers around the country published this incompetent AP "news" story, which: 1) describes a drop in the share of income received by the top fifth as a jump, 2) describes a jump in wages and benefits as a drop, and 3) asks readers to be shocked that half of all jobs pay less than the median wage. This is not a liberal bias but an ignorance bias. Does the AP have any editorial standards?

 There is no excuse for such appalling indifference to reality when purporting to report on readily available information about incomes. The true meaning of these politically tortured statistics was explained, for example, in my survey of the topic in "The New Promise of American Life," a 1995 anthology edited by Lamar Alexander and Chester Finn. There is also an excellent book, "Myths of Rich and Poor" by the Dallas Fed's chief economist, W. Michael Cox.

 I am writing this just before the release of last year's estimated poverty rate, but I have no doubt that this statistics will also be badly garbled by politicized reporting. The poverty rate probably rose a bit from 2002, yet nonetheless remained far below the 13.9 percent average from 1980 to 1998 and also lower than in any single year during that period. In 1998, for example, the poverty rate was 12.8 percent. Nothing but political bias can explain why last year's below-average poverty rate will soon be reported as if it was unusually awful. But I'm sure that is exactly what is about to happen.

 The quality of economic reporting in the United States is awful and getting worse. Perhaps some philanthropist should underwrite a few joint degree programs in economics and journalism. In the meantime, watch the bylines on the most ridiculous "news