When John Kerry rattles off his tiresome lists of complaints about the economy, it is sometimes challenging to even figure out what he imagines he's saying.

 At the Democratic convention, he was emphatic that "our great middle class is shrinking." Some misunderstood that to mean that middling incomes are shrinking. Factcheck.org said, "Kerry's description of a declining middle class is supported by new Census Bureau figures showing median household income failed to grow in 2003."

 But unchanged was not "declining," and this was not at all what Kerry meant by "shrinking." Kerry was not going through this banal exercise of feigning surprise that incomes are always weak for a while after recession (median income did not just fail to grow from 1989 to 1993, it fell 5.4 percent). By saying the middle class "is shrinking," he meant a smaller percentage of households is earning a middle-class income.

 Whenever Kerry makes any negative proclamation about the U.S. economy, the famously accommodating major media feel obligated to prove he's not making it up. The New York Times searched diligently for this shrinking middle class on July 29, when David Cay Johnston reported that "Internal Revenue Service information shows overall income among Americans ... shrank two consecutive years (2001 and 2002)."

 Despite brave efforts to convert this into a Kerry-like anxiety about "modest incomes," accompanying facts showed the 5.7 percent income decline was entirely confined to those earning more than $100,000. The reason is that the IRS, unlike the Census Bureau, includes capital gains. Realized capital gains were huge in 2000 because savvy investors liquidated quickly at the start of the market's collapse. But the market kept falling during the following two years. That, plus massive job loss among corporate managers, is the main reason why taxable income in 2001-2002 fell 22 percent among those who previously had incomes above $10 million and by 5 percent at incomes between $5 million and $10 million.

 That was serious income shrinkage, to be sure, but it wasn't exactly middle class. National Review contributor Don Luskin noted that the number of returns reporting incomes between $25,000 and $100,000 grew rapidly from 2000 to 2002, while the number earning higher or lower incomes (including capital gains) diminished. By this measure, the middle class was expanding rather than shrinking.

 In a valiant effort to narrow this widening gap between Kerry's campaign rhetoric and inconvenient facts, journalistic partisans quickly turned their attention to the Census figures, which exclude both taxes and the vanishing capital gains. The Washington Post devoted nearly three pages to a front-page feature on "The Vanishing Middle-Class Job" by Griff Wittee. This was supposed to be a really big deal -- "the first in an occasional series abut the changes roiling the middle of the American workforce."

 There were seven colorful graphs, not one of which made any point worth making. The first voiced alarm that, "The percentage of households earning close to the median income has fallen steadily over three decades."

 The second clarified the same point by saying, "In 1967 nearly a quarter (22.3 percent) of households made between $35,000 and $49,999 in inflation-adjusted terms. But that share was down to 15 percent by 2003."

 In reality, what is "close to the median" has changed, of course, because median income rose 30 percent since 1967.

 Anyone troubled by the fact that a smaller percentage of Americans now earn between $35,000 and $50,000 than in 1967 should be truly shocked to discover that the percentage earning less than $35,000 fell far more dramatically -- from 52.8 percent to 40.9 percent. Why isn't The Washington Post planning a series on the presumably horrible shrinkage of all those vanishing lower -class jobs? Continued...