President Bush recently announced that inequality has "been rising for more than 25 years." But what did the president mean by inequality? And when, during that 25- year period, was inequality rising?

Edward Lazear is chairman of the President's Council of Economic Advisers, so he seems a likely source for the president's impulse to be mistaken for a Democrat. In a speech last May, Lazear said, "There is little doubt that there has been a 25-year trend of a growing gap, sometimes called income inequality, between the wages of the skilled and the unskilled."

"College graduates," he explained, "earn about two-and-a-half times as much as high school drop-outs." If that weren't the case, why stay in school?

The "skill gap" between earnings of college graduates and dropouts is not what "income inequality" means. Economists who try to explain differences in household income by differences in earnings among full-time workers are ignoring huge differences in the number of such workers. There were 16.7 million full-time workers in the top fifth in 2005, but only 3.2 million in bottom fifth. The emphasis on earnings inequality also ignores the source of most income of the bottom fifth -- namely, more than $1.5 trillion of government transfer payments such as Social Security, unemployment benefits, the Earned Income Tax Credit, Medicaid and food stamps.

Even if wage gains are still growing much faster among college grads than among high school grads and dropouts (which has been unclear since about 1993), that certainly would not imply larger numbers of Americans are stuck in low-wage jobs today than in the past. In 1980, only 14 percent of the workforce had four or more years of college, while a third had less than a high school diploma. By 2005, a third had graduated college and only 10 percent had less than a high school diploma. The relative wages of college grads and dropouts are far less significant than their relative numbers.

Fed Chairman Ben Bernanke joined the presidential chorus in a recent speech. He too said, "Rising inequality ... has been evident for at least three decades." Like Lazear, he quickly changed the subject from income to some gap in "real weekly earnings" between the top and bottom 10 percent of "full-time wage and salary workers" in 1979 and 2006. Average earnings of those narrow slices of full-time workers cannot possibly measure actual income inequality, unfortunately, because very few low-income households contain anyone who works full-time.

Looking at data for 1979 and 2006 cannot tell us whether any change between those years happened recently or a long time ago. He even notes, paradoxically, that "wages of workers in the middle of the distribution have grown more slowly in recent years than those of workers at the lower end of the distribution."

Bernanke did mention disposable income, but provided data for only two years, and that data is incorrect. He said, "The share of income received by households in the top fifth of the income distribution, after taxes have been paid and government transfers have been received, rose from 42 percent in 1979 to 50 percent in 2004." The top fifth's share of income before taxes and transfers was 50 percent in 2004, so it could not possibly have been unchanged after taxes were subtracted from top incomes and transfers added to lower incomes.

The correct figure for the top fifth's share of disposable income was 44.9 percent in 2004 -- essentially unchanged from the 44.8 percent figure of 1993. The top fifth's share of disposable income did rise from 40.7 percent in 1979 to 43.4 in 1985 and 43.5 percent in 1988. But it still remained at 43.3 percent in 1992, and the apparent increase since then (to 44.9 percent) is a statistical illusion.

In 1993, Census surveys were computerized and the maximum amount of recorded earnings from one source increased from $299,999 to $9,999,999. As soon as that happened, the top fifth's share suddenly appeared to jump by 1.5 percentage points to 44.8 percent in 1993. But that was no increase in inequality. The Census Bureau was simply collecting better information on high incomes. Aside from the statistical aberration of 1993, in other words, the top fifth's share was essentially unchanged from 1988 to 1992 and unchanged from 1993 to 2004. The related Gini index for disposable income -- a broad measure of income inequality for all households -- has been virtually unchanged since 1985, aside from gyrations caused by the problematic inclusion of taxable capital gains.

Citing the Congressional Budget Office (CBO), Bernanke claims, "The share of after-tax income garnered by the households in the top 1 percent of the income distribution increased from 8 percent in 1979 to 14 percent in 2004." Two-thirds of that alleged "trend" in the top 1 percent's share ended in 1988 and the rest in one year -- 2004. In a Feb. 6 Wall Street Journal article with David Henderson, I demonstrated that the top 1 percent's share appeared to increase only because the CBO incorrectly adds an unbelievably huge and rising share of corporate profits to top incomes. Even with that overestimate included, however, the top 1 percent's share of after-tax income increased from 8 percent in 1979 to 12 percent in 1988 and was still 12 percent in 2003.

Nobody can possibly prove a "long-term tendency toward greater inequality" by drawing an imaginary straight line between two years such as 1979 and 2004. When officials in such high positions claim income inequality has increased "since 1979," are they not obliged to confide that the increase ended around 1988?