Widely used figures from the U.S. Bureau of the Census are misleading, according to a study from the Heritage Foundation's Center for Data Analysis. Researchers analyzed the Census Bureau's figures on the distribution of income and found that they overstate the degree of inequality and obscure important factors that determine it.

The trouble begins when, in order to measure income distribution, the Census Bureau ranks households from highest to lowest incomes, divides them into five quintiles, and determines the share of total income received by each quintile. Since the quintiles are based on households, they do not contain equal fifths of the U.S. population, and are in fact unequal in size. For instance:

• The top Census "quintile" contains not 20 percent of the population but 24.3 percent, while the bottom quintile contains only 14.8 percent of the population. (see figure).

• This is due to differences in household size -- for instance, 54.9 percent of the households in the bottom quintile consist of just one person, compared with only 7 percent of households in the top quintile.

• In addition to size, the households differ in other important respects -- for instance, more than twice as many people in the Census' top quintile are of working age, compared to the lowest quintile (see figure).

When the figures are adjusted to include a broader measure of income, and the income distribution is divided so that each quintile contains equal numbers of people, the share of income received by the bottom quintile is nearly three times higher than Census reports (see figure).

Furthermore, the relative inequality of the bottom half of the distribution to the top half shrinks substantially: while the Census shows the bottom half receiving roughly $1 of income for every $4 received by the top half, the difference shrinks to $1 for every $2 after the adjustments (see figure).

Source: Robert Rector and Rea Hederman, "Income Inequality: How Census Data Misrepresent Income Distribution," CDA Report No. 99- 07, September 29, 1999, Center for Data Analysis, Heritage Foundation, 214 Massachusetts Avenue, N.E., Washington, D.C. 20002, (202) 546-4400.

For text http://www.heritage.org/Research/Labor/CDA99-07.cfm