Opinion Editorial | |
| Monday, October 4, 1999 | |
Census Bureau's Squishy Income Distribution Numbers |
On September 30, the Census Bureau released data on poverty and income distribution in the United States for 1998. Once again, the data show a high degree of inequality. According to the Census figures, the share of total income going to the lowest 20 percent of households has fallen to 3.6 percent from 4 percent in 1985, and the share going to the top quintile has risen to 49.2 percent from 45.3 percent over the same period.
These numbers are always treated as if they are etched in stone -- objective, immutable and accurate. But in fact, they are as squishy as a wet sponge. No one is accusing the Census Bureau of manipulating their data for political purposes. However, there are serious problems with the income distribution data that are severely distorting our perceptions about rising inequality. A new report by Robert Rector and Rea Hederman of the Heritage Foundation details some of these deficiencies.
The most important area of concern is one that the Census Bureau has publicly acknowledged, and that is the failure to account for taxes paid or benefits received from government programs. At present, the official Census income data are based on before-tax money income and exclude in-kind government benefits, such as public housing and food stamps. Hence, much of what we as a society do to equalize incomes -- imposing high tax rates on the rich and spending billions of dollars on welfare -- is excluded from the Census figures.
Some years ago, the Census Bureau finally admitted that this was a serious drawback to the data and it began publishing supplementary data that take account of benefits and taxes. However, the official definition remains unchanged and it is only with some effort that the supplementary data can be found.
An equally serious problem identified by Rector and Hederman is the Census Bureau's failure to adjust the data for family size. Although each quintile by definition has the same number of households, those in the lower quintiles tend to have far fewer people than those in the upper quintiles. Indeed, the top quintile has 65 percent more people than the bottom quintile: 24.3 percent of the total population versus 14.8 percent.
When Rector and Hederman adjusted the 1997 Census data for taxes, benefits and population, it sharply reduced income inequality. It almost tripled the share of income going to the bottom quintile, while cutting the top quintile's share by 20 percent (see figure).
The Heritage study reveals many other weaknesses of the Census income data. It should be required reading for everyone writing about poverty and income distribution.
Source: Bruce Bartlett, senior fellow, National Center for Policy Analysis, October 4, 1999.
The National Center for Policy Analysis is a public policy research
institute founded in 1983 and internationally known for its studies on public policy issues.
The NCPA is headquartered in Dallas, Texas, with an office in Washington, D.C.
Julie Hillrichs, Dallas, TX 972-386-6272 Sean Tuffnell, Dallas, TX 972-386-6272 Joan Kirby, Washington, DC 202-220-3082 Internet: http://www.ncpa.org Home | Support Us | All Issues | Social Security Debate Central | Contact Us |
