The Heritage Foundation

WebMemo #1592 on Economy

August 29, 2007

August 29, 2007 | WebMemo on Economy

Census Report Adds New Twist to Income Inequality Data

On August 28, the U.S. Census Bureau released its annual report on poverty, income inequality, and the number of Americans with health insurance. The report announces good news: In 2006, a growing economy increased the median family income and lowered the poverty rate. Furthermore, income inequality did not statistically increase.  

More importantly, the Census Bureau used a new, more accurate measure of income inequality. The Census Bureau divides the number of households into quintiles which each contain a fifth of the total households. While these quintiles have an equal number of households, they do not have an equal number of persons. The new measure, called "equivalence-adjusted income," partially corrects for household size.

To further improve the accuracy of its annual report, the Census Bureau should: (1) ensure that quintiles have an equal number of persons instead of equal households, and (2) measure after-tax, after-transfer income instead of money income. Accounting for redistributive policies such as the income tax and Food Stamps would give a better idea of the resources available to different households.

Good News in the Numbers
The report shows that median family income increased by .7 percent in 2006-from $47,845 to $48,201. With the unemployment rate below 5 percent and more Americans working, the poverty rate fell from 12.7 percent to 12.3 percent, the lowest the poverty rate has been since 2002. The Census Bureau also determined that income inequality is statistically unchanged by measuring the amount of income going to each quintile.[1]

The average income for each quintile group increased. The bottom quintile increased by more than 3 percent and the top quintile increased by 2 percent. The amount of income needed to reach the top quintile increased as well-from $94,712 to $97,032 (in 2006-CPI-RS dollars).

A Different Measure of Inequality
The report acknowledged that not all households are the same size. It confirms Heritage Foundation analysis that households in the bottom quintile are smaller than households in other quintile groups.[2] The report states that the lower quintile has a higher share of single-person households as well as smaller families, which are usually single-parent families. Expenses differ for households based on their size and their number of children. To reflect the realities of these differing expenses, the Census Bureau calculated income inequality using a measure called "equivalence-adjusted income."

This improved measure partially accounts for the smaller households in the bottom quintiles and, therefore, indicates that there is less inequality than indicated by the previous method. According to the new measure, the bottom quintile earned 3.7 percent of total income, compared with 3.4 percent using the previous measure. The share of income earned by the top quintile, which has the most people, fell from 50.5 percent to 49.4 percent. The second and middle quintiles gained income while the fourth quintile lost income.[3] The findings are not surprising; the top quintiles are more likely to have married families with children-and thus more population and workers-than the bottom quintiles.

Taxes Not Included
The report looks only at money income, which leaves out taxes and also other benefits from government programs like Food Stamps. For example, income taxes take money away from a household, and programs like the Earned Income Tax Credit (EITC) increase other households' income. Some programs-like Food Stamps-offer in-kind benefits that a household can use to purchase specific items like food. Other programs assist low-income households in areas like housing, energy, and health. For these reasons, it would be more accurate to measure the resources available to a household by including these factors in the calculations.

Census studies that are released after this report will include some of these transfers and taxes, which will lower inequality in those reports. The Census Bureau should give more weight to this data in its annual report, because programs that give non-cash benefits are an increasingly popular way to battle poverty and redistribute income. EITC, one of the most expensive redistributive programs, costs $40 billion a year.

The 2005 Census report that measured disposable income lowered the Gini coefficient from .45 to .418.[4] The share of income going to the top quintile fell from 50.34 to 47.28, while the share of income in the bottom quintile increased from 3.42 to 4.42. These are large changes and show the importance of methodology in measuring and counting income.

The Census report brings good news: Poverty declined and income increased in 2006. Median family income increased to its highest level except for the peak of the previous business cycle in 1999 and 2000. The number of households reporting income of less than $10,000 declined, while the number of households reporting more than $100,000 increased. Families are becoming richer in today's economy.

The attempt by the Census Bureau to adjust for the different household sizes in the various quintiles is an important step in producing better analysis. The top quintile contains 20 percent of the nation's households but has approximately a quarter of the population. As a result, the top quintile has a larger share of income because it has more people and workers. The Census's use of a new measure to account for this ("equivalence-adjusted income") paints a more accurate picture of income inequality.

The Census Bureau should include post-tax and transfer calculations in its next report on income and poverty. By better capturing disposable income, these factors would reveal an even lower level of income inequality in the United States. The Census report refutes the notion that "the rich are getting richer and the poor are getting poorer." Instead, all quintiles are growing wealthier with more money income.

Rea S. Hederman, Jr., is Senior Policy Analyst in the Center for Data Analysis at The Heritage Foundation.

[1] Another measure of inequality is the GINI coefficient, which estimates the distribution of income across the entire population. The GINI coefficient increased slightly from .469 to .47 in the latest report.

[2] Robert Rector and Rea Hederman, "Two Americas: One Rich, One Poor? Understanding Income Inequality in the United States," August 24, 2004, at

[3] The effect on inequality as measured by the Gini coefficient is also quite large. The Gini coefficient decreased from .47 to .454 after accounting for taxes and transfers.

[4] U.S. Census Bureau, "The Effect of Taxes and Transfers on Income and Poverty in the United States: 2005," March 2007, at

About the Author

Rea S. Hederman, Jr. Director, Center for Data Analysis and Lazof Family Fellow
Center for Data Analysis