Skip Main Navigation Skip To Navigation Content


Skip top of page navigation
Census.govPopulation Poverty Main Poverty DataIncome, Poverty and Health Insurance in the United StatesPoverty in the United States: 2002 › Press Briefing

Poverty in the United States: 2002 - Press Briefing

Dr. Daniel H. Weinberg, Chief, Housing and Household Economic Statistics Division Press Briefing on 2002 Health Insurance Estimates from the Current Population Survey, September 26, 2003

Good morning. I'm Dan Weinberg, Chief of the Housing and Household Economic Statistics Division at the Census Bureau. The income and poverty data are from the Current Population Survey, the source of the official poverty estimates, not from the American Community Survey, from which data were released earlier this month [JPG - 41k] 1.

Present with me today are a number of analysts who worked hard on these reports; they will be available to answer your questions after the briefing. I'd also like to thank all the Field Representatives and telephone interviewers who work so hard to collect these data, and particularly to thank the households who answered our survey questions. Please hold your questions unless it's a technical clarification.

Let me first summarize the main findings from the income and poverty reports. These statistics come from the Annual Social and Economic Supplement to the Current Population Survey2.

Now let me give you some of the details.

Chart 2 presents the key estimates of median household income [JPG - 81k]. As I noted earlier, median money income for all households fell 1.1 percent between 2001 and 2002 to $42,400. These estimates reflect the effect of the recession that began in March 2001 and ended in November 2001. Real median household money income has declined 3.4 percent between 1999 and 2002, statistically the same as the decline in money income in a comparable period (1989 to 1992) covering the July 1990-March 1991 recession. Overall, real median household money income is up 30.1 percent since 1967, the first year median household money income was computed.

As you can see from the chart, we are also featuring four additional measures of household income this year. These four take account of the income and payroll taxes people pay, and the noncash benefits they receive4. The Census Bureau has been computing comparable measures for roughly two decades. I apologize for the acronyms, but that was the only way to make the chart readable. The acronyms are explained in the report.

While the basic pattern of growth and decline is more or less the same over time for all these measures, there are differences, especially over shorter periods. For example, between 2001 and 2002, money income after taxes fell by 0.8 percent, but there was no statistically significant change for the other three comprehensive measures shown.

The number in poverty in 2002 - 34.6 million people - was 1.7 million more than in 2001. As Chart 3 shows, this increase led to a higher poverty rate in 2002 - 12.1 percent compared with 11.7 percent in 2001 [JPG - 76k]. The poverty rate is 0.8 percentage points higher than its recent low point of 11.3 percent in 2000. Chart 3 also indicates what has happened to the percentage of people in "severe" poverty (those with money incomes half or less of their thresholds) and the "near poor" (those with incomes between 100 percent and 125 percent of the thresholds). The 14.1 million people with incomes less than half their thresholds represent 4.9 percent of the population (and 41 percent of the poverty population), percentages not different from 20015. The number and percent of the population of "near poor" remain unchanged in 2002, at 12.5 million and 4.4 percent, respectively.

The Current Population Survey implemented a new question in 2003 to collect information on race, allowing individuals to report one or more races. This implements the change mandated by the Office of Management and Budget in the way to collect and report race for individuals. There is no way of knowing how people who reported more than one race would have reported their race under the old question. As a result, I must emphasize that there is no single way to make comparisons for race groups between 2001 and 2002.

Chart 4 presents the changes in real median household income by race and Hispanic origin between 2001 and 2002 [JPG - 115k]. I realize that the chart is complicated, so I'll pause to give you a moment to review it. Median household money income did not change for households with a non-Hispanic White householder, or for those with a householder who reported a single race of Asian. However, income fell for Blacks by 2.5 to 3.0 percent, for Asians who reported more than one race by 4.0 to 4.5 percent, and for Hispanics by 2.9 percent6. Under the four alternative income measures, households with householders who reported their race as Asian or Native Hawaiian and Other Pacific Islander were the only ones to show a decline in real median household income between 2001 and 2002. The decline ranged from 2.7 to 4.1 percent, depending on the income definition used.

Chart 5 shows that the only racial or ethnic group to show an increase in its poverty rate was Blacks (1.2 to 1.4 percentage points) to 24 percent [JPG - 93k]7. Poverty rates did not change for non-Hispanic Whites, for Asians, or for Hispanics.

Chart 6 shows that median household income declined between 2001 and 2002 in only one region, the Midwest [JPG - 79k]. The South continues to have the lowest median household income of all four regions.

Chart 7 shows the regional pattern for poverty [JPG - 64k]. Consistent with changes for money income, the Midwest was the only region to have an increase in its poverty rate, to 10.3 percent, from 2001 to 2002. The South continues to have the highest poverty rate among regions.

As shown in Chart 8, child poverty stood at 16.7 percent in 2002, unchanged from 2001 but down from its recent peak of 22.7 percent in 1993 [JPG - 85k]. The poverty rate for those 65 and older - 10.4 percent in 2002 - also did not change from 2001. The poverty rate for those aged 18 to 64 was up, to 10.6 percent - an increase of one-half (0.5) percentage point from 2001 to 20028.

Differences by nativity and metropolitan residency status show that the declines in money income and increases in the official measure of poverty were not widespread. It is only households headed by a foreign-born non-citizen that showed a decline in income, of 3.9 percent, while it is only Natives who showed an increase in poverty, of 0.4 percentage points. Households in metropolitan areas (MAs) had declines in income, and people in those areas showed increases in poverty, while households and people outside MAs had neither.

Earnings are by far the largest component of income. The real median earnings for women who worked full time, all year, increased by 1.8 percent. Such women have not experienced an annual decline in earnings since 1995. The earnings of comparable men increased 1.4 percent between 2001 and 2002. As shown in Chart 9, the ratio of female-to-male earnings for full-time year-round workers was 77 percent in 2002, which matches the all-time high reached in 2001, up from 74 percent first reached in 1996 [JPG - 85k].

Based on a comparison of two-year moving averages for states, real median household income decreased for 10 states and the District of Columbia, and rose in only one, Oklahoma [JPG - 79k]. Also using a comparison of two-year moving averages, the poverty rate rose in nine states and fell in none. Five states had both a decrease in income and an increase in poverty - Florida, Hawaii, Illinois, Michigan, and Mississippi.

Both income and income inequality can be measured in a number of ways [JPG - 103k]. Let's look at money income first. According to the most widely used measure, the Gini index, household income inequality did not change for the ninth consecutive year. This result is confirmed by examining the shares of aggregate household money income received by each quintile of households. However, this story is different when the alternative income measures are examined. Income inequality as measured by all four alternative measures showed a decline between 2001 and 2002, with declines for the top 20 percent of households showing up as increases for the middle 60 percent.

Valuing noncash benefits and subtracting taxes also affects the estimated poverty rate. The poverty report provides two sets of alternative estimates of poverty. The first focuses on recommendations from the National Academy of Sciences (NAS) on how to measure resources (income) and how to change the poverty thresholds (the measure of need)9. The second presents the effects of changing the income measure in ways consistent with the income alternatives discussed earlier, as well as on how changes in the factor used to adjust the thresholds for inflation over the past several decades would affect poverty10. Some researchers think it is important to consider changes on the resource side and the threshold side together, whereas others focus on how to measure resources while retaining the historical poverty thresholds. Additional research on measuring economic well-being is underway and I will talk about one effort shortly; the Census Bureau does not choose which methods of measuring poverty are most appropriate. That responsibility rests with the Office of Management and Budget.

Chart 12 presents six alternative measures of poverty that have been developed by the Census Bureau in response to the NAS report [JPG - 82k]. The differences among the measures will not be discussed in detail; the measures are discussed in the report. Again, I apologize for the acronyms; the measures are described in the report. In contrast to the official poverty measure, none of these six measures show an increase in poverty rates between 2001 and 2002.

Chart 13 presents nine additional measures that use the four alternative income measures mentioned earlier along with two versions of poverty thresholds [JPG - 82k]. All ten, including the official poverty measure, show an increase in poverty rates between 2001 and 2002 of 0.3 to 0.4 percentage points.

The third report in your press package is Supplemental Measures of Material Well-Being. It sets the stage for work on an experimental series of expenditure-based or consumption-based measures of well-being to complement the income-based measures discussed earlier. The Census Bureau has worked with the Bureau of Economic Analysis, the Bureau of Labor Statistics, and the Energy Information Administration to develop this report. Section II of the report provides some background on the underlying concepts of defining and measuring consumption, including a discussion of some of the research and data requirements for calculating expenditure-based poverty measures. Section III includes currently available information on some direct indicators of material well-being from three surveys: the Survey of Income and Program Participation, the Consumer Expenditure Survey, and the Residential Energy Consumption Survey. Finally, Section IV describes research that is relevant to formulating all of these supplemental measures.

Let me conclude by summarizing the main findings. Median household money income in 2002 fell 1.1 percent or $500 from the 2001 level, to $42,400, and median household money income after taxes fell by 0.8 percent [JPG - 118k]. However, three other comprehensive measures of income that account for both taxes and noncash benefits showed no change between 2001 and 2002. The ratio of female-to-male earnings for full-time, year-round workers was 77 percent, matching an all-time high. The official poverty rate rose, from 11.7 percent in 2001 to 12.1 percent in 2002. The number in poverty increased also, by 1.7 million people, to 34.6 million in 2002. Several alternative measures of poverty show a mixed picture, with some showing increases between 2001 and 2002, and some showing no change. Several measures of household money income inequality indicate that inequality did not change from 2001 to 2002, but four alternative measures of income showed a decline in income inequality from 2001 to 2002.

I'll be glad to answer questions from the press at this time. Please wait for the microphone and identify yourself and your affiliation.


  1. You can obtain additional unpublished detailed tables from the Census Bureau directly or on our website (
  2. The Annual Social and Economic Supplement was administered to about 78,000 households in February through April 2003. As in all surveys, the data in these reports are estimates, subject to sampling variability and response errors, although the Census Bureau works hard to minimize these errors. All statements which compare two or more estimates in the reports and in this briefing meet the Census Bureau's standards for statistically significant differences. Other apparent numerical differences are, however, not necessarily statistically significant ones. All historical income data are expressed in 2002 dollars and were adjusted using the Consumer Price Index Research Series, which measured inflation at 1.6 percent between 2001 and 2002. The poverty thresholds are also updated each year for inflation.
  3. In 2002, the average poverty threshold for a family of four was $18,392; for a family of three, $14,348.
  4. The four additional comprehensive measures are After-tax money income (MI-Tx on the chart), Money income less taxes plus noncash benefits except Medicare and Medicaid (MI-Tx+NC-MM), Money income less taxes plus noncash benefits (MI-Tx+NC), and Money income less taxes plus noncash benefits plus an imputed return on home equity (MI-Tx+NC+HE).
  5. The number of severely poor is up from 13.4 million in 2001.
  6. The percentage declines for Blacks, Asians, and Hispanics were not different.
  7. The two percentage differences for Blacks cited are not different from one another.
  8. The poverty rates for those 18 to 64 and those 65 and older were not different.
  9. The six measures in the report represent two important dimensions for which the methods remain unresolved: how to account for medical out-of-pocket expenses, and whether to adjust for geographic differences in the cost-of-living. Another six measures that parallel these, but which use a different method for updating the thresholds over time, are presented on our web site.
  10. The income measures are money income plus the four alternatives discussed above. The two sets of thresholds are the official thresholds, and the thresholds derived by using a series of inflation estimates based on the methodology adopted by the Bureau of Labor Statistics in the early 1980s, but projected back to 1967 (termed the CPI-U-X1).

Source: U.S. Census Bureau | Social, Economic, and Housing Statistics Division: Poverty |  Last Revised: June 22, 2010