Dr. Daniel H. Weinberg, Housing and Household Economic Statistics Press Briefing on 2002 Health Insurance Estimates from the Current Population Survey, September 24, 2002
Good morning. I'm Dan Weinberg, Chief of the Housing and Household Economic Statistics Division at the Census Bureau. I am pleased to welcome you to the press briefing on the 2001 income and poverty estimates. Your press packets contain a press release, a copy of my remarks, a copy of the charts I will be using today, and the two reports we are releasing. You can obtain additional unpublished detailed tables from the Census Bureau directly or on our web site (www.census.gov).
Present with me today are a number of analysts who worked very hard on these reports; they will be available to answer your questions after the briefing. '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 answer our survey questions.
Please hold your questions unless it's a technical clarification. The main presentation should take about 25 minutes.
Let me first summarize the main findings. I will be discussing three major well-being indicators today. They are income, poverty, and income inequality [JPG - 43k].
After some background remarks, I'll be talking first about income and poverty trends overall and for regions. Then, I'll discuss racial and ethnic differences, followed by differences by age and nativity. Next will be a focus on earnings, the largest component of income. Then I'll discuss income inequality, followed by changes in income and poverty for states. I'll conclude by briefly discussing some experimental estimates of income and poverty. As you know, the official poverty measure does not account for noncash benefits or taxes. After that, I'll take your questions.
These statistics come from the Annual Demographic Supplement to the Current Population Survey, a sample survey conducted each month for the Bureau of Labor Statistics to compute official statistics on employment and unemployment. The estimates released today are based on an expanded household sample begun in 2001 that was designed to improve health insurance coverage estimates for children. The Annual Demographic Supplement was administered to about 78,000 households in February through April 2002. In addition, these estimates are weighted to reflect the results of Census 2000, as are the estimates for 1999 and 2000 that are used to compute year-to-year changes.
As in all surveys, the data in these reports are estimates, subject to sampling variability and response 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 2001 dollars and were adjusted using the Consumer Price Index Research Series, which measured inflation at 2.8 percent between 2000 and 2001. The poverty thresholds are also updated each year for inflation. In 2001, the average threshold for a family of four was $18,104; for a family of three, $14,128. These estimates reflect the effect of the recession that began in March 2001, according to the National Bureau of Economic Research (NBER), a private research firm. The NBER has not yet determined the end of that recession.
Turning to the first topic, Chart 2 presents the key estimates of median household income [JPG - 41k]. As I noted earlier, median income for all households fell 2.2 percent between 2000 and 2001 to $42,200. Overall, real median household income was up 31.6 percent since 1967, the first year median household income was computed. The 2001 level is higher than all but three years since 19671.
Chart 3 shows that median household income did not change in the Northeast between 2000 and 2001 [JPG - 39k]. Income did decline 3.7 percent in the Midwest, 1.4 percent in the South, and 2.3 percent in the West2. The South continues to have the lowest median household income of all four regions.
As Chart 4 shows, the number of poor in the U.S. in 2001 -- 32.9 million people -- was 1.3 million more than in 2000 [JPG - 41k]. This increase in the number of poor led to a higher poverty rate in 2001 -- 11.7 percent compared with 11.3 percent in 2000, which matched the lowest rate recorded, 11.1 percent in 19733. Except for 1999 and 2000, the last year for which the poverty rate was as low as in 2001 was 1979.
Chart 5 shows the regional pattern for poverty [JPG - 34k]. The South was the only region to have an increase in its poverty rate from 2000 to 2001. Its rate of 13.5 percent was the highest among the regions, although it was not different from the West in 1994 through 1999.
Chart 6 presents the changes in real median household income by race and Hispanic origin between 2000 and 2001 [JPG - 52k]. Median household income did not change for Hispanics, but fell for non-Hispanic Whites, by 1.3 percent; for Blacks, by 3.4 percent; and for Asians and Pacific Islanders, by 6.4 percent4. This was the first annual decline (that is, from one year to the next) for non-Hispanic Whites and Asians and Pacific Islanders since 1991, and for Blacks since 19815.
Chart 7 shows that the only racial or ethnic groups to show an increase in their poverty rates were Whites and non-Hispanic Whites [JPG - 51k]. The poverty rate for non-Hispanic Whites rose from 7.4 percent in 2000 to 7.8 percent in 20016. The poverty rates for Blacks, Asians and Pacific Islanders, and Hispanics, matched their historic lows -- 22.7 percent, 10.2 percent, and 21.4 percent, respectively.
Chart 8 shows that the 3-year-average median household income of American Indians and Alaska Natives, $32,100, was higher than for Blacks, not different from Hispanics, and lower than for non-Hispanic Whites and Asians and Pacific Islanders [JPG - 51k]. The 3-year-average poverty rate for American Indians and Alaska Natives, 24.5 percent, was not different from that for Blacks, and higher than the poverty rate for non-Hispanic Whites, for Asians and Pacific Islanders, and for Hispanics.
I now turn to differences by age and nativity.
As shown in Chart 9, child poverty stood at 16.3 percent in 2001, unchanged from 2000 but down from its recent peak of 22.7 percent in 1993 [JPG - 41k]. The poverty rate for those 65 and older -- 10.1 percent in 2001 -- remained not different from its all-time low. The poverty rate for those aged 18 to 64 was also 10.1 percent; an increase of 0.5 percentage point from 2000 to 2001.
Chart 10 shows median household income and poverty for people classified by their nativity [JPG - 44k]. Median household income declined 1.5 percent for households maintained by a native, compared with a 5.3 percent decline for households maintained by a foreign-born householder7. This decline, in turn, consisted of a 4.2 percent decline for households with non-citizen householders and a 5.4 percent decline for those with a naturalized householder8.
The poverty picture, however, was different. The 0.4 percentage point increase in the poverty rate for all people reflects a 0.3 percentage point increase in the poverty rate of natives, with no statistically significant change for the foreign born.
Earnings are by far the largest component of income. The real median earnings for women who worked full time, all year, increased for the fifth consecutive year. However, the earnings of comparable men did not change between 2000 and 2001. As shown in Chart 11, the ratio of female-to-male earnings for full-time year-round workers was 76 percent in 2001, a new all-time high, up from 74 percent first reached in 1996 [JPG - 46k].
Turning now to income inequality, Chart 12 shows the time series of changes versus 1967 as a base for several different measures of income inequality through 2001, including the Gini index, the most widely used measure [JPG - 50k]. According to the Gini index, for the eighth consecutive year, overall household income inequality did not change from year-to-year. However, several other measures, including some not shown here, showed an increase9. For example, the share of aggregate income going to the lowest household income quintile dropped from 3.6 percent in 2000 to 3.5 percent in 2001. The Gini index and most of the other indexes show that income inequality in 2001 was above their 1999 and earlier years' levels10.
Based on a comparison of two-year moving averages for states, real median household income increased between 1999 and 2001 for three states and fell in 12 [JPG - 39k]. In the same period, the poverty rate fell in four states and increased in two. Massachusetts had both an increase in income and a decline in poverty.
The Census Bureau also produces a series of experimental measures of how noncash benefits and taxes -- which are not considered in the official estimates -- affect income and poverty. The income report shows 14 experimental definitions of income. The Census Bureau's research shows that income inequality under a broadened definition of income -- one that takes into account the effects of noncash benefits and taxes -- was roughly 8 percent lower (that is, more equal) than under the official cash income definition. Government benefits did more than taxes to reduce income inequality.
Valuing noncash benefits and subtracting taxes also affects the estimated poverty rate. The Census Bureau has continued its research into experimental poverty measures, based on recommendations made by the National Academy of Sciences, and issued a report last October11. Those measures not only accounted for benefits and taxes in income, but they also used a new set of experimental poverty thresholds.
Chart 14 presents six alternate measures of poverty that have been developed by the Census Bureau in response to the National Academy report [JPG - 48k]. These six measures represent two 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.)
Over the period 1999 to 2001, the pattern of decline, then increase, shown by the official measure of poverty was corroborated by the numerical changes in all six alternate measures, although not all these changes were statistically significant. Four of the six experimental measures shown in the report exhibited a statistically significant increase in poverty between 2000 and 2001, while two did not.
Where the alternatives make the most difference is in what they tell us about who was poor. For the most part, the alternate measures show a poverty population that looks more like the total population than under the official measure. Among the groups with lower experimental poverty rates are female householders with no spouse present, children, and African Americans. Among the groups with higher experimental poverty rates are non-Hispanic Whites, the elderly, and, for measures that account for differing costs of living, those who reside in areas with high housing costs.
As a specific example of how our perceptions of poverty can change, let me focus on children. Children are 36 percent of the official poor even though they are but 26 percent of the total population. Using the six experimental measures presented here, the percentage of the poor who are children ranged from 30 to 32 percent, less than the percentage under the official measure but still more than their percentage in the population.
The decision on whether, when, and how to change the official poverty measure is the responsibility of the Office of Management and Budget, not the Census Bureau.
Let me conclude by summarizing the main findings. Median household income in 2001 fell 2.2 percent from the all-time high level first reached in 1999 and maintained in 2000, to $42,200 [JPG - 41k]. Along with that change, the poverty rate rose, to 11.7 percent in 2001 from 11.3 percent in 2000 (a level that matched the all-time low of 11.1 percent in 1973). Also, the number of poor increased -- by 1.3 million people -- to 32.9 million poor in 2001. The picture on household income inequality was mixed, with some measures showing a year-to-year increase while others did not. Income inequality was higher in 2001 than in 1999 or in earlier years.
I'll be glad to answer questions from the press at this time. I'll have about 20 minutes for questions. Please identify yourself and your affiliation.