CHAPTER 3
POVERTY

BACKGROUND

The federal government began measuring poverty in 1959. During the 1960s President Lyndon Baines Johnson declared a national war on poverty. Researchers realized that very few statistical tools were available to measure the number of Americans who continued to live in poverty in one of the most affluent nations in the world. In order to fight this "war," it had to be determined who was poor and why.

During the early 1960s Mollie Orshansky of the Social Security Administration suggested that the poverty income level be defined as the income sufficient to purchase a minimally adequate amount of goods and services. The necessary data for defining and pricing a full "market basket" of goods was not available then, nor is it available now. Orshansky noted, however, that in 1955 the U.S. Department of Agriculture (USDA) had published a "Household Food Consumption Survey," which showed that the average family of three or more persons spent approximately one-third of its after-tax income on food. She multiplied the USDA's 1961 economy food plan (a no-frills food basket meeting the then-recommended dietary allowances) by three.

Basically this defined a poor family as any family or person whose after-tax income was not sufficient to purchase a minimally adequate diet if one-third of the income was spent on food. Differences were allowed for size of family, gender of the head of the household, and whether it was a farm or nonfarm family. The threshold (the level at which poverty begins) for a farm family was set at 70 percent of a nonfarm household. (The difference between farm and nonfarm households was eliminated in 1982.)

Poverty Thresholds

The poverty guidelines set by the U.S. Department of Health and Human Services (HHS) are based on the poverty thresholds as established by the U.S. Bureau of the Census. The poverty thresholds, used for statistical purposes, are updated each year to reflect inflation. People with incomes below the applicable threshold are classified as living below the poverty level.

The poverty guidelines vary by family size and composition. For a family of four in 2004 the poverty guideline was $18,850 in annual income. A person living alone who earned less than $9,310 was considered poor, as was a family of eight members making less than $31,570. Notice, in Table 3.1, that the poverty level is considerably higher in Alaska and Hawaii, where the cost of living is higher than in the contiguous forty-eight states and the District of Columbia.

The poverty guidelines set by HHS are very important because various government agencies use them as the basis for eligibility to key assistance programs. HHS uses the poverty guidelines to determine Community Services Block Grants, Low-Income Home Energy Assistance Block Grants, and Head Start allotments. The guidelines are also the basis for funding the USDA's Food Stamp Program, National School Lunch Program, and Special Supplemental Food Program for Women, Infants, and Children (WIC). The U.S. Department of Labor uses the guidelines to determine funding for the Job Corps and other employment and training programs under the Workforce Investment Act. Some state and local governments choose to use the federal poverty guidelines for some of their own programs, such as state health insurance programs and financial guidelines for child support enforcement.

POVERTY—THEN AND NOW

Since the late 1950s, Americans have seen some successes and some failures in the battle against poverty. For the total population in 1959, 22.4 percent, or 39.5 million persons, lived below the poverty level. After an initial decline through the 1960s and 1970s, the poverty rate began to increase during the early 1980s, coinciding with

a downturn in household and family incomes for all Americans. The poverty rate rose steadily until it reached an eighteen-year high of 15.2 percent in 1983, a year during which the country was climbing out of a serious economic recession. The percentage of Americans living in poverty then began dropping, falling to 12.8 percent in 1989. After that, however, the percentage increased again, reaching 15.1 percent in 1993 (39.3 million people). It then dropped to 12.7 percent in 1998 (34.5 million) and continued to drop to 11.3 percent (31.1 million) in 2000; however, because the nation's economy slowed, the poverty rate took an upturn to 34.6 million in 2002, representing 12.1 percent of the population. Figure 3.1 provides a graphic depiction of the number of poor people and the poverty rates for years 1959 through 2002.

Analysts believe the overall decline in poverty is due to both the growth in the economy and to the success of some of the antipoverty programs instituted in the late 1960s; yet not all demographic subcategories have experienced the same level of change. For example, the poverty rate of those sixty-five years of age and older has dramatically improved from 35.2 percent in 1959 to 10.4 percent in 2002. For related children under eighteen years of age in black families, however, the improvement from 65.6 percent in 1959 to 46.3 percent in 2002 shows that antipoverty programs still have not reached many people in need. Table 3.2 gives the nation's poverty rates from 1959 through 2002 for people by categories of age, race, and ethnic background.

Race and Ethnicity

Poverty rates have been consistently lower for whites than for minorities. According to the U.S. Census Bureau in Poverty in the United States: 2000 (Washington, DC, 2001), in 1959, 18.1 percent of all whites, or 28.5 million people, lived below the poverty level. By 1970 the rate declined to 9.9 percent, about where it remained for the next ten years. In 1983 the percentage of whites living in poverty reached an eighteen-year high of 12.1 percent. By 2000 it had dropped to 9.4 percent (21.3 million people); in keeping with the overall rise in the poverty rate, the number of whites in poverty in 2002 increased to 10.2 percent (23.5 million people), according to the updated Poverty in the United States: 2002 (Washington, DC, 2003).

By contrast, according to Poverty in the United States: 2000, poverty rates for African-Americans declined from 55.1 percent (9.9 million) in 1959 to 33.5 percent (7.5 million) in 1970. In 1983 rates for African-Americans were 35.7 percent, still almost triple the rate for whites. In 2000, a year in which the American economy was very strong, over one-fifth (22.1 percent, or 7.9 million) of all African-Americans were poor. This percentage was the lowest recorded poverty rate ever for African-Americans. African-American children and older persons have been particularly affected by poverty. In 2002, 31.5 percent of African-Americans under eighteen years of age and 23.6 percent of African-Americans sixty-five years or older (as opposed to 19.7 percent of African-Americans aged eighteen to sixty-four years) were poor. (See Table 3.2.) In 2002 approximately 24 percent of African-Americans (between 8.6 and 8.9 million people) were poor, according to the U.S. Census Bureau.

It was not until 1972 that the U.S. Census Bureau added the ethnicity "Hispanic" to the categories for which it regularly records statistics. That year, according to Poverty in the United States: 2000, 22.8 percent of Hispanics in the United States lived below the poverty level. Between 1972 and 1985, the number of Hispanics living below the poverty level more than doubled, from 2.4 million to 5.2 million. Over this same period, the total Hispanic population rose from 10.6 million to 18.1 million. (Persons of Hispanic origin can be of any race.) The poverty rate among Hispanics reached 29.9 percent in 1982 and then dropped to 26.2 percent in 1989. The poverty rate reached 30.7 percent, its highest recorded point, in 1994, then dropped to 29.4 in 1996 and 21.2 (or 7.2 million) in 2000. In 2002 a large proportion (28.6 percent) of Hispanics under eighteen years of age and about one-fifth (21.4 percent) of those sixty-five years and older were poor. (See Table 3.2.) Overall, the U.S. Census Bureau reported that 21.8 percent of all Hispanics (approximately 8.5 million people) were poor in 2002.

According to the U.S. Census Bureau, the Asian and Pacific Islander (API) poverty rate in 2002 overall was 10.3 percent (or 1.3 million people). The rate was lower than it was in 1987, the first year that the Census Bureau kept statistics on Asians and Pacific Islanders, when 16.1 percent lived below the poverty level. See Table 3.2 for the breakdown of poverty rates by different age groups among APIs.

Although a far higher percentage of African-Americans and Hispanics than whites were poor, over half of the 41.8 million people who were poor in 2002 were non-Hispanic whites. In the same year, 55.9 percent of poor Americans were non-Hispanic whites, 20.5 percent were African-Americans, 20 percent were Hispanics, and 2.8 percent were Asians and Pacific Islanders.

Age

In 2002, 16.7 percent of children under eighteen years old were poor; this was a decline of 6 percentage points from 22.7 percent in l993. The percent of those ages sixty-five and older below the poverty line declined from 12.2 percent in 1993 to 10.4 percent in 2002. (See Table 3.3.) From 1959 to 2002 the number of people sixty-five years and older living in poverty dropped significantly, from about 35 percent to 10.4 percent. (See Figure 3.2.)

In 2002 more than one-fourth of the nation's poor were either under eighteen years of age (16.7 percent) or sixty-five years of age and older (10.4 percent). In comparison to the population above the poverty level, children were overrepresented among the poor, while the elderly were underrepresented. Children made up about 35 percent of the poor even though they represented only a little more than one-fourth (25.4 percent) of the total population. People over sixty-five years old made up 12 percent of the total population in 2002 but only 10.4 percent of the poor. (See Figure 3.2 and Table 3.4.) Most observers credit Social Security for the sharp decline in poverty among the elderly.

CHILD POVERTY. The child poverty rate, at 16.7 percent of the nation's children under the age of eighteen, was over twice the poverty rate for adults (7.8 percent) in 2002. Very young children were at the greatest risk of being poor. According to the National Center for Children

in Poverty, the United States has the highest rate of young child poverty of all Western industrialized nations.

In 2000 nearly 18 percent of children under the age of three lived in poverty, a decline from 27 percent in 1993. Nearly 45 percent of very young children living with a female householder (with no spouse present) were poor, compared to 9 percent of children under three years old living with families headed by a married couple. Poverty rates for African-American children under age three (35 percent) and Hispanic children under three (30 percent)

are three times higher than the rates for white children (10 percent).

Childhood poverty is a matter of great concern because strong evidence exists that poverty can limit a child's physical and cognitive development. According to the Children's Defense Fund report The High Price of Poverty for Children of the South (May 1998):

• Poverty is a greater risk to children's overall health status than is living in a single-parent family.
• Poor children are twice as likely as nonpoor children to be born weighing too little or to suffer stunted growth.
• Poor children suffer more mental and physical disabilities.
• Poverty makes children hungry. Hungry children are more likely to be hyperactive and to have serious behavior problems. They are four times more likely to have difficulty concentrating in school.
• Poor children score lower on reading and math tests and are twice as likely to repeat a year of school as nonpoor children.
• Poor children earn 25 percent lower wages when they become young adults.

Regions

In 2002 the Midwest had the lowest poverty rate (10.3 percent) among the nation's four regions, followed by the Northeast with 10.9 percent and the West with 12.4 percent. Poverty rates were highest in the South (13.8 percent). (See Table 3.3.)

Family Status

About one in every ten families (10.4 percent) in the United States was living in poverty in 2002. Families headed by married couples had the lowest poverty rate (5.3 percent). More than a quarter (26.5 percent) of all families with a female householder (no husband present) were living in poverty, a 1.8 percent increase since 2000. (See Table 3.3.)

BY RACE. According to the U.S. Census Bureau in Poverty in the United States: 2002, in 2002 non-Hispanic whites (8 percent) had a much lower poverty rate than African-Americans (24.1 percent) or Hispanics (21.8 percent). In addition, Hispanic and African-American female heads of household are at particularly high risk of poverty. While 20 percent of non-Hispanic white families headed by a female with no husband present were living in poverty in 2002, 38.2 percent of female-headed African-American families and 36.4 percent of female-headed Hispanic families were poor. Asian single mothers had the lowest poverty rate at 15.2 percent in 2002.

States

State poverty rates vary widely from year to year and should be used with caution when ranking the states for statistical purposes. However, on average, over the years 2000 through 2002, Connecticut, Maryland, Minnesota, New Hampshire, and New Jersey had the lowest poverty rates, while Arkansas, the District of Columbia, Louisiana, Mississippi, and New Mexico had the highest rates. (See Table 3.5.)

Work Experience

The probability of a family living in poverty is influenced by three primary factors: the size of the family, the number of workers, and the characteristics of the wage earners. As the number of wage earners in a family increases, the probability of poverty declines. The likelihood of a second wage earner is greatest in families headed by married couples.

About 5.7 percent of all Americans who worked in 2001 lived in poverty, compared to 39.7 percent of those who did not work that year. Approximately 42 percent of poor persons age sixteen and older worked at least some of the year, with 74 percent of them working twenty-seven weeks or more. Overall, 70.5 percent of people sixteen years old and over worked at least some of the year, and about 64 percent worked twenty-seven weeks or more. (See Table 3.6.)

Most poor children live in families where one or more adults work. However, millions of working parents are not able to earn enough to lift their families out of poverty—even those who work full-time all year. In the report A Hand Up: How State Earned Income Tax Credits Help Working Families Escape Poverty in 2001 (Washington, DC: Center on Budget and Policy Priorities, 2001), Nicholas Johnson reported that in 2000 approximately 4.4 million families with children in which the parents were not elderly or disabled had incomes below the federal poverty line. Of these families, three million (69 percent) had a working parent. (See Figure 3.3.)

Education

Not surprisingly, poverty rates drop sharply as years of schooling rise. The U.S. Bureau of the Census reported that in 2002 the poverty rate was 21 percent for persons who had not completed high school, 11.1 percent for those who had graduated from high school but had not gone to college, 8.0 percent for those with some college but less than a bachelor's degree, and 4.5 percent for those with a bachelor's degree or higher.

The general relationship between education and welfare applied for all races. However, there were some large differences among the races. For example, the 2002 poverty rates for African-American and Hispanic high school graduates age twenty-five and over who had not gone to college were 22.2 percent and 14.7 percent, respectively, while white graduates had a much lower poverty rate, 9.3 percent.

MEDIAN INCOME

In its surveys, the U.S. Bureau of the Census differentiates between households and families. A "household" is an individual living alone or a group of persons living together who may or may not be related, while a "family" is composed of two or more related individuals. All families are households, but not all households are families. In 2002 there were over 111.2 million households in the United States but only 75.5 million families. (See Table 3.7.)

Household Income

The median income (half the population earns less than this amount, and half earn more) for all households in 2002 was $42,409. African-American ($29,026) and Hispanic ($33,103) households had much lower median incomes than did non-Hispanic white households ($45,086) and API households ($52,626).

Family Income

In 2002 the median income for all families was $52,704. The median family income was 2.86 times the 2002 average poverty threshold of $18,400 for a family of four. Families headed by married couples of all racial and ethnic backgrounds enjoyed the greatest financial success (a median income of $61,254), with incomes 2.11 times that of female-headed households ($29,001). (See Table 3.7.)

Per Capita Income

In its annual Current Population Survey, the U.S. Census Bureau determines the per capita income in the United States. Per capita income is computed by dividing the total money income by the total population. In other words, it represents the amount of income that every man, woman, and child would receive if the nation's total earnings were divided equally among them. In 2002 the per capita income for all Americans was $22,794. The 2002 per capita incomes for different racial and ethnic groups were: non-Hispanic whites, $26,128; Asians and Pacific Islanders, $22,810; African-Americans, $15,441; and Hispanics, $13,487. Comparing the per capita income with the poverty level for a family of four in 2002 ($18,400) reveals that the poverty level for four people is less than the per capita income.

TAX RELIEF FOR THE POOR

Both conservatives and liberals hailed the Tax Reform Act of 1986 (PL 99-514) as a major step toward relieving the tax burden of low-income families, one group of Americans whose wages and benefits have been eroding since 1979. The law enlarged and "inflation-proofed" the

Earned Income Tax Credit (EITC), which provides a refundable tax credit that both offsets taxes and often operates as a wage supplement. Only those who work can qualify. The amount is determined, in part, by how much each qualified individual or family earned. It is also adjusted to the size of the family. To be eligible for the family EITC, workers must live with their children, who must be under nineteen years old or full-time students under twenty-four years old.

The maximum credit for 2003 was $2,547 for taxpayers with one child, $4,204 for taxpayers with more than one child, and $382 for persons with no children. Families get less if their income is very low because they are also eligible for public assistance. Working families with one or more children receive the maximum benefit if their earnings are at least $10,700, up to an adjusted gross income of no more than $14,600. Benefits phase down gradually when income surpasses $15,000 and phase out entirely for families with two or more children earning more than 33,000. (See Figure 3.4.)

The largest EITC benefits go to families that no longer need welfare. The gradual phaseout and the availability of the EITC at above-poverty income levels help to stabilize a parent's employment by providing additional money to cover expenses associated with working, such as child care and transportation. Research has found that the EITC has been an effective work incentive and has significantly increased work participation among single mothers.

Those who do not owe income tax, or who owe an amount smaller than the credit, receive a check directly from the Internal Revenue Service (IRS) for the credit due them. Most recipients claim the credit when they file an income tax form. The EITC lifted more than 4.8 million people, including 2.6 million children, out of poverty in 2003.

Although the Tax Reform Act of 1986 has helped ease the burden of federal taxes, most of the poor still pay a substantial share of their income in state and local taxes. To relieve this tax burden, seventeen states have enacted a state EITC; sixteen states base EITCs on the federal credit. These state programs boost the income of families that move from welfare to work and prevent states from taxing poor families deeper into poverty.

HOW ACCURATE IS THE "POVERTY LEVEL"?

Almost every year since the U.S. Bureau of the Census first defined the poverty level, observers have been concerned about the accuracy of the estimated poverty level. The figure had been based on the finding that the average family in the mid-1950s spent about one-third of its income on food. That figure was multiplied by three to allow for expenditures on all other goods and services. This represented the after-tax money income of an average family relative to the amount it spent on food. Since 1965 the poverty threshold has been adjusted each year only for inflation.

Since that time, living patterns have changed, and food costs have become a smaller percentage of family spending. For example, the U.S. Bureau of Labor Statistics, in its Consumer Expenditures in 2000 (Washington, DC, 2001), reported that in 2000 the average family spent 13.6 percent of its total expenditures on food, while housing accounted for about one-third (32.4 percent) of family spending. Based on these changes in buying patterns, should the amount spent on food be multiplied by a factor of seven instead of three? Or should the poverty level be based on housing or other factors? What about geographical differences in the cost of living?

The proportion of family income spent on food is not the only change in family living since the 1950s. Both parents in a family are far more likely to be working than they were a generation ago. There is also a much greater likelihood that a single parent, most likely a mother, will be heading the family. Child-care costs, which were of little concern during the 1950s, have become a major issue for working mothers and single parents at the beginning of the twenty-first century.

Critics of the current poverty calculations tend to believe that the poverty levels have been too low, since they are based on a fifty-year-old concept of American life that does not reflect today's economic and social realities. Most feel the poverty level should be raised, probably to about 130 to 150 percent of the current levels. A 1989 study prepared by the Joint Economic Committee of Congress cited the example of a working mother with two children earning an income at the poverty level who spends $50 per week on child care and no more than 30 percent of her earnings on housing. (This is the proportion that the U.S. Department of Housing and Urban Development [HUD] has established as the basic affordability standard.) In this particular example, 30 percent of her income would equal about $226 a month for rent and utilities. Unless this woman lived in public or subsidized housing (more than two-thirds of the poor do not live in public or subsidized housing), finding an apartment for her family under this budget would be very hard. If she spent a bare minimum on food, she would have about $30 left after taxes. This $30 would have to cover medical care, clothing, personal care items, and an occasional ice cream cone for her two children.

Some are concerned because the poverty threshold is different for elderly and non-elderly Americans. When the poverty threshold was first established, it was thought that older people did not need as much food. Therefore, the value of their basic food needs was lower. Consequently, when this figure was multiplied by three to get the poverty rate, it was naturally lower than the rate for non-elderly people. (The U.S. government, however, uses the poverty rate for non-elderly Americans when determining the eligibility for welfare services for all people, including the elderly.) Critics point out that while the elderly might eat less than younger people, they have greater needs in other areas, which are not considered when their food needs are simply multiplied by three. Probably the most notable difference between the needs of the elderly and non-elderly is in the area of health care. The Bureau of Labor Statistics, in Consumer Expenditures in 2000, found that while the total population interviewed spent about 5.4 percent of their income on health care, those over sixty-five years of age spent 12.2 percent. These critics feel that the poverty level should be the same for everyone, no matter what their age.

A 1995 report prepared by the National Research Council's Panel on Poverty and Family Assistance raised several important issues regarding poverty thresholds or measurement of need. The panel recommended that new thresholds be developed, using consumer expenditure data to represent a budget for basic needs: food, clothing, shelter (including utilities), and a small allowance for miscellaneous needs. This budget would be adjusted to reflect the needs of different family types and geographic differences in costs. Research and discussion continue on these issues.

How Should Income Be Defined?

The Panel on Poverty and Family Assistance also recommended that family resources be redefined to reflect the net amount available to buy goods and services in that budget for basic needs. Critics have pointed out that the definition of income used to set the poverty figure is not accurate because it does not include the value of all welfare services as income. If the value of these services were counted as income, they believe the proportion of Americans considered to be living in poverty would be lower.

In the 1990s the Census Bureau developed several experimental methods of estimating income for evaluating poverty levels, but the bureau has had considerable difficulty determining the value of many of these subsidies. For example, the bureau first tried to consider Medicare and Medicaid at full market value (this meant taking the total amount of money that the government spent on medical care for a particular group and then dividing it by the number of people in that group). The value often was greater than the actual earnings of the low-income family, which meant that, although the family's total earnings may not have been enough to cover food and housing, adding the market value of Medicare or Medicaid to its earnings put the family above the poverty threshold.

This did not make much sense, so the Census Bureau began trying a "fungible value" (giving equivalent value to units) for Medicare and Medicaid. When the bureau measures a household's income, if the earners cannot cover the cost of housing and food, Medicare and Medicaid are given no value. However, if the family can cover the cost of food and shelter, the Census Bureau figures the difference between the household income and the amount needed to meet basic housing and food costs. It then values the health services at this difference (up to the amount of the market value of the medical benefits). This is very complicated, but the Bureau of the Census believes it gives a fair value to these services. Similar problems have developed in trying to determine the value of housing subsidies, school lunches, and other benefits.

Still other observers point out that most income definitions do not include assets and liabilities. Perhaps the poor household has some assets, including a home or car, that could be converted into income. One experimental definition of income includes capital gains on earnings, although it seems to make little difference—about 90 percent of all capital gains are earned by those in the upper fifth of the earnings scale. Including assets generally means little, since the overwhelming majority of poor families have few financial assets. The Bureau of the Census reports that more than half the poor have no assets at all and about four-fifths have assets of less than $1,000, a relatively insignificant amount from which to earn income.

Another major issue is the question of income before and after income taxes. While the Tax Reform Act of 1986 (PL 99-514) removed most poor households from the federal income tax rolls, many poor households still pay state and local taxes. Naturally, some critics claim, the taxes paid to local and state governments are funds that are no

longer available for feeding and housing the family and, therefore, should not be counted as income.

Table 3.8 lists the various experimental definitions for income that the Bureau of the Census has considered. Table 3.9 shows the effects of selected definitions on the poverty rate.

HUNGER IN AMERICA

When thinking of the consequences of poverty, people often visualize the poor being badly housed or, in the worst case, homeless, instead of people not having enough to eat. While it may be hard to imagine some Americans not having enough to eat, many Americans go to bed hungry or experience times when there is not enough food for the family.

During the 1980s a growing number of studies found that Americans, especially children, were suffering from hunger. Many observers did not believe these reports or thought they had been exaggerated. In 1984 a Task Force on Food Assistance appointed by President Ronald Reagan found that it could not "report definitive evidence on the extent of hunger" because there was no agreed-upon way to measure hunger.

To determine the extent of hunger in the United States, the Food Research and Action Center (FRAC) in Washington, D.C., an advocacy group for the poor, released the findings of the Community Childhood Hunger Identification Project (CCHIP). In this first FRAC survey, interviews were conducted in 2,335 households with incomes at or below 185 percent of poverty and with at least one child under twelve years of age. The results of this survey, released in 1991, indicated that 32 percent of U.S. households with incomes at or below 185 percent of the poverty level were hungry. At least one child out of every eight under twelve years of age suffered from hunger. Another 40 percent of low-income children were at risk for hunger.

Between 1992 and 1994 FRAC sponsored the second round of CCHIP surveys in nine states and the District of Columbia (5,282 low-income families with at least one child age twelve or under). For the purposes of its report, FRAC defined hunger as food insufficiency—skipping meals, eating less, or running out of food—that occurs because of limited household resources. The results were reported in Community Childhood Hunger Identification Project: A Survey of Childhood Hunger in the United States (Washington, DC, 1995).

Based on the findings of the second CCHIP surveys, FRAC concluded that about four million children age twelve and under experienced hunger in some part of one or more months during the previous year. Another 9.6 million children were at risk of becoming hungry.

The CCHIP survey studied one child in each household (the child with the most recent birthday) and found that, in comparison with nonhungry children, hungry children were:

• More than three times as likely to suffer from unwanted weight loss.
• More than four times as likely to suffer from fatigue.
• Almost three times as likely to suffer from irritability.
• More than three times as likely to have frequent headaches.
• Almost one and a half times as likely to have frequent ear infections.
• Four times as likely to suffer from concentration problems.
• Almost twice as likely to have frequent colds.

Based on the findings from CCHIP, FRAC concluded that although federal food programs are targeted to households most in need, a common barrier to program participation is a lack of information, particularly about eligibility guidelines. FRAC believes that if federal, state, and local governments made a greater effort to ensure that possible recipients were aware of their eligibility for food programs, such as Women, Infants, and Children (WIC) and the School Breakfast Program, there would be a large drop in hunger in the United States.

In 1997 the Urban Institute conducted the National Survey of American Families (NSAF). Nearly half of low-income families (those with family incomes up to 200 percent of the federal poverty line) who were interviewed in 1997 reported that the food they purchased ran out before they got money to buy more or they worried they would run out of food. Four out of five of these families with food problems reported suffering actual food shortages, and one out of five worried about food shortages. More children than adults lived in families that worried about or had trouble affording food, so that 54 percent of low-income children experienced the problem. The NSAF was repeated in 1999, and families reported fewer problems affording food than in 1997. Four in ten low-income families were either concerned about or had difficulty affording food, down 10 percent from 1997. However, approximately half of all low-income children still lived in families with difficulties affording food or concern about lack of food. A third NSAF was conducted in 2002, with results to be released in late 2004.

Since 1995 the Food and Nutrition Service of the U.S. Department of Agriculture (USDA) and the U.S. Census Bureau have conducted annual surveys of food security, food insecurity, and hunger. (Food-secure households are those that have access at all times to enough food for an active, healthy life. Food-insecure households are uncertain of having, or unable to acquire, enough food to meet basic needs at all times during the year.) The survey is based on an eighteen-item scale:

1. Worried food would run out before (I/we) got money to buy more
2. Food bought didn't last and (I/we) didn't have money to get more
3. Couldn't afford to eat balanced meals
4. Adult(s) cut size of meals or skipped meals
5. Respondent ate less than felt he/she should
6. Adult(s) cut size or skipped meals in three or more months
7. Respondent hungry but didn't eat because couldn't afford
8. Respondent lost weight
9. Adult(s) did not eat for whole day
10. Adult(s) did not eat for whole day in three or more months
11. Relied on few kinds of low-cost food to feed child(ren)
12. Couldn't feed child(ren) balanced meals
13. Child(ren) were not eating enough
14. Cut size of child(ren)'s meals
15. Child(ren) were hungry
16. Child(ren) skipped meals
17. Child(ren) skipped meals in three or more months
18. Child(ren) did not eat for whole day

Figure 3.5 shows that the food insecurity (uncertainty over acquiring sufficient food) and hunger have risen steadily since 1999. The number of food-insecure households rose from about 9 percent to 11.1 percent of U.S. households, and the prevalence of hunger rose from about 3 percent in 1999 to 3.5 percent in 2002. In 2002, 11.1 percent of households reported food insecurity at some time during the year, and 3.5 percent reported being hungry. (See Figure 3.6.)

Low-income households were more likely to experience food insecurity and hunger during the year. More than 14 percent of households below the poverty line reported being hungry at some time during 2002, about 3.5 times the rate in the general population. (See Figure 3.7 and Table 3.10.)

Families headed by married couples are much less likely to experience food insecurity than families headed by single females; 10.4 percent of married-couple households and 32 percent of female-headed households reported food insecurity in 2002. Food insecurity is also more prevalent among African-American families, 22 percent of whom experience food insecurity, and Hispanic families, 21.7 percent of whom experience food insecurity, than among non-Hispanic whites, 8 percent of whom experience food insecurity. (See Table 3.10).

GROWING DEMAND FOR EMERGENCY FOOD ASSISTANCE

Second Harvest National Research Study

America's Second Harvest is the nation's largest charitable hunger-relief organization, serving 23.3 million persons per year. In 2001 Mathematica Policy Research conducted a landmark study of the Second Harvest network by means of interviews with 32,000 clients and 24,000 questionnaires from local agencies. The findings are reported in Hunger in America, 2001 National Report (Myoung Kim, Jim Ohls, and Rhoda Cohen, Princeton, NJ, 2001). The study found the following characteristics of recipients of emergency food assistance:

• More than 23.3 million people sought emergency food assistance in 2001: 21.3 million at food pantries, 1.3 million at kitchens, and 0.7 million at shelters. Of the adult clients who visited emergency food programs in 2001, approximately 61.6 percent were female, 38.3 percent male.

• More than a third (38.9 percent) of all emergency client households had at least one member working.
• Sixty-four percent of the households have incomes at or below the poverty level.
• Assistance from government welfare programs (TANF, general assistance, and Supplemental Security Income) was the main source of income for 15.9 percent of all clients. Other government assistance (social security, workers' compensation, and unemployment compensation) was the main source of income for 30.5 percent of clients.
• Some 44.9 percent of recipients at all program sites were white; 35.4 percent, African-American; 16.7 percent, Hispanic; and 4.8 percent, Native American/Alaskan Native.
• Almost two-thirds (62.8 percent) of clients are high school graduates.
• Almost two-thirds (63.9 percent) of clients have applied for and 29.8 percent currently receive food stamps. See Figure 3.8 for an analysis of the use of food stamps by type of food program.
• Ten percent of all clients are homeless. Thirty-six percent reported having to choose between paying for food and paying their rent or mortgage. Figure 3.9 shows the housing situation for pantry, kitchen, and shelter clients.
• Twenty-nine percent reported that at least one household member was in poor health.

Over half of the agencies surveyed reported an increase in the number of clients serviced at program sites since 1998.

U.S. Conference of Mayors Status Report

Since 1982 the U.S. Conference of Mayors has conducted an annual survey of hunger and homelessness in U.S. cities. The 2002 survey focused on twenty-five cities and indicated a growing demand for emergency food assistance. The Status Report on Hunger and Homelessness in America's Cities: 2002 (Washington, DC: U.S. Conference of Mayors, 2002) reported that demand for emergency food assistance had increased by an average of 19 percent from the previous year in all twenty-five of the cities surveyed. Demand for emergency food among families with children increased 17 percent. Thirty-eight percent of persons requesting emergency food assistance were employed. The causes of hunger, according to officials in the surveyed cities, included low-paying jobs, unemployment, high housing costs, and weakening of the economy.