First published in the Review of Social Economy in late 1999.
The Census Bureau has estimated the nation's annual poverty rate since 1963 using data from the Bureau's Current Population Survey (CPS) conducted in March of each year. Census deems a household "poor" if annual income falls below specified income thresholds. There are two problems with this methodology. First, the Current Population Survey dramatically undercounts household economic resources. Second, the fact that household income falls below a specific level reveals little about the nature of material deprivation within the household.
This paper will take an alternative approach to assessing poverty: examining the material living conditions of low-income Americans. Using data from various government surveys this paper examines ownership of property and consumer durables; housing space, and housing conditions; food and nutrient consumption; and the height, thinness and obesity of low-income persons. Finally, we attempt an overall assessment of material deprivation based on material living conditions.
Lyndon Johnson's War on Poverty and the federal government's current system of overty measurement were born together. Johnson announced his now famous Var on Poverty during his State of the Union address in January 1964. To accomany the President's~ announcement, the 1964 Report of the Council of Economic dvisors contained a chapter on the problem of American poverty, setting rntative poverty income guidelines. In May 1965 the White House Office of conomic Opportunity (OEO) adopted similar but more elaborate poverty thresholds as a working definition of poverty, based on the work of Mollie Orshansky at the Social Security Administration (Fisher, 1992).
Orshansky's methods still provide the framework through which the government annually estimates the extent of poverty. In devising this framework, Orshansky began with the basic cost of food for different size families as specified by the Department of Agriculture's "economy food plan." Using food expenditures as a foundation, she then established non-food allowances as a multiple of food expenditures based on the share of budget allocated to non-food items in the general population in the 1950's (Orshansky 1965). In general, this produced a budget plan in which food spending was expected to be one third of total household expenditure. 1 This analysis resulted in an overall set of poverty income thresholds adjusted for family size: for a four person home the poverty threshold was $3,130; for three persons it was $2,400, two persons at $2,050, etc.
Using family cash incomes reported in Census Bureau's Current Population Survey (CPS), Orshansky determined that 34.6 million persons were living in "poverty" in 1963 (Orshansky 1965). In subsequent years, the original income thresholds have been adjusted for inflation according to the Consumer Price Index (CPI-U), but little else has been changed in the framework. Official poverty estimates have been made for each year from 1959 to the present, comparing the poverty income thresholds to household income data reported in the CPS.
The chief merit of this income-based poverty measure is simplicity; it allows the numbers of the poor to be crisply reported, and permits the concept of poverty status to be readily incorporated into a wide range of surveys and studies. However, there are three critical problems with this method of poverty assessment:
1. The fact that a household's cash income falls below the specified threshold provides no clear indication that the household actually experiences a significant material deprivation and provides no information about the exact nature of the hardships when they do occur.
2. The government's determination of poverty is based only on annual income; all assets accumulated by a household in prior years are ignored. As a result the count of the "poor" includes a significant number of relatively well off households who would not be considered impoverished by any common sense definition.
3. A comparison of the CPS income data to Commerce Department data used to measure the Gross Domestic Product shows that the CPS dramatically and consistently under reports the economic resources of households. Even a modest undercount of the economic resources of low-income households will cause a major distortion in the measured poverty rate. Thus, even if the Orshansky/ current government method of determining poverty is accepted in principle, the income reporting in the Current Population Survey may well be too inaccurate to provide a suitable database for assessing poverty.
However, an alternative approach to measuring poverty can be devised based on a survey of actual material living conditions. While this approach lacks the formal simplicity of the government's income-based methods, it does circumvent the problem of income under-reporting. It also provides a manifold and concrete picture of poverty when compared to the one dimensional and abstract income-based method.
While the literature on the traditional income-based poverty measurement is voluminous, examination of the material living conditions of less affluent Americans has been rare. Only a small number of studies have explored non-income based assessments of living conditions. Jorgenson and Slesnick (1989) and Slesnick (1993) have employed an expenditure-based poverty measures, while Lino (1996) devised a measure which combined both expenditure and income data. Direct assessments of the material living conditions and material deprivation among the poor have previously been attempted in Rector (1995), Jencks (1984), Mayer and Jencks (1995), and most extensively in Federman, et al. (1996). Lebergott (1976) provides a masterful treatise on the improvement in American living standards over time but the information is now quite dated. Finally, Callan et al. (1996) provide an interesting estimate of poverty in Ireland based, in part, on a series of material deprivation indicators.
This paper will first examine the problems with income-based poverty measures; it will then examine the material living conditions of less affluent Americans directly. The second section explores the undercounting of economic resources in the Current Population Survey compared to other official sources. The third section provides an alternative definition of poverty based on physical living conditions. The next five sections examine the physical living conditions of Americans especially those who meet the official government definition of poverty. The fourth section examines ownership of property and consumer durables based on the American Housing Survey (AHS) conducted by the United States Department of Housing and Urban Development and the Residential Energy Consumption Survey (RECS) conducted by the United States Department of Energy. The fifth section will deal with housing space and conditions using data from the AHS and RECS. Comparative housing data from other nations is taken from United Nations and World Bank sources. The next section deals with food consumption, nutrient intake and body stature utilizing data from the Continuing Survey of Food Intakes by Individuals (CSFII) of the United States Department of Agriculture, the National Health and Nutrition Evaluation Survey (NHANES) and the Pediatric Nutrition Surveillance System (PedNSS) of the Centers for Disease Control. The seventh section examines utilization of medical care using data from the National Health Interview Survey. The last sections of the paper utilize comprehensive data from the Survey of Income and Program Participation (SIPP) to explore the extent of material hardship within the US population. The ninth section attempts an assessment of the overall rate of material hardship or "poverty." The final section provides a conclusion.
UNDERCOUNTING THE INCOMES OF AMERICAN HOUSEHOLDS
The accuracy of the official government poverty measurement performed annually by the Census Bureau must ultimately rest on the Census' ability to obtain a reasonable count of the income or economic resources available to individual households. But data from other government sources indicate that Census significantly undercounts household income. The potential magnitude of this undercount can be revealed by comparing Census income figures from the Current Population Survey (CPS) with the US Commerce Department's National Income and Product Accounts (N~IPA) which measure the Gross Domestic Product (GDP).
In 1996, Commerce Department figures from NIPA show that the total value of the US GDP was $7.6 trillion (US Department of Commerce, Nov 1997, D-6). Aggregate personal income received by American households including personal tax contributions to Social Security was $6.8 trillion by NIPA figures. However, the aggregate household "money income" figures (which form the basis for the Census' official poverty estimate) fall far short of the NIPA total. According to the Census Bureau's Current Population Survey, total money income of US households was $4.8 trillion in 1996.2 Thus the Census official or money income total was some $2 trillion short of the NIPA figures; this shortfall amounts to roughly $20,000 for each American household, and exceeds the entire economies of all but a handful of the nations in the world.
Table 1 provides a detailed breakout of the discrepancies between NIPA and CPS income figures. While comparison is hindered by the incompatibility of NIPA and CPS income sub-categories, the general picture is clear. The Census CPS money income figures significantly undercount interest, dividends, rents, and private health insurance, all of which would predominately accrue to the rich and middle class. But the CPS also undercounts self-employment income and government transfers, which provide significant economic resources to less affluent households.3
Part of the discrepancy between Census and NIPA figures can be explained by the fact that NIPA personal income totals include many forms of non-cash income, which are specifically excluded by the Census Bureau's official definition of income, which is restricted to cash or money income. Among the types of non-cash income which are included in the NIPA figures but are omitted from Census money or cash income are: the value of private health insurance coverage; interest earned in private life insurance funds; Medicare; Medicaid; Food Stamps and other government food programs; and the earned income tax credit.
To its credit, in the mid 1980's, the Census Bureau, recognizing the inadequacy of its official money income concept, began experimenting with alternative income concepts. Census currently estimates poverty according to some 14 "alternative income definitions," some of which include significant non-cash resources. These alternative measures provide an increase in accuracy and significantly lower the estimated poverty rate. (For example, inclusion of taxes, non-cash benefits, and capital gains combined with the use of a more accurate inflation measure (CPI-U-XI) reduces the poverty rate in 1997 from its official level of 13.3% down to 8.8%.) Unfortunately these alternative income measures are not widely known to the press or policy makers.
Moreover, as Table 2 shows, even using alternative definitions of income CPS figures still fall short of NIPA data by some $1.3 trillion. Significant underreporting of the incomes of less affluent households remains, including a shortfall of over $250 billion in the count of government transfers.4 Clearly, the difficulties of a precise comparison of CPS and NIPA figures are considerable, and the figures in Tables 1 and 2 are intended to give only a very rough approximation of potential underreporting. However, they do strongly indicate an extensive underreporting for all income definitions employed in the CPS.
Corroborative evidence on the problem of underreporting can also be found by comparing CPS money income figures with expenditure figures from the Labor Department's Current Expenditure Survey. in 1995 Census claimed that the lowest income fifth (or quintile) of US households had an average income of $8,350 (US Bureau of the Census 1996: xii.). In the same year, the Consumer Expenditure Survey of the Department of Labor showed that the average household in the same lowest income quintile spent $14,607 (US Department of Labor 1997). Thus the Labor Department survey shows $1.75 in spending for every $1.00 of income Census claims these same households possess. This is no fluke; a similar wide gap between spending and alleged "income" occurred throughout the 1980s and 1990s. But the picture is still incomplete. When counting household expenditures, the Labor Department's Consumer Expenditure Survey excludes public housing subsidies and health care subsidies provided through Medicaid, Medicare, and other government medical programs. If housing and medical subsidies were included, the total expenditures of the average household in the bottom income quintile rise to around $20,000.5 This means less-affluent households spend some $2.40 for every $1.00 of "income" reported by Census.
A DEFINITION OF POVERTY BASED ON MATERIAL HARDSHIP
Clearly, efforts to use income data to assess poverty are fruitless if the income data themselves are substantially inaccurate. The above comparisons indicate that poverty measurements using CPS income figures are, at best, suspect. In consequence, the subsequent sections of this paper will attempt to examine poverty from a different perspective, providing a careful examination of ownership of amenities, housing conditions, food and nutriment consumption and health care utilization among less affluent households.6
We will begin by proposing the following rough definition of poverty or material hardship based on physical living conditions. An individual should be deemed
to be "in poverty" or hardship if any of the following three conditions exist:
1. The individual frequently lacks sufficient food to eat or is significantly undernourished due to an inability to afford or obtain sufficient food.
2. The individual lives in housing that is severely overcrowded (with more than 1.5 persons per room); is severely dilapidated; or is unsafe.
3. The individual has a significant, health impairing, medical condition requiring treatment and cannot afford or otherwise obtain medical care.
The notion of poverty or hardship is strengthened if the individual suffers from multiple problems rather than a single isolated condition.
OWNERSHIP OF PROPERTY AND AMENITIES
Table 3 shows property and consumer durable ownership among all US households and among households who are "poor" according to the official government definition. Some 41 percent of poor households actually own their own homes. The median value of homes owned by these households is $65,000 or 70 percent of the median value of all homes owned in America. Some 900,000 households classified as poor, own homes worth over $150,000. The typical home owned by the poor is a three bedroom house with one and a half baths. It is in good repair, has a garage or carport, and was constructed in 1962. It has a porch or patio and is located on a half-acre lot (US Department of Commerce, American Housing Survey 1997).
Roughly 70 percent of poor households own a car or truck. Remarkably more than one quarter of poor households own two or more cars or trucks. Modern conveniences are very common and, in some cases, almost ubiquitous among poor households. Two thirds of the poor have air conditioning; a similar number have microwaves while almost 30 percent have automatic dishwashers.
It is not surprising that nearly all poor households have color television, but nearly half actually own two or more color television sets, nearly three quarters of the poor now have VCRs, and more than one in five own two VCRs. While these numbers do not suggest lives of luxury they also seem quite distant from conventional images of poverty.
In the US, both the overall population and the poor live, in general, in very spacious housing. As Table 4 shows, nearly 70 percent of all US households have two or more rooms per tenant. Among the poor overall, this figure is 60 percent.
Among the urban poor, it is 55 percent. Crowding is quite rare; only 0.5 percent of all households and 1.9 percent of poor households are severely crowded with 1.5 or more persons per room. Moderate crowding (1.01 to 1.50 persons per room) is slightly more common, characterizing 2.1 percent of all households and 5.6 percent of poor households. By contrast, social reformer Jacob Riis, writing on tenement living conditions around 1890 in New York City, described crowded families living with four or five persons per room and some 20 square feet of living space per person (Riis 1971:6, 41, 59).
Table 5 places current crowding figures in an international perspective. Measured by a room-per-person basis, poor US households have about as much living space as the general population in developed nations such as Germany, Sweden, and the Netherlands. But with two rooms per person, American poor have twice as much living space as the average household in Poland, Hungary, and Romania; three times as much living space as the average household in Iraq, Mexico, and Panama. Finally, they have four times the room of the average family in Peru, Honduras, and India who have only a half a room per person or less (United Nations 1995).
Housing space can also be measured by square feet per person. The Residential Energy Consumption survey conducted by the US Department of Energy shows that Americans have an average 718 square feet of living space per person. Poor Americans have 440 square feet (US Department of Energy 1993: table 3.4). Reasonably comparable, international square footage data are provided by The Housing Indicator Program of the United Nations Center for Human Settlements, which surveyed housing conditions in major cities in 54 different nations. This survey showed the US to have by far the most spacious housing units, 50 to 100 percent more square footage per capita than city dwellers in other industrialized nations (United Nations Centre for Human Settlements and the World Bank 1993).
America's poor also compare favorably with other nations in square footage of living space. Table 6 compares housing space available to poor Americans with data for the general urban population in the 54 nations covered in the United Nations Housing Indicators Program. Housing space for poor Americans appears to be comparable to the average person in major cities in industrialized nations such as the United Kingdom, France, and Japan. Poor Americans have nearly three times the living space of average urban citizens in middle-income countries such as Mexico and Turkey. Poor American households have seven times more housing space per person than the general urban population of very low-income countries such as India and China.
Of course, it might be possible that the housing of poor American households could be spacious but still dilapidated or unsafe. However, data from the American Housing Survey indicates that such is not the case. For example, the survey provides a tally of households with "severe physical problems." Only a tiny portion of poor households and an even smaller portion of total households fall into that category.
The most common "severe problem" according to the American Housing Survey is a shared bathroom, which occurs when occupants lack a bathroom and must share bathroom facilities with individuals in a neighboring unit. This condition affects about one percent of all United States households and two percent of all poor households. About one half percent of all households and two percent of poor households have other "severe physical problems" as shown in Table 7. The most common are repeated heating breakdowns and upkeep problems.
The American Housing Survey provides a count of households affected by "moderate physical problems," also shown in Table 7. A wider range of households falls into this category: nine percent of the poor and nearly five percent of total households. However, the problems affecting these units are clearly modest. While living in such units might be disagreeable by modern middle class standards, they are a far cry from Dickensian squalor. The most common problems are upkeep, lack of a full kitchen, and use of unvented oil kerosene or gas heaters as the primary heat source. (The last condition occurs almost exclusively in the south.)
HUNGER AND MALNUTRITION IN AMERICA
There are frequent charges of widespread hunger and malnutrition in the United States (Community Childhood Hunger Identification Project 1995 and Second Harvest 1997). Reliable survey data shows that while hunger definitely exists in the United States it is relatively restricted in scope and frequency. For example, the Third National Health and Nutrition Examination Survey conducted by the United States Department of Health and Human Services in 1988-1989 found that 96 percent of US households reported they had "enough food to eat." Some three percent reported that they "sometimes" did not have enough food and around a half percent said they "often" did not have enough food. Among the poor, 84 percent reported having enough food, while 13.2 percent reported shortages, "sometimes" and 2.7 percent "often" (Interagency Board for Nutrition Monitoring 1995: VA 95).
The "Food Security Measurement Project" conducted by the United States Department of Agriculture in 1995 provides similar but more detailed data. The report makes a clear distinction between actual malnutrition and the less severe problem of hunger. It provides a series of detailed questions to determine the scope of hunger, which it defines as: "The uneasy or painful sensation caused by a lack of food. The recurrent and involuntary lack of access to food" (US Department of Agriculture 1997: 5).
The study categorized 3.3 percent of all households as "food insecure with moderate hunger," a condition in which food consumption in a household is involuntarily reduced to the extent that adults experience hunger. Some 0.8% of households were found to have experienced "severe hunger," defined as a condition where children in the home had experienced hunger or adults had undergone "extensive reduction in food intake such as going an entire day without eating" (US Department of Agriculture 1997: 35). Among poor households 10.5 percent experienced "moderate hunger" and 3.6 percent "severe hunger." Although many poor families had difficulty stretching their food dollars, in 86 percent of poor homes hunger was not present.
The USDA food security report is particularly interesting because it provides a long series of questions revealing details of eating patterns and food shortages. One clear pattern which emerges from the data is that parents respond to food shortages by first cutting back their own food while continuing to feed their children. Thus parents are more likely to experience hunger than are the young. As Table 8 shows, a half of one percent of children had skipped a meal because of lack of food during the last month; one in a thousand had gone an entire day without eating during the month. Among parents the number was higher: during the month prior to the survey nearly one percent of all parents report going a whole day without eating due to a lack of money to buy food.
It is widely believed that lack of financial resources forces poor people to eat low quality diets that are deficient in nutriments and high in fat. However, survey data show that nutriment density (amount of vitamins, minerals, and protein per kilocalorie of food) does not vary by income class (Windham et al. 1983: 33).
Nor do the poor consume higher fat diets than do the middle class; the percentage of persons with high fat intake (As a share of total calories) is virtually the same for low income and upper middle income persons (Interagency Board for Nutirtion Monitoring, 1995: VA 167). Over consumption of calories in general, however, is a major problem with the poor as itis the general US population.
Poor adult men are slightly less likely than non-poor men to be overweight (30.4 percent to 31.9 percent), but poor adult women are significantly more likely to be overweight than are non-poor women; 47.3 percent of poor adult women are overweight compared to 32 percent of non-poor women (ibid: VA 219).
Examination of the average nutriment consumption of Americans reveals that age and gender play a far greater role than income class in determining nutritional intake. For example, the nutriment intakes of adult women in the upper middle class (above 350 percent of poverty) more closely resemble the intakes of poor women than they do those of upper-middle class men, children or teens (US Department of Agriculture 1995). The average nutriment consumption of upper-middle income pre-schoolers, as a group, is virtually identical with that of poor pre-schoolers; however, not with the consumption of adults or older children in the upper-middle class. This same pattern holds for adult males, teens and most other age and gender groups. In general children aged 0-11 have the highest average level of nutriment intakes relative to RDA followed by adult and teen males. Adult and teen females have the lowest level of intakes. This pattern holds for all income classes.7
Nutrition and Poor Children
Government surveys provide little evidence of widespread undernutrition among poor children; in fact they show that the average nutriment consumption among the poor closely resembles that of the upper-middle class. Example: children in families with income below the poverty level actually consume more meat than do children in families with incomes at 350 percent of poverty or higher (roughly $57,000 for a family of four in today's dollars).
Table 9 shows the average intake of protein, vitamins and minerals as a percentage of the recommended daily allowance (RDA) among poor and middle class children at various age levels8 (US Department of Agriculture 1995: Tables 10-1, 10-4). The intake of nutriments is very similar for poor and middle class children and is generally well above the recommended daily level. For example the consumption of protein (a relatively expensive nutriment) among poor children is, on average, between 150 and 267 percent of the RDA.
When shortfalls of specific vitamins and minerals appear (for example, among teenage girls), they tend to be very similar for the poor and the middle class. For example, while poor teenage girls, on average, tend to under-consume vitamin E, vitamin B-6, calcium, phosphorus, magnesium, iron and zinc, a virtually identical under consumption of these same nutriments appears among upper middle class girls.
It is important to note that the nutriment intake figures in Table 9 represent only the average nutriment intakes for children in various age categories. Thus the figures do not necessarily rule out the possibility that there could be pockets of undernutrition within each age group of poor children even while the average consumption of the group remains high. The data do, however, appear to refute any claim of widespread undernutrition and hunger among poor children in general.
Poor Children's Weight and Stature
The high average level of nutrient intake among poor children however, does not mean that all poor children have achieved optimal nutritional and health status. One well-published concern is that a disproportionate shortness occurs among low-income children (Lewitt 1997). The weight and height of poor children are generally measured relative to standard reference tables established by The National Center for Health Statistics (NCHS) at the Centers for Disease Control and Prevention (CDC). The NCHS reference standards represent the expected anthropometric distribution among children given very high levels of health and nutrition. The concept of "short stature" refers to a child having a height for age below the fifth percentile of the NCHS reference distribution. (It is important to note that under good conditions, five percent of children are still expected have normal heights below this cut off.) If more than five percent of a particular group of children have heights below the cut off level, it is an indication of potential nutritional or other health problems within the group. Similar data can be provided on thinness, (defined as weight for height below the fifth NCHS reference percentile), and obesity, (defined as weight for height above the 95th percentile).
Table 10 provides figures on short stature, thinness and obesity for low-income children under age five in 1996 in the CDCs Pediatric Nutrition Surveillance System (PedNSS). It is important to note the PeDNSS is not a representative sample of low-income children but consists primarily of children participating in the Women Infant and Children (WIC) program; the WIC program requires, as a condition of eligibility, not merely low income status but ancillary clinical conditions such as anemia, low birth weight, and other clinical medical problems. This raises the definite prospect that disproportionate problems within the PeDNSS children may reflect initial selection bias rather than the effects of low-income status per se.9 With this caveat, Table 10 shows that thinness occurs at less than the normal expected rate and is not a general problem. By contrast, obesity appears among 8.4 to 13.5 percent of the children, roughly twice the expected level. This is an indication that caloric intake exceeds expenditure among the affected children. The percentage of low-income children with short stature ranges from 5.8 to 10.3, and is above the expected level of five percent for each age group (Centers for Disease Control, 1997: 5). This disproportionate occurrence of shortness may be an indication either of current under nutrition, health problems or of the lagged effects of low birth weights among some of the affected children.
While possible disproportionate shortness among low-income American children is a cause for concern, it can be informative to place these data in a broader perspective. Table 11 provides international data on children's size based on the World Health Organization Global Data Base on Child Growth. For international data children are judged to be short or "stunted" if their height falls below two standard deviations or the 2.3 percentile level of the NCHS standard tables rather than the fifth percentile level conventionally used for US data.10 Table 11 shows the percentage of children under age five in developing nations who are judged to be "stunted" by this standard. In developing nations as a whole some 43 percent of children are stunted; in Africa more than a third of young children are affected; in Asia nearly half (de Onis, 1993). By contrast, in the US, some 2.6 percent of young children in poor households are stunted by a comparable standard, a rate only slightly above the expected standard for healthy well-nourished children.11 While concern for the well being of poor American children is quite justified, the data overall underscore how large and well nourished poor American children are by global standards.
This should not be too surprising. Throughout this century, improvements in nutrition and health have led to increases in the rate of growth and ultimate height and weight of American children. Poor children have not been unaffected by this trend. Poor boys, today, at ages 18 and 19 are actually taller and heavier than boys of similar age in the general US population in the late 1950s. Poor boys living today are one inch taller and some ten pounds heavier than Gis of similar age during World War II, and nearly two inches taller and 20 pounds heavier than American doughboys back in World War I (Karpinos 1961: 336 - 64).12
ACCESS TO MEDICAL CARE
Much of the discussion concerning health care in the US involves the 15.4 percent of persons who lack health insurance (US Bureau of the Census, March 1996 Current Population Survey). But lacking insurance is not the same as lacking health care; in fact most uninsured persons receive medical care when needed. Data concerning access to medical services is provided in the National Health Interview Survey in 1993 (Bloom etal. 1997; Cohen etal. 1997, Simpson, etal. 1997) and is reported in Table 12.
As shown, three percent of the population reported that at some point in the prior year they needed medical care but were unable to obtain it, among the poor the figure
was 6.4 percent. Some 2.8 percent of the whole population reported needing but not obtaining prescription medicines and 3.8 percent reported needing but not obtaining eyeglasses. Among the poor the figures were 6.3 and 7.3 percent respectively. Finally, 0.7 percent of the population, and 1.4 percent of the poor reported needing but not obtaining mental health care. To what degree, if any, the failures to obtain medical services resulted in significant health impairment or suffering remains unknown13 If less serious conditions, such as delay in seeking medical care because of worry about the cost, and needing but not being able to obtain dental care are added to the problems in Table 12, then the share of persons with any "unmet medical need" can be inflated to 17 percent of the total population and 27 percent of the poor. However, these data are highly problematic, again, because of the lack of any information concerning the severity of the underlying ailment.
COMPREHENSIVE ASSESSMENT OF MATERIAL PROBLEMS
The surveys discussed in the previous sections each examine a single aspect of economic life. For example, the Household Food Security Measurement Project examines hunger, while the American Housing Survey provides data on property ownership and housing conditions, and the National Health Interview Survey looks at medical care. None permit a comprehensive examination of living conditions in households. Fortunately this shortcoming was addressed in a special supplemental questionnaire in the 1992 Survey of Income and Program Participation (SIPP). This questionnaire is unique in addressing a wide range of topics including: housing conditions; ownership of conveniences; hunger; and medical care. While the SIPP questionnaire provides less detailed information on each specific topic than do other government surveys, the overall range of topics examined make it the best available instrument for a general assessment of material living conditions and hardship problems within the US.
This section will use the 1992 SIPP data to attempt an overall assessment of the extent of material hardship in the US population. In presenting the SIPP data we will use persons rather than households as the unit of measure.14 To make the assessment, we have selected 13 general problem conditions and 7 separate household upkeep problems15 This permits coverage of a wide range of problems including: financial difficulties; crowding; housing deficiencies; lack of basic household necessities; interruption of utilities; hunger; and lack of medical care. In order to make the data more comprehensible, we have provisionally ranked these problems into three categories: threshold indicators; moderate material problems and substantial material problems.
The threshold indicator questions cover financial rather than material difficulties. Families who answer affirmatively to these questions have clearly had difficulty making ends meet but have not necessarily suffered any concrete material deprivation. We have labeled these questions "threshold indicators" because they identify households under financial stress which are prone to experience, but have not necessarily experienced, concrete material difficulties. The three threshold indicator questions from SIPP cover the household's financial status during the prior 12 months: at any time did the household fail to meet its essential expenses including medical care, rent, mortgage, and utilities?; at any time did it fail to pay the full gas, oil or electric bill?; at any time did it fail to pay the full rent or mortgage?
As Table 13 shows, some 84 percent of the population reported being able to meet "essential expenses" at all times during the prior year, while some 16 percent report they were unable to meet essential expenses at some point during the year. Among the poor, over 60 percent reported they were able to meet essential expenses throughout the year, while 36 percent reported failing to meet essential expenses at some point. Among the general population during the sample year, some 9 percent at some point failed to pay their full gas, oil or electric bills; and 11 percent failed to fully pay the mortgage or rent bills. Among the poor these figures were 28 and 22 percent respectively.
Overall, some 81 percent of the general population and 58 percent of the poor responded negatively to all threshold questions. This means that well over half or the poor stated that they had been able to fully meet the cost of rent, mortgage,
On the other hand 1 8 percent of all persons and 42 percent of the poor answered affirmatively to at least one threshold question. Obviously, those families responding affirmatively had difficulties making ends meet and experienced financial stress; however, as suggested above, the fact that a family experienced financial stress does not necessarily mean it suffered from concrete material problems such as hunger or inadequate housing. Indeed, of those who answered at least one threshold question affirmatively, only half reported any moderate or substantial material problems.
Moderate Material Problems
Moderate problems include: having gas, electricity or oil cut off for non-payment during the last year; having phone service cut off for non-payment during the year; and moderate crowding (with more than one and less than 1.5 persons per room).16 Other moderate problems include: having three or four current housing upkeep problems; lacking a stove or refrigerator; and occasional hunger (characterized by the assertion that, during the last four months, members of the household "sometimes did not have enough food to eat"). The final moderate problem covers households without medical insurance or other health coverage who stated that, sometime in the last year, a family member needed to go to a doctor and did not go. (Unfortunately, this medical question did not specify the severity of the ailment nor whether the failure to visit a doctor was due to a lack of ability to pay.)
Table 13 shows the prevalence of these moderate material problems. Among the general population, the most common problems were: moderate crowding (4.4 percent) disconnection of the telephone (4.2 percent); housing upkeep problems (3.6 percent); lack of doctor's care among the uninsured (3.0 percent); and cut off of utilities (2.4 percent); and "sometimes" not having enough food (1.9 percent). Overall, as Table 14 shows, 85 percent of Americans had no moderate problems. Among the 15 percent of Americans who experienced moderate problems: 10.8 percent had one problem; 3 percent had two; and 1.2 percent had three or more.
Moderate material problems occurred roughly four times more frequently among the poor than the non-poor. Among poor persons the rates were: moderate crowding (14.0 percent) disconnection of the telephone (14.0 percent); housing upkeep problems (10.1 percent); lack of doctor's care among the uninsured (9.2 percent); and "sometimes" not having enough food (7.9 percent); and cut off of utilities (7.8 percent). Overall some 58 percent of the poor had no moderate problems; among those who experienced moderate problems roughly a quarter had one; and some 17 percent had two or more.
Substantial Material Problems
Substantial problems include: eviction during the prior year; substantial crowding (with more than 1.5 persons per room); five or more current house upkeep problems; and frequent hunger. The last condition is characterized by the assertion that members of the household "often" did not have enough food to eat during the last 4 months. As Table 13 shows substantial material problems were relatively rare: 1.9 percent of all persons suffered from substantial crowding; 0.5 percent were evicted sometime in the prior year; 0.7 lived in run down housing with substantial upkeep problems; and 0.5 percent said their family often did not have enough food. Overall 3.4 percent of Americans were shown to have a substantial material problem. Overlap of substantial problems was very rare, only 0.2 percent of the total population had two or more substantial problems. Overlap of moderate and substantial problems was also far less than expected; half of those with a substantial problem had no additional problems moderate or substantial (see Table 15).
The frequency of substantial problems was some six times greater among the poor than the non-poor. Among the poor 13.2 percent had substantial problems while among the non-poor the figure was 1.9 percent. By far the most common problem was substantial overcrowding (more than 1.5 persons per room) which affected 7.7 percent of poor persons. The other substantial problems, eviction, widespread upkeep troubles, and frequent food shortages, each affected about two percent of poor persons.
Total Material Problems
As Table 14 shows, overall most American homes were free of both substantial and moderate problems. Among all Americans, 83.3 percent had no problems; an additional 9.8 percent had no substantial problems and only one moderate problem. While the comparative absence of these problems within the general populace may not be surprising, the limited prevalence of problems among the poor is a considerable surprise. Over half of all poor persons had no problems of either category. Nearly another quarter of the poor had no substantial problems and only one moderate problem. Thus nearly three quarters of the poor were either free of problems entirely or had only one moderate problem. Less than a tenth of the poor had three or more overall problems.
Material Problems and Demographic Groups
Tables 16 and 17 provide data on material problems and threshold conditions among various demographic groups. The population is divided into four categories: persons in single parent families with children; persons in married couple families with children; persons in elderly-headed households without children; and those in non-elderly households without children. Considerable variation occurs across the four groups. Elderly households without children have the fewest problems followed by non-elderly households without children; next come married couples with children; finally, single parent families with children have the highest level of problems. The consistency of this prevalence pattern is striking: it obtains for each threshold indicator, for every specific upkeep problem; and every material problem except two (lacking a stove or lacking a refrigerator).
As the tables show, households without children were notably free of problems. Among elderly households without children only 6 percent reported any threshold financial condition, and less than one percent had any substantial material problem. Overall 98.8 percent of the elderly had either no material problems or only one moderate material problem. Similar conditions existed among the non-elderly without children. While some 13 percent reported at least one threshold condition, only 1.3 percent reported any substantial problems. Overall, 96.7 percent of these families had either no material problems or one moderate problem.
Problems were more common among married couple families with children. More than one person in five in these families experienced a threshold financial condition. Almost 13 percent lived in families that failed to pay their full gas, oil, or electric bills at some point in the year. (These numbers may, in part, reflect the high unemployment rate of 7.5 percent during the survey year of 1992-the highest unemployment level in the last 15 years.) Despite this, only 4.2 percent experienced any substantial material problems. Overall, 92 percent had either no material problems or one moderate problem.
Not surprisingly, single parent families with children had the highest prevalence of problems; over 40 percent answered affirmatively to at least one threshold question. Nearly nine percent experienced a substantial problem-most commonly over-crowding. However, even in this demographic group, the majority experienced no material problems; some 62 percent had no material problems and an additional 20 percent had only one moderate material problem.
MEASURING OVERALL MATERIAL HARDSHIP OR POVERTY
In a preceding section of the paper we provided a rough definition of poverty or hardship based on three conditions:
1. frequently lacking food or being significantly undernourished due to food shortages;
2. being severly overcrowded (more than 1.5 persons per room) or living in dilapidated or unsafe housing;
3. or having a significant health impairing medical condition and being unable to afford treatment.
We will no use the SIPP data on material problems to attempt to estimate the extent of poverty or overall hardship based on this definition. However, since the SIPP questions do not exactly match our definitional criteria, our efforts must be an approximation only.17
The four substainial problems listed in Table 13 - eviction, widespread upkeep problems, substantials overcrowding, and frequent hunger - correspond at least partially, to our definition of poverty. (We did not categorize the SIPP question concerning failure to visit a doctor/hospital as a substantial problem because no information was given on the severity of the untreated condition.) In 1992, 8.4 million persons or 3.4 percent of the population had one or more substantial material problems.18 However, a small share of households reporting substantial material problems have relatively high income levels, suggesting that the problem condition may not be primarily income-related. (Substantial overcrowding is by far the most common condition among this higher income population.) Finally, some lower income persons who experience no substantial material problems do suffer from an array of moderate problems.
To deal with these complications, we have created a somewhat expanded category which we shall term overall material hardship. Persons shall be considered to be in a condition of overall material hardship if they meet either of the following two problem conditions: a) they have one or more substantial problems; or b) they have three or more moderate problems. Finally, we set a third, limiting condition: that all persons in the overall material hardship population should reside in households with income below 200 percent of the official poverty thresholds.19 According to these criteria, in 1992, 8.7 million persons or 3.5 percent of the population were in a condition of overall material hardship. We shall term this group the hardship population.
As a whole the hardship population has a mean of 2.3 and a median of 2.0 material problems. Slightly more than half of the group are poor according to the official government definition; roughly a quarter of the hardship population were black and three quarters were non-black. Among those persons who were poor according to the official government definition, 15.2 were in hardship; among the non-poor the rate was 1.7 percent; among blacks 7.9 percent were in hardship; while among whites the rate was 2.8 percent.
Table 18 provides additional demographic information on the hardship population. Some 52 percent of the group lived in married-couple families with children; over a third were single-parent families with children; some 2 percent were elderly without children and 9 percent were non-elderly without children. Thus, almost nine out of ten of the hardship population were persons in families with children. The hardship prevalence rates among these groups were as follows: some 4 percent of persons in married couple families with children were in hardship; among single-parent families with children the rate was more than twice as high at roughly 10 percent. Less than one percent of elderly persons without children in the home were in hardship, while among the non-elderly without children the rate was 1.1 percent.
The hardship population was 26 percent of the size of the official poverty population in 1992. The racial composition of the hardship population was very similar to the official poverty population. However, in terms of other demographic characteristics, there were substantial differences between the hardship group and the official poor. When compared to the official poor, the hardship population had proportionately roughly twice the proportion of persons in married couple families with children; roughly the same proportion in single parent families with children; and significantly fewer elderly and non-elderly persons without children.
Finally, Table 19 shows the ownership or availability of amenities among the hardship population. The picture is a familiar one; modern amenities are very common although not quite as common as among the official poor. It is important to note that an individual in material hardship in the United States may have a standard of living well above that of the average population across most of the globe. For example, even intermittent free medical care received in the United States is likely to be superior to most medical care available in developing nations. An American who is judged to live in overcrowded housing according to the above standard actually possesses more living space than general populations throughout many developing nations. Similarly, an American who is hungry or "undernourished" may well have an above average nutrient intake by global standards.
Today the typical American, defined as poor by the government, has a refrigerator, a stove, a clothes washer, a car, air conditioning, a VCR, a microwave, a stereo, and a color TV. He is able to obtain medical care and his home is in good repair and is not over-crowded. By his own report, his family is not hungry and in the last year he had sufficient funds to meet his essential needs. While this individual's life is not opulent, it is equally far from the popular images of poverty conveyed by politicians, the press, and activists.
Many of the popular conceptions about poverty in this nation are inaccurate, particularly the image of poverty as a static and unyielding condition. In her original article defining poverty, Mollie Orshansky stated:
there are still many who must watch America's parade of progress from the sidelines, as they wait for their turn, a turn that does not come.
(Orshansky 1965: 3)
This statement was untrue when it was uttered and remains untrue today.20 In the more than two centuries of our nation's existence, material living conditions of Americans, including lower income Americans, have improved enormously. That progress continues today. We have not only triumphed over poverty as it was historically understood, but that triumph has been so great that we have difficulty remembering what it meant to be poor or even to be middle class in earlier generations.
Bloom, B, Simpson, G., Cohen, R. A., and Parsons, P. E. (1997) "Access to Health Care Part 2: Working-Age Adults" Vital Health Statistics, 10(197).
Callan, T., Brian N., Whelan, B. J., Whalen, C. T., and Williams, J. (1996) Poverty in the 1990c: Evidence from the 1994 Living in Ireland Survey, Dublin: Oak Tree Press.
Centers for Disease Control and Prevention, National Center for Health Statistics (1995) "Health Aspects of Pregnancy and Childbirth: United States, 1982-88", Vital and
Health Statistics, 23(18): Table 24.
Centers for Disease Control and Prevention, National Center for Health Statistics (1997) "Pediatric Nutrition Surveillance, 1996 Annual Summaries", unpublished tables,
Cohen, R. A., Bloom, B., Simpson, G. and Parsons, P. E. (1997) "Access to Health Care Part 3: Older Adults" Vital Health Statistics 10(198).
Community Childhood Hunger Identification Project (1995) A Survey of Childhood Hunger in the United States, Washington, DC: Food Research and Action Center.
de Onis, M. and Habicht, J. P. (1996) "Anthropometric Reference Data for International Use: Recommendations from a World Health Organization Expert Committee," American Journal of Clinical Nutrition 64: 650-8.
de Onis, M., Monteiro, C., Abe. J. and Clugston, G. (1993) "The Worldwide Magnitude of Protein-Energy Malnutrition: an Overview from the WHO Global Database on Child
Growth," Bulletin of the World Health Organization 7 1(6): 703-7 12.
Federman, M., Garner, T. I., Short, K., Cutter IV, W. B., Kiely, J., Levine, D., McGough D. and McMillen, M. (1996) "What Does it Mean to be Poor in America," Monthly Labor Review 119(5): 3-17.
Fisher, G. M. (1992) "The Development and History of the Poverty Thresholds," Social Security Bulletin 55(4):3-14.
Internal Revenue Service (1988) Income Tax Compliance Research: Supporting Appendices to Publication 7285, Washington D.C.: Department of Treasury.
Interagency Board for Nutrition Monitoring and Related Research (1995) Third Report on Nutrition Monitoring in the United States, Washington, D.C.: US Government Printing Office: volume 5.
Jencks, C. (1984) "The Hidden Prosperity of the 1970s," The Public Interest 77: 37-61. Jencks, C. (1994) The Homeless, Cambridge, Massachusetts, Harvard University Press.
Jorgenson, D. W., and Slesnick, D. T. (1989) "Redistributional Policy and the Measurement of Poverty" in Slottje, D. (ed.), Research on Economic Inequality, Greenwich:
JAI Press, volume 1: 1-48.
Karpinos, B. D. (1961) "Current Height and Weight of Youths of Military Age," Human Biology 33(4).
Lebergott, 5. (1976) The American Economy: Income Wealth and Want, Princeton, New Jersey: Princeton University Press.
Lewit, E. M. and Kerrebrock, N. (1997) "Population-Based Growth Stunting," The Future of Children 7(2), United States of America: The Center for the Future of Children.
Lino, M. (1996) "Income and Spending of Poor Households with Children," Family Economics and Nutrition Review 9(1).
Mayer, S. E. and Jencks, C. (1995) Has Poverty Really Increased Among Children Since 1970? Working Paper 94-14, Center for Urban Affairs and Policy Research, Northwestern University.
McCrohan, K., Smith, J. D. and Adams, T. K. (1991) "Consumer Purchases in Informal Markets: Estimates from the 1980's, Prospects for the 1990's," Journal of Retailing 67(1).
Orshansky, M. (1965) "Counting the Poor: Another Look at the Poverty Profile," Social Security Bulletin 28(1).
Rector, R. (1995) "How 'Poor' Are America's Poor?" in Simon, J. L. (ed.), The State of Humanity, Cambridge, Massachusetts: Blackwell Publishing: 24 1-256.
Rector, R. and Lauber, W. F. (1995) America's Failed $5.4 Trillion War on Poverty, Washington, DC: The Heritage Foundation.
Riis, J. (1971) How the Other Half Lives, New York: Dover Press.
Second Harvest (1997) "1997 National Research Study", in Hunger 1997: The Faces & Facts, Chicago, Illinois.
Simpson, G, Bloom, B. Cohen, R. A. and Parsons, P. E. (1997) "Access to Health Care Part I: Children" Vital Health Statistics, 10(196).
Slesnick, D. T. (1993) "Gaining Ground: Poverty in the Postwar United States," Journal of Political Economy, 10 1(1): 1-38.
United Nations (1995) Compendium of Human Settlements Statistics 1995, New York, NY: United Nations.
United Nations Centre for Human Settlements and the World Bank (1993)The Housing Indicators Program, Vol. III Preliminary Findings, New York: United Nations.
US Bureau of the Census (1997) Current Population Reports, Money Income in the United States: 1995, Washington D.C.: US Government Printing Office: 60-193.
US Bureau of the Census (1997) Current Population Reports, Money Income in the United States: 1996, Washington, D.C.: US Government Printing Office: 60- 197.
US Bureau of the Census (1996) Current Population Reports, Money Income in the United States: 1995, Washington D.C.: US Government Printing Office: 60-193.
US Department of Agriculture, Agriculture Research Service (1995) Food and Nutrient Intakes by Individuals in the United States, I Day, 1989-91, Nationwide Food Surveys,
NFS Report N. 91-2 ,Washington, D.C.: US Government Printing Office.
US Department of Agriculture (1997) Household Food Security in the United States in 1995: Summary Report for the Food Security Measurement Project, Washington D.C.: US Department of Agriculture.
US Department of Commerce, Bureau of Economic Analysis (1997) "BEA Current and Historical Data," Survey of Current Business, Washington, D.C.: US Department of
US Department of Commerce and US Department of Housing and Urban Development (1997) American Housing Survey for the United States in 1995, Washington, DC: US
US Department of Energy, Energy Information Administration (1995) 1993 Residential Energy Consumption Survey, Housing Characteristics 1993, Washington, DC: Energy
US Department of Energy, Energy Information Administration (1997) 1997 Residential Energy Consumption Survey, Preliminary data available http://www. eia. doe.gov emeu/consumption.
US Department of Labor, Bureau of Labor Statistics (1997) Report 911, Consumer Expenditures in 1995, Washington D.C.: US Department of Labor.
US Department of Labor, Office of International Economic Affairs, Bureau of International Labor Affairs (1992) The Underground Economy in the United States, Washington, DC: US Department of Labor: Occasional Paper No. 2.
Windham, C. T. et al. (1983) "Nutrient Density of Diets in the USDA Nationwide Food Consumption Survey, 1977-1978: Impact of Socioeconomic Status on Dietary Density," Journal of the American Dietetic Association.
1 In order to determine an allowance for non-food expenditure. Orshansky examined the household budgets among the general US population in 1955, finding that families of three or more persons spent, on average, a third of their budgets on food, while among smaller households roughly a quarter of spending went to food. After assuming that the non-food share of budget for the poor should be the same as for the general population, Orshansky set the poverty income thresholds for families of three or more persons at an amount equal to three times the cost of the family's economy food plan (allowing a third of spending for food and two thirds for other purposes). The poverty threshold for two person households was set at four times the household's economy food plan; for single person households the income threshold was set at 80 percent of the two-person level.
2 The degree of underreporting in the CPS, as measured by the ratio of CPS aggregate income to NIPA aggregates, has remained relatively constant since the 1950s; however, the type of income which is underreported has changed considerably. In the 1950s transfer payments were a small part of income and comprised a correspondingly small part of the underreporting; by the 1990s government transfers had risen to a quarter of all unreported income in the CPS. During the same period means-tested government transfers rose from around one percent of GDP to five percent (Rector and Lauber 1995), and also increased as a share of underreported income. The disproportionate growth in the underreporting of transfer payments means that the CPS has increasingly underreported the economic resources of poor and lower income households over time. This bias renders the CPS unreliable for measuring long term temporal changes in the living standards of lower income households.
3 The NIPA figures contain substantial adjustments for underreporting of self-employment income based on IRS research. For a discussion of the size of the informal sector of the US economy and under-reported self-employment income see: McCrohan et a!. (1991); US Department of Labor 1992; and Internal Revenue Service (1988).
4 Some 40 percent of the nearly $300 billion in unreported government transfer payments in Table 2 were means-tested transfers targeted to lower income groups. In addition, in 1996 there was over $20 billion in government expenditure on public housing programs which was almost totally unreported in the CPS and was also not included in the personal income data of the NIPA.
5 This estimate, based on CPS figures on the share of program recipients in each quintile, assumes that the lowest income quintile of households receive 74 percent of public housing benefits, 47 percent of Medicaid assistance (excluding nursing home care) and 29 percent of Medicare.
6 Information on changes in material living conditions over time can be found in Rector (1995) and Lebergott (1976).
7 Supporting tables available upon request.
8 Table 9 also provides the "mean adequacy ratio" for various groups. The mean adequacy ratio represents average intake of all the nutriments listed as a percent of RDA. However, in computing mean adequacy, intake values exceeding 100 percent of RDA are counted at 100, since the body cannot use an excess consumption of one nutriment to fill a shortfall of another nutriment.
9 By contrast, the third National Health and Nutrition Evaluation Survey (NHANES III) is a representative sample of the US population. NHANES III appears to give much lower rates of short stature among poor children than does the PeDNSS.
10 For a discussion of the use of NCHS anthropometric tables internationally see de Onis (1996).
11 Calculation by the authors using NHANES III data and WHO standard tables for shortness for age. WHO standard tables are based on the NCHS standards.
12 Recent data on young males in poverty provided by the National Center for Health Statistics of the US Department of Health and Human Services based on the second National Health and Nutrition Examination Survey.
13 Federal law provides a foundational medical safety net by requiring that emergency rooms and hospitals treat all persons, irrespective of ability to pay, in any situation where the failure to provide care would place the individual's health in "serious jeopardy." Hospitals provide some $18 billion in uncompensated care each year.
14 By contrast, Tables 3 through 8 in the prior sections use the household as the unit of measure.
15 The SIPP supplement also contained the following question: "during the past 12 months, has there been a time when your household had someone who needed to go to a dentist but did not go?" This item was only weakly related to other problem conditions and has been omitted from our list of material problems because of the lack of information concerning the severity of the underlying dental problem.
16 Much of the difference in the crowding data presented in Table 4 from the American Housing Survey and Table 13 from SIPP is due to the fact that the data in Table 4 uses households as the unit of measure while Table 13 uses persons.
17 One additional problem is that the SIPP survey is unlikely to contain persons who are currently homeless. While they should be counted as poor, the homeless at any given point in time are unlikely to exceed 400,000 or about 0.15 percent of the population (Jencks 1994: 13) so their numbers would not appreciably affect the calculations in this paper.
18 One caveat to this analysis is that roughly a third of those with substantial problems have only one such problem and also lack any accompanying moderate problems or threshold financial conditions. Among those experiencing a single isolated substantial problem, overcrowding is most commonly that problem. It is unclear to what extent this single problem group really represents a reasonable concept of impoverishment.
19 Slightly over 2 million persons were excluded from the overall hardship population because they resided in households with incomes greater than 200 percent of poverty. Multiple problems within this excluded group were rare; two thirds had only a single substantial problem and no moderate problems. The most frequent problem conditions within this group were substantial crowding and substantial upkeep problems.
20 For example, between the early 1940's and the early 1960's, the average income of lowest quintile of households nearly doubled in constant dollars. Lebergott (1976) also provides abundant evidence of rising living standards among the less affluent before 1960. To provide one striking example: in 1940,40 percent of all US households lacked an indoor flush loilet, by 1950 Ihe number was 29%, and by 1960 it was 13cc. Among blacks the figures were as follows: 74% in 1940, falling to 31% in 1960 (Lebergott, 1976: 276).