Defining Poverty

[Chapter Three of  The Conquest of Poverty.]

Any study of poverty should logically begin with a definition of the problem we are trying to solve. Precisely what is poverty? Of the thousands of books and articles on the subject that have appeared over the last two centuries, it is astonishing how few have troubled to ask this question. Their writers have taken it for granted that both they and their readers know precisely what is being discussed. Yet popularly the term is very vague. It is nearly always employed in a relative rather than an absolute sense. In Victorian England it became the fashion for some politicians to say that “the Rich and the Poor form Two Nations.” But as every family’s income, if arranged on a scale according to its dollar amount, would probably form a dot on a continuous smooth curve, the dividing line between the poor and the not-poor would be an arbitrary one. Is the poorer half of the population anywhere to be called the Poor, and the richer half the Rich? The discussion today is conducted dominantly in these comparative terms. Our reformers are constantly telling us that we must improve the condition of the lowest fifth or the lowest third of the population. This way of discussing the subject was made fashionable by President Franklin D. Roosevelt in his Second Inaugural Address in January, 1937: “I see one third of a nation ill-housed, ill-clad, ill-nourished.” (The objective standards on which this statement was based were never specified.)

It is obvious, however, that all merely relative definitions of poverty make the problem insoluble. If we were to double the real income of everybody, or multiply it tenfold, there would still be a lowest third, a lowest fifth, a lowest tenth.

Comparative definitions lead us, in fact, into endless difficulties. If poverty means being worse off than somebody else, then all but one of us is poor. An enormous number of us are, in fact, subjectively deprived. As one writer on poverty succinctly put it nearly sixty years ago: “It is part of man’s nature never to be satisfied as long as he sees other people better off than him self.”1

A discussion of the role that envy plays in economic and all human affairs can be deferred to another place. In any case we are driven to try to find an absolute or objective definition of poverty. This turns out to be more difficult than it might at first seem. Suppose we say that a man is in poverty when he has less than enough income, or less than enough in nutrition and shelter and clothing, to maintain himself in normal health and strength. We soon find that the objective determination of this amount is by no means simple.

Let us turn to some of the recent “official” definitions in the United States. In January, 1964, when President Johnson was launching his “war on poverty,” the annual report of the Council of Economic Advisers contained a long section on the problem. This offered not one but several definitions of poverty. One was relative: “One fifth of our families and nearly one fifth of our total population are poor.” A second was at least partly subjective: “By the poor we mean those who are not now maintaining a decent standard of living — those whose basic needs exceed their means to satisfy them.” Each of us might have his own conception of a “decent” standard, and every family might have its own ideas of its “needs.” A third definition was: “Poverty is the inability to satisfy minimum needs.”

The Council of Economic Advisers, basing its estimates on “low-cost” food budgets compiled by the Social Security Administration, decided that the poverty “boundary line”was established by “a family whose annual money income from all sources was $3,000 (before taxes and expressed in 1962 prices).” Yet on the very next page the Council report declared that in 1962 “5.4 million families, containing more than 17 million persons, had total incomes below $2,000.” How could these 17 million persons exist and survive if they had so much less than enough “to satisfy minimum needs”?

In a 50-page study published in 1965,2 Rose D. Friedman subjected these Council estimates to a thorough analysis. Using precisely the same data and the same concept of “nutritive adequacy” as the Council, she found that the dividing line between the poor and the not-poor would be not $3,000, but a figure around $2,200 as the relevant income for an on-farm family of four. Where the Council on the basis of its figure estimated that 20 percent of all American families in 1962 were poor, Mrs. Friedman found that on her adjusted calculation only about 10 percent were poor.

I must refer the interested reader to the full text of her study for the details of her excellent analysis, but two of her disclosures will be enough to illustrate the carelessness of the Council’s own estimates.

One astonishing error by the Council was to use its $3,000 a year estimate as the “poverty boundary” for all families of any size. Mrs. Friedman’s estimates ranged from $1,295 for two person households to $2,195 for four-person households to $3,155 for households of seven persons or more. (The official “poverty line” estimates now also specify a similar range of differences for families of different sizes.)

A second error of the Council was equally astonishing. Based on a previous official estimate that a poor family of four needed about $1,000 a year in 1962 for adequate nutrition, the Council multiplied this amount arbitrarily by three to get what the family needed for all purposes. But it is notorious that poorer families have to spend a larger proportion of their income on food than do richer families. Mrs. Friedman found that this multiple of three was much higher than the level at which three-fourths of the families concerned did get along on and still get an adequate diet. She found that the amount actually spent for food, on the average, by a family of four with an income of $2,200 was about $1,248 a year. In other words, the fraction of income spent on food at this level was about 60 percent and not 33 percent. Yet the official “poverty line” estimates, at this writing, are still kept unrealistically high by continuing to be implicitly based on this arbitrary multiple of three times adequate diet costs. What Is “Adequate” Nutrition?

One of the great problems involved in arriving at any objective standard of poverty is the constantly changing concept of what constitutes “adequate” nutrition. This was once measured in calories. As time has gone on, and scientific research has continued, it has been insisted that adequacy also requires certain amounts of protein, calcium, iron, Vitamin A, thiamine, riboflavin, niacin, ascorbic acid, etc. The newest insistence has been on the need for a multitude of amino acids. Recently a nutrition survey done at Pennsylvania State College concluded that “only one person in a thousand escapes malnutrition!”3 On this basis even affluence is no assurance of nutritional adequacy.

Yet compare this scientific ideal not only with the historic situation before the present century, when getting enough to eat was the major problem of the great majority of the populace of the world, but with the conditions that still prevail among that majority. Compared to a supposed subsistence-minimum of 3,500 calories, half the people of the world today still get less than 2,250 calories per day, and live on a diet primarily of cereal in the form of millet, wheat, or rice. Another 20 percent get less than 2,750 calories per person per day. Only the well-to-do three-tenths of the human race today get more than 2,750 calories as well as a varied diet which provides the calories that not only satisfy hunger but also maintain health.4

Official estimates of “poverty-threshold” income by Federal bureaus are still unrealistically high. I quote from a recent official bulletin:

The decade of the sixties has witnessed a sizable reduction in the number of persons living in poverty. Since 1959, the first year for which data on poverty are available, there has been an average annual decline of 4.9 percent in the number of poor persons. However, between 1969 and 1970, the number of poor persons increased by about 1.2 million, or 5.1 percent. This is the first time that there has been a significant increase in the poverty population. In 1970, about 25.5 million persons, or 13 percent of the population, were below the poverty level, according to the results of the Current Population Survey conducted in March, 1971 by the Bureau of the Census.5

Yet though the estimate of the poor was then only 13 percent of the population compared with about 20 percent in 1962, the government statisticians were still using their old high estimate for 1962 — and writing up the dollar amount year by year to correspond with increases in the Consumer Price Index. The same bulletin quoted above informs us: “The poverty threshold for a nonfarm family of four was $3,968 in 1970 and $2,973 in 1959.” If Mrs. Friedman’s more careful calculations had been used, the “poverty threshold” for an on-farm family of four would have been closer to $2,900 than to $3,968 in 1970 and the percentage of “the poor” would have been closer to 7 percent than to 12.6. In fact, an earlier bulletin of the Bureau of the Census,6 which had estimated that “about one ou of ten families were poor in 1969, compared with about one out of five in 1959,” informs us that if the Bureau’s various “poverty thresholds” for families of different sizes were decreased to 75 percent of its existing estimates (i.e., to approximately the levels suggested by Mrs. Friedman’s calculations), then “the number of poor persons would drop by 40 percent in 1969, and the poverty rate for persons would drop from 12 percent to seven percent.”

It is clear from all this that government bureaucrats can make the numbers and percentage of “the poor,” and hence the dimensions of the problem of poverty, almost whatever they wish, simply by shifting the definition.

And some of our American bureaucrats have been doing just that. On December 20,1970, for example, the Bureau of Labor Statistics announced that, as of the spring of that year, it took a gross income of $12,134 to maintain a family of four on a “moderate” standard of living in the New York-northeastern New Jersey area. The implication was that any family of four with a smaller income than that was less than “moderately” well off and presumably the taxpayers should be forced to do something about it.

Yet the median income of a typical American family7 was estimated by the Bureau of the Census to be only $9,433 in 1969. This means that half of the number of American families were receiving less than that. Clearly a good deal less than half of American families were lucky enough to be receiving the “moderate” income of $12,134.

Most of those who try to frame a definition of poverty no doubt have in mind some practical purpose to be served by such a definition. The purpose of the Federal bureaucracy is to suggest that any income below its definition constitutes a problem requiring government relief, presumably by taxing the families who earn higher incomes to supplement or subsidize the lower. If the present official U.S. definitions of poverty were applied to a country like India, we would have to label as poverty-stricken the overwhelming majority of its population. But we do not have to go to India for such an example. If we go back only a little more than forty years ago in our own country, we find that in the so-called prosperous year 1929 more than half of the people in the United States would have been labeled “poor” if the “poverty-threshold” income since developed by the Council of Economic Advisers had then been applied. (This is based on statistical comparisons that fully allow for the changes in the price level in the meantime.)8

Let us look at one more example of the consequences of establishing an excessive or merely relative definition of poverty.

The term poverty may connote hunger, but this is not what is usually meant in discussions about poverty in America. Consider, for example, the facilities available to the poor. Tunica County, Mississippi, is the poorest county in our poorest state. About eight out of every ten families in this county had incomes under $3,000 in 1960 [i.e., under the official “poverty threshold” level] and most of them were poor by national standards; yet 52 percent owned television sets, 46 percent owned automobiles, and 37 percent owned washing machines. These families might have been deprived of hope and poor in spirit, but their material possessions, though low by American standards, would be the envy of the majority of mankind today.9

To sum up: It is difficult, and perhaps impossible, to frame a completely objective definition of poverty. Our conception of poverty necessarily involves a value judgment. People in different ages, in different countries, in different personal circumstances, will all have different ideas of what constitutes poverty, depending on the range of conditions to which they themselves are accustomed. But while the conception of poverty will necessarily be to some extent relative and even individual, we should make every effort to keep it as objective as we can. Otherwise if, for example, our national income in real terms continues to rise as much in the next forty years as in the past forty years, our social reformers will tend to raise correspondingly their standard of what constitutes “poverty.” And if this happens, the paradoxical result will be that the problem of poverty will seem to them to be getting larger all the time when it is really getting smaller all the time.

One writer has seriously suggested that we “define as poor any family with an income less than one-half that of the median family.”10 But on this definition, if the wealth and income of all groups increased more or less proportionately, as in the past, and by no matter what rate or what multiple, the percentage of “the poor” would never go down, while the implied absolute amount of relief required would keep soaring.

Our definition obviously should not be such as to make our problem perpetual and insoluble. We must avoid any definition that implies the need of a level of help or any method of help that would tempt the recipient to become permanently dependent on it, and undermine his incentives to self-support. This is likely to happen whenever we offer an able-bodied adult in charity or relief more than or even as much as he could earn by working. What he needs is a level of subsistence sufficient to maintain reasonable health and strength. This subsistence level must constitute our working definition of the poverty line. Any relief program that tries to provide more than this for idle able-bodied adults will in the end do more harm than good to the whole community.

• 1. Hartley Withers, Poverty and Waste, London, Elder Smith, 1914; Second Revised Edition, John Murray, 1931, p. 4.
• 2. Poverty: Definition and Perspective. American Enterprise Institute, Washington, D.C.
• 3. Bulletin No. 1, July, 1968, Foundation for Nutrition and Stress Research, Redwood City, California.
• 4. Rose D. Friedman, op. cit.; M. K. Bennett, The World’s Food, New York: Harper & Bros., 1954
• 5. Consumer Income, Series P-60, No.77, May 7, 1971, U.S. Department of Commerce, Bureau of the Census.
• 6. Series P-60, No.76, December 16,1970.
• 7. Not necessarily a family of four. The term “family” as used by the Bureau for thiscalculation “refers to a group of two or more persons related by blood, marriage, or adoption and residing together; all such persons are considered members of the same family.” Economic Report of the President February, 1971, Table c-20, p.220.
• 8. Source: Jeanette M. Fitzwilliams, “Size Distribution of Income in 1962,” Survey of Current Business, April, 1963, Table 3; Herman P. Miller, Rich Man-Poor Man, New American Library, 1964, p.47
• 9. Herman P. Miller, Rich Man, Poor Man, New York, Thomas Y. Crowell Co.
• 10. Victor R. Fuchs, “Toward a Theory of Poverty,” in U.S. Chamber of Commerce, The Concept of Poverty, Washington, D.C., 1965, p.74.