John F. Cogan
Poverty statistics are the primary indicator of living conditions among people at the economic spectrum's lower end. These data are among the most important and the most politically sensitive data published by the U.S. government. That the method used to measure poverty has remained unchanged since its inception, despite well-recognized conceptual and methodological problems, is testimony to this sensitivity. In this environment, only a report firmly grounded in science can produce the kind of agreement among government officials that would lead to improvements in measuring poverty. The major recommendations and conclusions for changing the measurement of poverty by the Panel on Poverty and Family Assistance are not based on scientific evidence. They lie well outside the National Research Council's stated mission "to deliver science advice" to the government. Therefore, I have chosen to dissent.
There are parts of the report for which the panel should be commended. The sections that address problems with the current measure, alternative poverty concepts, and measuring poverty across families of different sizes are particularly illuminating. More analyses based on the scientific literature would have improved the report. Social science research has developed a vast body of scientific knowledge about issues relating to the measurement of poverty. Indeed, many panel members have been important contributors to this knowledge base. There exist, for example, well-developed studies for constructing efficient, meaningful indices to account for geographical differences in living costs. This literature identifies sampling procedures that can be applied to maximize the informational content of surveys at minimum cost and to develop appropriate weighting schemes to create a consumption bundle that
Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.
OCR for page 385
Measuring Poverty: A New Approach APPENDIX A Dissent John F. Cogan Poverty statistics are the primary indicator of living conditions among people at the economic spectrum's lower end. These data are among the most important and the most politically sensitive data published by the U.S. government. That the method used to measure poverty has remained unchanged since its inception, despite well-recognized conceptual and methodological problems, is testimony to this sensitivity. In this environment, only a report firmly grounded in science can produce the kind of agreement among government officials that would lead to improvements in measuring poverty. The major recommendations and conclusions for changing the measurement of poverty by the Panel on Poverty and Family Assistance are not based on scientific evidence. They lie well outside the National Research Council's stated mission "to deliver science advice" to the government. Therefore, I have chosen to dissent. There are parts of the report for which the panel should be commended. The sections that address problems with the current measure, alternative poverty concepts, and measuring poverty across families of different sizes are particularly illuminating. More analyses based on the scientific literature would have improved the report. Social science research has developed a vast body of scientific knowledge about issues relating to the measurement of poverty. Indeed, many panel members have been important contributors to this knowledge base. There exist, for example, well-developed studies for constructing efficient, meaningful indices to account for geographical differences in living costs. This literature identifies sampling procedures that can be applied to maximize the informational content of surveys at minimum cost and to develop appropriate weighting schemes to create a consumption bundle that
OCR for page 385
Measuring Poverty: A New Approach reflects true differences in living costs. There is also a rich literature on statistical properties of alternative imputation procedures that would be required to incorporate in-kind benefits and taxes into measures of family resources. To some people, these contributions may not be eye-catching; they may not be newsworthy; but they are scientific. Instead of focusing on these areas where science can make a contribution, the report is devoted to recommendations and conclusions that are driven by value judgements. According to the report, the poverty line should be raised from its current level, it should rise faster than inflation over time, and fewer resources should be counted when determining whether a family's income is above or below the poverty line. These recommendations are not scientific judgements. They are value judgements made by scientists—with a particular point of view. In essence, the panel has mostly eschewed the role of scientific panel and has instead assumed the role of a government policy maker. By so doing, the panel has not served well either the policy community or the scientific community. Although it can be difficult to establish a precise boundary between where science ends and policy making begins, this panel has ventured far afield in a desire to make a difference. Instead of using strong scientific research to produce recommendations that would compel a particular policy approach, the panel has made recommendations with little scientific bases. My dissent focuses on four major recommendations and conclusions: measuring the poverty line, choosing a range for the poverty line, updating the poverty line, and accounting for medical care in measuring family resources. This dissent is not intended to be a comprehensive critique of the panel report. Although there is considerably more in the report that I find objectionable, to avoid obscuring the central reason for my dissent, I do not address objections that are not germane to it. MEASURING THE POVERTY LINE The report recommends a new method for measuring the poverty line: The poverty threshold should represent a budget for food, clothing, and shelter (including utilities) and a small additional amount to allow for other needs … I focus first on this seemingly noncontroversial recommendation because it illustrates the lack of scientific basis that permeates the report's major recommendations for changing the measurement of poverty. My objection to this recommendation is that the choice of particular commodities is not based on science. The choice may appear to be quite reasonable, and the panel may be correct when it argues that these commodities "represent basic living needs with which no one would quarrel." But
OCR for page 385
Measuring Poverty: A New Approach what scientific basis exists for concluding that food, clothing, and shelter are basic needs and health care or personal care services are not? Is it a scientific proposition that designer tennis shoes are a basic need but that the services of primary care physicians are not? What scientific basis exists for concluding that all types of food, clothing, and shelter, rather than only a subset, are basic needs? The report provides no answers to these questions. It does not attempt to establish a scientific basis nor does it present scientific evidence to support its choices. The panel's primary rationale is that "the United States has major assistance programs to provide food and housing…[and] clothing allowances historically were separately identified grants under Aid to Families with Dependent Children." This argument is faulty on several accounts. First, given the broad array of government-provided benefits, the same argument could be used to support the inclusion of any number of other commodities as basic needs. Health care, education, transportation, and laundry services are all currently provided by the federal government to the poor. Second, the fact that the government provided medical care to the poor on an entitlement basis long before it established entitlements for either food or housing assistance might suggest that medical care is every bit as basic a need as the former set of commodities. Also, the fact that the U.S. government spends an increasingly substantial proportion of its budget to provide medical insurance for the low-income population is a strong indication that medical care is viewed as a priority commodity. The foregoing should not be taken to mean, however, that scientific study has no role in this choice. Scientific analysis can play a significant role by evaluating methods to improve the quality of existing consumption data. It can establish criteria for evaluating the statistical accuracy of alternative poverty budgets. It can evaluate alternative sampling methodologies to improve a survey's ability to count certain groups, such as the homeless. Scientific analysis can ascertain living conditions of families at different income levels so that policy officials can determine the levels of income that should qualify as poverty. UPDATING THE POVERTY LINE The panel report recommends updating the poverty line annually by the growth rate in the median level of expenditures on food, clothing, and shelter, rather than by the Consumer Price Index as is the current practice. If adopted, the recommendation would fundamentally change the concept of poverty from an absolute standard to a relative standard. Under the recommended method, the poverty line would rise about 8 percent faster per year than under the current method. This recommendation, like the previously discussed one, cannot be de-
OCR for page 385
Measuring Poverty: A New Approach duced from any set of scientific principles, facts, or arguments. Any updating method, be it one to ensure an absolute poverty threshold, a relative threshold, or one that falls somewhere in between, is a policy choice, not a scientific one. But unlike the previously discussed recommendation, this one would have a substantial impact on the level of poverty over time. At various points, the report forthrightly states that many of its recommendations are not made on the basis of scientific evidence alone, that they also involve the value judgements of panel members. But this recommendation is all judgement and no science. The choice of how rapidly the poverty line should rise over time derives from society's values. Judgements about these values are more properly made by elected officials charged with translating societal values into law rather than in reports issued by scientific bodies. CHOOSING A RANGE FOR THE POVERTY LINE The report's introduction argues correctly that the choice of a poverty threshold is not a scientific one. The panel then concludes that the appropriate range for the poverty line is between $13,700 and $15,900 for a family of four.1 This range is between 14 and 33 percent higher than the comparable current poverty line. In terms of consumption of the three basic needs—food, clothing, and shelter—40 to 55 percent of four-person families consume less than this amount. The report attempts to create an impression that this range lies within the scientific community's consensus about where the poverty line should be drawn. The policy-making community should be aware that there is no consensus within the scientific community. Furthermore, even if there were, it should carry no more weight among policy makers than a consensus among theoretical physicists that they prefer tofu to beef burgers. Choosing a poverty line or a range for that line is a policy maker's job, not the job of a scientific panel. Scientific expertise can inform policy makers' choices. For example, this expertise can be brought to bear on measuring and assessing living conditions at or near alternative poverty lines. Unfortunately, the report provides no information on the level of economic deprivation among persons at any of the poverty levels discussed. MEASURING FAMILY RESOURCES: THE ISSUE OF MEDICAL CARE For measuring family resources, the report recommends that out-of-pocket expenditures for medical care be subtracted from a family's income. This recommendation is troubling. It assumes that all medical care expenditures are 1 The report is vague about why the panel chose to label its range a conclusion instead of a recommendation. However, the distinction is immaterial since there is no scientific basis for recommending or concluding that a particular range is appropriate.
OCR for page 385
Measuring Poverty: A New Approach nondiscretionary. Within the field of economic science, the assumption that all medical care expenses are nondiscretionary runs contrary to three decades of economic research. From the early work of Pauly (1968) and Grossman (1972) to later work by Newhouse (1993) and others, economists have viewed health as an economic good, responsive to both income and price changes. This consumer choice approach has dominated economic analysis of health care and a greatly enhanced analysis of health care expenditures. Although this research does not offer any firm conclusions about how health care should be treated in the context of poverty measurement, its basic premise is at odds with the panel's rationale. The panel's recommendation is based on an approach suggested in a 1985 conference paper by David Ellwood and Larry Summers. In the decade since that paper was presented, there has not been, to my knowledge, a single critical evaluation or discussion of it in any major peer reviewed scientific economics journal. The paper's merits aside, its approach has not undergone the kind of assessment that science requires before a scientific consensus is reached. The report argues that deducting out-of-pocket expenses removes medical care entirely from the calculation of poverty. The argument is not correct, as the following example illustrates. Consider two healthy families—the Smith family and the Jones family. Suppose the Smith family has an income that is $2,000 higher than the Jones's. The Smith family purchases a $3,000 health insurance plan while the Jones family purchases no health insurance. Both families are fortunate enough to have no additional out-of-pocket health expenditures during the year. According to the report's recommended treatment, the Smith family would be poorer than the Jones family. And it would be so only because it chose to spend its higher income on health insurance. The panel also argues that, by excluding medical care from its list of basic goods, its treatment is consistent. However, for two reasons, this argument is less than satisfactory. First, the 15 to 25 percent add-on to the poverty threshold "for other needed expenditures" can be construed as building in an amount for medical care. In fact, the dollar value of this percentage—$1,800 to $3,200—is more than one-half the actuarial value of Medicaid for the noninstitutionalized population and close to the cost of a typical private insurance plan. Second, the panel could have obtained the same range for the poverty threshold by including medical care as a fourth basic commodity and basing the threshold on the 20th instead of the 30th percentile of the consumption distribution. One final point about the panel's treatment of in-kind benefits is in order. Much of the impetus for changing the way in which resources are counted comes from the fact that the current method ignores the value of billions of dollars in noncash benefits for food, housing, and medical care that are spent on low-income families. The reader will be surprised to see that the panel,
OCR for page 385
Measuring Poverty: A New Approach after making adjustments to countable income, concludes that families living near the current poverty line have fewer countable resources than they would have under the current poverty measure. CONCLUSION I dissent because the report's recommendations—to choose three particular commodities upon which to base the calculation of poverty and to exclude other commodities; to establish a normative range of values within which the poverty line should fall; to increase the poverty line over time to account for perceived improvements in the standard of living; and to exclude medical expenses from family resources—are the outcome of highly subjective judgements. These are judgements that do not result from scientific inquiry and, therefore, in my opinion, are improperly placed in this report. I do not believe that this report will be the basis for improving the measurement of poverty because its recommendations are not based on scientific evidence. To my disappointment, the panel has missed an extraordinary opportunity to enlighten and inform government officials about problems of measuring poverty and about the solutions to those problems. REFERENCES Grossman, M. 1972 On the concept of health capital and the demand for health. Journal of Political Economy 80(March/April):223-255. Newhouse, Joseph P. 1972 Free for All? Lessons from the RAND Health Insurance Experiment . Cambridge, Mass.: Harvard University Press. Pauly, M.V. 1968 The economics of moral hazard: Comment. American Economic Review 58:535.