This is the second post in a series of three focused on the measurement of U.S. poverty
As I described in my last post, virtually nobody is very happy with the official poverty measure in the United States. Commentators on the far right such as Robert Rector are concerned that too many people are counted as poor. As much as Rector dislikes the current poverty lines, he is even more critical of the Census Bureau’s Supplemental Poverty Measure (SPM). The SPM is intended to augment, but not replace, the existing poverty measure. Here is a very brief summary of how the SPM is computed:
- Threshold: The SPM poverty threshold is set on the basis of expenditures on a set of commodities that all families must purchase: food, shelter, clothing and utilities (FSCU). The recommended reference threshold is at the 33rd percentile in the distribution of FSCU in the population of families with two children over the previous five years.
- Adjust for taxes: Subtract taxes or add in tax credits estimated to apply to household types.
- Adjust for differences in housing prices: Using recent data on costs of housing across metropolitan statistical areas.
- Add in-kind benefits: the estimated dollar value of food stamps and other benefits would be added to family incomes, just as cash benefits are currently included.
- Subtract child care and out-of-pocket medical care: These medical out-of-pocket spending would be based on the estimated amount of premiums, copays, and deductibles different families are likely to pay.
The SPM would substantially change who is counted as poor, an issue that I may return to in a future post. Below is a table from a Census report that compares the SPM, the official poverty rate, and a third definition. Under the SPM a higher share of people would be counted as poor 15.8 versus 14.5 percent, but the composition changes substantially with fewer children and more adults.
Are the changes outlined in the SPM warranted? More generally, what criteria should we use to decide on a poverty threshold?
The rationale for setting a poverty line has always been twofold. First, to reflect a minimally adequate cost of living in the United States at a point in time, and second, to measure the effect of anti-poverty programs on the material resources of low-income households.
Relatively little data was available about the cost of living when a mid-level economist working for the federal government named Mollie Orshansky was tasked with developing a poverty threshold in 1963. This measure – it was hoped – could enable policymakers to gain a deeper understanding of the efficacy of large-scale national programs aimed at ending poverty. Orshansky began by looking at estimates of different budgets for nutritionally adequate diets and ended up focusing on the “economy budget.” This amount was multiplied by a factor of three to account for housing and other costs, and ultimately formed the basis for the Orshansky index. By May 1965, this measure was adopted as the Office of Economic Opportunity as the official measurement of poverty. Adjusted for inflation through the Consumer Price Index, this measure has basically remained unchanged since that time.
In 1995, an expert report commissioned by the National Academy of Sciences issued recommendations to change the official poverty line. Many of the recommendations outlined in this report form the basis of the SPM. Here is what the report authors say about the official poverty line: “From the beginning, the poverty measure had weaknesses, and they have become more apparent and consequential because of far-reaching changes in the U.S. society and economy and in government policies.”
The 1995 panel identifies three kinds of weaknesses with the official poverty line. First, the official poverty line does not sufficiently adjust for differences in the cost of living across households that are affected by region, size of the household, work-status of the adults, and medical expenses. Second, the poverty line does not capture the effect that government transfers have on the disposable income of families. Food stamps is the classic example – a very common program that has a substantial impact on how much families can consume, but which is not measured under the current threshold.
The third point made by the 1995 panel is essentially an argument about changes in how society should evaluate poverty:
“Changes in the standard of living call into question the merits of continuing to use the values of the original thresholds updated only for inflation. Historical evidence suggests that poverty thresholds— including those developed according to ‘expert’ notions of minimum needs— follow trends in overall consumption levels. Because of rising living standards in the United States, most approaches for developing poverty thresholds (including the original one) would produce higher thresholds today than the current ones.”
In other words, the authors of the report make the case that the standard of living that underlines our assessment of poverty has changed since 1965. The income needed to support a minimal standard of living in 1995 should reflect the “rising standard of living” from the intervening 30 years, and we should broaden our definition of poverty to accommodate this change.
A value free measure of poverty?
This highlights a central tension in the poverty measurement debate: on the one hand, experts have appealed to technocratic principles that are overtly aimed at providing more accurate adjustments for relative uses of resources needed to maintain similar standards of living. On the other hand, the 1995 report (and those that have followed) make an argument about the standard of living that should be used to measure poverty.
In the official dissent to the 1995 report, John F. Cogan states:
“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 judgments. These are judgments that do not result from scientific inquiry and, therefore, in my opinion, are improperly placed in this report.”
If Cogan is right, the 1995 panel’s recommendations are flawed but could be rescued by returning to more scientific standards of poverty. But what standards result from scientific inquiry alone, and can we attain a value-free measure of poverty? There are really two issues here. One is about whether there is a value-free conception of poverty. The second is whether we can use scientific criteria to define who fits within that conception. My answers are “no” and “yes.” If we are able to define a target standard of living, the tools of science can enable us to measure how certain people meet or fail to meet those standards. The first question is difficult to resolve because (1) there is no broad social consensus on the meaning of poverty, the way that society evaluates poverty is highly dependent on context and what reference point they use and (2) soliciting the perspectives of experts is likely to raise accusations that the experts have a “pro-poor” bias and are going to pick a criteria that gives more salience to the plight of low-income people.
If we could resolve the first question, we could make substantial progress on the second. We have an impressive armamentarium of social surveys that measure the living standards of the poor (although these too can be improved). The one unresolved issue is whether we should focus on the resources that sustain a given standard of living, the level of welfare that people can enjoy, or something even more abstract (such as the minimum set of capabilities that we should aspire to in our society). More on this next week.