On July 13, 2008, New York City’s poverty rate was 18 percent. Twenty-four hours later it had ballooned to 23 percent. How did more than 400,000 New Yorkers become impoverished overnight? The answer is that Mayor Michael Bloomberg adopted a new and more complex—and, he argued, more accurate—measure of poverty than the one the federal government uses. His action reignited a debate in Washington, D.C., and beyond about how America determines who is poor—a debate that many hope will be settled by the U.S. Congress this year.
Most people who care about measuring poverty—academics, policymakers, nonprofit leaders, and the like—agree that the way the federal government currently determines who is poor and who is not doesn’t work. The so-called “poverty line” was determined in the mid-1960s by calculating the amount of money it costs to buy a basic basket of food and then multiplying that amount by three. Each year the line is updated to account for inflation. (The current poverty line is $10,830 for a single person and $22,050 for a family of four.) If a person lives in a household whose income is less than that amount, he is considered poor. If the household’s income is that amount or more (even by one dollar), he is not poor. The measure does not consider other living costs besides food, and the federal poverty line is the same whether a person lives in New York City or McAlester, Okla.
The federal poverty line is used to determine eligibility and appropriations for all types of federal, state, and local aid, including food stamps, Temporary Assistance for Needy Families (TANF), and Medicaid. How the line is determined has real material implications for low-income families. The poverty line is also the most important way that America measures how well it is treating its most disadvantaged members. A large and growing percentage of people below the poverty line indicates that we are not doing enough. A small and declining percentage of people in poverty tells us that we might be on the right track.
Now is the time to improve our measures of poverty. But the current conversation around Bloomberg’s initiative and other proposals to update how we measure poverty falls into the same trap by reducing the complexity of poverty to a single figure, a line. If our goal is to achieve a better measure of well-being in order to diagnose human needs and design effective solutions, no line—no matter how thoughtful or sound—will do.
Given the dynamic nature of poverty, social service organizations and policymakers require more context, more nuance, and, quite simply, more data. The data we need already exist. The problem is that data are scattered across dozens of government and nonprofit organizations and require experts to access and interpret. We are living in the information age—it’s time that the socially minded community asked for more of it. It is time to move beyond the poverty line.
DEFINING THE POVERTY LINE
The current federal poverty line was created in 1964 by Mollie Orshansky, an economist working at the U.S. Social Security Administration.1 Tasked with setting a threshold for what it meant to be poor, she started by analyzing the cost of one of life’s basic necessities: food. Orshansky’s first step was to determine the cost of feeding a family on the “economy food plan,” the cheapest of the four food plans deemed nutritionally adequate by the U.S. Department of Agriculture (USDA). She then estimated that the average family spent one-third of its budget on food. The poverty threshold, then, could be set by multiplying the cost of the most basic food plan by three.
Because of the timing of Orshansky’s calculations—President Lyndon Johnson launched his War on Poverty in 1965—the federal government quickly adopted her threshold as the basis for policy. And so it was done: The poverty line was born. Except for annual adjustments for inflation, the poverty line has not been touched since.
For decades, critics have complained about the limits of Orshansky’s measure. Do we determine a family’s income by using pre- or post-tax earnings? Do we include food stamps and other transfers? What about the cost of other necessities like shelter, utilities, and transportation? Why have a single poverty line for the entire country when the cost of living varies widely across the nation?
After three decades of frustration and cries for improvement, in 1992 Congress asked the National Academy of Sciences (NAS) to organize a group of academics and policy thinkers on the issue. The Panel on Poverty and Family Assistance designed a better way to measure poverty.2 The group’s detailed recommendations were published in Measuring Poverty: A New Approach and sought to compensate for the shortcomings of Orshansky’s threshold by accounting for the full consumption needs of families, a more accurate measure of household income, and regional variation in cost of living.
The proposal was brilliant, so brilliant that the NAS panel’s recommendations were largely ignored for the next decade and a half. Why? Economist Rebecca Blank, President Barack Obama’s undersecretary of commerce for economic affairs at the U.S. Department of Commerce, gave a pragmatic answer when she wrote in the Los Angeles Times: “Unfortunately, no president (Democrat or Republican) has wanted to touch this political hot potato. If a new measure shows higher poverty, the president looks bad, but if a new measure shows lower poverty, he’ll be accused of dismissing the problem.”3
Without revision, the poverty line has become increasingly useless as a tool to target and evaluate public policy, especially at the state and local level. Federal, state, and local governments have largely stopped using the poverty line to determine eligibility for social programs. Program eligibility is now frequently set at 125 percent, 150 percent, or even 200 percent of the federal poverty line.
With the federal government unwilling to revise the way it measures poverty, the stage was set for Bloomberg to redraw the line largely on the basis of the NAS panel’s recommendations. In announcing the new poverty measure, Bloomberg said: “If we are serious about fighting poverty, we also have to start getting serious about accurately measuring poverty. Since the mid-’60s the economy has vastly changed. So has society and so have government benefits, but the poverty formula hasn’t adjusted in response. We can’t devise effective strategies for tackling poverty until we understand its full dimensions.”4
And what has come of Bloomberg’s new standard, which resulted in designating nearly half a million more New Yorkers as poor? Not much, according to Frances Fox Piven, distinguished professor of sociology and political science at the City University of New York Graduate Center. “Mayor Bloomberg proposed a more realistic measure of poverty that took some account of higher living costs in New York City. … Bloomberg was right: The official poverty line in the United States is unreal,” Piven is quoted as saying in a Gotham Gazette article. “It does not take account of the actual costs of basic subsistence. … If we thought a new measure would mean more generous policies, we were wrong.”5
Some nonprofit leaders have been similarly skeptical. “The new measures being floated, by New York for instance, definitely make more sense, but they still fail to come near to really understanding need deeply enough to then provide viable policy solutions to address poverty,” says Sondra Youdelman, executive director of Community Voices Heard, a New York City-based grassroots organization.
WAYS OF MEASURING POVERTY
Even if the new poverty line being used by New York City is inadequate, there must be a line that would adequately measure poverty. Or is there? Some scholars, advocates, and policymakers argue that the main problem with the American system of measuring poverty is that a line is an absolute measure.6
Absolute measures have clear shortcomings. It is difficult to establish an objective minimum level of goods necessary for an individual or a family. This is especially difficult in a country as heterogeneous as the United States, where there are large regional variations in the cost of goods and services. And the definition of what is necessary changes significantly over time. Food costs have declined in the United States, but the costs of other goods and services, such as health care, have increased. Thus, a measure based on the cost of food will gradually underestimate the actual minimum cost of living over time. And what other goods and services should be factored in? For example, in some cities (like Los Angeles) it is essential to have a car, whereas in other cities (like New York) most people use public transportation.
Measuring poverty in an absolute way also poses challenges because the purchasing power of a dollar changes not only over time, but also across space. One of the primary measures of poverty used in the developing world is also an absolute measure: the dollar-per-day standard, which is often cited by such organizations as the World Bank. Using this measure, the poverty rate in developing countries is determined by the percentage of people living on less than one dollar a day, or, in slightly more advanced developing countries, two dollars a day. Although appealing in its simplicity, the dollar-per-day measure has come under significant criticism because, among other things, it does not account for the fact that one dollar means different things in different places.7 Not only does the cost of basic necessities vary dramatically from one country to another, but the types of services that governments provide (such as health care and education) can also be quite different.
Alternative ways of developing a poverty line do exist. Chief among these is a relative measure of poverty that is based on a percentage of a country or region’s personal income or consumption. For example, the European Union defines individuals living in households whose income is less than 60 percent of a nation’s median income to be living in poverty. Proponents of relative poverty measures argue that they are more useful because they account for changing levels of inequality within a society. As the overall wealth of a society increases, so would the poverty level if lower-income households don’t keep up. Relative measures of poverty shift the definition of poverty from being about the material resources needed for survival to being about having less than others in a society.
One problem with the relative approach is that changes in the relative poverty rate may not fully capture the real changes in the material well-being of “the poor.” Ireland provides an illustrative example. When the Celtic Tiger took off economically in the late 1990s, paradoxically, so did the poverty rate. As the median income increased, many families who didn’t experience a loss in income were now found to be in poverty, although their absolute state remained unchanged.8 The current economic downturn may have the reverse effect—as median income falls, so too may the number of families in poverty, even though they still face material hardship. The strength of the relative measure of poverty—its consideration of the overall income distribution—is also one of its most important weaknesses. If policymakers are primarily interested in reducing the material hardships associated with poverty, relative poverty measures may not capture the information the policymakers are seeking.
In addition to the absolute and relative measures of poverty, some suggest using a subjective approach to the poverty measure. The absolute and relative poverty measures depend on an external agency determining the threshold for being poor. The best judge of whether someone is poor or not, however, may be the poor person himself. Subjective poverty measures rely on a person’s own reporting of whether or not he has an income that is adequate to meet his needs.9 Although a useful gauge of aggregate perceived well-being, subjective measures of poverty are often impractical when it comes to providing government support for the poor. If a person is able to determine for himself whether he is poor and deserves government help, there is an obvious incentive for misreporting.
WHAT A POVERTY LINE CAN’T TELL YOU
No poverty line, regardless of how well conceived or how well intentioned, can provide the information that nonprofit leaders and policymakers need to better serve their community. A line cannot provide information about the depth or intensity of deprivation. It cannot tell us about the duration of poverty. It does not provide direct information about actual deprivation, such as homelessness or hunger. In addition, a poverty line does not provide any information about the correlates or causes of poverty. Finally, a simple line limits our understanding of poverty to the economic realms, ignoring the social and political dimensions of exclusion and marginalization.
A poverty line does not provide information about the depth or intensity of poverty. Those people labeled as poor could all be concentrated just below the poverty line, or they could be concentrated near zero income. If most poor people are concentrated just below the poverty line, they are likely working and poor. To help those people, one would concentrate on work support programs, such as the Earned Income Tax Credit. A different set of policies and programs would be pursued if most of the people had little or no income.
A poverty line does not provide information about how long people have been poor. Episodic poverty is often precipitated by the loss of a job, a sudden illness, or another unexpected crisis. Chronic poverty may be the result of physical or mental disabilities, generations of cumulative deprivation, poor education, or lack of jobs. Episodic poverty might require short-term assistance, whereas chronic poverty could require job training and more holistic supports.
Although income is highly correlated with a person’s material circumstances, it provides incomplete information about actual deprivation. Is the person unable to afford rent and thus living on the street? Is the family unable to purchase enough food so the adults in the household are going hungry? Are they making ends meet by not purchasing badly needed medications? In each of these instances, different programs would be needed to alleviate the problem.
In addition, the poverty line does not provide important information about a person’s health, something that is intimately connected to poverty. People who are poor are more likely to have health problems and less likely to receive adequate care for those problems. At the same time, people who have health problems are more likely to be poor because of their inability to participate fully in the labor market. Poverty lines are unable to account for this important correlate of poverty, which is problematic because health interventions can play an important role in reducing poverty.
Last, knowing that a family is below the poverty line does not tell us the extent to which they are otherwise integrated into society. There is something qualitatively different about a family with an annual income of $15,000 where the father and mother have a high school education, vote, have health care, and live in a clean and modest home in a safe neighborhood, and another family at the same income level that does not have any of these things. The former is a working family that is making ends meet; the latter is the makings of an underclass. Social exclusion—a term commonly used in Europe that has failed to catch on in the United States—is the degree to which an individual (or group) is detached from the larger society.10 These covariates of poverty help to round out the picture of exclusion and give policymakers better information with which to design solutions.
THE POVERTY LINE IN PRACTICE
Most people would agree that New York City and Los Angeles are very different cities with distinct social problems, yet the two cities have virtually identical federal poverty rates—18.6 percent and 18.9 percent, respectively. The different social and economic characteristics of these two cities, however, indicate that the causes of poverty are likely to be quite different. It is also likely that the solutions required to alleviate or reduce poverty in each city need to be equally distinct.
One of the most striking differences between the nation’s two largest cities is their racial and ethnic composition. Whereas 25.1 percent of New Yorkers are black, only 9.9 percent of Angelenos are black. Los Angeles, on the other hand, has a much larger Hispanic population than New York City—48.4 percent compared with 27.5 percent. In addition, the percentage of people who speak only English at home is 52.2 percent in New York City, compared with just 40.2 percent in Los Angeles.
The large difference in food stamp participation rates between the two cities is also glaring. Even though the official poverty rates are the same, 13.9 percent of New Yorkers receive food stamps, compared with only 5.4 percent of people living in Los Angeles. The different participation rates in this program indicate that a program that works well in New York City might not have the same impact in Los Angeles because of their distinct demographics.
It is no surprise that only 11.2 percent of Los Angeles residents use public transportation to get to work, compared with 54.6 percent of New Yorkers who use public transportation to commute. New York City’s public transit system is ubiquitous and provides low-cost access to jobs throughout the city. One could infer from these data that an investment to improve Los Angeles’ public transit might help move people out of poverty by making it easier for them to get and hold jobs, whereas a similar intervention in New York City would probably have limited impact.
TOWARD A SOLUTION
To address the shortcomings of the poverty line, the Obama administration announced in March that it will begin publishing a “supplemental poverty measure” (SPM) that is similar to the one used by Bloomberg and is based largely on the recommendations of the NAS panel. The SPM will be calculated by determining the consumption spending of an average household at the 33rd percentile of income— well above extreme deprivation, but below the national median. After determining what this household spends on basic consumption—such as food, housing, and medical care—the U.S. Department of Commerce will determine what a family needs to subsist at a basic level. This new line will be adjusted regionally by housing cost.
The Obama administration deserves praise for trying to craft a better measure of poverty. But the plan has two fundamental flaws. First, the SPM has no teeth. According to Blank, who is leading the administration’s efforts, the supplemental measure will not replace the existing line when it comes to determining who is eligible for poverty programs or how poverty funding is allocated. Instead, the line will be an additional macroeconomic indicator that will provide a different way to assess the well-being of low-income households in America.11
The second problem is that the regional variation in the poverty line is based solely on the difference in the cost of housing, without consideration for differences in other important costs. Blank says that the reason is that the only good data created annually at the city level are the housing cost estimates derived from the U.S. Census Bureau’s American Community Survey. But why not use this as an opportunity to improve and expand data collection? There is no reason the American Community Survey cannot be expanded to include a consumer expenditure module. This would permit the calculation of a unique poverty line for every metropolitan area that is based directly on the spending needs, patterns, and capabilities of real households in that community.
Now is the time to develop better data on poverty’s causes and consequences. The production of such data should adhere to three principles: It should be local, comprehensive, and accessible. National-level statistics on everything from poverty to educational attainment are readily available on an annual basis from the American Community Survey. State-level statistics are also available for most measures, but often not as frequent or detailed as national snapshots. Local-level statistics, however, are sorely lacking. Government surveys should provide statistically representative samples of all major metropolitan areas when possible and, for the most basic and vital indicators, be sure to capture a representative sample of smaller cities, towns, and rural areas on a rotating basis.
Data on poverty, its determinants, and its consequences also need to be comprehensive. The census provides a decent way to access its data for users who want to generate tables of, say, educational attainment by race and income. The Centers for Disease Control and Prevention provide excellent measures of birthrates by age. But it is virtually impossible to marry these data to find out the birthrate of African Americans (race), in their early 20s (age), who are high school graduates (educational attainment). That is why universities and foundations spend millions of dollars annually to conduct separate surveys that capture all four variables. Government agencies should collaborate with nonprofit, foundation, and social services leaders to ensure that data collection is comprehensive.
Finally, data on poverty and its covariates need to be accessible to non-experts. Federal agencies sort data to suit their needs and address their policy research questions. The USDA, for example, publishes national rates of food insecurity by a host of characteristics including income, race, and household type. It also publishes rates of food insecurity by state. What if a nonprofit is interested in food insecurity by race in Alabama? The data exist; they are sitting in a micro-data file that is supposedly accessible to the public. Using these data, however, requires expensive software and sophisticated technical knowledge and programming skills to manipulate.
As David Dodge, who works for Right to the City in New York, points out: “When we need really local data, we have to rely on the Furman Center at New York University, which does a lot of work to collect local information. They put out reports that have specific information about the neighborhoods in New York. But it shouldn’t be a nonprofit’s responsibility to collect and analyze that information. It should be the government’s role.” Government statisticians should make the relevant data available through an easy-to-use point-andclick interface that allows the user to design and generate his own tables, customized to the lowest level of geography possible.
Congress is expected to hold hearings on the poverty line this fall. Our recommendation to Congress and the Obama administration is simple: Make it count. Use the retooling of the poverty line as an opportunity to change the way data are collected on low-income households and to improve the way that policy and programs use the poverty line in determining eligibility and allocating funding. The ball is in motion. Now it’s time to make sure that our efforts make concrete improvements in the lives of those struggling to make ends meet.
Measuring poverty accurately is a must, but alone it is not enough. We need to expand our understanding of poverty. We must move beyond the line.
Rourke L . O’Brien is a National Science Foundation graduate research fellow and doctoral student in sociology and social policy at Princeton University. He previously served as a policy analyst at the New America Foundation, where he continues as a nonresident research fellow in its asset-building program.
David S. Pedull a is a doctoral student in sociology and social policy at Princeton University. He previously served as a fellow with the Congressional Hunger Center and Office of the Mayor in New York City, and as a research associate at the Brennan Center for Justice at New York University School of Law.