It’s Time for a Better Poverty Measure

Mark Greenberg shows why the federal poverty measure is flawed and how it can be updated.

A homeless man is seen in Fresno, CA, which has one of the highest concentrations of poverty in the nation. An improved poverty measure would provide a more accurate picture of how many people are falling into poverty during the recession and who they are. (AP/Rich Pedroncelli)
A homeless man is seen in Fresno, CA, which has one of the highest concentrations of poverty in the nation. An improved poverty measure would provide a more accurate picture of how many people are falling into poverty during the recession and who they are. (AP/Rich Pedroncelli)

Download this memo (pdf)

Introduction and summary

The federal poverty measure shapes our understanding of how many people are in poverty, who is in poverty, and how much poverty goes up or down when economic conditions and policies change. But the official measure is deeply flawed. The dollar figures used to determine if families are in poverty are low and in many ways arbitrary. The rules don’t consider some resources, such as tax credits and food stamps, and some key family expenses that determine a family’s available income. As a result, the poverty measure often doesn’t show the impacts of important policies that are intended to improve the economic well-being of families. It needs to be updated and improved.

No alternative is perfect, but it’s possible to develop a substantially better measure than the current one. Pending legislation, the Measuring American Poverty Act, would take a significant step forward by directing the Census Bureau and Bureau of Labor Statistics to adopt a new measure drawing recommendations from the National Academy of Sciences, or NAS. The legislation would also charge a new NAS panel with developing a “decent living standard” measure that would likely be considerably higher than even an improved poverty measure. Moreover, the Obama administration could adopt a new poverty measure and “decent living standard” administratively without awaiting legislation.

Now is the logical time for the administration or Congress to act to improve the poverty measure. Doing so would provide a more accurate picture of how many people are falling into poverty during the recession and who they are, and it would ensure that the administration’s policies and performance can be gauged against a consistent measure that reflects the impacts of a broad range of policies.

How the current poverty measure is calculated and why it’s deficient

A poverty measure typically has two components: the thresholds used to determine whether an individual or family’s income is below the poverty line, and the resource-counting rules used in making that determination.

Current poverty thresholds were established in the 1960s. At that time, research indicated that the typical family spent about one-third of its income on food, so poverty thresholds were derived by multiplying a low-cost food budget by three. Since then the thresholds have only been adjusted for inflation. A family is considered poor if its pre-tax cash income falls below the applicable poverty threshold.

The current measure has received many criticisms, including the following:

  • The thresholds are low. Living costs and standards have changed in many ways since the 1960s. Food now comprises only about one-seventh of an average family’s expenditures. The poverty line represented nearly 50 percent of median income for a family of four in the early 1960s, but now represents only about 28 percent of median income. So the level at which a family is considered poor has fallen further and further outside the mainstream.
  • The thresholds are essentially arbitrary because they simply represent a number calculated more than 40 years ago and then adjusted for inflation, and they no longer represent anything in relation to family incomes or costs.
  • The resource-counting rules both understate and overstate resources. They fail to reflect the effects of policies such as refundable tax credits, near-cash benefits such as Supplemental Nutrition Assistance Program (formerly food stamps) or subsidized housing assistance. At the same time, they also do not consider the impact for family budgets of tax liabilities, work expenses, or health care costs.
  • The rules make no adjustment for geographical variation despite the large variations in costs across areas and regions of the country.

Conservatives sometimes assert that the poverty measure is flawed because it fails to count available resources. But in fact, the measure both undercounts and overcounts resources. Moreover, because the rules don’t reflect the impacts of tax policies, noncash benefits, child care assistance, and assistance with medical costs, the current measure doesn’t show a reduction in poverty when such policies are expanded or an increase in poverty when they are contracted.

The National Academy of Sciences recommendations

In the early 1990s, Congress provided funding for a panel of the National Academy of Sciences to examine the poverty measure. In its 1995 report, “Measuring Poverty: A New Approach,” the NAS panel recommended a new measure, seeking to address each of the principal concerns with the current poverty measure.

The NAS panel urged that there should be a logical relationship between how poverty thresholds are constructed and resource-counting rules. Any effort should start by deciding what the threshold is trying to measure and establish resource-counting rules consistent with it.

The panel recommended that a poverty threshold be an amount of money intended to meet a set of basic needs: food, clothing, shelter, and a little more. Accordingly, the panel recommended that cash and noncash resources available to meet food, clothing, and shelter should be counted. At the same time, resources not available to meet those needs because, for instance, they were going to pay taxes or meet health care or work-related expenses should not be counted.

Specifically, the NAS recommended that:

  • Poverty thresholds should be calculated by using current data about how much families actually spend on food, clothing, and shelter, with an adjustment for “a little more.” Operationally, the NAS panel recommended using an amount somewhere between the 30th to 35th percentile of family costs as measured in the Consumer Expenditure Survey, along with a multiplier of 15 to 25 percent for other necessary expenses.
  • Thresholds should vary geographically to reflect variations in the costs of meeting the needs in the thresholds.
  • All resources available to meet the needs in the thresholds should be counted, including tax credits, food stamps, the value of subsidized housing, and other benefits that can meet those needs.
  • Funds that are not available to meet the needs in the thresholds—because they are used to meet tax liabilities, pay child support, or pay out-of-pocket medical costs or work-related expenses—should not be counted as resources.

How would an NAS measure affect the poverty rate and who’s counted as poor?

As discussed below, it’s impossible to state the precise effects on poverty rates and numbers if the NAS recommendations were adopted, because in some cases the panel specified a range of acceptable approaches, and because Congress or the administration may wish to adopt an approach drawn from but not identical to the NAS recommendations—as would be the case in the Measuring American Poverty Act.

Still, several things seem clear, both from experimental work done by the Census Bureau and Bureau of Labor Statistics, and from the experience of New York City, which has recently implemented a local poverty measure based on the NAS recommendations:

  • Poverty thresholds would be higher than under the current measure. For example, in 2007, the official poverty threshold for a two-parent, two-child family was $21,027. For that year, according to the Census Bureau, an NAS-style threshold would have been in the range of $23,465 to $27,744, depending on how home mortgage principal and medical costs were treated. In New York City, the threshold for a family of four calculated under an NAS approach was $26,138 in 2006, as compared with the official threshold of $20,444.
  • Poverty rates would be higher. In 2007 the official rate was 12.5 percent, but the census experimental work suggested that the poverty rate would have been in the range of 15.1 to 16 percent under the approaches most similar to the NAS recommendations. Under New York City’s calculations, the poverty rate for the city would have been 23 percent in 2006 under an NAS approach, versus 18 percent under the official measure.
  • Elderly poverty would go up. How much it would go up would particularly depend on how households without mortgage costs were treated, but it seems clear that elderly poverty would rise, both because thresholds were increased and because medical expenses were given consideration.
  • Extreme poverty—having an income below 50 percent of the poverty line—would go down, because the measure would count noncash benefits such as SNAP and housing assistance.
  • Immigrant poverty would likely go up because of the higher thresholds and counting of work expenses, and because immigrant households are less likely to participate in some of the counted benefits programs than are other comparably poor households.
  • Poverty would likely go up in high-cost urban areas relative to other areas, all else being equal, though it doesn’t follow that poverty rates would necessarily go down in other areas, given the NAS thresholds and resource-counting rules.

Strengths, weaknesses, and issues in the NAS approach

An NAS approach would directly address many criticisms of the current poverty measure by:

  • Applying thresholds that actually reflect the costs families incur to meet a set of basic needs.
  • Ensuring a logical relationship between the thresholds and resource-counting rules.
  • Using resource rules that both better reflect family resources and expenses such as health care, work-related costs, and child support paid, and that do a far better job of showing the effects of key policies.
  • Providing for geographic variation in the thresholds to reflect variations in actual costs.

Each of these is important, but it’s particularly worth emphasizing how valuable it would be to have a measure that better reflects effects of policies. For example, the American Reinvestment and Recovery Act, or ARRA, contains a number of provisions intended to help poor and vulnerable groups, but key parts of that help will not show up in the current measure of poverty, because the expansion of tax credits—the Earned Income Tax Credit, Making Work Pay, and Child Tax Credits—as well as SNAP benefits and child care assistance under ARRA will all have no effect on the current measure. Similarly, if Congress allows the tax credits to expire or doesn’t sustain the law’s child care funding, there will be no effect on measured poverty. Adopting an NAS-type approach would fix this problem.

However, the overall conceptual approach and specific details of the NAS approach also face a number of criticisms. Among them:

  • The NAS recommends that a poverty measure should measure economic deprivation by assessing whether households can afford to meet a set of basic needs. But international comparisons and many other developed countries use a “relative” measure of poverty based on the share of families below 50 or 60 percent of median income, on the premise that in a developed society, measuring the number of families far from the median provides a better measure of whether families are outside of the social mainstream.
  • The NAS approach has also been criticized for “aiming too low,” because a body of work using self-sufficiency standards, basic living budgets, and family budgets often concludes that the amount of income a family needs for a reasonably decent life or similar formulation may be twice the current poverty line or higher.
  • Thresholds that only reflect food, clothing, shelter, and “a little more” do not adequately reflect the developmental needs of children.
  • Even for thresholds that seek to reflect food, clothing, and shelter, it’s not clear why using expenses in the range of the 30th to 35th percentile of what all families spend is the best measure to use.
  • Because health care and work-related costs are not included in the thresholds, the measure can be easily misunderstood as suggesting they’re not important, and by only subtracting actual expenses, the measure provides no recognition that some families have low or no expenses because they are going without needed health or child care.
  • In changing the relative poverty status of particular groups—for example, the elderly, people who live in high-cost and low-cost areas—there are questions as to whether the results of new measures would or should be consistent with other research about hardships and well-being for these groups.

Another concern relates to any substantial change in the poverty measure. Currently at least 82 federal or federally assisted programs use information about numbers of people in poverty in some way in formulas for allocating funds to states or localities, or use some percentage of poverty in calculating benefits eligibility. Any shift in the official measure could create “winners” and “losers” in states and could have disruptive effects on how benefits eligibility is determined.

The Measuring American Poverty, or MAP, Act discussed below has a number of provisions intended to build on the NAS approach while seeking to address many of these criticisms.

The Measuring American Poverty Act’s approach

The Measuring American Poverty Act, introduced by Rep. Jim McDermott (D-WA) in the House and Sen. Chris Dodd (D-CT) in the Senate, would direct the Census Bureau and Bureau of Labor Statistics to adopt a “modern” measure of poverty drawing from the recommendations of the NAS. Among the bill’s key provisions:

  • Thresholds: The Census Bureau and BLS would be required to adopt thresholds along the lines recommended by the NAS. The bill would also provide for geographic variation, and provide authority to further develop the thresholds to better reflect the needs of children, including young children. The bill would provide for lower thresholds for households owning their homes free and clear, and authorize additional threshold development for other subgroups if reliable data indicated substantial variation in the amounts of money needed by the subgroups to purchase similar quality shelter.
  • Resources: The bill would adopt the NAS approach of counting tax credits, noncash benefits such as food stamps, and housing subsidies if they are available to households to meet the needs in the thresholds. At the same time, per the NAS, the bill would provide for subtracting expenditures for health care, necessary work-related expenses, and child support paid.
  • The historical measure: The bill would treat the current official poverty measure as the “historical” measure, and require that calculation and reporting of poverty rates should be done for both the modern and historical measure.
  • Use of the new measure: The bill would specify that adoption of the modern measure would have no automatic effects on program funding formulas or eligibility rules that currently use the official poverty measure. Instead, Congress could over time make whatever adjustments it considered appropriate on a program-by-program basis.
  • Decent Living Standards and Medical Care Risk Measure: The bill would direct that new National Academy of Sciences panels make recommendations for Decent Living Standards and Medical Care Risk measures. The Decent Living Standard would be defined as “the amount of annual income that would allow an individual to live at a safe and decent, but modest, standard of living,” that is, an amount intended to be above that of the poverty thresholds. The Medical Care Risk measure would calculate the extent to which individuals are at risk of being unable to afford needed medical treatment, services, goods, and care, taking into account both uninsured and underinsured statuses.
  • Calculation of relative measure: While the bill would not mandate reporting of relative poverty measures using percentages of median income, it would require that public online tools be made available to allow members of the public to calculate poverty using alternative approaches, including calculations based on 50 and 60 percent of median income.

Evaluating the MAP Act approach

The MAP Act recognizes the value and importance of moving to an NAS approach. At the same time, it addresses a range of concerns that have been raised. In particular:

  • In addition to establishing a “modern” poverty measure the bill would lay the groundwork for developing a Decent Living Standard measure. This is intended to recognize that while both research and public opinion tell us that a family needs far more than to exceed the current poverty line in order to have a “reasonably” decent life, it would not be feasible or desirable to have a new poverty line that is twice as high (or higher) than the current one. Over time, a Decent Living Standard recognized in federal law could become an important vehicle for analyzing and talking about the need to increase the number of families that can attain a decent living standard, as well as reducing the number of families in poverty.
  • While the bill does not adopt a relative measure, it would make it far easier to calculate such a measure and to foster discussion over time about the strengths and limits of such measures.
  • The bill addresses several concerns about NAS thresholds by directing further threshold development to better reflect the needs of children and by recognizing the distinct status of households that don’t face substantial monthly housing costs because they own their homes free and clear.
  • The bill ensures that there would be no immediate effects on existing funding or eligibility rules by specifying no automatic effects on program funding formulas and benefits eligibility. Over time there may be good reason to adjust funding formulas and eligibility rules, but there are many reasons why this should happen after additional research and experience and on a program-by-program basis.

Moving forward

Either Congress or the administration could create a new poverty measure. The current measure is contained in a directive of the Office of Management and Budget, and the administration could change the measure without congressional action. In many respects the best result would be administration action, so that the measure could be developed and continually refined without locking in the detailed rules contained in parts of the MAP Act. Still, the introduction of the MAP Act is an important step forward in showing how the administration or Congress can build on the recommendations of the National Academy of Sciences and the subsequent learning and experience to develop a significantly better poverty measure and lay the groundwork for a Decent Living Standard.

Mark H. Greenberg is a Senior Fellow at American Progress.

Download this memo (pdf)

Also see:

The positions of American Progress, and our policy experts, are independent, and the findings and conclusions presented are those of American Progress alone. A full list of supporters is available here. American Progress would like to acknowledge the many generous supporters who make our work possible.