Conference on "Improving the Poverty Measure After 30 Years"
Introductory Remarks

April 16, 1999, Madison, Wisconsin

Constance F. Citro
Committee on National Statistics
National Academy of Sciences

The Basis of the Current Measure

We on this opening panel were asked to consider the basis of the current measure and the motivation for change. I worried about repetition when I was preparing for the session but concluded that some repetition was desirable. After all, we have not changed the current poverty measure in over 30 years-perhaps repeatedly pointing out its flaws will help push us along toward actually revising the measure.

I imagine most of you are familiar with the current poverty measure, which was developed by Mollie Orshansky at the Social Security Administration in the early 1960s. I will just briefly run through the concepts that underlie Mollie's poverty thresholds and her definition of family resources-which are compared to the thresholds to determine poverty status.

Oversimplifying, Mollie's poverty threshold concept was the cost of a minimally adequate diet-appropriate for the size and type of family-times 3 to allow for all other expenditures. The minimum diet costs came from work by U.S. Department of Agriculture analysts, who used data from a 1955 Household Food Consumption Survey on the food-buying patterns of lower income households together with nutritional guidelines to develop the so-called Economy Food Plan. The multiplier of 3 came from the same 1955 survey, which showed that average families spent one-third of their after-tax income on food. The threshold for a two-adult/two-child family that Mollie developed for 1963 was $3,100 dollars. I'm not sure what today's threshold for a two-adult/two-child family is, but whatever the amount, it is the same in purchasing power as the original 1963 threshold. The reason is that the thresholds have been updated over time simply by changes in the Consumer Price Index (CPI).

Mollie's definition of family resources, namely, before-tax gross money income, was about the only one possible at the time in terms of available data. This concept was reasonably appropriate then, given low taxes on the poor and the absence of large in-kind transfer programs. The data source for measuring gross money income was also the only one available-the March Current Population Survey (CPS).

Problems with the Current Measure

The official poverty measure has served well to track broad changes in the extent of the poverty population over time as our economy has experienced periods of growth and recession, so why would we need to change it? Almost from the beginning, the measure has not served well to assess the effects of government policies and programs on poverty among key subgroups of the population, including children, the elderly, and others. And the underlying reason for that has to do with the official measure's concept of poverty. That concept-gross money income to meet a specified level of need for food and all other expenditures-has not proven flexible enough to accommodate the variety of approaches the government has used since the 1960s to support the low-income population nor to recognize that not all income is available for consumption and that disposable income varies for different types of families-for example, workers have expenses that are necessary to earn money, in contrast to retirees.

Here is a list of problems with all three aspects of the current poverty measure: the thresholds that represent the standard of need; the definition of family resources, and the data source for measuring resources. These problems are well known to students of poverty, so I'll just run through them briefly.


  • Out of date with respect to changes in the standard of living.
  • Equivalence scale (adjustment for family size and type) has anomalies.
  • No adjustment for geographic differences in the cost of living.

Family resource definition:

  • Does not include the value of such near-cash benefits as food stamps.
  • Does not subtract payroll and income taxes and other expenses needed to work.
  • Counts child support payments as income both to the payer and the recipient.
  • Does not account for out-of-pocket medical care costs (MOOP).

Data source:

  • Does not use most complete income survey-the Survey of Income and Program Participation.

The National Academy of Sciences (NAS) panel on poverty and family assistance, for which I was study director, in its 1995 report recommended a revised poverty measure that would address these problems. The NAS proposed basing poverty measurement on the concept of money and near-money disposable income that is available to meet the basic needs of all families-food, clothing, shelter, and a little more for other common necessities.

Motivation for Change

There are two main reasons that I see to change the current measure. The primary reason, and here I repeat myself, is that the current measure, by its concept, only partly captures programmatic and societal changes that affect the low-income population. Programs that operate via in-kind benefits or through the tax system are not captured in the current measure, although they would be under the NAS proposed measure. Thus, the current measure will reflect changes in social security benefits and cash benefits under Temporary Assistance for Needy Families (TANF); it will not reflect changes in food stamp benefits, Social Security payroll taxes, the Earned Income Tax Credit (EITC), income taxes, health insurance coverage, child care subsidies, or non-cash benefits under TANF.

Because of the conceptual mismatch between the official poverty measure and the reality of government programs and families' circumstances in the real world, we do not have an accurate picture of the extent and composition of the poverty population at a point in time or over time. From recent research, the official measure and the proposed NAS measure do show broadly similar trends over time-it would be disconcerting if that were not the case, given the dominant effect of the economy on poverty. However, for policy analysis and planning, we need a more finely-tuned measure. The same research shows important differences for population groups between the two measures as the result of particular targeted policies.

For example, there are important differences in child poverty rates over the period 1991-1996 under the current measure and the proposed NAS measure (from Short, Shea, Johnson, Garner, 1998). From 1991-1993, poverty increased for children by 4.1 percent under the current measure and 5.5 percent under the NAS measure-not a big difference. However, from 1993-1996, poverty decreased for children by 9.7 percent for the current measure and by 15.7 percent for the NAS proposed measure. The major factor in the bigger decline for the NAS measure was the EITC, which is not counted in the official measure even though it is a cash benefit, because it operates through the tax system. I wish we had data, by the way, that would permit a similar comparison in the 1970s, when the Food Stamp Program became nationwide in operation.

A second important reason to change the poverty measure-and here I am not repeating myself-is so that we can change it again. I do not make this point facetiously but in all seriousness. Although the implementation of the NAS panel's recommended measure would go far to give us a much more appropriate and informative poverty statistic, no measure is without flaws and a continuing process of review and improvement is needed. On this point, I want to read a section from the NAS panel report.

Although we do not fully understand the reasons, it seems that the "official" standing of the U.S. measure and the fact that it is used to determine eligibility for a number of government assistance programs have made it almost impervious to change. Other statistical measures with equally great political and budgetary consequences, such as the CPI, are regularly reviewed and revised, but even obvious changes-such as defining income in after-tax terms once the Census Bureau had developed reasonably good procedures for estimating income and payroll taxes-have not been made to the poverty measure. Although maintaining a key concept over time is desirable to facilitate analysis of trends, it is dangerous to let a key social indicator become so frozen in place that, when societal conditions change, it can no longer adequately reflect what it was designed to measure. (p. 43)

Two of the panel's recommendations address the process of revising the official poverty measure:

Recommendation 1.3 The U.S. Office of Management and Budget should adopt a revised poverty measure as the official measure for use by the federal government. Appropriate agencies, including the Bureau of the Census and the Bureau of Labor Statistics, should collaborate to produce the new thresholds each year and to implement the revised definition of family resources.

Recommendation 1.4 The Statistical Policy Office of the U.S. Office of Management and Budget should institute a regular review, on a 10-year cycle, of all aspects of the poverty measure: reassessing the procedure for updating the thresholds, the family resource definition, etc. When changes to the measure are implemented on the basis of such a review, concurrent poverty statistics should be run under both the old and the new measures to facilitate the transition.

The NAS panel report also noted that the poverty measure, unlike the CPI or unemployment rate, does not have a clear "home" within the federal government. The Census Bureau publishes the official poverty statistics but has never been empowered to change the measure. OMB issued directives implementing the minor changes to the thresholds that were adopted in 1969 and 1981, but does not itself have research or operational capabilities to implement or revise a new poverty measure. This fragmentation of responsibilities for the poverty measure is perhaps another reason why the measure has been changed in only minor ways since it was first developed.

As we work toward a revised measure, we may want to think about the bureaucratic structure for implementation and further revision. In this regard, one might separate the issue of the level of the thresholds, which is sensitive given the implications for government spending on income-support programs, and the definition of family resources within a given threshold concept. A possible scenario is that the Census Bureau could be empowered to make appropriate changes to the definition and measurement of resources within an agreed-upon concept. As an example, under a money and near-money disposable income concept, if transportation vouchers became widespread for an aging population that is dispersed in suburban areas and needs assistance in reaching doctors and other services, the Census Bureau could measure those vouchers and add them to income appropriately-similar to the treatment of food stamps-without having to go through too many hoops.

To sum up: I believe that first, the need to revise the current official poverty measure is clear. Secondly, we should move as expeditiously as possible toward that goal and toward institutionalizing arrangements so that periodic revision is possible and we are not again in the situation of having a key statistic that no longer serves our needs.