Back to previous page

Is the poverty rate even higher than we think?

By ,

Our poverty rate -- which is now at record highs -- was originally developed to show how many people lacked enough money for food and basic sustenance. But critics on both right and left contend that the official rate uses an outdated measure of poverty, painting a distorted economic portrait that affects everything from financial aid to Medicaid benefits. In response, the government has developed alternative ways to quantify the poor. And one approach that the Census Bureau has embraced under Obama seems likely to show that the U.S. poverty is even higher than the official rate.


The Census is experimenting with an alternative poverty threshold based on 1995 recommendations from the National Academy of Science. It will release these new figures in late October, and the guidelines they’re relying on suggest that we’re actually underestimating how poor Americans are.

In 1995, the NAS was commissioned by Congress to find a more accurate way to define poverty than the original measure developed in the 1960s and used ever since. Back then, the Johnson administration decided that a family of three or more was considered below the poverty threshold if they spent more than a third of their income on food. Inflation was eventually factored in as well. That’

The measure has barely changed over the past half century, even as the economic pressures on Americans have shifted dramatically, along with the way the government has responded to such pressures. Food stamps aren’t factored in, for instance. But neither are out-of-pocket medical expenses, transportation, or child-care costs part of the equation.

The NAS panel agreed. It recommended an alternative measure that factored in the benefits of public assistance, as well as cost of payroll taxes, transportation, and health-care expenses and the impact of geographic differences. And things start to look different when you start to factor all this in. Here’s a comparison between the official poverty rate and two experimental measures of poverty based on the NAS recommendations, which the Census publicly posts:


The yellow line (NAS2) shows that poverty is significantly higher than the official rate. But the red line (NAS1) is mostly lower. What’s responsible for the difference? Both experimental rates factor in the benefits of public assistance, as well the cost of out-of-pocket medical expenses, work transportation, etc. But the lower rate is adjusted to the Consumer Price Index, while the higher rate is based on the Consumer Expenditure Survey--an average for actual consumer spending on food, clothing, utilities, and shelter by those at the 33rd percentile of income.

The NAS itself endorsed the latter approach, arguing that the inflation index “ignores changes in living standards over time.” That is, our needs necessarily adjust as society advances. What’s more, the CPI reflects inflation in goods far outside the bounds of basic needs. But the NAS didn’t technically redefine the new poverty threshold, only issuing broad guidelines that prompted the Census to produce eight experimental models with varying results, as the graph above shows.

Under the Obama administration, however, the Census Bureau has tried to consolidate the 1995 NAS findings into a single alternative yardstick. The new Supplemental Poverty Measure uses the Consumer Expenditure Survey, which suggests its poverty rate could be even higher than the official level for 2010.

The Census remains mum about the new figures, but its own research confirms as much. In a June working paper, a Census researcher found that “the SPM resulted in slightly higher poverty rates than the official measure for most groups” in 2009. Similarly, a state-level Urban Institute study found that there’s a “slight increase” in the poverty rate for non-elderly adults under the SPM, though child poverty is generally lower, according to senior research associate Laura Wheaton.

Critics counter that the SPM’s use of actual consumer expenditures instead of an inflation-adjusted bundle of goods is a relative, not absolute, measure of poverty. They say this “increases in income result in equivalent increases in expenditures on necessities, and thus, the poverty threshold will always rise together with median income,”as a Brookings briefing notes. In other words, poverty isn’t likely to decrease even when Americans have objectively higher standards of living, because people will spend more on food as their incomes rise. If you adjust for inflation instead of consumer expenditures, the 2009 poverty rate is actually lower than the official level when you factor in the other NAS recommendations, as the earlier graph shows.

Expect this debate to reignite when the Census releases its alternative poverty figures next month. One thing remains clear, however: the recession has made Americans poorer, no matter which of these yardsticks you use.

© The Washington Post Company