There’s no getting around the fact that the new poverty numbers are dismal. Yet it is worth remembering the fact that food stamps and EITC benefits are not included, as Conor Dougherty and Sara Murray explain in the Wall Street Journal:
The threshold for poverty in the U.S. in 2009 was a family of four earning $21,756. But this only takes into account monetary income, while omitting the many benefits that now form the backbone of the government efforts to lift the poor. Such programs include subsidized housing and the Earned Income Tax Credit.
The poverty rate “misses most of the programs that have been added or expanded in the last 20 years to reduce poverty,” says Bruce Meyer, an economist at the University of Chicago.
For instance, the government estimates if the food stamp program was counted, it would have lifted 3.6 million people above the poverty threshold last year.
Of course, 3.6 million would still leave us with almost 40 million Americans in poverty. According to Elizabeth Kneebone, a senior research analyst at Brookings writing in 2009:
The EITC delivers over $40 billion dollars a year in wage supplements to lower-income workers and their families and lifts more than 4 million people out of poverty each year.
This still leaves 36 million in poverty. And as Ron Haskins notes, EITC spending last year was $50 billion. In a paper on EITC expansions, Bruce Meyer writes [PDF]:
The credit targets resources at those below the poverty line,particularly families with children. It raises more than 4.0 million people above thepoverty line. While it is especially aimed at people around the poverty line, it also raises1.4 million people above half the poverty line. The expansions to the EITC under theAmerican Recovery and Reinvestment Act of 2009 were targeted at large families thatare especially likely to be low income. I estimate that the ARRA will raise over 400,000 additional individuals above the poverty line. All of these numbers are likely to be sharp underestimates because of survey mis-reporting, and probably should be raised by almostfifty percent. [Emphasis added.]
Intrigued by that last sentence, I check out a 2009 paper Meyer co-authored with Wallace K.C. Mok and James X. Sullivan on “The Under-Reporting of Transfers in Household Surveys: Its Nature and Consequences“:
We find that under-reporting is common and has increased over time. Less than half of Workers’Compensation benefits are typically reported, and only about two-thirds of Food Stamp Program,TANF, WIC and Unemployment Insurance benefits are commonly reported. Three-quarters ofSSI benefits and a much larger share of SSDI and OASI benefits tend to be recorded. There is substantial variation across surveys, with the CE Survey typically having the lowest reporting rate and the SIPP having the highest rate for most programs.
Over time, the reporting of many programs in the surveys has sharply deteriorated. We have also seen a noticeable rise in the share of responses that are imputed. This rise inimputation and under-reporting is part of an overall pattern of decline in the quality of data from U.S. household surveys. Other papers have shown a rise in survey nonresponse and item nonresponse and a drop relative to alternative sources (Atrostic et al. 2001, Meyer and Sullivan2007b, 2009).
Nicholas Eberstadt has been written extensively on how we measure poverty, and in particular about the gap between income and consumption among the very poor:
What is wrong with the official poverty rate? It measures the wrong thing–and always has. That thing is income. But poverty is a matter of consumption, and a huge gap has come to separate income and consumption at the lower strata of our income distribution. In 2006, according to the annual Bureau of Labor Statistics Consumer Expenditure Survey, reported purchases by the poorest fifth of American households were more than twice as high as reported incomes.
For reasons still only partly understood, the surfeit of spending over income among poorer U.S. households has increased dramatically since the 1970s–making income an ever less dependable predictor of living standards for the disadvantaged. Indeed, while the official poverty thresholds are meant to be constant over time, a whole host of data confirm the (welcome) fact that material conditions for our population in “poverty” have been steadily improving. The official statistic is incapable of documenting–or even recognizing–any changes in living standards among America’s poor.
The world has changed since 2006, when Eberstadt wrote his piece on “The Poverty of the Official Poverty Rate,” but this has to be part of the picture.
So why does any of this matter? The poverty rate is often used as an indicator of deprivation, and as a spur for more transfers to the poorest Americans. I think there is a strong case for more generous transfers to, for example, non-custodial parents, among others, particularly if we finance these expansions by paring back entitlement spending on middle-income and affluent individuals. Yet it would be useful to factor in the public assistance that already flows to the poor to have a sense of how much further we need to go.