On The Prevalence Of "Extreme Poverty" In The United States

The World Bank has a definition of “extreme poverty,” which, in their formulation, means “living on less than $1.90 per day.” I have no idea how they picked the figure of $1.90 for the cutoff, but that certainly is a very low figure. Parsing the definition a little, notice that they use the term “living on” as the standard, rather than saying, for example, that a person has less than $1.90 per day in “income.” As you will see, the difference is significant. Anyway, the WB in its most recent (April 2019) report on this subject gives a figure of about 10% of the world’s population, or 736 million people, as “living on” resources below that very low $1.90 per day level per capita. That actually represents a large decline in the percent of people in this “extreme poverty” condition over the past couple of decades; but it still comes to a very large number of people. Whether that 10% figure is accurate or even substantially exaggerated, I have no doubt that there are still many, many people in the world who have less than $1.90 per day in resources to “live on,” and who therefore live in what could only be called deep, grinding poverty.

But the question for today is, is there any substantial number of people in the United States who live in conditions meeting this World Bank test of “extreme poverty”? And if so, how could that be, and how could such a thing be allowed to persist? Some call this the most important question out there in the field of poverty studies.

Driving this debate has been a pair of well-known researchers who have for many years been making a career pushing large numbers as being the supposed count of those in the U.S. in this “deep poverty” condition. The researchers are sociologist Kathryn Edin of Princeton and social policy guru Luke Shaefer of the University of Michigan. In 2015, Edin and Shaefer published a book titled “$2.00 a Day: Living on Almost Nothing in America.” The $2.00 per day was an arbitrary rounding up of the World Bank’s equally arbitrary $1.90 per day “extreme poverty” cutoff, although besides doing that rounding the authors also switched the definition of the measurement metric from “living on” that amount or less, to having “cash income” of that amount or less. The book famously came up with a figure of some 1.2 million households with children in the United States, and 3.6 million households total, supposedly meeting the test of having cash income less than the $2 per day per capita figure.

You will not be surprised that Edin and Shaefer’s 2015 book drew excited praise — and a notable dearth of skepticism — from the usual progressive suspects. (Example — William Junius Wilson in the New York Times: “This essential book is a call to action, and one hopes it will accomplish what Michael Harrington’s “The Other America” achieved in the 1960s — arousing both the nation’s consciousness and conscience about the plight of a growing number of invisible citizens.”) Less excited was the Manhattan Contrarian, who, in a post on May 26, 2016, had this to say:

This book is completely preposterous. . . . [I]s there really anyone in America who lives on less than $2.00 per day, at least as averaged over some reasonable period of time?  The idea is completely ridiculous, and frankly an insult to the hundreds of millions of people in the world who actually do live in grinding poverty at this very low level.  Edin and Schaefer just intentionally confuse two very different concepts, which are, first, having "income" of only $2.00 per day, and second, only having $2.00 or less of resources per day available on which to live.  The second is what the World Bank is talking about when it talks about serious poverty out there in the world.  The first is a statistical artifact of the particular definition of "income" that you choose to use.

I proceeded to make the obvious comment that the $2 per day per capita “cash income” definition of extreme poverty was essentially an intentional deception driven by the arbitrary exclusion from the calculation of the well-more-than $1 trillion annually of in-kind and tax-related anti-poverty distributions in the United States (e.g., food stamps, EITC, public housing, Medicaid, etc.). Those arbitrary exclusions cost the taxpayers some $25,000 per year per person for each person in the U.S. deemed to be “in poverty” (a much larger number of people than the ones deemed in “extreme poverty”). I did not even get into in that post the whole subject of under-reporting of income by low-income people when they answer Census surveys, let alone the question of the multi-trillion-annual-dollar underground and illegal economy. I was not alone in criticizing the Edin/Shaefer effort, with many of my co-skeptics emphasizing the under-reporting point; but needless to say the praise coming from the progressive side completely drowned out our few dissenting voices.

Which only encouraged Edin and Shaefer to double down. In early 2018 they came out with a new Report titled “Welfare Reform and the Families It Left Behind.” That Report claimed that the number of households with children living below the $2 per day per capita “cash income” threshold had continued to increase from the 1.2 million at the 2011 data cutoff for their 2015 book, peaking at close to 1.4 million in 2014 before falling back a little in 2015. The Report also responded to critics who had accused the authors of ignoring the issue of under-reporting of incomes at the low end, making the following claim:

Are these results driven by underreporting in survey data? No. When we control for underreporting, we find that the downward spiral since 1995 is even more dramatic than previously reported.

OK then. They say that they have “controlled” for under-reporting of income.

That brings us to May 2019. A group of very serious researchers, led by Bruce Meyer of the University of Chicago, has just come out with a new academic paper under the aegis of the National Bureau of Economic Research, titled “The Use and Misuse of Income Data and Extreme Poverty in the United States.” You will only find an abstract at that link; to get the full paper, you will need to click through to the SSRN website and pay five bucks.

It would be fair to say that this new paper eviscerates the work of Edin and Shaefer. Meyer et al. do the evisceration in much more polite language than I would have used. The single biggest issue identified is that Edin and Shaefer rely for their data on surveys distributed and collected by the Census Bureau, particularly something called the Survey of Income and Program Participation, where Census then reports as income whatever the respondent says without any kind of follow up or double check. Large numbers of people simply report their income as zero (or maybe just leave that box blank) and Census blindly says that their income is zero. Meyer et al. do the heavy lifting of going and getting government data from other serious sources (IRS, social security, food stamp program) for the same people, and making several other obvious and major corrections. Lo and behold:

We re-examine the rate of extreme poverty by linking 2011 data from the Survey of Income and Program Participation and Current Population Survey, the sources of recent [ed. — e.g., Edin and Shaefer] extreme poverty estimates, to administrative tax and program data. Of the 3.6 million non-homeless households with survey-reported cash income below $2/person/day, we find that more than 90% are not in extreme poverty once we include in-kind transfers, replace survey reports of earnings and transfer receipt with administrative records, and account for the ownership of substantial assets.

Of course, when someone puts down their “income” as zero on the SIPP form, we don’t know the reason. Maybe large numbers are just making an error. Others could be embarrassed about receiving federal welfare benefits. Still others could be unwilling to reveal income received in the illegal or underground economy. But whatever the reason, Meyer et al. make clear that some 90+% of the supposed “extreme poverty” of the Edin/Shaefer work comes from systematic refusal to account for ascertainable receipt of known income and/or government benefits. A couple of other choice quotes from Meyer et al.:

[F]ocusing on such low-income thresholds as $2/person/day in the United States is likely to yield a group filled with more gross errors than households that are truly impoverished.


[T]he households receiving means-tested in-kind transfers — who appear to be among the most materially deprived Americans — are almost all not in extreme poverty simply by virtue of the extreme poverty income threshold being lower than benefit amounts.

Not stated is that the “extreme poverty” threshold of $2/person/day (which would be $730/person/year) is not just a little less than the level of available means-tested in-kind transfers, but wildly less. As stated above, the amount of means-tested in-kind transfers to those in “poverty” in the U.S. is in the range of $25,000/person/year. And, if you truly have no income, all you need to do to get the approximately $25,000 per person in in-kind annual transfers is go and ask.

Being data geeks, Meyer et al. can’t resist the temptation to conclude with a supposed quantification of how many people are left in seeming “extreme poverty” after all their corrections and adjustments:

In the end, our best estimate of the extreme poverty rate is 0.24% among households and 0.11% among individuals, . . . .

That would still constitute several hundred thousand people. But my question is, who could these people possibly be, when anybody finding themselves in this “extreme poverty” condition can just go ask for benefits and get something in the range of $25,000 per person annually (admittedly, almost all in-kind) just for the asking? I would like to see someone actually bring some of these people forward so that we can evaluate if there is any actual deprivation or suffering going on, versus some kind of artifact of the statistics. My suggestion in my May 26, 2016 post was that people meeting the criterion would not at all be the people you are thinking of when you think of “extreme poverty.” For example:

. . . students living off scholarships, fellowships and loans (scholarships, fellowships and loans don't count as "income"); kids supported by their parents while they look for their first job (family support doesn't count as "income"). . .

There could be many other such categories.

Meanwhile, the Edin/Shaefer work continues to be used by advocates demanding huge increases in cash welfare spending to alleviate the suffering.