The Bottom 1%

A lot of research and writing explores the dynamics of the richest one percent, .1 percent, and even the .01 percent. But researchers only have the fuzziest understanding of America's poorest.

A homeless shelter in Cranston, Rhode Island (Brian Snyder/Reuters)

How is it different to be poor—very poor—in a developing country than in the richest country in the world? That's the question asked in a new paper from Brookings researchers Laurence Chandy and Cory Smith.

To answer it, Chandy and Smith needed to know more about that very bottom of America's wealth distribution—those living below the global poverty line of $2 per day. Not so easy: In contrast with how well-documented, how legible, the lives of the richest Americans are, those of the poorest are opaque, having fallen between the cracks in researchers' databases. The poorest are "the least explored and understood part of the distribution."

But that's only true for rich countries. In the developing world, precisely the opposite is the case: In such places, they write, "tax systems are insufficiently developed to provide an accurate portrayal of the rich so the upper tail remains largely unknown; by contrast the concentration of incomes near zero means
that surveys capture the lower tail relatively well."

The result of this is that "we know least about the top of the distribution in poor countries and least about the bottom of the distribution in rich countries." Unfortunately, in America, that means we have the worst understanding of "those whose lives are most precarious."

I spoke with Chandy about their paper and why it's so important that we improve our understanding of the dynamics of poverty in America. A lightly edited transcript of our conversation follows.

Rebecca Rosen: To start with, I know that your research typically looks at global poverty, particularly poverty in the developing world. What brought you to look at U.S. poverty? How does poverty here look to someone who spends his time thinking about poverty in much poorer countries?

Laurence Chandy: My co-author Cory Smith and I were drawn to this topic last summer after reading an excellent paper on U.S. poverty by Luke Shaefer and Kathryn Edin. Their report was the first to apply the $2 a day poverty line, a definition of poverty traditionally used in the developing world, to the U.S., and therefore to draw an analogy between the levels of destitution that prevail in these very different settings.

When Cory and I first heard about this paper, we were frankly quite dismissive. Our priors were that the depth of poverty was of an altogether different magnitude in developing countries, and that people who drew this kind of comparison were probably motivated by parochial ideas such as a desire to cut off foreign aid to poor countries. But after reading Luke and Kathryn’s paper, it was clear that this was a serious and careful piece of research by informed U.S. poverty scholars. Their paper motivated us to examine the topic ourselves, but to approach it from our perspective as analysts of global poverty.

R.R.: Can you tell readers a bit about their findings, and then how those findings look in a global context? How poor are the poorest Americans? Where do they fit in with the global picture?

L.C.: Well, the first and most striking finding of theirs was simply that $2-a-day poverty exists in the U.S. (Note that the official U.S. poverty line works out at around $16 per person per day, so we’re talking about people living at a tiny fraction of this threshold.) Luke and Kathryn focused their attention on families with children, and estimated that between one and five percent of these families lived under the $2 threshold, depending on whether a narrow or broad (i.e., including various benefits and tax credits) definition of income is used.

If we crudely compare this one-to-five-percent range with estimates of $2 poverty in the developing world, it ranks the U.S. among various middle-income countries: countries like Russia, West Bank and Gaza, Albania, Argentina. I would stress that such a comparison is indeed crude. To make a fair comparison, you would want to use a similar methodology for measuring poverty across these different settings.

One of the things we wanted to do in our paper was to attempt a more rigorous comparison. So we took the approach used by the World Bank to estimate poverty in developing countries and tried to replicate this as faithfully as possible using data from the U.S. Most importantly, this meant using a different data source than Luke and Kathryn—a survey of reported consumption as opposed to a survey of reported income. (Luke and Kathryn’s estimates drew from the Survey of Income and Program Participation, or SIPP; to replicate the Bank’s approach, we drew from the Consumer Expenditure Survey.)

When we did this, the estimate of $2-a-day poverty dropped all the way to zero. But this still left us with lots of questions: What did the range of estimates of $2 poverty in the U.S. tell us about how poverty is manifested differently in the U.S. and the developing world? And is the methodology used by the World Bank for estimating poverty in the developing world the best way to analyze the condition of America’s poorest people?

Different Estimates of a $2-a-Day Poverty Rate
A figure from Chandy and Smith's paper showing the different estimates for a $2-a-day poverty rate using different data sets

R.R.: We see a lot of articles and research on what the top one percent looks like, the top .1 percent, even the top .01 percent, but our understanding of the very bottom of the income distribution is much less differentiated. Why is that? What can we know about America’s very poorest?

L.C.: Our main tool for assessing people’s welfare is representative surveys, where we take a sample of the population, ask them lots of questions, and extrapolate their answers to estimate the condition of the population as a whole. The trouble with this approach is that it does a much better job of telling us about people nearer the middle of the distribution than about those at the extremes or “tails.”

In the last decade, pioneering work by a group of economists including Thomas Piketty, Tony Atkinson, and others has enabled us to obtain a much clearer understanding of the top of the distribution, by combining traditional survey data with tax data (i.e., individuals’ tax returns). But we have no equivalent solution for revealing what is going on at the bottom of the distribution.

There are plenty of reasons for being skeptical about what surveys tell us about the bottom of the distribution. Two concerns loom large: Are surveys interviewing the right people, and are respondents giving the right (truthful) answers? Some of the poorest people in the U.S. (unregistered migrants, institutionalized populations, people living in informal housing, people on the move, etc.) are missed by surveys, resulting in sampling bias. And there are several reasons why people at the bottom of the distribution may provide inaccurate responses to survey questions, especially those regarding their income. For instance, respondents may be reluctant to declare earnings from informal work out of fear that this would prompt a visit from the IRS.

What we try to do in our paper is to discern what different estimates of $2-a-day poverty in the U.S. can tell us about the characteristics of the country’s poor. For instance, we obtain lower estimates of poverty when we extend the reporting period over which income is assessed. This suggests that a significant amount of $2 poverty appears to be temporary. Like Luke and Kathryn, we obtain lower estimates when we include various government benefits and tax credits in our definition of income. This tells us that these programs are making a critical difference for millions of individuals—at the very least, the difference between living above or below the $2 threshold.

R.R.: What might we gain from further research into such extreme poverty? Why is it important to improve our understanding of what this poverty looks like?

L.C.: First and foremost, I think there’s a moral obligation to get a better handle on the condition of the poorest people in the country. If we accept a shared responsibility for the poor in our society, then surely that responsibility is greatest for the poorest of the poor.

There’s a lively political discussion underway, sparked in part by the 50th anniversary of LBJ’s declaration of a War on Poverty, as to whether the government’s approach to tackling poverty is working and how it might be strengthened. There is some suggestive evidence that welfare reforms over the past two decades did most to assist those living just above and below the poverty line, while doing little for those furthest below the line; and that the deepest poverty in the U.S. is becoming more prevalent. If there is to be another round of welfare reforms, it is imperative that these are informed by good empirics and a thorough understanding of the condition of America’s poorest people.

One of the points we make in our paper is that even if these people have their most basic material needs met, they may be deprived in other important ways. A reliance on in-kind transfers can leave people virtually excluded from the cash economy and with little agency to address unexpected needs. We describe this in our paper as a state of purgatory. These are the sorts of challenges that need to be understood and addressed if poverty is to be overcome.