In 1962, socialist, activist, and writer Michael Harrington published The Other America: Poverty in the United States. The book has sold over a million copies since. A review in The New Yorker made it to the desk of President Kennedy, and is considered to have been the catalyst of LBJ’s War on Poverty. Matthew Desmond’s Poverty, by America may be one of the most celebrated books on the subject published in the six decades since. Desmond’s book echoes Harrington’s themes: the appalling degree of material want in the world’s richest nation and the need for muscular government programs to fight it. Margaret Talbot, in another glowing New Yorker review, writes that Poverty, by America “deserves to be one of those books you see people reading on the subway, or handing around at organizing meetings, or citing in congressional hearings. Its moral force is a gut punch.” But crafting policy to address poverty, as President Johnson did in the ’60s, requires more than moral force. While there is much to like in Poverty, by America, significant empirical flaws undercut the strength of its argument. 

Unsurprisingly, Desmond rejects the idea that poverty is a result of individual decisions or personal failures. But he is just as skeptical of attempts to explain poverty through historical forces such as deindustrialization. Instead, Desmond places the blame on the exploitative nature of economic relations in a neoliberal system. His central theses are that poverty springs from deliberate effort by the affluent to redirect society’s resources toward themselves, and that the government’s response to this exploitation has been ineffective in helping those most in need. Desmond offers many examples of economic exploitation: landlords in poor neighborhoods who charge high rents for substandard housing; large corporations that decimate unions by outsourcing jobs; payday loan stores and check-cashing outlets located in poor neighborhoods that charge exorbitantly high fees. This material is the heart of Desmond’s first book, Pulitzer Prize-winning Evicted: Poverty and Profit in the American City, and it is where Poverty, by America is at its strongest.

The latter thesis — that the United States government has failed to adequately help those in need — rests on much shakier ground. Here, Desmond selectively chooses nonstandard measures of poverty to support his case. He argues that the poverty rate has shown no discernible long-term trend since the 1960s despite increased spending on anti-poverty programs. The poverty rate increases during recessions and declines during economic expansion, but essentially has remained flat over decades. 

But this depends on what data you look at. Desmond’s case relies primarily on the official poverty measure (OPM). The OPM, originally devised in the 1960s, compares family income against a threshold of what would be needed to keep a family out of poverty. In 2022, for example, a family of five earning below $35,495 was considered poor. But a crucial problem is that the OPM only counts cash income. It doesn’t include noncash and near-cash income, such as the Earned Income Tax Credit, housing subsidies, and Medicaid benefits. These benefits have all grown substantially over the past several decades, making the measurement problem worse over time. Spending on the EITC, for example, increased from $5 billion in 1975 to $61 billion in 2010 (in 2010 inflation-adjusted dollars). Likewise, spending on Medicaid increased from $90 billion in 1980 to $772 billion in 2020 (in 2020 inflation-adjusted dollars). As a result, the OPM undercounts the resources families have to meet their basic needs and thus overcounts the number of people who are poor. This problem is compounded by the fact that people underreport even their cash income in surveys. One study that matched administrative records to survey responses in New York found that poverty estimates would be 2.5 percentage points lower with a fuller accounting of benefits from food assistance, cash welfare, and housing assistance alone. Income underreporting also has gotten worse over time. 

Poverty measures that account for these and other deficiencies show a marked downward trend in poverty. One National Academies of Sciences, Engineering, and Medicine report shows how much the choice of measure matters: While child poverty remained flat from 1967 to 2016 when using the OPM, it declined by over 40% when using the more complete Supplemental Poverty Measure (SPM), which includes government noncash benefits. A report by researchers with the National Bureau of Economic Research that set the poverty rate in 1963 at 19.5% ¹ found that, when the full range of benefits that families receive were counted, poverty fell to 1.6% by 2019. The Center on Budget and Policy Priorities, a progressive think tank, likewise reports that the safety net’s effectiveness at reducing child poverty has grown dramatically since 1967: Using the SPM, nearly 40% of children were lifted above the poverty line in 2016. And if we measure poverty by how much families report consuming rather than income they receive, we see the same trend: large declines in the percentage of people who consume below a poverty threshold. One study that linked survey responses to administrative data to correct for income underreporting found that the percentage of households living in “extreme” poverty over the year—or under $2 a day—was just 0.18%, rather than 2.08% before the adjustments. 

Desmond critiques the use of the SPM measure in a lengthy footnote. His main objection is that many research reports use an “anchored” threshold in their implementation of the SPM. Anchored thresholds are only updated for inflation over time. This is in contrast to the semi-relative thresholds of the SPM, which are updated to reflect how much people spend on a basic bundle of goods, like housing, food, and childcare. Because of rising living standards, spending on these items has increased above and beyond inflation — the median American lives in a substantially larger house, for example, than they did a generation or two ago. Anchoring the poverty line in, say, 1963 means anchoring our standard of living in 1963: A poor person by 2023 standards is materially better off than their counterpart 60 years ago, which is what the anchored measure reports. 

There is a reasonable argument to be made for relative poverty measures, and both are informative. Even setting this debate aside, an unanchored SPM shows substantial declines in poverty over time. According to data from the Center on Poverty & Social Policy at Columbia University, the unanchored SPM rate for the total population halved from from 19% in 1967 to 9% in 2020. As might be expected, over the same period, the anchored SPM fell by even more — from 25% to 8%. ²

There are also issues with some of the book’s less central claims about poverty. Desmond downplays the association between family structure and poverty by stating that “A study of eighteen rich democracies found that single mothers outside the United States were not poorer than the general population.” However, the study he cited, which approvingly argues that universal social policies can significantly reduce single-mother poverty — and that they do so much more in other countries than in the United States — nevertheless notes that single mothers are disproportionately poor in all of the countries. 

One of Desmond’s principal arguments is that the biggest beneficiaries of federal aid are affluent families. The implication is that the lifestyles of the rich are paid for by perverse transfers of income from the poor and middle class. However, according to the independent Tax Foundation, the U.S. tax and transfer system is “highly progressive” when accounting not only Federal taxes and transfers but state and local taxes and transfers as well. As they report, “The lowest [income] quintile experienced a combined tax and transfer rate of negative 127.0 percent, meaning that for each dollar they earned, they received an additional $1.27 from the government, netting transfers (gains) and taxes (losses), while top quintile had a rate of positive 30.7 percent, meaning on net they paid just under $0.31 for every dollar earned.” The report further notes that the top quintile funded 90% of all government transfers in 2019. So while it is true that many programs, such as homeowner subsidies, help middle-class families rather than poor ones, these programs are still paid for primarily by affluent Americans. We may of course discuss whether wealthy Americans should pay even more than they currently do, but we should do so with a complete understanding of current tax and expenditure patterns. 

Elsewhere, Desmond states that “the bulk of the evidence suggests that the employment effect of raising the minimum wage is inconsequential.” It’s true that a debate on this issue is far from settled. Some of the papers he cites do indicate that the evidence supporting a connection between the minimum wage and employment is weak — but weak evidence for a connection is not the same thing as evidence against one. Even the studies he cites are cautious in their pronouncements. One states that “The weight of that evidence points to little or no employment response to modest increases in the minimum wage” (italics added), and concludes that firms may respond to higher wage costs not necessarily by cutting jobs but cutting hours, reducing fringe benefits, and raising prices for customers. Moreover, the most recent (2022) meta-analysis of the effect of minimum wage increases concludes that “(i) there is a clear preponderance of negative estimates in the literature; (ii) this evidence is stronger for teens and young adults and the less educated; (iii) the evidence from studies of directly affected workers points even more strongly to negative employment effects; and (iv) the evidence from studies of low-wage industries is less one-sided.” In this case, and throughout the book, Desmond’s emphasis on the moral imperative to reduce suffering tends to obscure the fact that policy choices often require trade-offs. 

Poverty, by America is a noteworthy contribution to the long-standing conversation about the problem of poverty. Desmond uses moral persuasion to spur us to do more about it. The book’s strength is in describing how people and institutions often exploit the poor, causing them to pay more for many goods and services than other Americans. However, Desmond undermines his call to action through his imbalanced treatment of trends in poverty and the effectiveness of economic growth and government programs in reducing it. This is a significant weakness: One needs a full understanding of a problem in order to prescribe the most efficient and effective solutions.