Princeton sociologist Matthew Desmond’s previous book, Evicted, offered a compelling account of poverty in America. Illuminating and thought-provoking, its ethnographic accounts of deep struggle spurred new research and increased policy focus on the links between poverty and housing instability. Though it didn’t get everything right, it was an important book. Sadly, Desmond’s latest offering, Poverty, by America, provides little of value to readers looking to learn more about poverty in the United States and how to reduce it.
The book’s original sin is its confused understanding of poverty’s causes. Opting for tropes over scientific research, Desmond’s explanations range from irrelevant to backward. While the book reads like a narrative strewn together from talking points borrowed from political advocates, it contains a surprisingly extensive set of notes that reference serious research. Unfortunately, these largely serve as window dressing. Desmond often fails to engage seriously with this research, and as a result, his policy recommendations are unproductive.
Desmond opens with a failed attempt to define and measure poverty. In a chapter titled “The Kind of Problem Poverty is,” he starts with two anecdotes. The first features a hardworking 55-year-old father who snorts speedballs (a mixture of cocaine and other drugs) to stay awake during his double shifts. The second recounts the struggles of a homeless and severely mentally ill woman. Though harrowing, such stories do not represent “the kind of problem poverty is.” The vast majority of poor people don’t abuse illegal drugs, don’t work double shifts or even full-time jobs, and aren’t homeless. Among poor adults aged 18 to 64, less than 10 percent worked full-time, year-round in 2021. Less than 2 percent of poor people are homeless at any point in time.
Desmond proceeds to misrepresent the poverty population with incorrect data. He generally relies on the official poverty measure, which does not account for most of the government programs that provide aid to low-income Americans, a fact that most serious poverty researchers recognize. He fails to acknowledge the progress made in reducing poverty over the past several decades. And he wrongly believes that the United States has significant levels of extreme poverty.
Consider Desmond’s claim that the U.S. is “the richest country on earth, with more poverty than any other advanced democracy.” In fact, the U.S. does not have the highest rate of absolute deprivation among advanced democracies. Though he doesn’t note it explicitly, Desmond’s source for this claim sets the poverty line at half of the median income in each country. Because the U.S. has among the highest median incomes in the world, such relative poverty measures make it appear poorer. Using an absolute measure of deprivation, the National Academy of Sciences found that the child-poverty rate in the U.S. is slightly lower than the United Kingdom’s and modestly higher than Canada’s.
What about the statement that “more than a million of our public schoolchildren are homeless, living in motels, cars, shelters, and abandoned buildings”? Based on the literal homelessness definition, approximately 95,440 children are homeless on any given night. Desmond’s number relies on a much broader definition of homelessness, including students whose families at some point during the school year shared housing with others for economic reasons.
Another false claim: “In the years following the end of guaranteed cash welfare, the United States has witnessed a shocking rise in extreme poverty, one that tracks with other grim indicators.” Over 90 percent of the reported level of extreme poverty results from the underreporting of income in surveys, along with the complete exclusion of in-kind transfers and disregard of substantial assets in making these determinations.
The insinuation that the United States harbors “a kind of extreme poverty” characterized by “bare feet and swollen bellies” is absurd. The U.S. Department of Agriculture’s annual report on food insecurity does not even mention the words “starvation” or “malnourishment,” nor the types of nutritional deficiencies that lead to swelling of the stomach. Fortunately, Desmond’s claim is decreasingly true for “faraway places” as well, with worldwide extreme poverty falling by 76 percent since 1990. In the United States, study after study after study after study has shown deep declines in poverty over the past half-century. The fallacy, again, is Desmond’s use of trends in the official poverty measure, which omits the programs that have lifted millions of families out of poverty.
After providing this misleading account, Desmond tries to explain why the nation is so impoverished. But one can’t explain something without correctly defining it first. Desmond instead leans on a collection of rehashed talking points about poverty in America without demonstrating or quantifying their importance.
Desmond’s first explanation is that employers don’t pay their workers enough money because the minimum wage is too low and labor unions have been weakened. Yet less than 10 percent of poor working-age adults actually work full-time, year-round. If nobody in a family works full-time, or nobody works at all, then paying workers more money will typically be insufficient to lift them out of poverty.
Even for the minority of poor people who do work full-time all year, Desmond’s explanation of a low minimum wage falls short. Increasing the minimum wage is unlikely to move a significant number of people out of poverty. It is more likely than not to reduce employment, so that the small drop in poverty due to wage gains is offset by families who fall into poverty due to job loss. Bizarrely, Desmond writes that most studies on the effects of the minimum wage find no employment loss—and he does this while citing the most recent and authoritative literature review on the topic, which states that 79 percent of the most credible studies find significant employment loss.
Even if the minimum wage did not reduce employment, it would still have little effect in fighting poverty. Household incomes depend on many factors beyond the hourly wage of a given worker, including how many hours that worker works, the number of other workers in the household, and other sources of income. That is why almost all studies have found that raising the minimum wage has little effect on poverty. A 2019 Congressional Budget Office report found that more than doubling the federal minimum wage from $7.25 to $15 per hour would reduce poverty by under 3 percent.
Low-wage workers, especially those with kids, receive substantial assistance to ensure that they stay out of poverty. For example, a single parent with two children who works full-time and earns $10 per hour will make $20,000 during the year, below the $24,860 poverty line for a family of three. But that parent will get $6,604 from the Earned Income Tax Credit, $2,800 from the refundable portion of the Child Tax Credit, and likely more than $7,000 from food stamps, increasing total income to more than $36,000—well above the poverty line. This excludes other supplements the family may receive, such as other nutrition programs, Medicaid, and subsidies for energy, child-care assistance, and housing.
Desmond next points to alleged exploitation of poor people by the banking industry. Payday lending, in which short-term loans are taken out at high interest rates, is a prime culprit. While fees for short-term loans can certainly be high, Desmond makes no attempt to quantify the extent to which they lead families into poverty. Notably, some research suggests that banning payday lending would ultimately hurt the poor.
Landlords are another target. Desmond points to rents that are nearly as high in poor neighborhoods as elsewhere, even though housing and neighborhood conditions are worse. He ultimately acknowledges the reason why rent might be higher for some low-income tenants: they may present a higher risk of nonpayment of rent, which erodes landlords’ ability to cover their costs. Still, he points to his own research, which finds that landlords in poor neighborhoods enjoy profit margins $50 higher per month than in non-poor neighborhoods. Assuming that’s more or less correct, poor renters would pay $600 per year more on rent than they “should”—amounting to approximately 2 percent of the poverty line for a family of four, hardly enough to be a major explanation for poverty.
Next, Desmond argues that the government provides more assistance to rich people than poor people. In reality, government taxes and transfers increase income for a two-person household in the bottom 20 percent of the income distribution by an average of $20,800; the taxes and transfers decrease income for a two-person household in the top 20 percent of the income distribution by an average of $56,000. Americans in the bottom 20 percent of the income distribution receive far more net assistance from government than any other quintile.
Desmond defends his backward logic through discussions of tax provisions that lessen tax liability for some taxpayers relative to others. Some of these provisions, like the deductions for mortgage interest and state and local taxes, can indeed lead to arguably inefficient outcomes. The tax reform passed in 2017 substantially weakened both of those. Yet tax distortions do not change the basic fact that the rich pay much more into government than they get back, while the poor receive much more from government than they put in.
Desmond’s final explanation for poverty is that affluent people hoard opportunity, echoing the arguments popularized by Richard Reeves’s book, Dream Hoarders. Here, Desmond is right that excessive regulations on housing construction, segregated public schools, and lack of connection between people of different classes are damaging to society and play a role in poverty. When people can’t access high-paying jobs in high-opportunity places or afford the housing needed to attend better public schools, poverty is likely to rise. Despite these helpful points, however, Desmond fails to address other productive policy avenues, such as school choice, that could help expand opportunities for low-income children.
That Desmond has mischaracterized the problem of poverty means, naturally, that his proposed solutions tend to be unserious. He argues that the nation could afford to end poverty without increasing the deficit simply by cracking down on tax cheats; that it is wrong to ask whether the government can afford the price tag of programs that would ease poverty; that we could afford to make a dent in poverty if we didn’t use the welfare system to guard the fortunes of the rich; and that people who seek to explain the perils of wealth taxation are “defeatist and boring.”
To his credit, Desmond tries to present concrete policies to eliminate poverty—but he gets the answers terribly wrong. He claims it would cost $177 billion to end poverty in America, but that number is far too low, being based on a total sum of the money it would take to bring each poor family exactly to the point above the poverty line.
Were the government to implement a policy ensuring that no one would fall below the poverty line, it would create a major disincentive to work. The government would provide any working family below the poverty line enough income to reach it. Working families just above the poverty line would see little incentive to work, since they could fall no lower than the poverty line if they stopped working altogether. The reduction in work by families below and above the poverty line would need to be compensated for with still more government transfers. Funding such a policy would be made even more expensive by the need to raise taxes to pay for it, which would discourage work by middle- and higher-income Americans as well.
The government assistance that Desmond ignores by adhering to the official poverty measure exceeds the amount of expenditure that he claims would be needed to eliminate poverty. Add up the pre-pandemic cost of non-cash transfers, such as food stamps ($60 billion), housing assistance ($49 billion), the earned income tax credit ($65 billion), and the refundable portion of the child tax credit ($36 billion), and the U.S. spent $210 billion on these programs in 2019. And that excludes the $627 billion spent on Medicaid.
How would Desmond spend his couple hundred billion? He seems to favor a guaranteed income to families with children (along the lines of the temporary child tax credit passed in 2021), more publicly funded construction and ownership of housing, and more spending on childcare, education, and transportation. He advocates a higher minimum wage, stronger unions, more banking regulation, and less reliance on the private market for housing. He scorns the earned income tax credit, which, he suggests—contrary to a large body of evidence—has not reduced poverty since its rollout and expansion. The lack of specificity in Desmond’s proposals makes them hard to evaluate. But eroding work incentives and relying more heavily on government production of goods and services is unlikely to be a recipe for success.
To solve a problem, one must first understand it. Poverty, by America, offers an egregiously incorrect understanding of the nature and causes of poverty in the United States, rendering its proposed solutions unproductive.
Serious research on poverty must contend with basic facts. Poverty has fallen significantly over the past 50 years. Government-assistance programs do a lot to diminish material hardship. Poverty is less likely when adults work, especially full-time, year-round; it is even less likely when adults are married. A serious debate about the causes and policy implications of these facts can be had—but only once these facts are recognized. Desmond’s book ignores them, in the process confusing the public conversation.