Princeton sociology professor Matthew Desmond has spent his academic career studying American poverty and public policy. In his latest book, Poverty, by America, he explicitly states his purpose as answering the question, “Why is there so much poverty in America?” This is an important question, and it can be answered—at least in part—by careful analysis of the data, so I was eager to read what he found.

Who is poor? / One would expect the analysis to begin with some clarity around just how much poverty is “so much”? The official American definition of poverty, underlying the Federal Poverty Level (FPL), is a family lacking the resources required to satisfy its minimum economic needs. That accords well with the meaning of “poverty” in everyday speech. Desmond seems to agree with that measure, but never gives his reader the number of people that he believes is “so much.” At the time he was completing his book in 2022, 37.9 million people (11.6 percent of the population in 2021) fell below the FPL, but he never says that 37.9 million people is his “so much.”

His only explicit answer to the question is a claim that America has more poverty than any other advanced democracy, based on data from the Organisation for Economic Co-operation and Development (OECD). But the OECD definition of poverty is different from the American definition, which Desmond adopts throughout the book except when he is justifying his “so much” claim. The OECD defines the poor in a country as families with incomes below one-half the median income in their country, but this measure describes poverty in terms of income distribution, not material condition. Consider that Americans whom the OECD deems poor have between 40 percent and 100 percent more income than people it identifies as poor in other advanced OECD countries. This relative measure leads to the paradox that a family in the United States with an income of $30,685 would be counted as poor, while families with income of only $14,141 in Italy or $21,904 in France would not be counted as poor. Using the same income standard for all countries shows the United States has at least 60 percent less poverty than other developed democracies.

After citing the deceptive OECD data to justify “so much” poverty, Desmond returns to using the FPL and a “supplemental” measure of poverty from the Census Bureau, neither of which shows any significant change over the last 50 years. He ignores at least two major research studies that show substantial declines in U.S. poverty over the last 50 years, with recent poverty rates between 1.1 percent and 4.0 percent, compared with the official 11.6 percent. These research studies argue that official American measures of poverty are overstated primarily because the Census Bureau does not count 88 percent of government transfer payments (subsidies) given to poor families. The book’s failure to even mention these two studies looks like an intentional exclusion to avoid the uncomfortable possibility that most of Desmond’s claims are wrong. (The author’s endnotes reference a journal article that demonstrates the superiority of one of these consumption-based measures, but he never even acknowledges the existence of that measure in the body of the book.)

Why is there poverty? / The book promises to identify the causes of poverty, but its “evidence” consists of anecdotes about individuals who were substance abusers, or sought jobs in the underground economy where they avoided Social Security taxes and thus were not eligible for disability benefits when injured, or lost their free housing because they physically threw their neighbors out of windows. These are the types of cases that constitute the 2–4 million people who are identified as poor by the improved methods that Desmond ignores. But his cases do not explain most of the 37.9 million who implicitly constitute his “so much” poverty.

The book offers several causes for why “so much” poverty exists. First, it claims poor people are distracted by scarcity, so they make bad decisions. This claim flies in the face of history. In 1949, 35 percent of Americans were poor, but before Lyndon Johnson’s War on Poverty started spending any money (1965), more than half of them had earned their way out of poverty. How is it that folks in the mid-20th century were not too distracted to raise their families out of poverty, but in the 21st century they are?

Next, he claims that working Americans look down on the poor as lazy and that somehow causes their poverty. He ignores the real evidence that government transfer payments to low-income households increased by 369 percent in constant dollars from 1967 to 2017, which enticed more than twice the proportion of low-income prime-working-age adults to choose the increased government subsidies rather than work. Almost half of the poor in Desmond’s universe are prime-working-age adults and 55 percent of them did not work at any time during the year. Those who did hold a job worked, on average, only 24 hours per week. This withdrawal from the labor force is about incentives, not labels.

Desmond objects to proposals that would ask poor individuals to behave differently, such as adopting the “success sequence” of (1) graduating from high school, (2) taking a full-time job, and (3) marrying before having children. Only 2 percent of people who observed all three elements experienced poverty, while 76 percent of those who violated all three were poor. He correctly points to some debate in the literature over the exact magnitudes involved, but there is no serious dispute about the importance of high school graduation and full-time employment in avoiding poverty. Desmond’s objection is not fact-based; it is metaphysical. He says that expecting a person to observe productive behaviors is extreme and “asking the person to just get a different life.” But millions of people have made the decision to “get a different life” and do what it takes to rise from poverty. Those who haven’t are the exceptions.

Another claim is that discrimination against Black people causes “so much poverty.” Discrimination is an important social issue, but it is largely irrelevant in the context of poverty. White people account for 67 percent of the poor, while Black people account for 23 percent. The White poverty rate has not declined over the last 50 years, while the Black poverty rate has systematically declined by almost half over the same period. If the most accurate measures of poverty are used, the difference between the Black and White poverty rates is now less than 1 percent.

Another proposed cause is that only one-third of applications for federal Disability Insurance under Social Security are approved, compared with one-half 20 years ago. But that ignores the full set of facts. Many more applications are being submitted for review today because the Social Security Administration loosened its standards and promoted the program. The net result is that, since 1972, the number of people receiving Social Security Disability payments has grown five times faster than the workforce that is paying the bills, causing the proportion of working-age people receiving disability subsidies to be five times greater, thereby reducing poverty.

Another putative cause of poverty is that rents have risen “too much,” consuming half of poor households’ spending. That is factually wrong; the reliable data show it is only about 25 percent. But even if true, his argument is irrelevant for explaining the amount of poverty defined, as he does, by income, not spending.

One final villain in the author’s list of causes is the decline in union membership. He argues that union discrimination against Black workers caused the union movement to lose its momentum and not enlist all the oppressed workers in the country. He then makes an unsupported leap to conclude that businesses sensed this weakness and removed worker protections. Yet Desmond fails to list a single protection that was removed. He merely imagines that union membership could have fallen only because government was not forcing workers to join unions.

He completely ignores two characteristics of labor economics. First, inefficiencies that unions forced on employers often led to economic failure or offshoring of work. At the same time, more workers voted against unionizing in free and open elections. Desmond tries to avoid that discussion by citing the factoid that, between 2016 and 2017, “the National Labor Relations Board (NLRB) charged 42 percent of employers with violating federal law during union campaigns.” There are three misleading components to this claim:

  • The wording sounds like a huge number of violations, but it was not 42 percent of the 10 million employers in America, but 42 percent of union organizing elections, of which there were fewer than 3,000.
  • The 42 percent figure is almost exactly the proportion of elections that unions lost. The standard operating procedure for unions is to file an unfair labor practice complaint in every lost election, and the NLRB in those years routinely converted each complaint into a charge.
  • The figure concerns charges, not proven violations. More than 90 percent of NLRB charges are “settled” without a finding. For the remaining less-than-10 percent without a negotiated settlement, the Board does not publish how many were proven to be in violation.

The book summarizes its kitchen sink full of causal theories by claiming that poverty springs from systemic exploitation of workers. It claims that all “privileged” people—which seems to be everybody who is not poor or near-poor—are guilty of exploiting workers and must adopt “poverty abolitionist” policies and behaviors. The rhetoric and sweeping generalization of this theory place it firmly in the company of other grand schemes that cannot be described rigorously or proven with data: systemic racism, existential climate change, and so forth.

Combating poverty / Desmond praises Johnson’s War on Poverty, writing, “Ten years after the first of these programs rolled out in 1964, the share of Americans living in poverty was half what it was in 1960.” But that is a distortion of history. Between 1960 and 1974, the poverty rate did decline from 22.2 percent to 11.2 percent, a drop of 11.0 percentage points. But during the previous 11 years, from 1949 to 1960, it had declined by 12.6 percentage points without any increases in government transfer payments, just normal healthy economic growth. Even within Desmond’s selected time period, poverty declined by 0.80 percent per year from 1960 to 1964 and by only 0.78 per year from 1964 to 1974 during the War on Poverty. Had poverty merely maintained its pre-1964 trend, it would have been 9.2 percent in 1974, not 11.2 percent. Furthermore, 1972 and 1973 were the low points for poverty during the entire 57 years after the War on Poverty was announced. Since 1973, poverty rates have simply oscillated with the business cycle, between 11.1 and 15.2 percent.

The official poverty rate failed to fall after the 1960s because government failed to count as income all the new money it gave to poor families and used upwardly biased price indexes to calculate poverty thresholds. If those data failures had been fixed, the measure of poverty would have fallen to a mere 1.1 percent. Desmond ignores those facts, so he then proceeds to claim that far more money needs to be transferred to low-income households. We can be reasonably sure that, like the extra money transferred in the last 50 years, these additional transfer payments will also not be counted by the Census Bureau, and Desmond and his progeny will continue to bewail “so much poverty.”

In addition to increased transfer payments, Desmond’s proposals constitute a patchwork of traditional progressive policies.

One proposal is to increase the minimum wage and index it to inflation. That is a complete misunderstanding of the facts. Some 40 percent of minimum-wage earners live with their parents. The median income for households with minimum-wage earners is almost the same as the median income for all households. The minimum wage is an entry to the job market, with two out of three minimum-wage workers getting a raise within their first year. The current federal minimum wage of $7.25 would create $15,080 in annual income if one worked full time. That is more than the $14,580 federal poverty level for a single individual, so the existing federal minimum wage is not a poverty wage. The principal cause of poverty is the choice not to work, not the pay from working. Among prime work-age adults in poor families, 55 percent choose not to work any hours during a year, and among those who do work, they work only 24 hours per week. Those workers earn $11.25 per hour, well above the minimum wage. Instead of making a fact-based case for why the minimum wage should exist at all, not to mention be raised, Desmond simply engages in an emotional tantrum, claiming, “Congress should outlaw undignified, even dangerous, poverty wages.” Where are the data?

The book also recommends using government to force employees to accept unionization by empowering the secretary of labor to compel sector-wide collective bargaining. Antitrust laws prevent companies in the same industry from colluding to set their prices, so why should it be okay for monopoly unions backed by bureaucrats to do the same? But the perversity of this proposal runs even deeper, abridging the equal rights of workers. Why should workers in a large New York plant be able to force conditions on workers in a smaller Texas plant in the same industry who do not wish to pay union dues or support union political positions?

One of Desmond’s solutions concerns housing. Here, he is onto something. Although none of these proposals would reduce poverty, one could improve the poor’s quality of life by eliminating many of the exclusionary zoning and building ordinances that limit the ability of property owners to build the housing of their choice, especially housing for low- and middle-income families. Unfortunately, Desmond takes this good idea and transforms it into a terrible one by proposing “inclusionary zoning”—that is, new housing legally mandated to go to low-income households. He claims, “While exclusionary zoning makes it illegal to develop affordable housing, inclusionary zoning makes it illegal not to.” His argument here is internally contradictory. When arguing for the value of removing regulatory barriers to affordable housing, he points out, “Once those plans [for affordable housing] are inked, it doesn’t take long for developers to bid on the job because they can make more money on multifamily complexes than stand-alone homes, even when they rent out a share of their units to low-income families.” If removing the barriers is so effective on one page, why does it fail on the next page?

The underlying structure of Desmond’s proposals is more government spending and higher taxes because he believes that income is a zero-sum game. He explicitly claims that the poor are poor because they have been oppressed by other people who are rising up the income distribution to become middle or upper class. He says, “Those who have amassed the most power and capital bear the most responsibility for America’s vast poverty: … corporate bosses who have spent and schemed to prioritize profits over people.” But economic prosperity is not a zero-sum game. Economic prosperity comes from hard work and innovation to create value for which people are willing to pay. The only part of income that is not earned by work or saving is the money that government forces us to disgorge through taxes for the benefit of government’s favored redistribution groups.

Despite his wide-ranging search to blame almost everybody for poverty, Desmond devotes but a single sentence to one powerful opportunity to help not only the poor but other low-income households as well: removing excessive government occupational licensing. In the 1950s, only one in 20 jobs required a license, but today it is more than one in four. There should be a clear call to repeal most licensing requirements, but his single sentence on the issue is tepid, and it seems to blame the businesses that hire licensed people rather than the government that created the licensing barriers in the first place.

Conclusion / Desmond promises his readers an answer to one question: “Why is there so much poverty in America?” He fails to do this on three counts. He provides no data for how much is “so much.” His theories of poverty’s cause are mostly warmed-over statist recitations of blame, with no supporting empirical evidence. And he demonstrates no understanding of the broad and significant research that has been done to answer these important questions. If he does not agree with that research, he should have offered evidence and logic to the contrary, but he simply ignores it.

In the end, his case is strictly emotional. Two pages from the close of the main text, Desmond sums up his whole approach: “We can feel it, the emotional violence we inflict on ourselves, knowing that our abundance causes others’ misery.” That is not helpful in addressing an issue for which we have large amounts of objective evidence, most of which he has ignored.