Americans, you may have heard, are dying "deaths of despair": suicides, drug overdoses, and alcoholic liver disease that are together driving down the nation's life expectancy. The concept has been bandied about since it was coined by Princeton economists Anne Case and Angus Deaton. After five years, Deaths of Despair and the Future of Capitalism is the pair's book-length offering on what they believe is driving the crisis.
Case and Deaton are accomplished social scientists. But although their book highlights real issues faced by America's working class, the pair do not overcome the burden of proof in explaining why the country has experienced a mysterious death spike. They did the world a service by identifying the trend; their elaboration, however, is deeply unsatisfying.
Deaths of Despair can be conceptually divided into three sections: the what, the who, and the why. The "what"—the empirical phenomenon around which the book is organized—is likely to be the concept most familiar to readers, but a rehearsal of its history may be useful.
Five years ago, Case and Deaton noticed a disturbing trend. Mortality rates, which usually fall steadily in developed nations, had plateaued or increased in the United States beginning in the late 1990s. They looked more closely at the data and identified that the trend was primarily driven by an increase in deaths among middle-aged, non-Hispanic whites. The largest share of that increase was deaths from suicide, drug overdose, and alcoholic liver disease—a proxy for drinking oneself to death. Because all of these deaths are in a sense "self-inflicted" and can indicate an unhappy mind, Case and Deaton labeled them collectively "deaths of despair."
Deaths of Despair is a book about what caused this increase, so it's worth asking—do all of these causes of death empirically go together? As I've written previously, the answer is probably no. The three distinct causes of death started rising in different parts of the country and have different proximate causes. Most "deaths of despair" today are drug overdose deaths, and most drug overdose deaths are the result of a shock to the illicit drug supply in the form of hyper-potent synthetic opioids, particularly fentanyl. Rising suicide rates are not obviously related to a general increase in despair, and exhibit a number of peculiar traits that Case and Deaton largely ignore.
The demographic dynamics of the crisis are also more complicated than the authors let on. While rising deaths among middle-aged whites drive the overall mortality decline, suicide and OD death rates are up across age groups, and drug ODs are up among blacks as well as whites. A truly parsimonious theory would explain these deaths as well, but Case and Deaton largely ignore them. They also underspecify the problem. Statisticians Andrew Gelman and Jonathan Auerbach noted back in 2015 that breaking down white middle-aged all-cause mortality trends further revealed that the trend line was only going up for white women in the southern United States. (Case and Deaton have acknowledged this criticism, although it goes unmentioned in the book.)
The book's second section more closely examines who, among middle-aged whites, is dying deaths of despair. In short, the answer is: those without a college degree. Case and Deaton show that over the last several decades, not only have those without degrees borne the brunt of the rise in mortality, but they have also faced worsening health outcomes, increasing chronic pain, and declining access to both economic and social capital. Their point is that something about the conditions of the "working class" has changed to cause the sudden spike.
There is an important elision made here, insofar as Case and Deaton more or less synonymize "no college degree" with "working class." Knowing the data they are working with makes clear why: Death records list highest degree achieved, but no other indicator of socioeconomic status. In other words, Case and Deaton's conclusion is partially a product of the constraints of their data.
Yet they largely fail to account for the fact that the average 30-year-old without a college degree in 2019 is qualitatively different from a 30-year-old without a college degree in, e.g., 1969. As the share of individuals with a college degree grows (16 percent of 30-year-olds in 1969 vs. 40 percent in 2019), the population of non-degree holders becomes more "negatively selected": As it becomes easier to get a degree, the population without one has a lower average skill level, which in turn predicts lower social and economic outcomes. One paper from last year actually attempted to account for this and found that deaths of despair among whites are driven primarily by high-school dropouts—a group receiving little to no particular attention from Case and Deaton.
This said, it is likely that the declining economic position of the working class has contributed to the rise in suicides, drug overdoses, and alcohol-associated deaths. In particular, communities' exposure to trade normalization with China, especially through a decline in manufacturing jobs, is tightly related to an increase in suicides and accidental poisonings (although less so to alcohol deaths). But that increase is not necessarily causal—while counties experiencing economic decline saw bigger increases in OD deaths, for example, the relationship is "mostly explained by confounding factors."
This question of causation—did declining working class standards cause the rise in deaths of despair?—is, or at least ought to be, at the heart of Deaths of Despair. Case and Deaton, however, waive the standards of causal inference now in vogue among economists, dismissing rigorous mathematical techniques as "of little use to us here in describing a slowly evolving and large-scale disintegration that involves a historically contingent set of forces, many of which interact." This sentence is, in effect, an abdication of responsibility: All that goes after need not be causally linked to deaths of despair, only conceptually related in the authors' minds.
The fundamental "why," Case and Deaton argue, is a system of "upwards redistribution" from the poor to the wealthy. Unlike either poverty or inequality—which are summarily dismissed as being the real problem—"upwards redistribution" seems to name a mélange of objections, rather than one single, measurable thing. The sudden spike in deaths of despair are caused, variously, by the rising cost of health care, too many lobbyists, concentration of the airline industry, not enough unemployment insurance, automation, Facebook, noncompete clauses, and so on.
Case and Deaton do not attempt formal analysis linking any of these causes to rising deaths of despair, but rather seem happy, in many cases, to draw sweeping conclusions based on single incidents of malfeasance. They freely acknowledge that, for example, "there is currently no smoking gun that links deaths of despair to a lack of safety nets, either within countries or between them," but nonetheless pin much of the blame for America's unique deaths of despair problem on our lack of universal health care and, in particular, the wickedness of the medical industry.
Their solutions are similarly broad. They call for universal insurance, in particular arguing that "further expansion of Medicaid would be helpful, for drug problems as well as other medical care"—ignoring the evidence that states which took Medicaid expansion under Obamacare actually saw higher rates of drug OD deaths than those that did not. They recommend further crackdowns on prescription opioids, even as deaths involving those drugs have fallen—the book is entirely silent on what to do about fentanyl ODs. Some prescriptions are distressingly vague, as in a call to "reform corporate governance" with few details as to what that would mean.
It is entirely possible that the condition of the working class in America is exacerbating the deaths of despair problem. The issue is that it is also entirely possible, even likely, that other factors—like the surge of illicit fentanyl from China and Mexico—play a much bigger role. Case and Deaton could have ruled out these possibilities by seriously grappling with the numbers, but they do not.
There are already a number of other, better books on the state of America's working class. It is good that Case and Deaton identified the "deaths of despair" phenomenon in the first place. But in order to tread new ground, a book on the topic would have needed to take the numbers far more seriously and address the problems outlined above. That Deaths of Despair and the Future of Capitalism does not do so makes it a disappointing contribution to solving a mystery that grows deadlier every year.