This is a troubling book. In large part, that is the nature of the subject: the murder rate in the United States, what Roth calls America's Homicide Problem, is one that has vexed historians and criminologists alike. The problem itself is straightforward: the United States, historically, has had a murder rate far higher than comparable countries in Western Europe. But a convincing explanation for the problem has been hard to find, notwithstanding the fact that many scholars have tried to explain why that is case. The American Historical Review, and Social Science History have had symposia on the question. Roger Lane, Eric Monkkonen, Jeffrey Adler, Jack Kenny Williams, Giles Vandal, and others have written monographs and articles exploring the issue. These studies have suggested any number of causes, including cultural and social forces, regional, ethnic or racial proclivities, gender norms, and the structure of the legal system. But none of the explanations has been completely convincing, with the result that the problem of homicide in the United States continues to trouble scholars as much as it confounds public officials.

Roth's study American Homicide, which spans the nation's entire history and all of its regions, is the latest and most comprehensive approach to the problem. Yet the book is troubling because of the answer it gives. Roth argues that homicide rates are low where “government is stable and where legal and judicial institutions are capable of redressing wrongs and protecting life and property” (469), and they are lowest where “citizens have confidence in government and the social hierarchy and are bound to one another with a feeling of patriotism, empathy, or kinship” (ibid.). That paints a disturbing picture of the United States, suggesting that the “American problem” reveals that goodly portions of the American people have never had trust in the government, respect for its institutions, or charity toward one another. The problem of American homicide becomes, in Roth's telling, the problem of America. After reading the book, the reader is no longer surprised that Americans kill one another as much as they have and continue to do, so much as he or she is surprised by the fact that the United States has managed to exist as a nation given the extent to which the American people hate the government, its institutions, and one another.

In his conclusion (469–74), Roth offers little hope that much can be done. Although he notes that murder rates have fallen from their highs, he also notes that homicides in the United States remain higher than in other Western countries. And he asserts that the reasons for these high rates will continue to exist. Tensions between husbands and wives, which led to spousal homicides throughout the nation's history, have different sources, but remain alive, promising that domestic violence and death will continue. So too, divisions between social, religious, and cultural groups will continue to separate American citizens from one another, and drive people to homicidal violence. None of the solutions offered, not continuing expenditures on the war on crime, not more and better policing, not bigger and stronger prisons, will prevent killings that hearts and minds determine are necessary. Roth argues that public officials can neither change these poisonous emotions with policy, nor stop hatred or distrust, and the murderous impulses they provoke, with rhetoric or popular appeals. Homicide rates may decline sometimes, as they have in the past, but ultimately, Roth suggests, the American problem with homicide is permanent and insoluble.

It is a deeply troubling conclusion. But I am not convinced.

I do not doubt that the homicide rates in the United States have been high, historically. Nor do I doubt the evidence Roth marshals on behalf of his claim. The homicides he describes, and the dynamics they reflect, are familiar enough to people familiar with homicide studies, or with urban histories, immigrant history, history of black-white relations, or studies of domestic relations and violence.

But in three respects I am not persuaded. First, Roth's own evidence suggests that racial, ethnic, religious, and gender tensions are neither permanent nor, inevitably, deadly. Minority groups, like African Americans, who were subject to (and sometimes perpetrators of) homicide in some regions were able to live peacefully in others. Religious tensions that marked the Irish and Italians for homicidal violence at some points in time and places have ceased to be salient. Even domestic violence patterns show the potential for decline, giving rise to the hope it may ultimately disappear. The wide spread of Roth's data can be read collectively or in terms of specific patterns. Read collectively, as Roth does, it suggests a country without hope; read specifically, his data seems to me to suggest that hatred and distrust are neither universal nor permanent. If those are the causes of the American problem, one reading of Roth's own data is that the problem can be solved.

Second, I am not convinced that hatred and distrust play the role that Roth gives them. To be sure, demonizing the other or despising the government and its works are probably neither healthy nor conducive to a nonviolent society. But other factors play a role in Roth's story as well. Roth notes, for example, that homicide rates increased when there was economic instability. The homicide rate rose in the early years of the Great Depression (440), declined during periods of high employment in the 1950s and 1960s (452), and then rose again as the economic troubles led to job loss and declining employment opportunities from the 1970s through the 1990s (451). The resurgent economy of the 1990s, even if it was based on unstable bubbles, was marked by a steep decline in homicide rates (451). That suggests what histories of race and ethnicity in the United States have posited: racial and ethnic tensions often are symptoms of larger economic issues. Perhaps, the reason the United States has a higher homicide rate than most countries in Western Europe does not reflect greater ethnic tensions, something that few European countries have avoided over the past thirty years, but the greater uncertainty and anxiety caused by the stronger free market and weaker welfare state in the United States.

Finally, Roth's evidence suggests that the American problem may not be a high national homicide rate. Instead, his evidence raises the possibility that the cause of the American problem is the need to turn an episodic problem reflecting very specific local tensions into a national issue. Homicide rates are not high everywhere in the United States, nor have they been. Roth's data suggests that homicide rates are not consistently high within racial or ethnic groups regardless of income, place, occupation, or expectation. Nor does his study demonstrate that Americans of all races, classes, statuses, or residences distrust the government or one another equally, or consistently turn to homicidal violence when they feel threatened. Maybe the reason we have not been able to come up with a convincing explanation of the American problem is that there is no American problem to solve.