Jonathan Caverley’s Democratic Militarism is an ambitious, original, and thought-provoking book, whose goal is no less than to question some of our most basic assumptions about the relationship between democracy, wealth, and war.

More than two centuries ago, Immanuel Kant argued that by empowering ordinary citizens, representative government tends to be a force for peace. Caverley turns that conclusion on its head. He argues that democracy can actually lead to increased pressure for military spending and war, because the citizens who hold power in democracies can avoid most of the costs of war (particularly when the military is heavily-capitalized, there is no conscription, and the war is small). Instead, it is the rich who pay the costs of arming a military and fighting a war, because the rich pay the most in taxes. For this reason, Caverley argues, ordinary voters are more hawkish than the wealthy elites who will foot the bill. Because democracy empowers ordinary citizens, democracy can actually lead to high levels of spending on a capital-intensive military, which in turn leads to high rates of war. These high-cost militaries, moreover, are often ill-suited to the situation on the ground, blasting the enemy with expensive weaponry rather than emphasizing human-intensive strategies, since human life is more precious to the median voter than elite-funded hardware. These pressures should be especially strong when the country also features high levels of inequality, because as inequality increases, the median voter bears a smaller and smaller proportion of the financial burden.

The book tests the implications of these arguments about wealth, democracy, strategy, and war using a variety of methods and different levels of analysis. Analyzing public opinion data, Caverley finds that wealthy individuals are less supportive of military spending than their poorer counterparts. He also finds that the effects of income on support for defense spending are greater at higher levels of income inequality, and that countries with high levels of inequality are more likely to build capital-intensive militaries. Caverley next carries out three in-depth case studies to illustrate and evaluate his argument in various historical contexts. Each case study also explores numerous additional extensions of the theory.

Caverley’s arguments and evidence are provocative and important; if Caverley is right, then “the role of democracy in international politics is likely to be much less benign” than we might have thought (p. 43).

Perhaps the most important theoretical claim of the book, described earlier, is that the poor are more likely to favor hawkish security policies than the rich because they can typically evade the costs of arming and fighting. Caverley takes issue with the Kantian assumptions about who bears the burdens of war. Because of redistribution, the median voter can build an army and fight a war without feeling much pain, economic or otherwise. As Caverley puts it, “a suboptimal, militaristic grant strategy can result from rational calculations on the part of the average voter, and no marketplace of ideas will cure it” (p. 6).

The insight that the average voter does not pay much of the cost of funding and fighting a war is important and plausible. However, where I disagree with Caverley is that this would lead the median voter to favor higher rates of military spending than a wealthier counterpart. Caverley sidesteps the fact that the level of defense spending that voters choose is a product of two factors: the tax rates for different economic strata, and the proportion of those taxes that are devoted to military spending (“guns”) rather than direct forms of social welfare (“butter”). (Though some note that modern wars are typically financed through debt or expanding the money supply, as opposed to “war taxes,” making the costs less visible, a general point that Caverley acknowledges. ^{Footnote 1} ) The median voter might have economic incentives to raise taxes on the wealthy, but as scholars such as Kenneth Scheve and David Stasavage have shown, concerns about efficiency and fairness also drive attitudes about how much the wealthy should pay in taxes. These attitudes create constraints on support for raising taxes, and limit the size of the government’s budget. ^{Footnote 2}

In light of constraints on government spending, voters must decide what kinds of expenditures to prioritize. Should strong social-welfare programs take priority over an expensive military, or should guns take precedence over butter? Caverley does not adequately consider the possibility that if the government budget is finite, middle- and low-income voters have material incentives to prioritize social spending from which they should expect to benefit—such as health care, education, infrastructure, unemployment benefits, retirement, etc.—over a capital-intensive military that will fight far-away insurgencies. High-income voters would have reason to prefer a lower top marginal tax rate, but might favor a higher proportion of government spending to be funneled toward defense rather than social programs for the poor. Moreover, income could affect a person’s moral views about the proper role of government spending. Lower-income voters might find it particularly morally objectionable to spend government funds on far-away insurgencies when friends or family members are suffering from problems like poverty, poor health, and lack of access to education. For these reasons, the net effect of personal income on preferences about levels of defense spending do not seem as clear as Caverley maintains.

Among the book’s many great strengths is the wide array of evidence it considers. Caverley devotes significant effort to evaluating his central claim that low- and middle-income voters are more hawkish on defense spending than the wealthy, looking both cross-nationally and within the United States. Another chapter explores variation in military capitalization as a function of inequality. Three case study chapters subject the hypotheses to further tests by exploring how wealth and inequality contributed to British imperialism, the U.S. war in Vietnam, and Israeli foreign policy.

While the book’s empirical portions are creative and at many points persuasive, there are some questions about the evidence that future research should address. Chapter 3, for example, assesses the hypothesis that the poor are more enthusiastic about defense spending. As Caverley acknowledges, disentangling the effects of wealth and inequality on attitudes about military spending and force structure is difficult. Caverley’s solution is to control for a host of possible confounders, including education, political party (coded from left to right), gender, and age. I applaud the attention to possible omitted variable bias, but some of the variables raise questions. Recall that Caverley’s goal is to assess the effect of personal/ household income on attitudes about defense spending. As scholars such as Joshua Angrist and David Pischke have pointed out, if we are trying to estimate the causal effect of a predictor or “treatment” variable (in this case, the person’s income), we should not control for other variables that are themselves a function of the treatment, and only control for variables that are fixed at the time the treatment is determined. ^{Footnote 3} The problem is that a variety of studies have suggested that the treatment, wealth, affects at least one of the control variables, right-wing political views. Moreover, if partisan affiliation is chosen in part on the basis of its stance on defense spending, then part of the dependent variable is already baked into the control variable. I would be more confident in the results if Caverley could show that the results hold even when partisanship is omitted from the analysis.

Caverley also concedes, in Chapter 4, that the relationship between inequality and capitalization is more complicated than predicted by his theory; inequality predicts greater capitalization up to a certain level, but at high levels of inequality, the relationship reverses. When Caverley does not include a square term for inequality, the positive relationship he hypothesizes disappears. While Caverley is admirably transparent about this in the chapter, at times he downplays this issue in the broader interpretation of his results.

These issues, however, should not detract from the overall conclusion that Caverley has written a remarkable and stimulating book. I encourage scholars and policymakers to read it carefully, and to take seriously the claim that inherent in democracy is a pressure for an expensive military and belligerence. This book is a significant achievement and deserves to be read and debated throughout the field.