Joshua Goldstein has written a provocative book that distills a great deal of information about war, peacekeeping, the United Nations, mediation and diplomacy, and the peace movement into an exceptionally readable and accessible format for a lay audience. This is difficult to do, and it is done well here.

The book makes two important arguments—one about the decline of war over the long sweep of history, the other about the effectiveness of peacekeeping—both of which I agree with. However, it also suggests a causal connection between them that leaves me quite skeptical. It is hard to see how peacekeeping could explain the long-term downward trend in war given its relatively recent invention. Moreover, while peacekeeping has been shown to be effective at preventing the resumption of specific wars, its effects on the overall level of war in the world are neither straightforward nor yet known, as I explain in the following.

As its subtitle suggests, Winning the War on War argues persuasively that deaths from war have declined dramatically over time.Footnote ¹ Because we do not tend to examine war from a long historical perspective,Footnote ² we fail to appreciate this truly monumental fact enough, and Goldstein thus provides an important corrective. The book (Chapter 10 in particular) is to be commended for dispelling key myths that have taken hold in the discussion of war and the human catastrophes that accompany it. For example, the notion that civilian deaths used to make up about 10% of deaths from warfare but now make up about 90% is, as the author rightly points out, oft repeated but factually unsubstantiated. The idea that war is somehow now much more brutal than it once was is simply not historically accurate. Indeed, as Goldstein shows, war has become less brutal (in terms of overall numbers killed) and less frequent.

The book probably overstates the claim a bit because it does not take into account advances in health care.Footnote ³ War has become less deadly not (only) because there is less of it but (also) because all sorts of things have become less deadly than they once were, including tuberculosis, measles, gangrene, childbirth, and so on. Advances in health care mean that many fewer people die from war; those wounded in battle are much more likely to survive than they once were. To determine how much of the decrease in war deaths is due to a decline in war and violence, and how much to penicillin and other medical advances, we really need data on war casualties, including those wounded as well as deaths. But my guess is that we would still see a marked decline in violence, and so this is a quibble about degree, rather than an outright rebuttal to Goldstein's first argument.

Chapter 2 describes human existence, from prehistoric times to the present, as a series of periods, each less violent than the last (a few ups and downs notwithstanding). This represents a fundamental change in a key facet of international relations, and of the human condition. And it raises a critically important question: Why is war declining in frequency and deadliness?

Goldstein's second argument is that peacekeeping is much more effective than many think. Here, it is not at all surprising that I agree, as the author bases his argument in part on my own research showing that peacekeeping significantly reduces the chance that war will resume.Footnote ⁴ Chapters 3–6 provide a highly readable primer on the invention of peacekeeping during the Cold War and its development since then.

My only quibble here is the repetition (pp. 103–4) of the oft-made argument that the difference between the success and failure of peacekeeping is determined by the “consent of the parties” or their will for peace. If this were true, it would be hard to argue that peacekeeping had any causal effect; peacekeeping would “work” when it was not needed and would “fail” when it was needed most. The key to establishing a causal effect of peacekeeping is determining how the presence of peacekeepers alters the incentives of the peacekept.Footnote ⁵

Fewer people are dying from war than at any time in human history. And international society has created a well-developed (if far from perfect) system of diplomacy, mediation, and especially peacekeeping. But is there a causal relationship between these two things? Goldstein notes that “we cannot prove that peacekeeping caused the greater peace of the 1990s compared with the Cold War years, nor even that peacekeeping played the most important role in it” (p. 108). But he dismisses several possible alternative explanations, including the evolution of civilization and norms against violence, nuclear weapons, economic prosperity, the democratic peace, and the end of the Cold War. And he states that “the kind of explanation we want would kick in mainly after 1945, and would accelerate after 1989” (p. 44). The strong implication is that peacekeeping fits the bill. I am skeptical for two reasons.

First and foremost, the UN system of peacekeeping might be able to explain the decline in violence after World War II (when the UN was created and peacekeeping invented) and especially after 1989 (when peacekeeping was applied to civil war in earnest), but it cannot possibly explain the decline from prehistoric times to World War II. The system of peacekeeping and peacemaking is thus likely endogenous to the longer trend documented by Goldstein in Chapter 2. Peacekeeping may now be contributing to the continued decline, but it is simply too recent to explain the long sweep of history. We thus still require an explanation for this longer trend.

Second, it is doubtful that peacekeeping can explain all of the decline of violence even after 1945. Goldstein talks about the emergence of zones of peace in East Asia, China, and Europe (pp. 289–90). But it is very hard to claim that peacekeeping is responsible for peace in these zones, since there has been very little peacekeeping in those regions. Something else must be going on as well.Footnote ⁶

None of this is to say that peacekeeping has no effect in terms of reducing the level of violence in the system overall, only that its effects are necessarily limited in both time and space. Moreover, the net effect of peacekeeping, even recently and in zones where it has been used with frequency, is not straightforward and remains to be tested empirically. We have strong evidence showing that peacekeeping prevents the resumption of wars that have already occurred,Footnote ⁷ but it is less clear what prevents the outbreak of war in the first place, or whether peacekeeping shortens war once under way. Peacekeeping also appears to have changed the decisiveness of war in ways that have contradictory effects.

Studying the effects of preventative diplomacy is difficult—successful cases are by definition dogs that did not bark. But to my knowledge, there are no empirical studies suggesting that the preventative efforts of the international peacemaking operations that Goldstein credits have gotten more effective over time.Footnote ⁸ It is possible that what he dubs the “unarmy”—the nonmilitary forces that work for peace, such as diplomacy and mediation, humanitarian assistance, and women's role in peacemaking (Chapter 7)—is better than it used to be (or than its predecessors) at preventing the outbreak of war, but we simply do not know.

Whether peacekeeping shortens wars once they have started is also debatable. Michael Gilligan and Ernest Sergenti find that peacekeeping interventions are ineffective at ending wars and that the arrival of UN peacekeepers has no significant effect on the duration of war.Footnote ⁹ However, it may not be the arrival of peacekeepers during ongoing wars that shortens wars but, rather, the prospective availability of peacekeeping to maintain peace once achieved that allows some wars to end that would otherwise last longer. Another long-term temporal trend in warfare, one that is more clearly attributable to peacekeeping, suggests that this might be so. Let me elaborate.Footnote ¹⁰

Wars have become much more likely than in the past to end in a draw rather than a clear victory for one side and a loss for the other. This is true for both interstate and civil wars, although the timing of the shift is different for the two types of war. It occurs after World War II for interstate wars and after the end of the Cold War for civil wars, lining up exactly with the invention of peacekeeping after 1945 and its extension to civil wars after 1989.

This trend shows up in multiple data sets, but one suffices to make the point here. In the latest version (v.4) of the Correlates of War (COW) data, there is only one war that ends in a tie or stalemate from 1816 to 1945: the Franco-Turkish War (1919–21). This represents under 2% of all wars (and only 0.3% of all dyads) in this period. After World War II, 29% of all wars (and 37% of all dyads) end in a tie.Footnote ¹¹ Among civil wars, under 3% end in a tie in the period from 1816 to 1945. There is an uptick after World War II, with ties representing 15% of wars from 1946 to 1988, but the dramatic rise occurs after the end of the Cold War, with 44% of post–Cold War civil wars ending in a compromise or stalemate.Footnote ¹²Figure 1 shows the time line.

Figure 1

War outcomes over time

How might we explain the decline of victories in both interstate and civil wars, and the timing thereof? For war to end in a draw, two conditions must obtain: that neither side can defeat the other outright and that belligerents can agree to stop fighting, settling for a draw rather than pushing on for a decisive outcome. There is little evidence that belligerents' ability to defeat each other outright has changed over time. What has changed is that there is now a mechanism, peacekeeping, that can help solve the well-known commitment problemFootnote ¹³ that otherwise leads belligerents to continue fighting, even when the war has stalemated. This is the only explanation that can account for the change in outcomes in both types of war, as well as the timing of those shifts.

Alternative arguments, including changes in the relative capabilities of belligerents, war-fighting strategies, the offense–defense balance, patterns of intervention and balancing behavior, terrain, population growth, the issues over which wars are fought, or regime type, cannot explain the change in war outcomes adequately. All of these explanations fail either because the purported explanatory variable does not change over time in a way that fits with the shift in outcomes or because it does not actually help predict war outcomes. Peacekeeping is the only potential explanation that passes both tests, for both civil and interstate wars.

What are the implications of this finding for the relationship between peacekeeping and the overall level of war and violence in the international system that concerns us here? On the one hand, if this argument is right, there is a set of wars that end sooner than they would have otherwise, implying that the availability of peacekeeping prevents a great many war-related deaths.Footnote ¹⁴ Another war-reducing effect of this change may follow, although this is admittedly more of a stretch. Wars that end without a clear victory for one side are not particularly useful as a decision mechanism “to resolve issues that cannot be reconciled by other means.”Footnote ¹⁵ A decline in the decisiveness of war could thus conceivably contribute to war's obsolescence over time.Footnote ¹⁶ On the other hand, we know that wars that end with a clear victory lead to a more stable peace than do draws.Footnote ¹⁷ So peacekeeping, by allowing wars to end sooner, but less decisively, may leave issues to fester that would otherwise be resolved, eventually leading to repeated fighting that would not otherwise occur.Footnote ¹⁸ Peacekeeping makes war less likely to recur in the hardest cases, where there is no clear winner, but it may make war more likely to end with no clear winner, adding to the number of hard cases.

The net effect is thus hard to calculate. The short-term benefit of ending wars more quickly, along with the very long-term (possible) effect contributing to the obsolescence of war, may well outweigh the negative effect of indecisive, and therefore unstable, outcomes, particularly when one considers the well-established effect of peacekeeping in stabilizing otherwise unstable peace. But much of this is quite speculative. We know that peacekeeping makes particular wars less likely to resume, all else equal. And there is good (though less well-confirmed) evidence that peacekeeping makes wars more likely to end in a draw. But the purported effects on the duration of stalemated war, on the overall stability of peace in the international system, let alone on the obsolescence of war, have yet to be shown empirically.

Winning the War on War gets many things right: Fewer people are dying because of war; the purported vast increase in the ratio of civilian to military deaths is a myth; and peacekeeping is effective at preventing the resumption of war. I agree with most of its recommendations (Chapter 12): We should fund the UN and peacekeeping operations better than we do; the UN should create a standing army of peacekeepers ready to deploy quickly (though I do not necessarily agree that it should be staffed with troops from the Permanent Five members of the Security Council—some of the legitimacy of peacekeeping comes from fact that it is not staffed by Great Powers); we should continue to develop tools of conflict prevention; and we should fly the UN flag on May 29 for international peacekeeping day. I have mixed feelings about the book's argument about the responsibility to protect (R2P). Goldstein presents it as part of an ongoing trend that will contribute to a more peaceful world (pp. 322 ff). I support R2P on ethical grounds, and think it will lead to a more just world, but there is a tension here that goes unremarked. Another word for military intervention, even if its motive is humanitarian, is war. R2P may well increase violence rather than reduce it.

While I agree with the book's two core empirical claims, I remain quite skeptical of the causal story that (appears to) hang them all together: that UN peacekeeping and peacemaking can explain the long sweep of history toward greater peace. Peacekeeping cannot explain the decline in armed conflict from prehistoric times through 1945, and it is not yet clear how much of the decline since then can be attributed to peacekeeping. We may be winning the war on war, but we do not yet know why.