In this welcome edited volume on nuclear nonproliferation, the contributing authors develop the concept of security “assurances,” that is, “promises to respect or ensure the security of others” (p. 1). Most generally, the authors ask whether “negative” assurances—here, a commitment to refrain from nuclear attacks upon a country—or “positive” assurances—a commitment to come to the aid of a country under nuclear attack—can convince that country not to acquire or retain nuclear weapons. Thus, the authors recognize importantly that preventing proliferation, like preventing the use of nuclear weapons, is not simply a coercion problem. As editor Jeffrey Knopf notes, assurance also undergirds deterrence: A party cannot deter another unless the latter believes—is assured—that its actions will prevent the former from acting on its threat.

The volume gives emphasis, then, to the (too often neglected) point that “cooperation” is required to halt an incipient nuclear-weapons program, and backs it with convincing evidence. The case evidence is taken to suggest that negative assurances helped halt the weapons program in Libya and, maybe, the retention of nuclear weapons by Ukraine, and that positive assurances helped halt nuclear proliferation by South Korea, Japan, and perhaps Sweden. The volume correctly acknowledges, however, that the effect of assurances is typically “modest” rather than “decisive” (p. 6); “that when states are faced with a critical threat from a third party, negative assurances seem to fade into insignificance” (p. 279); that various confounding factors mute the effects of assurances (including domestic politics, broader strategy, and the credibility of actions); and that profound trade-offs are required when negative assurances are provided to one conflicting party with positive assurances to another (p. 7).

In some ways, the volume is a model of inquiry. To judge when assurances might work, the authors employ 13 hypotheses that are drawn from relevant theoretical literatures. These hypotheses include whether the assurances are explicit or legally binding; backed by forward-deployed troops; “strong enough to overcome cognitive biases”; “tailored to take account of unique features of the target state's culture, decision-making procedures, and leadership concerns”; and used “in a way that alters internal debates in the target in a favorable direction” (p. 32). The chapter authors stay on message when evaluating their respective cases; they conclude by commenting on the validity, invalidity, or partial validity of the hypotheses, and they supply testable (inductive) hypotheses when the initial hypotheses prove deficient. James Wirtz, a proliferation expert, renders the final judgment about whether the hypotheses have held.

That said, Security Assurances and Nuclear Nonproliferation is not without limitations, as the authors—to their credit—do often recognize. Scholars should take note of these limitations when seeking to advance this important program of research.

Both the hypotheses and the testing strategy are problematic in some respects. First, the hypotheses are, at times, tautological. Because some of the independent variables are not clearly operationalized, the hypotheses are too easily validated or invalidated on the basis of whether proliferation did or did not occur in a given case. This is true, for example, of the hypothesis that effective assurances must overcome cognitive biases. What exactly are these biases, and could we know that they are overcome should a country retain its weapons capabilities? Likewise, the test of whether assurances are tailored “to take account of unique features” appears to come down to whether assurances actually work. A valid test of the hypothesis requires a standard for judging the adequacy of tailoring apart from the value of the dependent variable.

Second, the binary dependent variable—countries either did or did not pursue nuclear weapons—sacrifices some explanatory leverage. What about cases—like Iran, perhaps—of a country choosing to keep its options open by stopping just short of a nuclear weapons capability or possessing but not testing a nuclear weapon?

Third, the hypotheses are given equal standing when the independent variables in some of them arguably stand as necessary or permissive conditions in others. Although the authors hypothesize (Hypothesis 1) that “assurances are more likely to be effective when a target state's interest in nuclear weapons is driven to a significant degree by security concerns” (p. 32), the remaining hypotheses seem to assume the presence of a security threat. For that matter, the host of potential causal factors that the authors identify in their case studies could conceivably contribute to the outcomes in others. For instance, part of the story about Libya is that it lacked requisite indigenous nuclear weapons capabilities, suffered severely under a multilateral embargo, had a leader who felt vulnerable to the threat of a US-imposed regime change, and might have seen the United States as a potential ally against a radical Islamic threat. Might these same variables explain the workings of the key independent variables in other cases?

Fourth, the hypotheses center narrowly on assurances that relate to nuclear weapons use, and nonuse, although the authors recognize that positive nuclear weapons assurances are often used to allay concerns about conventional threats and that critical negative assurances often have nothing to do with nuclear weapons: for example, threats to regime survival preoccupied Muammar Gadhafi, not a US nuclear strike on Libya.

Fifth, as the authors recognize, seemingly validated relationships are potentially spurious. Because positive assurances often come with economic and military aid, it might only appear that the assurance, provided by aid, led a country to renounce the nuclear option. As the author of the Swedish case notes, “the threat that the United States might withdraw assistance was more important than the promise of new assistance” (p. 239).

Various issues also arise in the case studies. First, because authors of a case-by-case analysis tend to think of their cases in isolation from others, the findings rest primarily on temporal rather than cross-national variation. Thus, the author of the Iran case concludes, “In the main, the hypotheses reviewed here are not sustained…. One cannot help but suspect that even if Iran had faced no credible external threat, it still would have pursued a nuclear program, be it under the Shah or under the Islamic Republic” (p. 127). But does not the case provide strong support for Hypothesis 1, linking the success of assurances to security concerns? Second, as Knopf acknowledges, the cases betray a selection bias—apparently toward failure: They were selected for success and failure but “all involve countries that were at some point deemed serious proliferation risks” (p. 6). Whether or not that bias is mitigated by the editor's purported selection of cases, as well for the relevance of positive and negative assurances (and, thus, cases in which assurances might work), the selection strategy clearly oversamples proliferation. The study concludes, then, that assurances are only moderately effective, although nuclear weapons status remains relatively rare among contemporary states. Implicitly, the analysis downplays the role of multilateral negative and positive assurances that could sway some states—indeed, might reinforce a “nuclear taboo” that dissuades most states—from acquiring these weapons.

Despite these deficiencies, which reflect the challenges of conducting research in this field, the contributors deserve substantial credit for developing a useful concept, exploring its implications (in negative and positive forms), pulling the concept into explanatory propositions drawn from various theories, and testing the hypotheses on a set of relevant cases. They also deserve credit for producing a work that will provide a valuable reference on a topic with enduring scholarly and policy relevance.