As Matthew Kroenig correctly observes early in this book, political science scholarship on the causes of nuclear proliferation has focused almost entirely on why states seek nuclear weapons, to the exclusion of careful analysis on why other states help potential proliferants in their quest. This is a critically important question, because without outside help, most states cannot acquire nuclear arms. Kroenig focuses in particular on state-sponsored transfers of uranium enrichement or plutonium reprocessing technologies to states that do not already have them, an act he calls “sensitive nuclear assistance.” He is skeptical of the conventional wisdom in policy circles, which tends to emphasize economic drivers in explaining why states provide such sensitive assistance. Instead, he argues, strategic considerations are more important: States provide sensitive nuclear assistance when it hurts their enemies but spares themselves.

Three specific hypotheses reflect this rule. First, a state is more likely to provide sensitive nuclear assistance to another state if it is incapable of projecting conventional military power over the recipient. This makes intuitive sense since such cases allow nuclear providers to empower recipients without weakening themselves. Second, “states are more likely to provide sensitive nuclear assistance to states with which they share a common enemy” (p. 4), or, as the author writes on several occasions, “the enemy of my enemy is my customer.” This makes intuitive sense since such actions allow states to impose costs on their enemies. Third, “states that are vulnerable to superpower pressure are less likely to provide sensitive nuclear assistance” (p. 4), since superpowers are presumed to dislike nuclear proliferation in general as it weakens their global power-projection abilities.

The heart of Kroenig's book is a careful statistical analysis that attempts to show that these three factors, rather than economic incentives or other variables, explain why states provide sensitive nuclear assistance. This is an impressive feat, particularly in a field that does not normally make good use of statistical analyses; in this case, the analysis utilizes an extensive new data set developed by the author. For those reasons alone, Exporting the Bomb makes an important contribution, and deserves the widest possible readership among nuclear specialists and students of international security. Kroenig finds that states that are capable of providing sensitive nuclear assistance are three times more likely to assist others over which they cannot project conventional military power than those over which they can; seven times more likely to assist states with which they share a common enemy than states with which they do not; and five times less likely to provide sensititve nuclear assistance if they are party to a security pact with a superpower (p. 62).

Yet there are three problems with the analysis. The author defines states as being capable of providing sensitive nuclear assistance as soon as they themselves have acquired sensitive nuclear technology. There is no reason, though, to believe that this is how things work. A state can receive a uranium enrichment or plutonium reprocessing facility and still have no ability to provide similar technology to others. Indeed, no state appears to have transferred either type of technology to others within five years of receiving that technology itself.

The author also excludes cooperation on nuclear reactors from his definition of sensitive nuclear cooperation, arguing that “there is a widespread scientific consensus that sensitive fuel-cycle facilities, such as uranium enrichment facilities, represent a direct nuclear proliferation threat, while other less sensitive, civilian technologies are relatively resistant to proliferation” (p. 12). This leads him to exclude, for example, Soviet assistance to North Korea and Canadian assistance to India in building plutonium production reactors. Yet building a reactor—not constructing a reprocessing facility—is the most difficult step in producing weapons-usable plutonium. By excluding these cases, Kroenig misses a large set of important transfers.

Finally, while Kroenig's analysis of possible strategic causes of nuclear cooperation is strong and careful, his assessment of alternative economic motivations is much weaker. The author uses per capita GDP and GDP growth rates to measure economic need on the part of potential suppliers. But states with low per capita GDP are less likely to have the capacity to provide sensitive nuclear assistance (even if they technically qualify as capable suppliers under his definition.) And low GDP growth rates say little by themselves about economic need: The United States, for example, has had a lower growth rate than North Korea in recent years. Kroenig also entertains the interesting possibility that trade-dependent states might be more inclined to engage in sensitive nuclear assistance, but finds little explanatory power there. A more careful analysis of economic motivations would require the inclusion of sensitive reactor sales (which are far more lucrative) and consider the political influence of the nuclear lobbies in supplier countries.

Kroenig follows his statistical analysis with several meticulous and engaging case studies (including cases where sensitive nuclear cooperation did not happen). The Pakistani case is particularly interesting, in part because it treats the A. Q. Khan network honestly for what it was: a state-driven enterprise, rather than the work of a rogue scientist. So is the long look at why France helped the Israeli nuclear program while the United States did not.

Toward the end of the book, Kroenig redeploys his new data set to tackle another question: “Does sensitive nuclear assistance contribute to the international spread of nuclear weapons?” (p. 151). His answer, which is intuitively appealing if not at all surprising, is yes. States that receive sensitive nuclear assistance, he concludes, are three times more likely to acquire nuclear weapons that those that do not (after controlling for other factors that might motivate states to seek sensitive assistance in the first place). Yet this is far from the strongest factor: States with a threshold level of industrial capacity are 23 times more likely to acquire nuclear weapons that those that lack such capacity, and states that are involved in enduring rivalries are four times more likely to acquire weapons than states that are not. At the same time, nonsensitive nuclear cooperation (essentially assistance in building a nuclear reactor) is found to be a poor predictor of proliferation.

This all adds up to one of the more interesting analyses of nuclear proliferation in recent years. But does it hold lessons for policy? That is more questionable. Kroenig argues that the United States should be on the lookout for sensitive nuclear assistance from states like India, France, and Israel, even going so far as to suggest that India might assist Taiwan or Vietnam, but he never looks seriously at whether any of the recipients would want these suppliers' help. He also claims that security specialists should stop worrying about the spread of civil nuclear power, since the spread of sensitive fuel-cycle technologies is a far stronger predictor of proliferation. Yet far more states are currently acquiring nuclear power technology; even low odds that any particular one of them will subsequently lead to proliferation can add up to a significant problem.

Perhaps the most important lesson for U.S. policymakers, though, comes from Kroenig's insight into the ways that other countries' strategists think about proliferation. U.S. policymakers, he notes, usually start from the assumption that everyone is as threatened as the United States is by the spread of nuclear technology. Reality, he observes, is much messier, with far more states standing to benefit from proliferation, or at least not to be harmed by it. Recognizing that uncomfortable fact—and adjusting U.S. strategy to reflect it—could be invaluable.