Dr. Martin Hellman, Professor Emeritus of Electrical Engineering at Stanford, has recently initiated a major project to rigorously analyze the risks associated with the possession and use of nuclear weapons. This project, known as Defusing the Nuclear Threat (website here), is meant to catalyze research in this field and to motivate public action for nuclear policy changes based on the results of such studies. Last spring, Dr. Hellman published an article detailing the need for research on nuclear risk along with a preliminary risk analysis concerning the likelihood of the failure of nuclear deterrence and a descent into nuclear war.
Dr. Hellman’s effort is certainly praiseworthy in its goals. Dialogue on nuclear deterrence oftentimes fails to reach a high level of precision – discussions on our nuclear deterrent often do not refer to whom and what exactly we are deterring and gloss over rigorous analysis of those factors that make deterrence functional. A quantitative analytical approach, such as that suggested by Dr. Hellman, stands to make a major contribution to our understanding of how deterrence works. Moreover, deterrence is incredibly difficult to assess since it is measured in the absence of events (namely major wars and nuclear war in particular), and it is hard to make definitive, analytical conclusions based on convincing statistical analysis as to why a hypothetical event has not occurred. Dr. Hellman’s project, if it comes to fruition, may provide a means for developing robust metrics for deterrence, with risk of nuclear war being a much more nuanced barometer for evaluating deterrence policies than the current war/no war model that we rely on now. Consequently Dr. Hellman’s efforts ought to be encouraged.
That being said, there are some challenges facing Dr. Hellman’s current approach. First, a potential pitfall is in phrasing his risk analysis in terms of general “deterrence failure.” Analyzing deterrence requires a concrete identification of who is deterring whom. Thus, a deterrence failure should be considered in terms of the risks associated with various dyadic relationships (US-Russia, India-Pakistan, US-China, etc.) and the interaction of the two interacting deterrence policies. Some dyads may turn out to be more risk-prone than others, and thus contribute to the overall nuclear threat in different ways. The preliminary analysis, however, looks at world events irrespective of parties involved (granted his three examples are US-Soviet interactions, but they are treated as indicators of worldwide risk, not just the risk of nuclear war between those two countries).
Another pratfall lies in coding for all the variables that go into determining the success or failure of deterrence policies. Dr. Hellman’s preliminary analysis focuses on the incidence of nuclear war “trigger events,” referencing US-Russian relations, proliferation, terrorism, and other variables that are characteristics of the overall geopolitical system. Deterrence though is also dependent on variables such as the psychology and perceptions of individual leaders involved, particularly since policymakers have to come to their own conclusions as to whether or not their potential adversary’s deterrent is credible enough to dissuade a nuclear strike, should one ever be considered.
Finding a way to assess global risk probabilities based on how specific, unique individuals think and perceive the world is a formidable challenge. Overcoming these difficulties would prove a tremendous advancement in our understanding of nuclear deterrence – and we look forward to seeing this project’s results.