Challenges of Quantitative Nuclear Proliferation Modeling

There has been voluminous body of literature on the causes of nuclear proliferation. Based on this, the quantitative modeling efforts of state’s nuclear proliferation risk have been developed. Such quantitative tools could prove to be useful to provide early warning by processing very extensive nature of information involved in a systematic way. However, use of predictive models has had mixed success with significant probabilities of producing false alarms. In addition, some works suggest the fundamental limitations that decrease the robustness of the model. They can be categorized as: 1) Uncertainties in proliferation history datasets, 2) The limitations of regression and survival analysis, 3) Poorly considered correlations among the proliferation determinants, 4) The dynamics aspects of decision making surrounding nuclear weapons development, etc. Based on examining how a country specific situation leads to specific scenarios of nuclear proliferation, this paper reviews past efforts to find the determinants for the nuclear proliferation decision, evaluates the influence of the topics on the modeling results, and projects future research agenda to overcome the issue.