In 1983 the United State (U.S.) Intelligence Community (IC) conducted a study of intelligence judgement prior to “significant historical failures” of the IC in predicting world events affecting U.S. national security. As the documents below reveal, one of the failures studied was related to judgments of China’s nuclear program and first nuclear test in 1964.
Interestingly, while the IC happened to correctly predict the general timing of China’s first nuclear weapon test, this was accidental. Judgments of China’s nuclear weapons program were based on limited, uncertain information that was filtered through preconceptions favoring certain technical pathways. Specifically, the IC predicted that China would develop a plutonium (Pu) device based on an incorrect identification of a plutonium-generating reactor at a facility in Baotou (referred to in this report according to the Wade-Giles Romanization system as “Pao-tou”), and subsequent analysis of this facility predicted it could generate enough Pu to produce a testable nuclear device by 1964 or early 1965. Although this timing was generally correct – China’s first nuclear test occurred in October, 1964 – the nuclear device tested contained uranium-235 generated from a gaseous diffusion plan in Lanzhou. Given this, predictions about China’s nuclear weapons program were based on incorrect interpretations of sparse information that easily could have led to different (and additional) incorrect judgments and predictions.