CIA Cold War China Forecasting
What a prediction from 1973 tells us about predicting China today
Want to help figure out how to manage risks from frontier AI systems? The Centre for the Governance of AI is currently hiring for junior and senior research positions. They’re a non-profit organization based in Oxford with team members based in the US and all over the world. They do deep applied research on issues like responsible AI lab behavior, the regulation of the most capable AI models, and compute governance.
I (Jordan) have been super impressed by GovAI’s work. Their analysis has really helped ground the global discussion on AI with serious and informed analysis. Plus, their researchers make great podcast guests! Lennart Heim’s show on the geopolitics of GPUs was one of my favorites this year, and interactions I’ve had with their staff give me every indication they’re good people you’d enjoy working with. Personally, I think this field really needs more China expertise, and I would love to see someone from my audience tackle these questions full-time.
For more information, go to governance.ai/opportunities. Apply by August 9.
Dylan Levi King guest-authored the following piece.
In the summer of 1973, the CIA’s Office of Political Research was tasked by the Directorate of Intelligence to assess China’s future. They produced a brief report in 1974 titled, “CIA Analysis of China in 1980-85, and in the Year 2000.” Prepended with a memo by Secretary of State Kissinger, it appeared on the desk of President Ford toward the end of August. In 2009 it was declassified and made available through CREST (CIA Records Search Tool), where I discovered it on a routine crawl.
The once-top-secret report is a window into CIA futurology in the middle of the Cold War. Unlike some of the more abstract theorizing or academic vivisection of ongoing events (check out CREST’s China Collection), this memo was aimed at policymakers. Moreover, its analysis attempted the audacious feat of predicting events years beyond the retirement of everyone involved.
The final result is a halfway-reliable picture of what came to pass! Maoism did turn out to be flexible; a military alliance with the United States was possible; and future leaders in fact pursued what could be called a depoliticized authoritarianism. Analysts and policymakers today would do well to understand both what they got right and, more importantly, why they were able to.
To be sure, we already know where China ended up in the early 1980s and ultimately by 2000. And Kissinger noted in his memo to Ford that “some of the conclusions reached by this type of ‘futurology’ inevitably are controversial”; the analysts themselves enumerated many limitations in the introduction; and the report plainly gets some material predictions wrong.
For instance, the analysts struggled to discern who was going to assume power after Mao Zedong — they figured that Zhang Chunqiao 张春桥, Wang Hongwen 王洪文, or Li Desheng 李德生 would compete among each other after Mao’s death:
But the CIA analysts still got a lot right.
For starters, they were correct about the flexibility of Maoism and the importance of nationalism:
Forseeing future modifications of Maoism, forces of “great-Han nationalism,” and depoliticized authoritarianism between 1980 and 1985 was not off the mark. Though they don’t pick the pragmatists (like Deng Xiaoping) or the industrialists (like Hua Guofeng 华国锋) as the ones in power, the analysts were nonetheless aware of the waning power of the Cultural Revolution ideology and the alternatives that had gained traction since the late 1960s. Whoever assumes power in the 1980s, the analysts concluded, may have a radically different vision for the future, but will still work toward strengthening the country before implementing them.
The predictions for the year 2000 — then twenty-six years into the future — were flouted, as they did not consider the possibility of the Soviet Union’s collapse. (To be sure, very few within or without the CIA took this possibility seriously, especially in 1973. It is now uncharitably considered — along with the Cuban Missile Crisis, the Iranian Revolution, and 9/11 — one of the great American intelligence failures.)
But even while failing to see a rupture in the Sino-American alliance and the full-tilt marketization of the Chinese economy, they correctly imagined China by the year 2000 transitioning to a mixed economy in part due to competition with the revisionist Soviet Union.
They also predicted that a “hatred and fear of the USSR” will push China into a military alliance with the United States, negating the possibility of a conflict with Taiwan:
I can’t imagine we could have done any better than the Office of Political Research did in 1973. But should we be able to do better today? After all, we have far more data and computing power. True, machinations within the top leadership of the Communist Party of China remain opaque, but we have access to a much more detailed picture of the country than analysts did fifty years ago — and yet popular predictions about China over the previous decade have often proved nowhere near as reliable.
I would submit that our predictive capabilities today suffer because ideological statements and opinion have, in most contexts, replaced analysis. Unlike in 1973, China is the chief economic and military rival of the United States. Politicians, think tank intelligence assessors, private interest groups, and intelligence agencies, therefore, see the value of broadcasting extreme opinions on the future of the country. Prophesying collapse is a safe, patriotic choice; prophesying future Chinese omnipotence is a good way to further domestic goals — if, for example, you want to end regulation on and secure funding for AI projects, describing an arms race with China is a great option.
To attempt the same trick that CIA analysts pulled off in 1973 — broadly predicting China’s economic organization, foreign-policy orientation, and military capacity — requires both cynicism as well as optimism. Neither the Sino-futurist boosters or collapsarians have the intellectual sovereignty to balance the two.
It’s not that the Chinese-language graduate in business casual at Langley in 1973 was more capable, and he definitely didn’t have access to more impressive data. Indeed, based on declassified material in the CIA archives as well as the CIA’s own history of their Office of Strategic Research, a lack of SIGINT coverage and limited satellite imagery meant that intelligence from China seems to have been based on the atmosphere of ongoing negotiations, People’s Daily articles, and rumors from diplomats.
But access to impressive data was not why their reports won. These reports were reliable because they didn’t have to be anything else. Even if CIA competency in the Cold War is up for debate, they had most of us beat — and I bet they still do.
If you were one of the authors of this intelligence report, please reach out! We’d love to have you on ChinaTalk. Shoot a note to jordan@chinatalk.media.
Paid subscribers get access to more incisive analysis about this CIA report, including:
US assessments of China’s nuclear capabilities and willingness to cooperate in international arms-control efforts;
How the CIA thought Taiwan’s and Korea’s dictatorships could shed light on the direction of China’s authoritarian system;
And China’s appetite for territory — with some prescient analogues to today (Hong Kong, India, South China Sea…).