Rebuilding Trust in a Prisoner's Dilemma
Or, a Marriage Counselor's Advice to Beijing and Washington
For several years now, I’ve thought of the relationship between the U.S. and China in terms of that well-known staple of game theory, the Prisoner’s Dilemma.
In case you’ve forgotten how the Prisoner’s Dilemma works, a quick refresher: You have two people nabbed by law enforcement — let’s call them Joe and Jinping — for a crime they committed together. They’re kept in separate rooms and questioned, unable to communicate with one another. The cops don’t have the goods on them for the serious crime they committed — a crime that could very well send either of them away for, say, five years. Both Joe and Jinping know this: they’re aware that there’s weak evidence on the big felony count.
But they also know the cops have them each dead to rights for a lesser crime that would result in, say, a one-year term for each. If both stay shtum and refuse to rat out the other, they’ll each do just that one-year sentence — and for our Joe and Jinping, being considerably younger than certain others of those names you may be thinking of, a year isn’t all that bad. If Joe rats Jinping out (or “defects,” in the language of the game) while Jinping refuses to talk, then Jinping goes to the big house for five while Joe skates. Same the other way: Joe will do five if he keeps his trap shut but Jinping sings. And if both of them snitch, they both do three years: the worst outcome in terms of combined jail time — what’s known as the Nash equilibrium.
You can see where I’m going with this. No, it’s neither a particularly original1 nor an especially insightful comparison. There are, after all, important and obvious ways the U.S.-China relationship doesn’t fit: Beijing and Washington are not, for instance, locked in separate interrogation rooms: they can signal to one another their intention to act in good faith or to defect, they have histories and well-documented public pronouncements that analysts at the respective countries’ intelligence agencies constantly pore over. There’s also a significant disparity in power and wealth — and the fact that the U.S., invested as it is in a status quo where the U.S. enjoys primacy, has arguably more to lose.
All this complicates the calculation for both sides. Still, I do find the Prisoner’s Dilemma a useful way to frame the situation because it highlights what, for me, is the big question in the bilateral relationship: How can we get to the optimal outcome when, at present, there’s a huge deficit of trust? In other words, how can we get to a place where it isn’t automatically assumed by each side that the other will defect and act in bad faith?
I get asked that question, or variations on it, a lot. I remain enough of an optimist to believe that it’s possible to do so, but I’m alarmed that so few people seem to be thinking — at least out loud, or publicly — about how exactly to go about it. Naturally, I’m not privy to policy proposals that come out of, say, the State Department’s Policy Planning staff, and it could well be that they’re trying to tackle exactly this problem. Let’s hope so.
There’s only one book I’ve come across that lays out a concrete scheme to actively build trust between China and the U.S., and that’s Lyle Goldstein’s Meeting China Halfway: How to Defuse the Emerging US-China Rivalry (2015). Damien Ma, who heads the Paulson Institute’s think tank Macro Polo, first put me onto Goldstein, describing him as a highly original thinker whose ideas are tragically ignored in the Beltway. He’s right on both counts. I was struck by not just the originality of his ideas, and by the thoroughness and specificity with which they were presented, but also by the eye-rolls I’ve gotten when I’ve brought up his name or his ideas in “polite company.” In the spring of 2016, not long after formally joining what was then still called SupChina and before things really got bad between the U.S. and China, Jeremy Goldkorn and I interviewed him at the U.S. Naval War College in Newport, Rhode Island, where Goldstein was teaching. (He’s since left, and is now Director of Asia Engagement at Defense Priorities). He’s been on the show a couple of times since that first appearance — most recently to discuss a Taiwan wargame that CSIS staged. You can check out that show here, or read the transcript here.
The trust-building approach Goldstein presents in his book is to negotiate what he calls “cooperation spirals,” which are the inverse of the escalation spirals that got us to where we are now — and could take us to an even darker place if allowed to continue. Cooperation spirals are a series of explicit steps made in the expectation of reciprocity, requiring increasing levels of trust. Each additional step taken by one side is conditional on the other side undertaking specific reciprocal measures. These go from relatively easy — easier, to be sure, back in 2015 when Goldstein was writing than today, when not one of them sounds easy or even politically feasible — to much more difficult, requiring much greater foundations of trust. Below, for example, is what he suggests on Southeast Asia:
To get to a place where ideas like Lyle Goldstein’s can even get a hearing in DC, or for that matter in Beijing, we need to pull some narrative weeds that have put down deep roots. In the U.S., we would have to start by interrogating, challenging, and eventually discrediting narratives that insist Beijing is inherently deceitful or has in mind some long-term, sinister plan for global domination. Some of these views — over-quoting Sun Tzu is usually a giveaway — are frankly orientalist nonsense. But even those who don’t trade in fiendish plots will sometimes insist, for example, that “the Chinese only understand force,” or make other sweeping generalizations that do not conduce to the rebuilding of trust.
Decidedly unhelpful narratives are taking root in Beijing, too. A growing number of people seem convinced that Washington is successfully waging “cognitive warfare” (认知战)against China, mobilizing and coordinating its powerful media outlets, its NGOS, its public diplomacy efforts, and its popular culture in a campaign to discredit China.2 And as I’ve said elsewhere, the last five years of American technology policy toward China have pretty much ended what was, previous to that, still a live debate about American intentions: You’d find few people in Beijing who aren’t now convinced that the U.S. — irrespective of what party is in power in Congress or the White House — now seeks to actively stymie China’s rise and see the nation on its belly.
I know that we’re always on shaky ground when we analogize international relations to human relationships. All the same, I do it a lot, and I think that if we’re aware of the limitations, it can yield insights: strategic empathy, practiced by a strategic class and applied to foreign policy, isn’t all that different from the cognitive empathy we’d use, say, in trying to better understand a new friend or romantic partner. To learn what makes them tick and to manage the relationship better, you try and glean what their big formative experiences were, to learn something about their upbringing and socialization, their education, their childhood traumas, their relationship with parents, their friendships, their romantic histories, their tastes and their idiosyncrasies.
Perhaps unsurprisingly, one of the analogies I often reach for when it comes to the U.S. and China is marriage. I’ve half-joked many times over the last seven or eight years that I feel like a child whose parents are fighting, and neither one of them is behaving like an adult: actually, they’re both being complete jerks. I’m sure I’m not alone in having thought about the U.S. and China as a marriage on the rocks, with an assortment of problematic codependencies, deep-seated resentments, injured pride, and allegations of manipulation, bullying, coercive control or psychological abuse, and of course, betrayal. There’s no longer any trust left.
So let’s follow the thread of this — again, with that grain of salt always at the ready — and see where it takes us. Let’s assume that in this relationship, the couple wants to stay together: they can both remember a time not so long ago when at least in some regards, they both found happiness in the relationship. Sure, even in those happier times some of the ground was being laid for future pathologies, but neither denies that there were some good times. They want to stay together for the sake of the kids, too. Perhaps, if they can rebuild some of the trust, at least they can keep from blowing up at each other, not inflict more injury on one another, and not go broke hiring expensive divorce lawyers.
So what do marriage counselors say to couples whose relationships have been completely drained of trust? Having no firsthand experience, I asked ChatGPT 4 for some of the most common recommendations that family or marriage counselors offer to couples for whom trust is gone — a sense of betrayal felt, to a greater or lesser extent, by both parties — but who want to salvage the relationship. It gave me a list of ten items, amply caveated and described in that banal, harmless prose characteristic of our generative AI: Acknowledgment of the Issue, Open Communication, Professional Guidance, Patience and Time, Transparency, Rebuilding Intimacy, Forgiveness, Setting Boundaries, Accountability, and Joint Goals.
What surprised me was how applicable — if you squint a little — much of this is to our troubled bilateral relationship. A few of these have no salience at all: Let’s toss out “professional guidance” as impossible and “accountability” as mere wishful thinking, something of which great powers, with few exceptions, seem fundamentally incapable. Forgiveness strikes me as more an end state than part of the process of rebuilding trust, so let’s ignore that one too.
But the rest do suggest steps that might be taken — or even have been taken. Both the U.S. and China acknowledge that trust has been breached (though that “acknowledgment” more often takes the form of accusation, each side unhelpfully assuming the other to be the main offender), but at least there’s open acknowledgment from actors on both sides that the absence of trust is the issue.
Communication between the U.S. and China may still fall short of “open,” but the flurry of visits by U.S. officials to China in the summer of 2023 by, among others, John Kerry, Janet Yellen, and Gina Raimondo was a meaningful improvement that laid the groundwork for the Xi-Biden APEC sidelines meeting in November in Woodside, California. This was followed by Jake Sullivan’s Thailand meeting with Wang Yi, and many more planned meetings in the pipeline. Military-to-military communications have been restored, at least in part. Sufficient communication can short-circuit the Prisoner’s Dilemma: If our Joe and Jinping can coordinate, they can escape the Nash equilibrium.
Patience and time? Beijing will certainly need some patience, as we’re truly entering silly season, and don’t have much hope of emerging from it until after November 5th. It’s encouraging, though, that Xi appears to have signaled patience on the issue of Taiwan, both in his meeting with Biden and in speeches since the 20th Party Congress, where the usual formulation — “time is on our side” — while not at all encouraging to the Taiwanese at least doesn’t sound like a prelude to precipitate action.
Three of ChatGPT’s suggestions strike me as especially useful to think about when we analogize them to the U.S.-China relationship. One is rebuilding intimacy, which translates in our context to those extensive people-to-people ties, academic exchange, and commerce. This is of obvious importance, in light of how much trust evaporated during the pandemic when, for obvious reasons, Americans essentially stopped traveling to, working in, and studying in China. One could easily imagine any number of Goldsteinian cooperation spirals built around things like the re-opening of Confucius Institutes, bringing back the Peace Corps in China, Fulbright scholarships, removing caps on the number of Chinese journalist visas for the U.S. and eventually the re-opening of the Houston consulate on the U.S. side, and on the Chinese side, relaxing restrictions on non-political NGOs, increasing the number of journalist visas, further easing visa requirements for U.S. students and tourists, and finally re-opening the Chengdu consulate. There could be commitments to bilateral exchanges at the state-to-province level, funding commitments from both sides for the sort of cultural exchanges that we all used to take for granted — you get the idea.
Setting boundaries is another one that would seem to address one of the fundamental issues we face. I’ve often remarked that American suspicion about China, especially when it comes to Chinese technology companies, often centers not on what they have done, but rather on what they might do. The same point is made well in Kerry Brown’s 2023 book, China Incorporated, which I’ve been reading ahead of an interview for Sinica with him on February 17th. “[M]ost discussions,” he writes, “are about what China currently threatens to do because the evidence we have is ambiguous.”3 The threat of potential Chinese malevolence justifies actions that, in an irony not lost on some of us, actually increase the odds of malevolent action by Chinese state or non-state actors. We can see plainly enough the flaw in the logic of a preemptive first strike against another state just because it possesses the potential to do harm, especially if that state has second-strike capability. Instead, a policy of deterrence seems the wiser choice in this case.
Our worries about Huawei and other Chinese tech giants acquiring what would amount to a devastating first-strike capability are not entirely unfounded. I’m not going to argue with those who insisted that Huawei networking equipment should be kept out of core networks; they’ve already prevailed, and that ship has sailed. But that lethal first-strike potential cannot reasonably be invoked in every live debate over U.S. policy toward Chinese tech companies: not in renewable energy, not in EV batteries or vehicles, not in the case of Tik-Tok. Instead, boundary-setting — clear red lines that would specify serious sanctions for any actor that demonstrably crosses those lines — seems more the equivalent of deterrence here, and both morally and strategically a better approach than preemption. Vigilance would be required, and some process for verification, but if the goal is to ensure that Beijing doesn’t decide to weaponize Bytedance to steal data or interfere in elections while preserving the openness of American society, boundary-setting with serious teeth strikes me as a sound approach, especially because it could rebuild trust.
To end this on a positive note, one thing that the U.S. and China have in abundance is joint goals. The litany is, I hope, by now familiar: pandemic prevention, nuclear nonproliferation, security on the Korean Peninsula, international peacekeeping, addressing transnational crime, and above all, taking action to reduce greenhouse gas emissions and slow global warming. Just last week, one of my best friends — David Foster, a former bandmate from my college days — published a book on how to address political polarization in the U.S. Over the last few years, he’s steeped himself in the different approaches various organizations take to trying to bridge the enormous divide. One organization David told me about, Braver Angels, actually “uses family therapy principles to foster constructive dialogue between conservatives and liberals, [and] consider how to improve its effectiveness and reach.”4 David’s approach is actually quite different, but read the book to learn about that — or wait until he comes on Sinica, which he may well do, to talk about the parallels between polarization and bipolar competition! Another newer group David told me about, an NGO called Starts With Us, takes a different tack: They coordinate Democrats and Republicans to work together on specific local community projects as a way of recognizing that common ground and building trust: joint goals.
Here, again, Goldstein’s cooperation spirals might be applied. The White House has already made positive statements on China’s actions to curb precursor chemicals used in the production of fentanyl, signaling a belief that Beijing is acting in good faith on that front. Who knows? Maybe in their long chat in that Woodside mansion, Joe and Jinping mapped out some expected reciprocal steps — and we just might be taking tentative steps up that spiral ladder to the rebuilding of trust.
Illustrations by OpenAI. (2024). ChatGPT (4) [Large language model]. https://chat.openai.com
Shannon Tiezzi published a short piece in The Diplomat almost 10 years ago making the comparison; Gavyn Davies wrote an op-ed in the Financial Times noting the same with specific reference to the trade war; and Anna Wu published a paper in December 2023 in Advances in Economics and Management Research along these same lines.
See for example this piece translated by Daniel Crain from the Journal of United Front Science by Guo Yonghu and Zhang Hanyu, in the Sinifaction Substack.
Brown, Kerry. China Incorporated (p. 66). Bloomsbury Publishing. Kindle Edition. Emphasis in the original.
From the abstract of a Harvard Business School case study, “Braver Angels: A Grassroots Effort to Depolarize American Politics” by Francesca Gino, March 2020.