Sixty-Two Cases and Four Clean Wins

Published October 4, 2026

Across sixty-two cases spanning twenty-five years, the chapter's ledger is not the one either camp wanted. Fifteen are genuine blunders. Four are strategic wins. Ten are camouflaged strategies, cases where an apparent failure hid a plan already moving. Eight are adaptive improvisations, a first plan that failed and a later pivot that pulled value out of the wreckage. Eleven are genuinely contested. The blunders outnumber the clean wins by almost four to one. The wins are large enough in scale that they outweigh many of the blunders in strategic impact.

Fifteen blunders are 24 percent of the set. The chapter says these are not close calls. They are episodes where Beijing's own objectives were undercut by its own actions, with costs that were visible and measurable, and that were acknowledged, at least internally, as failures. The technology crackdown of 2020 and 2021 wiped trillions in market value from Chinese firms and sent private-sector confidence into a spiral that has not fully recovered. The first handling of COVID-19 information in Wuhan, the suppression of early warnings, cost Chinese lives and also cost Beijing's stated goal of showing that its governance was the more capable kind. The One Child Policy's demographic consequences, the overhang of local-government financing vehicle debt, and the repeated misses on semiconductor self-sufficiency timelines sit in the same pile. A local-government financing vehicle, an LGFV, is a company a city or province sets up to borrow for roads and buildings outside the official budget. These are errors, the chapter says, by a system that had the resources and, in most cases, the data to have acted differently.

Four strategic wins are 6 percent. The chapter says that number should pause anyone who calls Chinese strategy a masterclass. Four clean wins out of sixty-two attempts is not a record that supports a ten-dimensional chess story. The four are heavy. Accession to the World Trade Organization, the body that sets rules for global trade, reshaped the global labor market for a generation. The island-building campaign in the South China Sea created permanent military facts on reefs that diplomatic protest has not reversed. The displacement of Western telecom infrastructure in developing markets is the third. The fourth is a domestic surveillance and social-control architecture that, whatever its moral weight, did what its designers intended. All four succeeded, the chapter says, in domains where Beijing could control the tempo of escalation and where the other side faced higher costs for a fight than for accommodation. A separate analysis the chapter calls Report 060 describes a three-phase displacement: blunting American power from 1989 to 2008, building from 2008 to 2016, and global expansion after 2016. The wins clustered where the first two phases were finished before Western countermeasures could come together.

Ten camouflaged strategies are 16 percent. Huawei's preparation for American restrictions is the clearest example in the chapter. Preparation began in 2012, six years before the first major restrictions landed. Under the full weight of American sanctions, Huawei's global market share rose from 29 percent to 31 percent. The difference between a blunder and a camouflaged strategy is the difference between a state that stumbles and a state that absorbs a blow it saw coming. Another analysis, Report 057, found that many cases first coded as blunders in Western media later showed evidence of pre-positioning that changed the verdict.

Eight adaptive improvisations are 13 percent. The post-2024 stimulus debate is a live example. Whether the current round of economic intervention is a blunder piled on earlier errors, or a necessary adaptation to property deleveraging and demographic decline, remains contested. What is not contested is that Beijing has repeatedly changed course when the cost of staying the course became visible enough. The Belt and Road itself moved this way. Early money went into high-profile megaprojects. After write-downs and sovereign-debt fights, the portfolio shifted toward smaller, faster-returning digital infrastructure and green energy. The cumulative figure in this chapter is 1.308 trillion dollars, across more than sixty countries. That total hides the internal pivot. The change in what kind of project got built after 2020 is visible in the data. The chapter calls this institutional resilience, not brilliance: a system that can absorb the cost of its own errors and redirect the money.

Eleven contested cases are 18 percent. Andrew Batson has argued that many Western readings of Chinese economic policy fail because they judge results against the wrong baseline, an idealized market economy rather than the options actually in front of Chinese policymakers at the time. Whether debt exposure in Pakistan or Sri Lanka is a blunder, a deliberate debt trap, or something between depends, the chapter says, on the baseline and the time horizon.

The distribution by itself does not say when competence shows up. Across the preceding chapters a pattern formed around three conditions. Institutional alignment: when the bureaucracies, the state-owned firms, the military commands, and the Party organs pull the same way, execution is formidable. The South China Sea campaign worked in part because the navy, the maritime militia, the state construction firms, and the foreign-policy apparatus shared the objective and the timeline. The tech crackdown failed in part because it set the regulators against the innovation ecosystem with no coherent theory of how to punish and incentivize the same firms at once. Incentive coherence is the second. When the people carrying out a strategy are rewarded for the strategy's actual goals, performance improves. When they are not, it degrades. The LGFV debt crisis is the textbook. Local officials were rewarded for GDP growth. Infrastructure was the fastest path to growth you could see. The debt that paid for it was someone else's problem. The incentives were coherent for the person and incoherent for the system. Tyler Jost's research on institutional fragmentation maps onto this. China is more prone to strategic error when the institutions for a domain are split or siloed, because fragmentation is both a symptom and a cause of incentives that do not line up. Theory of mind is the third. In this use it means Beijing's ability to model how other actors will answer. Wolf warrior diplomacy assumed that an aggressive posture would cow smaller nations. It drove a wave of balancing that pushed countries from Australia to Lithuania into closer alignment with Washington. The model was wrong in a direction that made the problem worse.

The three conditions are not fixed. The chapter's most important finding about them is that the trajectory of all three has been negative since about 2013.

The case set splits into 2001 to 2012 and 2013 to the present, the passage from Hu Jintao's collective leadership to Xi Jinping's centralized authority. In the earlier period, alignment came through negotiation among factions. It was slower and produced a more robust consensus. Promotion, for all its corruption, at least forced local officials to compete on measurable economic outcomes. Theory of mind, never the strongest suit, benefited from a foreign-policy establishment that kept diverse channels to the outside world and was not yet fully under one leader's worldview. The Hu era was not a golden age. It mishandled the first response to the Sichuan earthquake, failed to anticipate the backlash against rare-earth export restrictions, and repeatedly could not turn economic leverage over North Korea into a change in behavior. The ratio of improvisations to blunders was still notably higher than under Xi. The system recovered faster, corrected earlier, and tolerated a wider range of internal dissent about strategy, so failed approaches were more likely to be named and dropped before they spread.

Under Xi, centralization improved coordination at the expense of flexibility. Campaign-style governance can move resources fast when the objective is clear and the opposition is internal. COVID lockdowns, the anti-corruption campaign, and the militarization of the South China Sea are achievements of concentrated authority. The same concentration disables the feedback that catches errors early. When the top leader is personally identified with a policy, the cost of reporting failure goes up and the quality of information at the top goes down. Minxin Pei has documented the paradox: the leader who concentrates the most power also concentrates the most vulnerability to bad information. Industrial subsidies surged sevenfold between 2007 and 2018, a span that covers both eras. The correlation between subsidy spending and productivity gains turned negative over the same years. More money, less strategic return. That is not a system learning. It is a system whose link between spending and outcomes is degrading.

The chapter's uncomfortable conclusion is that strategic competence is not improving. It peaked sometime in the window from 2008 to 2012, when there was enough centralization to act and enough pluralism to self-correct, and it has declined since. The decline is not precipitous and not called irreversible. Report 060's scoring of a twenty-five-case subset looked much more favorable to Beijing: 36 percent strategic wins and only 4 percent genuine blunders. That subset was weighted toward the years before 2013 and toward domains where the system's advantages are strongest. Expanded to sixty-two episodes, with the time weighting equalized, the picture darkens. The blunder rate roughly doubles. The strategic-win rate falls by a factor of six. The gap between the two distributions is itself a finding about how the choice of years and cases can change the verdict.

Of the three variables, theory of mind may be structural rather than cyclical. A state that does not hold competitive elections, does not tolerate a free press, does not permit independent polling of its own people, and suppresses the open debate that would show how preferences differ, cannot easily model systems that do all of those things. Democratic decision-making is noisy and often irrational by the standards of a system that prizes coherence. Modeling it takes an intuitive grasp of how plural institutions add up preferences, and that intuition is hard to grow from inside a system built on the opposite principle. The evidence the chapter lists is a pattern of underestimating democratic backlash. Australia's willingness to take economic costs rather than accommodate. The European Union's capacity to build a coordinated China policy after decades of member-state fragmentation. The bipartisan durability of American hawkishness, treated as a Trump-era aberration rather than a structural shift. And, most consequentially, Taiwan's domestic politics: support for formal independence read as a fringe position manipulated by outsiders, rather than as a deeply held preference of a significant and growing share of the population. Military activity data from 2025 shows PLA presence near Taiwan on all but two days of the year, a tempo driven by Chinese planning cycles rather than by specific Taiwanese provocations. A reactive posture would mean Beijing is answering signals it receives. An internally driven tempo means it is operating on a model of the situation that may not match the situation. Wolf warrior diplomacy did not only miss its objectives. It reduced the willingness of foreign counterparts to hold the frank talks that might have explained the miss. Failure produces isolation. Isolation degrades information. Degraded information produces more failure. Batson has noted the economic parallel. Officials who know that reporting failure will be punished report success, and the center doubles down because the data says the policy is working. Pei's analysis of authoritarian information problems applies: the response to being surprised is typically to tighten control of information, which makes the next surprise more likely. Blunders from bad incentives can be fixed by changing incentives. Blunders from misaligned institutions can be fixed by reorganizing. Blunders from a failure to model democracies may sit in the structure itself.

The chapter is explicit about what it cannot see. The sixty-two episodes are assembled from open sources, defector accounts, leaked documents, and inference from behavior. They are not assembled from Standing Committee minutes, internal war-gaming, or the decision memos. The verdicts are probabilistic. The line between a blunder and a camouflaged strategy sometimes hinges on pre-positioning that may be invisible from outside. Huawei was reclassified from blunder to camouflaged strategy only when the 2012 preparation timeline became available. How many other cases are misclassified the same way is, the chapter says, unknown, and unknowable from the information on hand. Batson has argued that the field of Western China analysis has a bias toward overconfidence, because admitting deep uncertainty about 1.4 billion people under an opaque system is professionally and psychologically uncomfortable. The eleven contested cases are a refusal to force a verdict. The fifteen blunders and four wins also carry uncertainty this format cannot fully show. The tools of judgment are imperfect in the other direction too. Western analysis has been shaped by incentives that reward confident prediction, that confuse access with insight, and that treat fluency in Mandarin as a stand-in for understanding how decisions get made. The intelligence community's repeated failures to anticipate major Chinese actions, the speed of island-building, the scale of the Belt and Road, the depth of the tech crackdown, suggest the apparatus studying the system is not reliably more accurate than the system. Report 058 separates competence, which is data and language, from expertise, which is context and judgment. The community has a great deal of the first. Whether it has enough of the second is left open. The three variables explain the existing cases reasonably well and are derived from those same cases, so they can be overfitted, bent to the examples that produced them. Jost's fragmentation thesis was built mainly from security cases. Whether it transfers cleanly to economic or diplomatic cases is an open empirical question. The model says nothing about black swans, the discontinuous events, a financial crisis, a pandemic, a military accident in the Taiwan Strait, that matter most because they fall outside any conditional frame.

The conditions that produce competence, as the chapter states them, are high alignment, incentives that match from the top of the chain to the bottom, and adversaries who behave in ways the decision-makers can accurately model. When all three hold, the results have rivaled or exceeded anything in the modern great-power record. The South China Sea is the clearest example. The other parties were ASEAN states, the Association of Southeast Asian Nations, whose responses were constrained by economic dependence on China and by their own divisions. The institutions were aligned. The incentives were clear. The model of the other side, while imperfect, was adequate. The failure conditions are the mirror. The tech crackdown combined all three failures. Several regulators chased overlapping mandates without coordination. Regulators were rewarded for showing loyalty through aggressive enforcement. Firms were rewarded for innovating and growing. Those rewards conflicted. The international reaction, especially the speeding up of supply-chain diversification, caught Beijing off guard, because the theory that foreign firms would accept regulatory uncertainty in exchange for market access was wrong.

Tang Shiping, noted in Report 060, warned that Chinese elite talk about national destiny and inevitable rise resembled Japanese boosterism in the 1980s, and that the same mix of real achievement and narrative overreach could produce the same kind of miscalculation. The parallel is imperfect. Japan in the 1980s was an American ally inside American security guarantees. China is a rival operating against them. The psychological dynamic, success breeding overconfidence and overconfidence breeding error, is visible in the cases. The genuine blunders of the Xi era concentrate not in periods of weakness but in periods of perceived strength. The tech crackdown came at a moment of peak confidence that the private sector could be reshaped without consequence. Wolf warrior diplomacy came from a genuine belief that the COVID response had shown the governance model was superior.

Western framing has made the assessment harder. The mask-diplomacy narrative of 2020 treated China's provision of medical supplies to countries hit by the pandemic almost only as cynical influence-buying. Some of it was influence-buying. Some of it was humanitarian assistance. Some of it was bureaucratic momentum from a supply chain that had ramped up and needed somewhere to send the output. The reality was all of these at once. A mental model that codes every Chinese action as either brilliant or incompetent misallocates attention. It overreacts to camouflaged strategies, seeing chess moves that are actually scrambling adaptations, and underreacts to genuine strategic wins, dismissing structural shifts as temporary provocations.

The chapter's forward claim is narrow, and it says the common prediction, that current trajectories will continue, is the prediction that missed the Soviet collapse, the Arab Spring, and the speed of China's own rise. Trajectories change. Leadership transitions happen. Shocks reshape incentives overnight. The subsidy surge meant to speed self-sufficiency may have slowed the adaptive capacity it was meant to build. Whether that reverses under pressure is a question the data cannot yet answer. Belt and Road investment in the first half of 2025 reached 124 billion dollars, a pace that, if sustained, would be a significant acceleration. Whether that is doubling down on diminishing returns, or an adaptation meant to lock in relationships before a potential second Trump administration further fragments the global order, is, as of the chapter's writing, contested. The data supports both readings. Pei's work on authoritarian resilience says the critical variable is not any one policy but the capacity to recognize and correct mistakes before the costs become unsustainable. Authoritarian systems do not typically collapse because they make mistakes. All systems make mistakes. They collapse because they lose the ability to correct them. Whether the imperfect but functional correction of 2001 to 2012 can survive the concentration of the Xi era is not settled by twelve years of data. Twelve years is not enough to tell a temporary degradation from a permanent shift.

The record the chapter leaves on the table is sixty-two cases, twenty-five years, fifteen blunders, four wins, ten hidden strategies, eight adaptations, and eleven question marks. Huawei's share went from 29 percent to 31 percent after a preparation that started in 2012. Subsidies rose sevenfold from 2007 to 2018 while the link to productivity turned negative. In 2025 the navy and air force were off the Chinese coast of Taiwan on all but two days. The first half of that year put 124 billion dollars more into the Belt and Road. The chapter says the scoreboard does not say what happens next.