What this covers

Victor Shih, a political economist of China, speaks with Dwarkesh Patel about how the CCP's overriding drive to preserve its own power shapes China's economic and technological trajectory. The conversation traces how Xi Jinping consolidated control by centralizing decision-making into Leading Small Groups—bodies where subordinates cannot meaningfully dissent—and how this structure produces a system that is simultaneously technocratic and self-sabotaging. Shih argues that because the Party's survival instinct trumps technical optimality, China accepts large costs in growth, efficiency, and even public health rather than cede control over banks, scientists, or key industries. The interview moves across China's debt crisis, its approach to AGI, succession politics, and the structural roots of both occasional policy successes and repeated large-scale failures.

Shih traces the output maximization embedded in China's socialist banking system—which funds production regardless of profit—back to how financial repression and capital controls lock domestic savings into state-directed investment. He examines why private-sector companies like Xiaomi, BYD, and DJI outperform state-backed competitors despite the latter's access to cheaper capital, and describes how rent-seeking at the provincial level once drove infrastructure investment but now produces buildings in shrinking cities. The conversation turns on China's distinctive AGI strategy: unlike the West's capability race, the Party treats AI primarily as a control and security problem, prioritizing "brakes" to halt any deployment that might threaten Party dominance. Shih discusses Xi Jinping's political rise as a serial coalition purge paralleling Stalin's, why there is no functioning succession plan, and why the next leadership transition risks being more brutal and destabilizing than previous ones. He closes on how content-review bottlenecks and the state's paranoia about information flows will constrain AI deployment in China, even as the system directs resources to AGI development.

Sharpest takeaway

Shih argues that China's political economy is driven by the Party's overriding instinct to preserve its own power, which makes the system simultaneously technocratic and self-sabotaging: it pours capital into output-maximizing industrial policy at huge waste, treats AI as primarily a control/security problem rather than a capability race, and faces a debt overhang and succession vacuum that constrain its long-run growth.

  • Decision-making has been centralized into Xi-led Leading Small Groups where no one can credibly dissent, so Party self-preservation trumps technical optimality.
  • The socialist banking system maximizes output not profit, producing both occasional successes and massive waste financed by financial repression.
  • The CCP's approach to AGI prioritizes 'brakes' and control over racing to capability, and there is no functioning succession plan after Xi.

The claims · ranked40 claims · weighted by value

This asset isn't compiled yet

You're seeing its claims, ranked. Compile it to build the argument threads, weight them, and check each claim against your library — the full view.

0.60

Corrupt political equilibria can produce either too much or too little construction depending on structure: in China the Party command structure lets a city/province secretary cut through red tape and benefit from getting projects built, while in the US (and historically Robert Moses-era deals) regulations create stakeholder groups whose interest is to prolong processes and absorb rent, discouraging building.

causalcontestednovelty 4/4durability 4/4· Victor Shih

“all the regulations in the US—environmental and so on—build their own stakeholders and lobbying groups. Then it's in the interest of those groups to prolong the process so they can absorb a higher share of the rent.”

0.59

The Chinese government's distinctive approach to AGI is that developing the 'brakes' is as important as developing the AI itself, because the Party fears any actor—foreign or internal—could use AI to usurp Party power, so unlike America's private-sector race to capability, China wants the ability to stop everything if needed.

factualcontestednovelty 4/4durability 3/4· Victor Shih

““We need to invest in AI, but we can’t go all out in investing in it without knowing what the brakes are. We have to develop the brakes at the same time.””

0.58

Survivors of the Cultural Revolution drew two opposite lessons: one cohort concluded the Party was too dictatorial and China needed to liberalize (many emigrated), while Xi Jinping's cohort concluded you must never be on the losing side of a political struggle, since winners can do terrible things to enemies—a lesson Xi acted on by building coalitions, notably cultivating military officers in Fujian who later became useful generals.

causalcontestednovelty 3/4durability 4/4· Victor Shih

“But for someone like Xi Jinping himself... the lesson was, “Just don’t be on the losing side. In any political struggle, make sure you’re on the winning side. Because if you’re on the winning side, then you can do terrible things to your enemies.””

0.55

China's $3 trillion foreign exchange reserve is less than 5% of its money supply, so even a rational reallocation of just 10% of assets abroad would exhaust the reserve, force massive devaluation, and require interest rates around 20% to stop the outflow, triggering mass bankruptcy—meaning controls work only because a credible command structure deters outflows.

causalcontestednovelty 4/4durability 3/4· Victor Shih

“Right now, it’s less than 5% of the money supply in China. So it doesn’t even take a panic. Even if people just want to rationally reallocate 10% of their assets outside of China, the foreign exchange reserve is gone.”

0.55

Because of financial repression and capital controls limiting individuals to moving $20,000 a year out of China, Chinese investors cannot invest abroad and must place money in domestic high-tech or in banks paying ~1% deposit rates; the state banking system then uses those deposits to finance local government seeder funds for industrial policy.

causalestablishednovelty 3/4durability 3/4· Victor Shih

“because of capital controls, they legally cannot move more than $20,000 a year out of China. They have to find domestic investments.”

0.54

Suboptimal Chinese policy arises not from lack of expertise but from the Party's instinct to preserve its power: even when a policy would let the Party survive politically, if it would reduce the Party's control (over banks, scientists, etc.) they will reject it, accepting costs like deaths or slower growth in exchange for retaining power.

causalcontestednovelty 3/4durability 4/4· Victor Shih

“what time and again leads to suboptimal policy is the Party’s instinct to preserve itself, and in a way worse than that, to preserve its power.”

0.54

Xi Jinping's rise parallels Stalin's: both were low-key, reliable bureaucrats who avoided early infighting, then on gaining power formed successive coalitions to purge the most threatening rival at each stage (Xi allied with Hu Jintao to purge Zhou Yongkang, as Stalin purged Trotsky), repeating the pattern until achieving absolute power.

causalcontestednovelty 3/4durability 4/4· Victor Shih

“you form a whole series of coalitions to get rid of your most threatening enemy at any given time. For Xi Jinping, it was this guy Zhou Yongkang... So Xi convinced Hu Jintao to join with him to purge Zhou Yongkang”

0.54

A fundamental difference between socialism and capitalism is that socialism cares only about output (more grain, metal, robotics) while capitalism maximizes profit; because China's financial system is socialist, profit-seeking companies are forced into socialist behavior—the state banking system will fund a strategic-sector producer indefinitely even if it never makes money, as long as it can produce the desired output.

causalcontestednovelty 3/4durability 4/4· Victor Shih

“For socialism, they only care about output... So then the state can use the state banking system, which they control, to allocate huge amounts of capital to maximize the output... But when you maximize output, you don’t necessarily make money doing that.”

0.54

Provincial leaders kept building not because investment-driven GDP directly earned promotion, but because large infrastructure projects generate rent-seeking: contractors pay kickbacks that officials pass up to superiors; James Kung's work shows that selling land cheaply to politically connected princelings, who then lobby on the official's behalf, statistically raises promotion chances.

causalcontestednovelty 3/4durability 4/4· Victor Shih

“My argument is that it’s not the actual investment, it’s the rent-seeking that comes along with the investment... You can give your superior $50 million to thank that person and increase your chance of getting promoted.”

0.50

Moving most decision-making into Leading Small Groups headed by Xi removed the peer-level debate that existed in the Politburo Standing Committee—where members were notionally equal and could even overturn the General Secretary—because LSG members are all lower-ranked than Xi and, given his power, none dare debate him, so decisions effectively become Xi's alone.

causalcontestednovelty 3/4durability 3/4· Victor Shih

“But when most of the decision-making was pushed into all these different LSG… Usually there’s one other person who’s similarly ranked as Xi Jinping... But everyone else would be these Vice Premiers, ministers, or provincial governors, who are definitely ranked lower than Xi Jinping. In no scenario will they debate a particular policy with him.”

0.50

Because socialist banks can't assess company quality, China uses bureaucratic expert panels (e.g. in the Ministry of Industry and Information Technology) to vet projects in sectors like semiconductors and robotics and sign off on loans; this system is nominally neutral but riddled with corruption, producing a few successes amid many failures and large-scale waste, exemplified by jailing dozens of MIIT officials over fraudulent semiconductor deals.

factualcontestednovelty 3/4durability 3/4· Victor Shih

“there’s an expert group in the Ministry of Industry and Information Technology. They assess all these different projects. The banks will literally send something into them”

0.50

Economists recommend boosting welfare rather than simply ending financial repression because savings deposits are highly concentrated in the top 10% of households; raising deposit rates to 4% would mainly benefit wealthy net-savers, while the median household has little savings beyond their home and would keep saving heavily for precautionary reasons absent better medical insurance—so welfare, not higher rates, is what raises consumption.

causalcontestednovelty 3/4durability 3/4· Victor Shih

“if you just get rid of financial repression, it will only benefit the net-savers, which is the top 10-20% of households. If you look at the median household, they have very little savings besides the home they live in.”

0.50

Chinese top leadership listens to and respects experts (unlike some US cases), but interest groups present competing experts to shape Xi's gut calls, and at times the leadership acknowledges experts are correct yet refuses to act for political reasons—e.g. knowing mRNA/US vaccines were better but only selling domestic vaccines to support the home pharmaceutical industry, and denigrating Western vaccines in propaganda.

causalcontestednovelty 3/4durability 3/4· Victor Shih

“Sometimes even though they know for sure the experts are right, because of other political reasons they’ll say, “Okay, you’re right, but we’re not going to do what you want us to do”

0.50

Unlike the 1976 transition managed by Long March veterans who trusted each other after decades together, today's Politburo members worked in disparate bureaucracies with little mutual trust—probably by Xi's design to prevent them ganging up—so without that social capital the next succession is likely to be more ruthless, brutal, and disruptive, though military-industrial members who just want their sectors subsidized could be stabilizing.

forecastcontestednovelty 3/4durability 3/4· Victor Shih

“many of them worked in such disparate bureaucracies that there’s not a lot of trust between them. That’s probably by design. Xi doesn’t want people in the Politburo to trust each other”

0.50

Xi sincerely wants unification but his desire is not strong enough to take a very risky gamble—otherwise he'd have acted in 12 years—and rather than a fixed self-sufficiency threshold triggering invasion, the threshold is conditional on exogenous factors; notably Russia's failure in Ukraine, where Putin wrongly expected a quick win, likely raised Xi's threshold for attacking Taiwan.

forecastcontestednovelty 3/4durability 3/4· Victor Shih

“his desire to achieve that is not so strong that he would take a very risky gamble to achieve it. If that were the case, he would’ve done it already. He’s been in power for 12 years.”

0.50

China's education system tracks students into STEM or humanities early, so STEM-trained officials may never learn supply and demand; the one universally required subject is government ideology (e.g. the 'Situation and Policy' class), which has been intensified recently and conveys the government's preferred views rather than economic or practical skills.

factualcontestednovelty 3/4durability 3/4· Victor Shih

“If you get tracked into STEM, you may never learn anything about supply and demand. It’s just not required.”

0.50

China's flagship technology and manufacturing successes—Xiaomi, BYD, CATL, DJI, and High-Flyer—were mainly privately funded at the outset, whereas the state-financed industrial system produces a few successes (like Huawei, which received substantial state funding) amid far more failures wasting billions, so the part of the system that works best is the private sector.

factualcontestednovelty 3/4durability 3/4· Victor Shih

“the success cases actually are not that... they do not include Xiaomi or BYD. Those were mainly privately funded at the beginning.”

0.49

Whereas being a Politburo member used to grant autonomy in one's domain, under Xi any policy preference he expresses must be followed exactly, because officials who drag their feet or pursue slightly different agendas get purged.

causalestablishednovelty 2/4durability 3/4· Victor Shih

“If Xi Jinping expresses a preference over a policy direction, that is the direction you have to go toward, no matter what. Because if he observes that you're dragging your feet, or that you're pursuing a slightly different agenda, you’ll be purged.”

0.49

China was highly fiscally decentralized from the mid-1970s to mid-1990s, which incentivized localities to attract FDI and investment and produced private-sector-driven growth, but the 1994 tax centralization moved the value-added tax and eventually most tax categories to the central government, steadily reducing local fiscal autonomy since then.

causalestablishednovelty 2/4durability 3/4· Victor Shih

“But then there was a tax centralization in 1994, where the central government basically said, “Okay, this is ridiculous. We don't want to end up like the Soviet Union and fall into different pieces. We need to control fiscal income.””

0.49

Content creation will be a major roadblock for AI in China because the Party is so paranoid about online content that it embeds human beings and algorithms to double- and triple-check material; even a well-trained, well-aligned model will still require human review of everything, slowing AI deployment in services and governance.

forecastcontestednovelty 3/4durability 2/4· Victor Shih

“content creation is going to be a big roadblock for China. Because China is so paranoid about the content that flows on the internet that they put human beings all over the place to slow things down”

0.49

Although Chinese leaders are knowledgeable technocrats, protecting the Party and the state apparatus—including keeping state-owned enterprises like pharmaceutical companies alive with national brands and cash flows for future research—will almost always be a higher priority than any technical consideration in a given decision, which reconciles their technocratic competence with disasters like Zero-COVID.

causalcontestednovelty 2/4durability 4/4· Victor Shih

“protecting the Party—protecting the state apparatus and all the important tools of the Chinese government—will almost always be a higher priority than any kind of technical consideration”

0.45

Infrastructure investment was highly productive in the 1980s-90s when China lacked infrastructure, but building successive high-speed rail lines yields rapidly diminishing returns, and with the population shrinking—especially in northeastern and southwestern cities—there is no longer need for rail connecting depopulating areas.

causalestablishednovelty 2/4durability 3/4· Victor Shih

“You build the second one and then the third one and there’s very rapidly diminishing returns.”

0.45

He Lifeng, the chief negotiator with the US and Vice Premier in charge of finance, is trusted despite a poor record—he is known for creating and perpetuating an enormous real estate bubble in Tianjin, including a Manhattan-sized district of empty office buildings—suggesting some officials are promoted precisely because their weakness makes them dependent on Xi.

factualcontestednovelty 3/4durability 2/4· Victor Shih

“he’s well known for starting and perpetuating the largest real estate bubble the world has ever seen in Tianjin... There’s this "New Manhattan"—literally the size of Manhattan—filled with empty office buildings. He made that happen.”

0.45

China keeps central debt looking modest (60-70% of GDP) by pushing obligations down to local governments, which it authorizes to issue ever more debt without funding them; Shih estimates local government debt at 120-140% of GDP, pushing total government debt toward 200%, and last year's ~10 trillion renminbi special local debt issuance was merely an accounting exercise swapping high-yield for low-yield debt without reducing the total.

factualcontestednovelty 3/4durability 2/4· Victor Shih

“the overall size of local government debt in China, I would estimate to be 120-140% of GDP.”

0.45

Because China's financial system is geared to fulfill the Party's—i.e. Xi's—priorities, AI will receive substantial resources regardless of the local debt situation; meanwhile local conditions are so dire that civil servants and teachers go unpaid for months yet still work because outside options (like food delivery) are worse and government jobs retain healthcare and other benefits.

forecastcontestednovelty 3/4durability 2/4· Victor Shih

“the financial system is geared toward fulfilling the priorities of the Party, i.e. the priorities of Xi Jinping. So, even if they can’t do anything else, they will do the things that Xi Jinping wants them to do. And AI, apparently, is one of the higher priorities”

0.45

Shih forecasts that, absent an AI breakthrough, China's economy will be roughly 1x to 1.2x the US economy by 2040 (he distrusts PPP comparisons, putting current real income near 70% of the US), a notably bearish view versus the common expectation of Chinese dominance, driven by slowing growth, trade pressure, no consumption stimulus due to high debt, and limits on how much the world will absorb Chinese supply.

forecastcontestednovelty 3/4durability 2/4· Victor Shih

“I think it’s 1x or 1.2x the US economy.”

0.45

Ding Xuexiang, a Standing Committee member who heads the Central Science and Technology Commission and runs the administrative office of the Central Commission on Cybersecurity (which Xi nominally chairs), would oversee any large-scale Chinese AI effort, and his extraordinary trusted status with Xi—despite only directly working under him for one year in Shanghai—makes him the key figure for AGI.

factualcontestednovelty 3/4durability 2/4· Victor Shih

“another very important position that he holds is head of the Office of the Central Commission on Cybersecurity... So he’s been steeped in cybersecurity for a couple of years.”

0.44

Many Politburo members have genuine elite technical training (e.g. Tsinghua engineers, military-industrialists), but specialized technical expertise does not necessarily make a good governor; what determines advancement is political acumen, which can't be known ex ante except among princelings who were trained in it by their high-official parents.

factualcontestednovelty 2/4durability 3/4· Victor Shih

“So they do know a lot about science. But does that mean they know how to govern? That’s a somewhat different question.”

0.44

Financial repression functions as a tax on savers, and combined with currency devaluation (a tax on consumers), it may meaningfully reduce average living standards despite China lacking a major income tax; this helps explain why, after decades of 5%+ growth, many migrant workers barely get by, homelessness is rising, and elderly care is abysmal for those outside the government/SOE system.

causalcontestednovelty 2/4durability 3/4· Victor Shih

“it’s all financed, as you say, with this financial repression which is basically a tax on savers... if you factor in the currency devaluation which is a tax on consumers, and then this financial repression which is a tax on savers, it might actually be a meaningful decrease in the quality of life”

0.44

An economically sound (but politically impossible) solution to China's debt is to cut unnecessary spending on national defense and many industrial policies—China already has the world's largest navy and faces no invasion threat—and redirect savings plus higher taxes on export firms toward welfare to bolster domestic demand and gradually pay down local debt; Shih is somewhat sympathetic to AI chip spending given US export restrictions but not to subsidies for batteries or solar.

normativecontestednovelty 2/4durability 3/4· Victor Shih

“it would be to reduce a lot of the other unnecessary spending: national defense, many of the industrial policies. And then to use the savings to bolster domestic demand with welfare policies and gradually decrease the amount of local government debt.”

0.44

Xi and the Chinese leadership spend an astonishing amount of time—roughly 270 days a year—in policy meetings, with Xi personally making decisions across a vast range of policy areas based on briefing books, such that his choices effectively become law, a micromanaging style comparable to Stalin.

factualestablishednovelty 2/4durability 2/4· Victor Shih

“Xi Jinping and his colleagues are having meetings about policies almost every day, like 270 days out of the year.”

0.44

Anyone who has been a mid-level or higher official can no longer leave the People's Republic of China, though some former vice ministers can travel to Hong Kong, reflecting an atmosphere where people fear deviating from the Party line because doing so can bring serious trouble.

factualestablishednovelty 2/4durability 2/4· Victor Shih

“Anyone who’s been a mid-level or higher official can no longer leave the People’s Republic of China. There are some former vice ministers who can go to Hong Kong.”

0.43

There is currently no succession plan after Xi, and his sudden death could trigger a financial crisis because financial repression depends on bureaucrats fearing Beijing's oversight: if there's any sense for even a week or two that no one is minding the store, officials approving large outflows for bribes could let capital flee.

forecastspeaker onlynovelty 4/4durability 3/4· Victor Shih

“Right now, the bureaucrat thinks: “If I approve this, someone in Beijing is going to notice it and I’ll be in jail in a week. So I will not approve it.” But if that bureaucrat thinks, “There’s nobody in Beijing...” then he’ll go ahead and approve it.”

0.43

Historically, designating a successor gets that person killed (as with Stalin and Mao), whereas a workable approach uses a transitional figure everyone trusts to represent the leader but who can never become number one—like Mao's wife Jiang Qing—which Xi could replicate via his wife or rising daughter, given the Party's ingrained sexism and these figures' lack of official positions.

factualcontestednovelty 3/4durability 3/4· Victor Shih

“The model we know doesn’t work is designating someone: “You’re going to be the successor.” That person is going to die for sure, in a horrible way. We’ve seen this with Stalin. We’ve seen this with Mao.”

0.40

Rather than building new top-level institutions, China will develop AI governance at the lower end by designating human beings in every government agency and commercial entity using AI to 'put their foot on the brake' if needed, and Shih is all but certain DeepSeek and every major Chinese internet company has a team that can pull the plug, for instance if AI started rapidly generating uncontrollable Falun Gong content.

forecastspeaker onlynovelty 4/4durability 2/4· Victor Shih

“They will want to designate human beings in all the government agencies, in all the commercial entities that are using AI or AGI, to put their foot on the brake if it comes to it.”

0.40

DeepSeek—originally built by hedge fund High-Flyer to trade Chinese markets—appears to have been trained to detect important Chinese government policy documents and meetings, because even simple prompts return high-quality links to the latest policies and official statements, in contrast to other Chinese models like Baidu's that are trained more on social-media content; this is useful for trading because Chinese government policy strongly moves stocks.

factualspeaker onlynovelty 4/4durability 2/4· Victor Shih

“part of what they really trained the model for was to detect important policy documents and meetings within the Chinese government.”

0.36

From 2000 to 2020 localities gained autonomy through land sales while the real estate market boomed, but the central government effectively killed the land market in 2022, leaving local governments highly dependent on the central government.

causalestablishednovelty 2/4durability 2/4· Victor Shih

“But then the central government basically killed the land market in 2022. Now I think localities in China are highly dependent on the central government.”

0.35

The advantage of the American system over the Chinese one is that, at least in theory, it can survive a leader who is not up to snuff.

normativeestablishednovelty 1/4durability 3/4· Dwarkesh Patel

“The advantage of the American system, of course, is that, at least in theory, it can survive a leader who is not up to snuff.”

0.32

Large AI investments may require an AGI-specific leading group because building power and data centers falls under the NDRC and figures like Li Qiang or He Lifeng, and there is a foreign dimension where Gulf-state cloud computing capacity would need China as its paying customer, effectively making AI a 'Belt and Road v2' requiring foreign-policy coordination.

forecastspeaker onlynovelty 3/4durability 1/4· Victor Shih

“You build up all this cloud computing capacity in the Gulf States. Who’s going to be the customer?... China, right? So then the foreign policy apparatus will have to coordinate that”

0.30

China already has the largest navy in the world, some of the best jet fighters, and some of the best missile systems, so it has more than enough for national defense and no one is going to invade it.

factualestablishednovelty 1/4durability 2/4· Victor Shih

“It has some of the best jet fighters in the world, some of the best missile systems. No one is going to invade China or come close to it.”