Keyu Jin on China’s Economy, Industrial Policy, and State Capitalism

Keyu Jin, an LSE-trained economist and author of The New China Playbook, sets out the mechanics behind China’s economic model for Lex Fridman: a politically centralised, economically decentralised state that runs local officials on a growth tournament, a state that seeds new industries before withdrawing to let the market pick winners, and a savings-heavy, production-heavy growth pattern now colliding with a real-estate crisis and a stalled consumption shift.

Key ideas

1. The mayor economy. China’s central government is politically consolidated but economically decentralised: local mayors and provincial officials compete against one another on a yardstick set by Beijing — first GDP growth, later environmental compliance and technological innovation — with promotion, demotion, or dismissal as the incentive. The mechanism explains both China’s speed (rapid industrialisation, then real-estate expansion, then EV and solar buildouts) and its waste (roughly 80 cities each backing their own EV brand). See Mayor Economy.
2. State-guided industrial policy: a big push, then withdrawal. Jin’s account of Chinese industrial strategy is a two-stage model: the state mobilises capital and coordinates supply chains to launch a strategic sector faster than the market would alone, then retreats once competition can allocate resources among rival firms. She credits this for China’s EV, solar, and (via crisis-driven catch-up) semiconductor progress, while conceding it wastes capital on companies that ultimately fail.
3. Capital is subordinate to politics, not the reverse. Where she argues US politics is shaped by capital, Chinese capital must answer to the political class. The Jack Ma/Ant Group case is her clearest illustration: the IPO halt reflected genuine financial-stability concerns, but the deeper signal to entrepreneurs was reputational — stay prominent in business, stay invisible in politics.
4. A production model that never learned to spend. China’s local-government incentive structure has always rewarded production and investment, never household consumption. Jin ties the resulting high savings rate partly to the one-child policy (concentrated per-child investment raises precautionary saving) and argues the recent real-estate crackdown compounded the problem by hitting both local-government finance (historically funded by land sales) and household wealth (concentrated in property) at once.
5. Tariffs are bad economics; export controls can backfire. As an economist, Jin calls tariffs ‘distortionary’ and notes the US trade deficit widened, not narrowed, after 2018-era tariffs began — a macroeconomic imbalance (the US saves less than it invests), not a trade-policy problem. She frames US chip export controls similarly: they may have accelerated, rather than prevented, Chinese domestic semiconductor catch-up, epitomised by DeepSeek’s emergence under supply pressure.

Content

The mayor economy: political centralisation, economic decentralisation

Jin opens by naming what she considers the single biggest Western misconception about China: that a small group, or even one person, runs the entire economy. In her account the reality is close to the opposite — extreme political centralisation sits over an economy that is, functionally, ‘more decentralised than the US’s.’ The mechanism is that Beijing holds a single, decisive lever over local officials’ careers — promotion, demotion, dismissal — and uses it to run mayors and provincial governors against each other on whatever yardstick it sets. The first-stage yardstick was pure GDP growth, which produced a predictable escalation: industrialisation and export manufacturing, then land sales and real-estate development once officials discovered fiscal revenue could be generated that way, then a wave of urbanisation funded by the same channel.

The yardstick is not fixed, and Jin uses environmental policy as her sharpest illustration of how it moves. For years, environmental protection went nowhere as an official priority, because taking it seriously would have cost local officials GDP growth under a GDP-only metric — until the central government made it an explicit penalising factor. Compliance, in her telling, sped up within a few years to the point of visibly bluer Beijing skies. The same logic later pulled in innovation: EVs, solar panels, semiconductors, and (most recently) DeepSeek became implicit yardsticks, and the result was heavy duplication — by her estimate, some 80 Chinese cities each backing their own EV brand. She is candid about the inefficiency this produces (wasted capital, misallocated resources when politically favoured firms rather than market signals pick early winners) while crediting the same mechanism with China’s genuine speed advantage in launching new sectors. Her sharpest structural criticism of the model is that it has never rewarded consumption: the local-government payoff has always run through production and investment, so nothing in the incentive structure pushes officials to build the social security, healthcare, and job security that would let ordinary households feel secure enough to spend rather than save. See Mayor Economy for the fuller mechanism and its contested framing.

State-guided industrial policy: the big push and its costs

Jin’s account of Chinese industrial policy runs against the training of most Western-educated economists, herself included — ‘if I talk to my Harvard professors, they’ll be very, very, very skeptical’ — but she defends it as empirically successful in a specific, bounded way. Her model has two stages: in a new, emerging strategic sector with no incumbent advantage anywhere in the world, the state mobilises capital, coordinates supply chains, and drives a ‘big push’ that a purely market-led process would take much longer to assemble, or might never attempt. Once the sector reaches a baseline level of market competition, the state’s job, in her telling, is to withdraw and let private actors — venture funds, market competition — decide which firms survive.

The evidence she cites for the model’s success is instructive precisely because of its selectivity: China’s internal-combustion-engine industry, where no comparable state push occurred, produced nothing internationally competitive, whereas EVs, solar panels, and (following the shock of US export controls) semiconductors have. She does not treat this as costless. The downside, by her own account, is real: capital wasted on companies that ultimately fail, and a misallocation problem inherent to letting political rather than market judgement pick early winners. Her summary verdict — ‘on balance, it’s been positive’ — is a bet that the launch benefit outweighs the allocation cost, not a claim that the mechanism is efficient in any conventional economic sense.

State, capital, and the private sector — the Jack Ma test case

Jin rejects a common Western framing that the Chinese state suppresses private enterprise. Her account of the incentive structure is that local officials generally want to help the strongest private firms, because a thriving company adds to local GDP, employment, and investment — the same officials, in her telling, work ‘tirelessly, day and night’ brokering relationships between struggling private firms, their creditors, and banks. She also describes China’s regulatory posture as innovate-first, regulate-after — contrasted explicitly with a European regulate-first default — which let financial innovations such as peer-to-peer lending platforms emerge before rules caught up, for better and for worse.

The clearest statement of where the state does draw a line is the Jack Ma case. Jin frames the underlying principle as a reversal of the US relationship between capital and politics: in the US, she argues, capital can shape politics; in China, capital must answer to it. Ant Group’s 2020 IPO halt, in her reading, reflected a genuine regulatory concern — an unregulated entity conducting bank-like lending without bank-like oversight — but the broader signal to China’s capitalist class was about visibility and prominence, not business ambition as such. Her formulation is a Chinese saying: ‘the tallest tree gets the most wind.’ The lesson she draws for entrepreneurs is not to abandon ambition but to keep it out of the public political sphere — stay collaborative with the state, avoid outsized personal influence or social-media prominence, and the space to build and profit remains wide open. She extends this to argue that Ma’s post-2020 travel abroad reflects personal choice rather than exile, and that younger entrepreneurs remain undeterred by his case — an interpretation offered as her own reading, not independently verified in the conversation.

Savings, consumption, and the real-estate crisis

Two threads converge, in Jin’s account, on the same structural problem: China produces extraordinarily well and consumes comparatively little. She traces part of the country’s historically high household savings rate to the one-child policy: with only one child to invest in, Chinese families concentrate spending on that child’s education and eventual housing, raising the total cost — and therefore the precautionary saving — associated with having children at all. A related consequence she highlights is the ‘six wallets’ phenomenon: young Chinese couples typically afford urban housing only by pooling savings across both sets of parents (and sometimes grandparents), a family-finance structure she says standard individual-consumption economic models simply fail to capture.

The real-estate crackdown compounds the imbalance rather than correcting it. Jin describes a policy shift — ‘housing is to be lived in, not speculated’ — intended to curb runaway property speculation, but one that struck at both pillars of the local political-economy model simultaneously: local-government finance, historically dependent on land sales to property developers, and household wealth, concentrated overwhelmingly in real estate rather than financial assets. The result, in her telling, is a slowdown that depresses government spending capacity and consumer confidence at once — a self-reinforcing drag she distinguishes clearly from the recurring Western prediction of outright economic ‘collapse’, a framing she considers a category error: the fundamentals (human capital, physical capital, macroeconomic and political stability) remain, in her assessment, well above what China’s current roughly $10,000 per-capita income level would predict, leaving unrealised potential rather than a structural ceiling.

Trade, tariffs, and the export-control backfire

Asked directly whether tariffs are good economic policy, Jin answers as an economist rather than a diplomat: no. She points to the US trade deficit widening, not narrowing, since 2018-era tariffs began, and locates the actual imbalance in macroeconomics — the US saves less than it invests — rather than in trade policy, meaning tariffs address the wrong lever entirely. Her preferred alternative for building domestic competitiveness (in chips, in any strategic sector) is direct investment in R&D, university research, and skilled immigration, citing 1980s US-Japan semiconductor competition as the precedent: competitive pressure, not protectionism, is what restored US leadership after Japan temporarily took the lead.

She extends the same logic to export controls. DeepSeek’s emergence, in her framing, is a case of ‘crisis innovation’: while Chinese firms could comfortably import US chips, there was little incentive to build costly domestic semiconductor capacity; sanctions and export restrictions removed that comfort and produced the existential pressure that accelerated catch-up. She draws a historical parallel — Napoleon’s Continental System indirectly spurring Britain’s Industrial Revolution, Portugal’s naval expansion following a Spanish blockade — to argue that technological blockades tend to motivate, not permanently disable, the blockaded party (see the parallel account in Dylan Patel and Nathan Lambert on DeepSeek and China AI). On China’s own negotiating posture with the Trump administration, she names three governing principles — equivalence, reciprocity, and realism — with political issues (Taiwan, Hong Kong, the state/private economic model) explicitly walled off from any trade discussion. She is careful to caveat the DeepSeek-as-verdict reading: technological leverage, whether in chips or rare earths, has a ‘half-life’, losing force once the constrained party develops substitutes, so she treats the US-China technology contest as an ongoing interdependent negotiation rather than a race with a single finish line.

See also

• Keyu Jin — speaker
• Lex Fridman — host
• Mayor Economy — the decentralised local-government incentive mechanism this episode introduces
• What Makes Economies Grow — theme; the structural-transformation and institutions camps this conversation extends with a concrete mechanism
• Yasheng Huang on the Development of the Chinese State — a contrasting institutional account of the same state
• Dan Wang on What China and America Can Learn from Each Other — related account of Chinese industrial policy and state capacity
• Dylan Patel and Nathan Lambert on DeepSeek and China AI — a technical account of the same export-control/DeepSeek dynamic