China’s Hard-Tech Dividend Over the Next Five Years

This note does not pose the binary “will China crash” question. It places the viewpoint inside the next 3–5 years of the policy elite’s objective function and constraints, and asks where capital will be channeled and where it will be suppressed. Framework: facts (data) → style (objective function) → path (foreign analogues) → policy (forecasts) → assets (mapping). The conclusion is not “bullish or bearish” — it is a structural allocation toward policy-favored directions, an avoidance of sacrificed directions, and a hedge against tail risks.

Tag conventions in this note: 【Fact-Official】= official Chinese sources (NBS / MoF / PBoC / MOFCOM); 【External cross-check】= third-party / overseas estimates (Rhodium / IMF / World Bank / academics / think tanks); 【Judgment】= reasoning based on public information; 【Reference】= historical analogues from other countries. External and official numbers diverge significantly on GDP / population / local debt / capital outflows — annotated case by case below.

Quick read · three core judgments

1. Structural deflation + balance-sheet repair is the baseline for the next 3–5 years, not a “V-shape recovery.” The closest analogue is Japan 1990–2010, but China starts from a much lower GDP per capita than Japan did then, with a thinner middle-class buffer.
2. “Hard tech + national security” is the policy-dividend mainline; “property + platforms + consumer finance” is the sacrificed direction. SOE high-dividend (“China Special Valuation”) is the “policy-guaranteed yield” instrument for the transition.
3. Three tail risks: Taiwan Strait (2027–2032 window), further leftward policy turn (anti-market / anti-wealth), escalation of US pressure. None of the three is fully priced into assets yet.

Snapshot of official vs. external divergence (each detailed below): Rhodium estimates 2025 real GDP growth at 2.5–3.0%, roughly half the official 5.2%; the IMF estimates broad local-government debt at 84% of GDP (the official figure has no broad measure), while the DZH database puts LGFV interest-bearing debt at $12.1 trillion (above the IMF’s $9.04 trillion); on SAFE basis, 2024 net FDI was −$168 billion, the largest capital outflow since the 1990 series began; Brad Setser estimates the official current account is understated by about $500 billion; Yi Fuxian estimates the real population is overstated by 130 million+. The divergence between official and external accounts is itself an explanatory variable behind foreign-capital discounts on Chinese assets, gold strength, and capital outflows from Hong Kong.

Systemic risks — a six-dimensional scan

1. Demographics: collapsing faster than expected

【Fact-Official】 Total population peaked at 1.412 billion in 2022. 2023 −2.08 million, 2024 −1.39 million, 2025 −3.39 million — the slope is accelerating, not converging (NBS 2026-01). 2025 births 7.92 million (YoY −1.62 million), birth rate 5.63‰, death rate 8.04‰, natural growth rate −2.41‰. The working-age population (16–59) shrank by 6.62 million in 2025 alone, now 60.6% of the total. Aged 60+ stands at 323 million (23.0%); aged 65+ at 224 million (15.9%). The official total fertility rate is estimated at about 1.05 (SWUFE Institute 2025).

【External cross-check】 Yi Fuxian of the University of Wisconsin–Madison has long challenged the official population count, arguing the real population is overstated by about 130 million+ (slightly more than one-third of the US population); his core evidence is the mismatch between newborn BCG vaccinations and reported births — for example, 2018 reported 6.21 million BCG doses implies at most 9.9 million births, versus the official 15.23 million. Other overseas demographers estimate 2025 real total fertility rate near 0.9 (rather than 1.05), placing China alongside Korea’s 0.7 in the bottom tier globally. Even setting aside Yi’s extreme estimate, 2025 births of 7.92 million are less than 56% of the 14.33 million projected when the two-child policy was launched in 2016 — the failure of population policy is itself settled.

【Judgment】 Demographics is the most irreversible and most certain variable for the next 30 years — the question is not “if it collapses” but “how fast, and what offsets it.” The key signal in 2025 is that the slope is accelerating, not converging — a single-year drop of 1.62 million births and 6.62 million workers turns “within 15 years China cannot replicate Japan’s path” from an estimate into reality. Direct implications: ① long-run pressure on aggregate demand, especially durables, new homes, education; ② the savings rate is lifted by aging, not released by consumption; ③ the policy “long-termism” narrative is forced forward (pensions, healthcare, social security).

2. Property: into year 5 of the unwind, new starts still collapsing

【Fact-Official】 Rogoff & Yang (2020, NBER) estimated property’s full-pipeline contribution to China’s GDP at a peak of 25–29% (including upstream / downstream). The PBoC’s 2019 urban household balance-sheet survey showed that ~70% of household total assets are housing. Full-year 2025 (NBS 2026-01-19): new commodity housing sales area 881 million m² (−8.7%); residential −9.2%; sales value RMB 8.39 trillion (−12.6%); new starts area 588 million m² (−20.4%); completions −18.1% — new starts are now less than 1/3 of the 2021 peak. 70-city new-home prices remained YoY negative throughout 2025; NBS official Q2 figure was −6.4% YoY.

【External cross-check】 Beike data (Caixin 2025-12): second-hand listings have reached 6.5 million units, up 60% from four years ago; buyers now view an average of 17 homes before purchase (vs. 10 in 2021) — severe oversupply, full negotiating power has shifted to buyers. Second-hand prices in the top four tier-1 cities fell for 7 straight months by November 2025, with November MoM −1.1% — a clear gap from the official “the bottom has held” line. A Reuters poll projects 2025 new-home prices −3.8% and 2026 −0.5%, more bearish than the official “stabilize in 2026” expectation. S&P Global Ratings (2025-01) argued the second-hand surge “would stabilize China property in 2025,” but conditioned on developers halting the new-starts collapse — given actual 2025 new starts at −20.4%, that condition has not been met.

【Judgment】 The current path is not “hard landing” — it is “controlled deflation + advance of the state, retreat of the private,” a slow marathon: private developers are allowed to unwind gradually (with “delivery of pre-sold homes” as the floor), and state-affiliated enterprises (Poly, China Resources, COLI) consolidate market share. Property is being redefined from “growth engine” to “essential public infrastructure.” Key read: sales −9% but new starts −20% — developers remain deeply pessimistic on the next 2–3 years; this active supply contraction means property will remain a drag on GDP through 2026–2027. The difference from Japan in the 1990s is that China can use sovereign credit to directly absorb the industry, but the cost is that property-related employment (agents, white-collar, middle managers) and upstream / downstream (building materials, appliances tied to new homes) get compressed for 5–10 years.

3. Local-government debt: halfway through the cleanup

【Fact-Official】 In November 2024 the NPC Standing Committee passed a 12 trillion “6+4+2” debt-resolution package: 6 trillion of one-off new local-debt quota to swap hidden debt + 4 trillion (5 years of 800 billion special bonds) + 2 trillion (post-2029 shantytown hidden debt). By end-2024 hidden local debt to be resolved fell from RMB 14.3 trillion to RMB 10.5 trillion (MoF); by end-2025, more than RMB 5 trillion of swap bonds had been issued, with cleanup more than halfway done; expected to fall to RMB 2.3 trillion by 2028. 2025 LGFV net financing was only about RMB 36.2 billion — less than 20% of 2024 (Yicai / Cailianshe); over 60% of financing vehicles have exited.

【External cross-check】 The numbers diverge widely: the IMF estimates LGFV debt at end-2024 at about $9.04 trillion (~RMB 60 trillion), and adding explicit local debt puts broad local-government liabilities at 84% of GDP (up sharply from 62% in 2019). Aggregating interest-bearing debt of ~4,000 LGFVs in the DZH database puts end-2024 at $12.10 trillion (IndexBox / third party), about 1/3 higher than the IMF. The Atlantic Council headlines it “extends and pretends” — its core challenge is that the cleanup just “swaps the creditor and stretches the tenor” without truly reducing total leverage in the economy; Carnegie / Pettis (2025-08) warned that “using the central-government balance sheet to clean up local debt is a bad idea” because it ultimately converts local credit risk into sovereign credit risk. Read: the official “halfway done” is real, but the headline debt stock itself differs sharply between the IMF / third parties and the official count — that is the root of the long-standing valuation gap on A-share bank stocks and LGFV-related high-yield credits between foreign and onshore investors.

【Judgment】 Cleanup is not deleveraging — it is stretching tenors + cutting coupons + changing the creditor structure — from city banks / trusts / shadow lenders to the central government and big banks. This essentially substitutes central-bank / fiscal credit for local credit, at the cost of further rising central leverage and drastic compression of local fiscal autonomy over the next 5–10 years — slower infrastructure pacing, pay cuts for local civil servants, contraction of public services in tier-3/4 cities; 2025’s collapse in LGFV net financing has already confirmed this path.

4. Deflation: CPI flat for 2025, PPI negative for 40+ consecutive months

【Fact-Official】 Full-year 2025 CPI YoY 0.0% (flat, NBS 2026-01), with food −1.5% and energy −3.3%; full-year 2025 PPI −2.6%, marking 40+ consecutive months of negative readings; the GDP deflator has been negative for 11+ consecutive quarters since 2023 Q2 — a state without precedent in the reform era. M2/GDP exceeds 230% (BIS). Q1 2026 GDP +5.0% (NBS 2026-04-16), with industrial value-added +6.1%, retail +2.4%, fixed-asset investment +1.8%.

【External cross-check】 Rhodium Group (2025-12) estimates 2025 real GDP growth at only 2.5–3.0%, roughly half the official 5.2%. The biggest divergence is in investment: the official Fixed Asset Investment (FAI) was −11% YoY nominal for Jul–Nov 2025 — a cliff drop — yet “Gross Capital Formation (GCF)” in official GDP still contributed +0.9 pp in Q3 — that internal inconsistency between the two series is Rhodium’s main critique. Rhodium further notes that the $1 trillion 2025 trade surplus accounted for 57–68% of China’s GDP growth — implying that without exports, the domestic side would already be dragging GDP into negative territory. Atlantic Council’s “Same overstatement” (2025) reaches the same conclusion as Rhodium. Read: even without fully adopting Rhodium’s estimate, the Q1 2026 scissors-pattern of “industrial +6.1% / retail +2.4” already validates a structure of “supply pulled by state narrative + independently weak demand” — which itself deepens deflation. Overseas investors’ discount on A-shares / Chinese assets partly reflects mistrust of official GDP, not pure pessimism.

【Judgment】 This is a balance-sheet recession (Richard Koo framework) — households and firms aren’t unable to borrow, they are actively deleveraging. Central-bank rate cuts cannot rebuild confidence because expected investment returns have been reset by housing-price expectations. Despite multiple structural cuts in 2024–2025 by the PBoC, the LPR floor has not yet hit zero — but the real rate (nominal minus the GDP deflator) is in fact lifted by deflation — which is why monetary policy looks loose but households and firms don’t feel it. The Q1 2026 scissors of industrial +6.1% / retail +2.4% expose more starkly the binary structure of “supply pulled by state narrative vs. independently weak demand” — and that structure itself deepens deflation.

5. External pressure: trade cools, tech containment continues

【Fact-Official】 MOFCOM basis: 2024 actual FDI −27.1%; Jan–Nov 2025 −7.5% (decline sharply narrowed but still negative). New foreign-invested firms registered in Jan–Apr 2025 totaled 18,800, +12.1% YoY, but actual FDI utilized was still −10.9% — the “more new firms but less capital injection” split suggests foreign capital has shifted from strategic bets to exploratory entry.

【External cross-check】 More bearish: on SAFE basis, China’s net FDI inflow in 2024 was −$168 billion, the largest capital outflow since data began in 1990 (Bloomberg 2025-02). The World Bank “China Economic Update” 2025-06 notes: in early 2025, net capital outflows (including errors and omissions) exceeded the current-account surplus, causing the PBoC’s net reserves to fall by $31 billion (−0.7% of GDP); Mainland–HK Stock Connect saw $57.3 billion net outflow in Q1 2025. Former NY Fed economist Brad Setser further points out: starting in 2024 China switched to an “internal payments dataset” to compute the current account, halving the surplus on that basis, with the obscured amount potentially as high as $500 billion — he calls this “using a lower surplus to mask capital outflows.” Read: the consensus offshore is that the official count systematically understates the scale of capital outflows, and the implied pressure on RMB and the HKD peg is more severe than official figures suggest — which is why 2025 saw persistent strength in “grey demand” for gold, dollars, and crypto inside China.

【Fact-US-China tariffs】 In 2025 China and the US went through five rounds of negotiations: Geneva → London → Stockholm → Korea, ending with a “one-year pause” framework: the US scrapped 91% tariffs, paused the 24% tariffs for 1 year, and paused port fees and the high-tech company look-through rule for 1 year; China scrapped 84% tariffs and paused the new rare-earth export controls for 1 year. But on 2025-12-22 the US implemented its 301 finding, imposing up to 50% tariffs on Chinese semiconductor products (with specific exemptions for Taiwan capacity) — technology containment continues to escalate. The China–US share of total Chinese trade has fallen from 21% in 2017 to about 13% in 2024.

【Judgment】 The dominant logic of external pressure has shifted from “trade” to “tech + capital” — a structural and irreversible change. The 2025 “one-year pause” is a tactical détente; the December 2025 50% semiconductor tariff is evidence of strategic escalation — markets should not misread the pause as a reversal. Two implications: ① exports pull less hard on low-end labor-intensive manufacturing, but the “forced domestic substitution” in high-tech industries is reverse-funded; ② capital markets systematically lift the risk premium on Chinese assets, and A-share / HK PEs cannot return to the 2020-style valuations.

6. Social mood: not unrest, but low desire and “downsizing”

【Fact-Official】 Youth unemployment (16–24, urban-survey, excluding students): August 2025 hit 18.9%, a record high since the December 2023 methodology revision; December dropped to 16.5%, February 2026 16.1%, March rebounded to 16.9% (NBS). The 2026 graduating class is expected to reach 12.22 million (Ministry of Education); supply-side pressure remains. The household savings rate was ~38% in 2023 (vs. pre-COVID ~30%), reflecting precautionary savings. “Lying flat,” “rotting,” “full-time children” have become cultural memes. Marriage registrations fell to 6.10 million couples in 2024, the lowest since 1980 (Ministry of Civil Affairs).

【External cross-check】 Columbia sociology professor Yao Lu’s 2024–2025 research: China’s “college diploma bubble is bursting”; the official youth unemployment definition excludes “full-time exam takers, flexible employment, the unmotivated job-seekers” — actual youth unemployment / under-employment is materially higher than the reported 16–19%; “flexible employment” among PhD / master’s students has surged in recent years, which is effectively disguised unemployment. Yao Yang at Peking University has likewise long questioned the methodology adjustments and the resulting “understatement.” The June 2023 methodology change (which lowered the figure from 21.3% to the new measure) came right after the prior measure hit a new high, and is widely read overseas as politicization of statistics — the new measure’s August 2025 reading of 18.9% setting a fresh high actually confirms the underlying trend is more severe than the official count implies.

【Judgment】 This is not “the eve of revolution,” it is the early form of a Japan-style low-desire society — young people don’t consume, don’t protest, they exit the system. For the policy elite this cuts both ways: short-term social stability, but long-term consumption engine flameout, irreversible birth-rate decline, shrinking tax base. The current response is fertility incentives (third child, subsidies) + consumption release (trade-in subsidies, vouchers) + information control (suppressing pessimistic narratives) — but the first two have weak effects, and the third produces backlash.

Governance style — priority of three objectives

【Judgment】 The objective function of the current leadership, in order of priority:

1. Regime security (top priority) — the Party’s leadership, internal concentration of decision-making power, defense against any systemic challenge. The precondition for everything else.
2. Social stability (binding constraint) — avoid large-scale mass incidents, control information flows, manage expectations. Slow deflation is tolerated; a rapid crash is not.
3. Long-term rejuvenation / great-power competition (strategic objective) — technological self-reliance, military modernization, supply-chain resilience. Ranks behind the first two, but is the steering wheel for long-term resource allocation.

A few observations on style:

Core judgment: this governance style resembles an updated version of late-Cold-War USSR + Japan-MITI-style industrial policy, not the 2000–2015 “market-leaning technocrat” style. That means many of the investment paradigms from 2000–2020 (platform economy, consumption upgrade, secular property bull) are no longer effective — not a cyclical problem but a structural one.

Reference paths — who the elite is learning from, who it is guarding against

Primary reference: Japan 1990–2010

【Reference】 The closest historical sample — after the asset bubble burst, Japan “rode it out” through controlled deflation, industrial upgrading, and social stability. Core features of the Japan path: ① house prices took 25 years to stop falling (some second-tier locations still not recovered); ② the banking system spent 10 years digesting bad debts, with mega-bank consolidation (Sumitomo Mitsui, Mizuho, Mitsubishi Tokyo); ③ the BOJ entered ZIRP in 1999 and QE in 2001; ④ manufacturing climbed up the value chain (robotics, semiconductor equipment, autos, precision machinery); ⑤ society entered a stable state of low desire, long life, low growth.

【Judgment】 The policy elite is systematically studying Japan (multiple internal-report-style papers from CITIC, CICC, Renmin University, CASS). What can be learned: industrial upgrading, deleveraging without triggering a financial crisis, social steady state. What cannot be copied: Japan’s collapse came at GDP per capita of $25,000 (the bottom of the developed world); China was at $13,400 in 2024 — China is more fragile at the starting point, “old before rich.” That is why the elite dare not replicate Japan’s “let the market clear,” and must instead use sovereign credit to backstop.

Secondary reference 1: Korea’s developmental-state model (1997 onward)

【Reference】 After the 1997 Asian Financial Crisis, Korea climbed out via chaebol consolidation + hard-tech localization + export orientation: Samsung, LG, Hyundai, SK made long-term, capital-intensive bets backed by state subsidies, eventually achieving global positions in semiconductors, displays, autos, and batteries. The cost: high household debt (~100% of GDP in 2024), collapsing fertility among younger cohorts (0.7), and a “hyper-competitive” society.

【Judgment】 China’s “new quality productive forces” narrative broadly replicates Korea’s developmental-state model, but at a larger scope with heavier subsidies — semiconductors, batteries, EVs, AI, robotics, biomedicine, commercial aerospace, low-altitude economy. Investment logic: don’t pick winners, pick lanes — because the elite will keep an entire lane alive, but individual companies can be iterated out.

Secondary reference 2: the Singapore governance model

【Reference】 Singapore’s core components: HDB public housing (95% of citizens in public housing, prices deeply intervened by the state); Temasek / GIC sovereign-fund led strategic industries; authoritarianism + technocrats + an efficient civil service; a limited but functional social safety net.

【Judgment】 The ceiling of China’s “new housing model” is a Singapore-style public-housing path — a three-tier structure of subsidized housing (policy rental) + commodity housing (price-capped) + upgrade housing (market high-end), where housing ceases to be a wealth vehicle and becomes essential infrastructure. That is why the three major projects (urban-village renewal, subsidized housing, “dual-use in peace and emergency”) were rolled out densely in 2024–2025. Implication for property investors: returns on commodity housing as an asset class will be long-term suppressed by the state, until a Singapore-style steady state is reached.

Anti-reference: the Soviet collapse (1989–1991)

【Reference】 The elite’s deepest fear is the USSR — believed to have collapsed because of relaxed ideological control, abandonment of Party leadership, and Western “color-revolution” penetration. That is why “political security” and “ideological frontline” are repeatedly emphasized, and why controls over civil information, education, culture, and religion will only tighten, never loosen.

【Judgment】 This means any market expectation of “loosen controls to unlock vitality” is in the wrong direction — the boundary of political control will only push outward, never retreat. Direct implication: confidence repair for the private economy cannot be achieved via property-rights documents alone; it can only be achieved by making private capital politically non-threatening to the elite — which is exactly why, after the platform crackdown, Tencent / Meituan / Alibaba all went through major compliance and “political alignment” overhauls.

Policy forecasts — five-dimension directions

【Judgment】 The likely policy trajectory over the next 3–5 years:

1. Fiscal and monetary: central government adds leverage to backstop, refuses “welfarization”

• Special government bonds + ultra-long-tenor bonds become routine; central-government leverage moves from ~22% toward 30%+.
• Debt-resolution path extended: LGFV hidden debt swapped into low-coupon long bonds, with the banking system absorbing the spread loss.
• Limited room for rate cuts (LPR already in the 3.0% zone), more reliance on structural tools (PSL, re-lending, targeted).
• No US / Europe-style household cash transfers: the elite has repeatedly opposed “feeding the lazy” — a path-dependent ideological constraint that won’t bend even when economically warranted.
• Fiscal deficit moves toward 4–5% (the traditional 3% red line is effectively breached).

2. Industry: subsidies poured into “new quality productive forces”

• Semiconductors (mature nodes prioritized for domestic substitution; leading-edge moves slowly under sanctions): NAURA, AMEC, Hwatsing, SMIC, CXMT, YMTC are representative.
• AI / compute: Huawei Ascend ecosystem, Cambricon, Hygon, SenseTime; data centers and power supporting (UHV transmission, energy storage).
• EVs + batteries + storage: BYD, CATL, Sungrow already global leaders.
• Robotics + industrial automation: Inovance, Estun, the humanoid-robotics chain (Unitree, Zhongqing, Galaxy General, AgiBot).
• Innovative drugs: licensing out (BeiGene, Innovent, Akeso, Keymed).
• Commercial aerospace + low-altitude economy: new priorities in the 14th Five-Year Plan.
• Defense: Shenyang Aircraft, Chengdu Aircraft, CSSC etc.
• Tools: government guidance funds, Big Fund Phase III (RMB 344 billion added in 2024), industrial-chain subsidies, tax breaks.

3. Finance and assets: control deflation, control house prices, support SOEs

• Property: maintain “delivery as floor”-style controlled unwind; commodity housing prices to keep grinding lower for 2–3 more years before approaching a bottom; subsidized housing supply expanded; tier-1 city cores may stabilize first.
• A-shares: the “national team” (Central Huijin) backstops at key levels (repeatedly verified in 2024) but will not push valuations sharply higher — the goal is stability, not exuberance.
• “China Special Valuation” SOE high-dividend: a policy-dividend instrument offering 5–7% yield — effectively “policy-guaranteed bond substitutes” in a deflationary environment.
• Hong Kong stocks: valuations anchored low long-term, but AI overseas + internet-platform regulation bottoming offer structural opportunities.
• Capital controls: not to be loosened, may tighten further covertly (QDII, overseas insurance, crypto).

4. Society: gentle patching, but no relaxation of control

• Delayed retirement is law (passed 2024): men 63, female cadres 58, female workers 55.
• Hukou: essentially open in tier-3/4 cities, marginal loosening in tier-1, with core access still controlled by “points/talent.”
• Fertility incentives: subsidies (pilots in Hubei, Sichuan, Hainan), longer maternity leave, expanded childcare — but expected effects are weak, because the root cause is economic insecurity.
• Social-security expansion: incremental improvements to rural pensions and urban-rural medical insurance, but far from enough to release consumption.
• Information space: further tightening (individual VPN use, self-media filings, labeling of AI-generated content).

5. External: strategic competition + tactical détente (one round completed in 2025)

• Toward the US — one round already done: five rounds of negotiations in 2025 (Geneva → London → Stockholm → Korea) culminating in a “one-year pause” framework — the US scrapped 91% tariffs, paused the 24% tariffs for 1 year, paused port fees and the high-tech look-through rule for 1 year; China scrapped 84% tariffs and paused new rare-earth controls for 1 year. But on 2025-12-22 the US implemented its 301 finding, imposing up to 50% tariffs on Chinese semiconductors — strategic containment continues. Read: tactical trade truce, strategic technology escalation — the long-term direction remains “non-decoupled but de-dependent,” with renewed friction possible any time in 2026 if a political window or event triggers.
• Toward Europe: courting Europe against the US backdrop, but ongoing friction on EV anti-subsidy, medical-device procurement; the core question is whether European industries can absorb China’s mature-tech capacity exports.
• Toward Russia: deeper strategic entanglement (energy, food, security), while avoiding deep entanglement in sanctions exposure.
• Toward the Global South: BRI 2.0, more selective, more financial tools (digital RMB settlement experiments).
• Toward Taiwan: grey-zone pressure keeps escalating (drills, economic pressure, information warfare), but the probability of avoiding active conflict within the 2027–2032 window still dominates — because military modernization is incomplete, economic stress is heavy, and the elite is risk-averse. This is the baseline, not certainty.

Sacrifices and dividends — who gets squeezed, who gets lifted

【Judgment】 The direction is already in the data. Below is the “sacrificed / dividend” mapping across populations, regions, and industries.

Industry

Regions

Populations

Bottom line: the policy-dividend direction ≈ “industries that can be written into the state narrative”; the sacrificed direction ≈ “industries belonging to the last growth paradigm but not in the new state narrative.” The key isn’t absolute good or bad — it’s the relative direction of resource allocation.

Investment mapping — four layers of allocation

【Judgment】 Investment opportunities sliced across “instrument-lane-time” into four layers:

1. A-share core allocation (policy-positive beta)

2. Hong Kong (valuation gap + internet repair)

3. Overseas indirect beneficiaries (hedge for China restrictions / migrations)

4. Hedges and tail risks

Allocation framework (principle, not advice)

• Core (50–60%): SOE high-dividend + hard tech + new energy + innovative drugs — anchored to the policy-dividend direction.
• Satellite (20–30%): HK internet repair + overseas indirect beneficiaries — alpha sources.
• Hedge (10–20%): gold + dollar assets + tail hedges — to dampen portfolio volatility.
• Cash / short-term bonds (5–10%): ready for rebalancing and black swans.

Bottom line: this isn’t a “bullish-on-China” or “bearish-on-China” portfolio, it’s a “long policy-positive, short policy-negative” portfolio. The elite’s objective function guides allocation direction more reliably than GDP growth.

Three biggest uncertainties — must be flagged honestly

【Judgment】 The three largest tail risks to the above framework — any one materializing would significantly reshape allocation:

1. Taiwan Strait (2027–2032 window)

Baseline (~70–80%): grey-zone pressure escalates (drills, economic pressure, information warfare, normalization of coast-guard enforcement) but active firing avoided — military modernization incomplete, economic stress, elite risk aversion. Risk scenario (~10–20%): miscalculation or incident escalating to military conflict.

Overseas wargame references: CSIS The First Battle of the Next War 2023 wargame, 24 runs: in most scenarios, US / Taiwan / Japan forces can repel a Chinese landing and preserve Taiwan’s autonomy, but at the cost of “bloody, catastrophic losses of personnel and equipment” on both sides. CSIS 2025-07 Lights Out?: 26 simulated blockade runs — any form of blockade produces uncontainable escalation pressure, and may be forced to escalate into large-scale war. Former INDOPACOM Commander Adm. Davidson’s “Taiwan 2027” is a capability timeline (when China possesses the ability), not necessarily an action timeline; most think-tank analysts view the more realistic window as 2030–2035. CFR / RAND share similar judgments.

Asset implication: any escalation triggers a re-rating of Chinese assets, HK equities, Taiwan equities, global tech, and semiconductor upstream; gold, the dollar, and select defense names benefit at the extreme. Read: the overseas consensus is “low probability, enormous cost” — that is the rationale for holding gold / select defense / dollar assets as a tail hedge, not a bet that it will happen.

Monitoring signals: ① frequency of US–Taiwan military contacts; ② scale and duration of PLA exercises; ③ semiconductor export-control escalation; ④ US debate on financial sanctions vs. China (SWIFT / FX); ⑤ normalization of coast-guard enforcement, economic blockade and other “near-war” measures by China toward Taiwan.

2. Further policy leftward turn

Baseline: currently marginally loosening (the late-2024 Central Economic Work Conference returned to “growth stabilization” language; warmer tone toward private business). Risk scenario: after a political event or external crisis, ideology re-tightens — anti-market, anti-wealth, anti-foreign escalation. Common Prosperity 2.0 may target property, financial assets, and overseas assets with progressive tax or restrictions.

Monitoring signals: ① new individual-income / property tax rules; ② public-opinion / judicial events involving private entrepreneurs; ③ new capital-control tools; ④ overseas-asset disclosure and tax look-through.

3. US-China policy path

Baseline: Trump 2.0 tariffs oscillating around 60%, tech containment continues to expand (BIS entity list, AI chip controls, TikTok-style asset divestitures). Risk scenario: financial decoupling (SDN sanctions on major Chinese entities), partial SWIFT exclusion, forced delisting of Chinese ADRs — that would reset valuations another notch lower.

Monitoring signals: ① OFAC SDN list changes; ② SEC audit requirements on Chinese ADRs; ③ US secondary sanctions on Chinese banks / energy companies; ④ congressional China-related legislation progress.

Core risk-management recommendation: always retain a 10–20% hedge (gold + dollars + some tail options) — never let the portfolio become a bet on a single country or a single scenario. Even if the baseline plays out, an all-in levered return will not compensate the loss when a tail materializes.

Bottom-line summary:

Structural deflation + expansion of state capacity + hard-tech import substitution is the mainline for China over the next 3–5 years; property + platforms + high-end consumption + low-end exports is the sacrificed direction; Taiwan / leftward policy turn / US pressure are three under-priced tail risks.

The investment corollary: core allocation in hard tech + new energy + SOE high-dividend; satellite allocation in HK repair + overseas indirect beneficiaries; tail hedge in gold + dollar + hedging tools.

The elite’s objective function guides your allocation more reliably than GDP growth — understanding the objective function matters more than predicting GDP.

References — think tanks · academia · official data

Think-tank analysis

• Rhodium Group, “China’s Economy: Rightsizing 2025, Looking Ahead to 2026” — China growth revision and structural outlook. rhg.com
• Atlantic Council · China’s “extend and pretend” debt strategy — local-government debt extension as cover. atlanticcouncil.org
• Carnegie · Michael Pettis China macro-imbalances column. carnegieendowment.org
• CFR · Brad Setser, “China’s Data Still Doesn’t Add Up” — balance-of-payments and statistical reconciliation. cfr.org
• CSIS, “The First Battle of the Next War: Wargaming a Chinese Invasion of Taiwan”. csis.org
• CSIS, “Lights Out? Wargaming a Chinese Blockade of Taiwan”. csis.org

Academic research

• Yi Fuxian (UW–Madison) · Project Syndicate China demographics column. project-syndicate.org
• Yao Lu (Columbia) · China youth labor market research. sociology.columbia.edu
• Richard Koo, “The Holy Grail of Macroeconomics” — balance-sheet-recession framework. amazon.com
• Rogoff & Yang (NBER 27697), “Peak China Housing” — estimating the peak of China property. nber.org/papers/w27697

Official data

• IMF · People’s Republic of China Article IV consultation. imf.org
• World Bank · China Economic Update. worldbank.org
• National Bureau of Statistics of China — official data releases. stats.gov.cn
• State Administration of Foreign Exchange (SAFE) — balance-of-payments data. safe.gov.cn