Vol.214 Macro Talk 104 | Q1 GDP Growth of 5%: Who Is Supporting China’s Economy (Recorded 4.25)
Summary
- The real information in Q1’s 5% growth is not the integer itself, but that its composition now differs from the “5” of the past 2-3 years. The Two Sessions loosened the full-year target to 4.5%-5%, so the actual figure landing exactly at 5% reignited the debate over its credibility; more important is that consumption and investment contributed more than 80% of growth, while foreign trade, despite rising about 15% in Q1 and imports approaching 20%, contributed less than one-sixth to GDP. “This 5 carries a slightly different meaning now.”
- CPI and PPI turning positive year on year in Q1 ahead of schedule means the early-year call for the “double decline” to end in Q2 has already been realized, but it does not prove that domestic demand has stabilized. CPI has been positive year on year for 5 consecutive months, while PPI turned positive in March; the problem is that the US-Iran war, which began in late February, pushed up energy and commodities. With the war continuing and prices still elevated, it is difficult to separate the contributions from cyclical recovery and external shocks.
- Investment’s return to positive growth after a historic contraction in 2025 was driven by government projects and high-tech manufacturing, not a renewed property upswing. Investment grew by just over 1% in Q1, with property in major cities at best approaching stabilization and still dragging on growth; infrastructure, major equipment installations and high-value-added manufacturing filled the gap, suggesting that China is at least beginning to keep investment above the zero line without relying on property.
- Three RMB50B national mother funds could leverage early-stage technology investment to roughly RMB600B, while also filling the investment shortfall and the gap left by shrinking dollar funding. One fund each will cover the Beijing-Tianjin-Hebei region, the Yangtze River Delta and the Pearl River Delta, with capital coming from ultra-long special treasury bonds; the mother fund can account for no more than 30% of a sub-fund and cannot be its largest investor, while 70% of each sub-fund’s capital must go to projects valued below RMB500M. The structure is clearly tilted toward Series A and earlier. “Invest early, invest small, invest in technology” is no longer just a slogan.
- State capital’s heavy bets on large models and AI chips during the 2022-2024 capital winter have produced paper gains ranging from several times to several dozen times, but the exit loop is not yet complete. A group of companies valued at roughly RMB10B-RMB40B before listing rose to several hundred billion yuan after listing; by investment amount, deal count and company coverage, state capital was generally the largest source of funding. The bets validated the principle of buying when nobody cared and prompted local governments to shift from property investment toward full-cycle technology investing, but returns remain locked up, and policy support for listings is also a meaningful pricing factor.
- The core risk in Hong Kong technology stocks is shifting from valuation expansion to share supply: known unlocks this year amount to less than HK$1.6T in market value, excluding companies newly listed this year. The cost of borrowing shares in one large-model stock reached 35%, before financing costs, reflecting either scarce lendable stock or market pricing for a roughly 40% decline. Meanwhile, some foreign investors have begun gradually allocating cash realized from risk-asset sales after the US-Iran war to China; this money has a longer horizon than hot-money rotation, but it is not yet a major migration out of other countries’ equities and into China.
- The common variable for the next cycle in consumption and foreign trade is whether technology capitalization can create a broader wealth effect, and whether China can capture incremental demand from developing countries. Property stabilization is only the floor for consumption; opening a listing route on ChiNext for new services and consumer businesses could raise the ceiling. Foreign trade is shifting from standardized processing for Europe and the US toward a model in which developing countries buy primary and finished goods, China exports intermediate goods and may sell high-value-added products to developed markets, supported by dim sum bonds, CIPS and wider RMB settlement. “If you want to capture more incremental growth, you may have to go to developing countries.”
Deep dive
1. The Integer 5% Obscures a More Important Structural Shift
李丰 first separated the debate into 2 layers: in March, the Two Sessions set a full-year target of 4.5%-5%, effectively acknowledging the challenges facing growth and abandoning “holding the line at 5%” as the sole benchmark; Q1 nevertheless came in at exactly 5%, prompting the market to keep asking how credible the number really is.
He argues that if the result had been 4.9% or 5.1%, there might actually have been less discussion. Investors should focus less on the coincidence of the round number than on the fact that “this 5 carries a slightly different meaning now”: the factors supporting growth have changed from those in the 5% growth phases of the past 2-3 years.
2. Deflation Indicators Turned Positive Early, but the War Complicated the Attribution
Earlier episodes had forecast that CPI and PPI would both turn positive year on year in Q2. The actual timetable was faster: CPI has been positive year on year for 5 consecutive months, PPI turned positive in March, and month-on-month improvement has already continued for several months.
李丰 did not interpret that directly as a full repair of domestic demand. The US-Iran war pushed up energy and commodities from late February and may have helped PPI turn positive; the US also returned to relatively high inflation in March, mainly because of energy. “It is already very difficult to distinguish” how much each factor contributed.
Another 1-2 months of data may not immediately resolve the attribution problem, since the war has not ended and energy prices remain high. What can be confirmed is that the double decline has ended; what cannot yet be confirmed is how much of the turn came from cyclical recovery versus endogenous demand.
3. Foreign-Trade Growth Beat Expectations Most, but Contributed Less Than One-Sixth
March last year was a front-loading window before trade tensions escalated after Trump took office, leaving a high base; March this year was also affected by the war, shipping and the Strait of Hormuz, with exports up only about 2.5% year on year. Even so, 李丰 puts Q1 total import and export growth at about 15%, with imports closer to 20%.
A shifted Lunar New Year may distort a single month, but the effect should largely wash out across the quarter, making the strength of imports particularly surprising to him. Higher imports of precious metals, resources and agricultural products are key to understanding the trade structure that follows.
The most counterintuitive figure is that consumption and investment contributed more than 80% of GDP growth, while trade contributed less than one-sixth. The same fact can be read pessimistically—as evidence that even rapid export growth cannot move the overall economy—or optimistically, as evidence that China’s dependence on exports has fallen materially.
4. Investment Turned Positive, but Not Because of Property
Retail sales grew by less than 5% and investment by just over 1%. With the two together accounting for more than 80% of growth, investment could provide only a limited incremental contribution; most of the remainder still came from consumption that did not look especially strong.
李丰 calls 2025 one of the very few years of negative investment growth since reform and opening. The cause was not only property: infrastructure and major projects also came under fiscal pressure during the local-government debt resolution process, leaving local governments temporarily stretched. Last year’s effort to resolve RMB12T in government operating debt, together with long-term bonds and other debt-resolution arrangements, also front-loaded government borrowing.
When investment turned positive in Q1, property was still a drag. Beijing and Shanghai were only approaching stabilization, not returning to growth; Shanghai may already have stabilized in some sense. The main offsets were infrastructure, major equipment installations and high-tech manufacturing investment, the last of which also brings physical spending on land, factories and equipment.
The structural significance is that China is at least beginning to push investment above the zero line with other projects while property continues to weigh on growth. The replacement sources are also no longer entirely traditional infrastructure projects whose efficiency is easy to question.
5. Three RMB50B Mother Funds Push Investment Support Toward Early-Stage Technology
The state approved 3 RMB50B funds, one each for the Beijing-Tianjin-Hebei region, the Yangtze River Delta and the Pearl River Delta. They will be funded by the government, with the NDRC allocating and managing the capital, and will mainly operate as mother funds investing in sub-funds. 李丰’s institution also applied to participate, and one of its sub-funds is nearing launch.
The funding comes from ultra-long special treasury bonds, whose maturities of 10-plus to 20-plus years match the duration of technology equity investments. About 70%-80% of each mother fund will be invested in sub-funds, with an explicit mandate to invest early and invest small.
The constraints are strict: 70% of a sub-fund’s capital must go into projects valued below RMB500M. At current funding valuations, that generally means Series A, and some companies reach that valuation even before Series A. Many funds accustomed to late-stage investing will therefore need to shift their capabilities.
The mother fund cannot be the largest investor in a sub-fund, and its stake is capped at 30%. Using an average contribution of 20%-25%, each RMB50B fund could create a sub-fund pool of roughly RMB200B, or about RMB600B across all 3. The NDRC has also indicated that this RMB50B program could be deployed for several consecutive years.
6. Multiple Sources of Capital Are Filling the Gap Left by the Capital Winter
The 3 funds have another objective: filling the dollar-funding gap in early-stage technology investment from 2022 to 2024. US LPs faced policy restrictions, causing dollar funds to shrink sharply; the impact was especially acute for early-stage internet and chip companies.
The new mother funds have not yet invested at scale, but the primary market already feels materially warmer. State capital has started investing directly, listed companies are more active, financial investors are returning, and some dollar funds that had stopped investing for several years but still had available capital have shifted from caution to activity.
Dollar capital has not returned across the board. Robots and hardware face relatively fewer restrictions and have attracted more recent participation; chips have attracted less, as have projects involving genetic data. Whether individual large-model companies accept foreign capital also depends on their customers, data and corporate background.
7. State Capital Bought When Nobody Cared and First Captured a Paper Feedback Loop
Large-model, GPGPU and other AI-chip companies were valued at roughly RMB30B-RMB40B in their more expensive pre-IPO rounds and RMB10B-RMB20B in cheaper ones; after listing, some rose to several hundred billion yuan. Based on the original investments, paper returns ranged from several times to several dozen times.
By investment amount, deal count and number of companies covered, the largest investor was money managed directly by the state, not state capital deployed through market institutions as LPs. The reason is straightforward: from late 2022 through the first half of 2025, foreign investors were absent and private funds could not raise capital. State capital was the only source still able to provide several hundred million yuan, and sometimes high hundreds of millions, when companies needed RMB1B or more to fund compute construction, IP purchases, tape-outs, trial production and team expansion.
李丰 retains 2 qualifications. The gains have not yet passed their lockup periods and therefore remain paper wealth; and the pace of STAR Market and Hong Kong listings was also supported by policy, so the returns cannot all be attributed to pure market outcomes.
The market lesson nevertheless holds: buy when nobody cares and sell when everyone is talking. These projects were both aligned with policy—“hot”—and backed by genuine upside in model and compute demand—“specialized.” Policy and industry judgment together produced the returns.
8. Four Years of Positive Feedback Are Pushing Local Governments Toward Full-Cycle Technology Investment
Local governments’ growth playbook over the past 2-3 decades centered on property and industrial development. They now have to manage land-finance debt, public-service satisfaction, new quality productive forces and technology investment attraction simultaneously. 李丰 argues that the first obstacle to this shift is willingness, followed by capability and fiscal capacity.
李翔 suggests that officials may already have been replaced by a more technology-oriented generation. 李丰’s response is that officials at bureau-director level and above in major cities are typically around 50 years old, meaning their professional development still took place under the old model. Even with personnel changes, the old experience base will not disappear overnight.
The 2022-2025 cycle changed that. Local governments supported key local industries when the market was at its coldest and then saw the paper gains of listed companies, while building the ability to screen, debate and make investment decisions. “Capability, feel and the deployment of capital have all shifted.”
State capital has consequently expanded from support focused on post-Series B companies and investment attraction to overseas talent returning to start companies, early incubation, mid-stage expansion and Pre-IPO. The participants are no longer just 1 or 2 star cities, but multiple economically strong provinces and key cities.
9. Changing Direction Does Not Mean Finding the Right Position; Risk Governance Is Next
李翔’s objection is important: under a Drucker-style division of labor, government should provide infrastructure, social security and public services, not take on high-risk equity investments. Even if market funds all invested badly and shut down, the damage would not necessarily extend to government functions or social stability.
李丰 agrees with the principle but argues that the sequence cannot be reversed. Government must first understand new industries and build the willingness and capability to act; only then should it ask where in the lifecycle to participate, in what form and under what controls. The key question over the next 5 years is how much state capital should occupy the early, middle and late stages, and how exits and accountability should be designed.
李翔 also raises the mismatch between officials’ terms and funds’ return periods, as well as the immense fiscal commitment and decision risk involved in Hefei-style projects. Lifetime accountability, responsibility extending across administrative terms and the low efficiency of investing in small projects could all hinder local implementation of “invest early, invest small.”
Banks provide a useful comparison. From the 2018 asset-management rules through AIC in 2024, policy has repeatedly pushed banks from lending toward direct financing. Banks, however, still mostly work with governments through AICs and as fund LPs; 李丰 calls this “only halfway through the transition.” Local governments may already be 60%-70% or even 80% of the way there, but the boundary of participation remains unsettled.
10. Dividends and ETFs Show That Policy Needs Market Feedback to Close the Loop
When policy pushed listed companies to increase dividends in 2023, the idea of “giving ordinary people returns in the capital market” was widely mocked. By 2024-2025, with deposit rates falling to the 1%-plus range, dividend yields of roughly 3% had become a comparable source of attraction for households.
Households first entered high-dividend ETFs through market-based choices, buying exposure to banks and large companies. Regulators simultaneously increased penalties for fraud, market rigging and malicious manipulation; combined with changes in the external environment and market rules, early buyers began making money on a year-by-year basis.
李丰 describes this as another positive feedback loop. ETFs absorb incremental buying, hold the top 50, 200, 300 or 500 companies, and gradually take on a function similar to a market-stabilization fund. “The actors have to receive positive feedback” for the loop to continue.
His policy summary is that roughly one-third came from policy pushing and removing obstacles from different angles, one-third from changes in participants’ understanding, and one-third from actually making money under market rules. The insurance-capital loop has completed more than half a circuit, but product innovation and other links have yet to connect fully.
11. Local Capabilities Are Diverging, and Industrial Policy Is Becoming City-Specific
李翔 warns that not every locality has the fiscal capacity or team needed for full-cycle technology investment, which could widen regional disparities. 李丰 agrees: a direction set by the central government does not mean the whole country moves in sync. “Policy is policy, but localities are localities”; central and local implementation is always a dynamic process.
As some policy latitude has been devolved, early examples of local specialization are emerging. Some cities favor culture and consumption, some are building around the pet economy, some are competing in aerospace, while others are betting on new-energy vehicles and advanced manufacturing. Everyone wants chips and biopharma, but marginal preferences are beginning to diverge.
李翔 cited a newly listed company in Hangzhou as an example. Its team was founded by overseas returnees; when venture capital was unwilling to support it, the local government provided money through talent or enterprise awards without taking a direct stake, then supported it all the way to listing. Local leaders also attended the listing and celebration. The city’s ambition to become “China’s No. 1 AI city” captures the full-cycle model of local participation.
12. Lockups and a 35% Borrowing Cost Expose the Fragility of Large-Model Stock Valuations
Two listed large-model companies rose together for a time, with combined market capitalization of roughly HK$400B, even surpassing Baidu. 李丰 used Alibaba’s Qwen daily active users and Token call volume as a reverse check: if the valuation logic for the standalone model companies were entirely sound, it would imply that Taobao, cloud computing and the other businesses were worth nothing. At least one side of the pricing needs to be reassessed.
Comparing them with OpenAI and Anthropic, valued at roughly $800B, does not solve the problem because liquidity conditions differ between China and the US. The 2 Chinese companies may genuinely differ in technology, products and customer expansion, but after initially moving in lockstep their share prices diverged by about 40%, and lockups may also have entered the pricing.
Hong Kong companies using a domestic-investor structure are typically locked up for 1 year, while many red-chip structures allow shareholders to sell after 6 months. Both companies listed in early January; the one closer to a 6-month unlock had only about 2 months left, prompting the market to price in potential supply early.
李丰 asked a colleague to inquire about borrowing shares for a short position and was quoted a 35% cost, excluding stock-loan interest and funding costs. That may simply reflect the scarcity of lendable stock, or it may incorporate short demand and expectations of a decline of more than 40%. He repeatedly stressed that the figure reflects multiple forces of supply and demand, not a single, definitive forecast of the downside.
13. Less Than HK$1.6T of Unlock Supply Is Forcing Hot Money to Rotate Faster
Last year, the market value of Hong Kong shares coming out of lockup was less than HK$700B. This year, the portion already confirmed to unlock and originating from companies listed last year is less than HK$1.6T, excluding companies newly listed this year that can unlock after 6 months. Including this year’s new listings could push the full-year figure higher; 李丰 has used roughly HK$2T as an example.
Hong Kong’s average daily total turnover is just over HK$200B, making the entire unlock volume appear to represent fewer than 10 trading days. But after primary shareholders sell, the proceeds are distributed to LPs and leave the market; what is needed to absorb the supply is new money, not a simple turnover of existing shares.
Not every unlocked share will be sold, but 2x supply means the market may not have enough money to trade every company and every stage at once. Capital is more likely to cluster in high-attention stocks that have not yet unlocked and in themes outside unlock windows, making prices prone to large swings in both directions.
Model, chip and other technology themes still drive A-shares and Hong Kong stocks, but the hot spots are rotating rapidly from optical-related companies to models, applications, US high-valuation proxies and AI drug discovery. 李丰 attributes the speed of rotation to insufficient capital and the dominance of short-term money, not sustained buying by long-term investors.
14. Foreign Investors Are Shifting From Watching China to Deploying Cash After Clearing Risk
After the US-Iran war, foreign investors’ willingness to actively reallocate toward China became clearer than last year. Last year, the prevailing posture was to “revisit China” and consider whether to allocate; in the limited sample of investors 李丰 has met this year, some have already taken concrete steps to increase their China exposure.
This is not yet a direct sale of one country’s stocks to buy China. The war reduced risk appetite, so investors first sold equities and other risk assets, then gradually directed part of the resulting cash into China. “They are not selling A to move into B”; that distinction means the inflow will not be rapid.
These investors have somewhat longer holding periods than the money chasing locked-up themes and rotating between them, with signs visible in both primary and secondary markets. The program did not extrapolate its limited contacts into a comprehensive, rapid global reallocation.
15. Property Provides the Floor for Consumption, While New-Consumer Listing Channels Raise the Ceiling
The early-year view was that property stabilization was the foundation for household consumption expectations to move from low to mid-range. Transaction volumes and values in major cities were relatively stable year on year and month on month in March and April, but another half-quarter of observation is needed; stabilization does not mean prices are rising.
李丰 summarizes consumption policy as having 2 ends. Stabilizing property is the floor; using existing technology capabilities to open new scenarios and create new business formats is the ceiling. Only if new formats can also produce a wealth effect through the capital market will entrepreneurs and capital have a durable incentive to keep pushing them forward.
Of the 4 new arrangements introduced on ChiNext, the standalone provision encouraging listings by new-service and consumer businesses is an important signal in his view. Part of Hong Kong’s role has already shifted from consumption toward technology; ChiNext may begin to take on the pricing function for technology-enabled consumption and services.
Whether Honor ultimately lists on the STAR Market or ChiNext is only his hypothetical example. But companies such as DJI, Bambu Lab and Insta360—combining manufacturing, technology and global consumer brands—represent a longer-term shift in the asset base. China’s listed assets may not still be dominated by traditional, low-value-added manufacturing 10 years from now.
16. How Broad the Technology Wealth Effect Will Be Remains the Real Point of Disagreement
李翔’s rebuttal is that AI and hard-tech companies are fundamentally creating more value with fewer people. Their employee and option-holder base may never approach that of property or manufacturing, and especially not that of consumer-internet platforms such as Alibaba, Didi and Meituan; a well-known technology company may have only 100-200 employees.
李丰 first distinguishes between 2 types of wealth effect. A rise in housing prices is a paper gain in household asset values, but selling one home still requires buying another home that has also risen in price. His point concerns how a company’s market value is distributed among founders, management and employees. Traditional property and manufacturing companies have generally shared a smaller portion of company-level gains with employees.
On scale, he does not deny that individual companies are smaller; he broadens the statistical boundary. The analysis should include not only 2 model companies but also chips, biopharma, industrial robots, smart hardware and technology services. About 1,000 employees at the Hangzhou company hold options; 40 such companies could correspond to the tens of thousands of employees at one large internet company. If there are 400 or 4,000 such companies, even a few hundred employees at each could cover hundreds of thousands of people.
李翔 still points out that not every employee will receive comparable gains, and that the employment and option scale of consumer internet will be difficult to replicate. 李丰’s deeper point is not that there must be more employees, but that the organization of companies is changing. Scarce knowledge workers need equity incentives: “If you do not do this, you may not be able to hire them.” Huawei is simply an unusual early adopter of a similar distribution mechanism.
17. Global Incremental Growth Is Moving Toward Developing Countries, and Africa Is No Longer Just a Marginal Market
After adjusting for inflation, real GDP growth in developed economies such as the US and Europe is already very low. If global growth is below 3%, most incremental growth will come from developing countries. Developed markets remain large in absolute terms but increasingly resemble a competition for existing demand.
李丰 puts the trade choice bluntly: “If you want to capture more incremental growth, you may have to go to developing countries.” This can also reduce exposure to tariff and protectionist risks between China and the US and China and Europe. China has been restructuring toward these markets through the Belt and Road and related arrangements for years.
He notes that beginning in early May, China should apply unilateral zero tariffs to more than 50 African countries with which it has diplomatic relations. Based on the 2025 growth rankings he cited, about 12 of the 20 fastest-growing economies in the world are in Africa. Low starting bases matter, but they also mean greater incremental demand relative to each country’s existing scale.
The comparison is China after joining the WTO in 2001. GDP, income and disposable income were not high then, yet ABB, Nike, Adidas, Suning, Gome and the first wave of consumer internet companies captured some of their fastest growth from roughly 2002 to 2012. Taken together, multiple developing countries could amount to “most of, or even an entirely new, China.”
18. China’s Foreign Trade Is Building a New Division of Labor: “Buy Twice, Sell Twice”
Beyond precious metals, higher imports are also coming from resources and agricultural products; 李丰 does not have a definitive answer as to whether gold is being bought by households or the central bank. For developing countries, China must first expand purchases of primary goods such as energy and agricultural products before those countries can build income and industrial capacity.
China will also move some final assembly to Southeast Asia and may eventually shift some of it to Africa, while exporting chips, electromechanical goods, chemicals and specialized textile materials as intermediate inputs, then buying the finished products back. 李丰 summarizes the cycle as “each side buys twice and sells twice.”
At the highest-value-added end, China sells new-energy vehicles, smart hardware and biopharmaceuticals to developed countries, and may eventually sell large-model Token services as well. This segment does not depend on high growth in those markets, but on their industrial structures and consumption patterns being able to absorb high-priced products.
The changes are already visible at the Canton Fair. Social-media accounts disagree over the number of attendees, but customer types are clearly changing. Alongside traditional appliances, the Pearl River Delta now has large exhibition areas for smart hardware and robots, while customers from developed countries are looking for products that are too costly or impossible to manufacture domestically.
19. Chip Trade Shows China Has Climbed to the Final 1-2 Rungs of the Value Chain
李丰’s estimate is that China imported more than $400B of chips last year. At the Q1 pace, chip exports could reach $200B-$300B this year. Chip imports first exceeded oil imports around 2014 and have roughly doubled since then.
Large imports and exports at the same time are not contradictory. Imports are concentrated in products that remain critical, are more advanced or have not yet reached comparable cost-performance levels domestically; exports include the control chips used in new-energy vehicles, batteries and electromechanical equipment, among other medium- to medium-high-value-added products.
His analogy is that China has climbed the grapevine to within 1-2 steps of the top. Companies in GPGPU, AI chips and high-value-added memory components continue to move upward; Micron’s push for the US government to tighten restrictions is also, in his reading, evidence that China is approaching the high end of memory.
For foreign-trade professionals, the old model of selecting Chinese factories and setting production specifications for European and US customers will shrink. New opportunities will center on selling China’s high-value-added products to developed markets, bringing Chinese capacity and intermediate goods to developing countries, and importing fish, meat, eggs, dairy and other consumer staples—and eventually finished goods—from those markets.
20. Dim Sum Bonds and CIPS Expand RMB Demand Without Full Convertibility
Imports rose by more than 20% in March while exports grew only about 2.5%, yet China still ran a surplus of roughly $50B, albeit sharply narrower than in the first 2 months. 李丰 argues that sustained strong imports can both help balance the surplus and expand the use of the RMB in trade.
He distinguishes between 2 types of RMB bonds. Panda bonds are RMB-denominated bonds issued inside China by offshore entities; dim sum bonds are RMB-denominated bonds issued by onshore or offshore entities in offshore markets such as Hong Kong.
Dim sum bond issuance exceeded RMB1T last year, a sharp increase from prior levels. Issuance in the first 3 months of this year was already close to RMB500B and could set another annual record. The market allows overseas governments and institutions to raise, hold and use RMB, making it an important tool for expanding RMB demand while the capital account remains not freely convertible.
CIPS is also scaling up. Average daily turnover on one day in March exceeded RMB1T. On 李丰’s figures, total turnover was below RMB200T last year and could reach roughly RMB400T this year; the figure includes multiple types of RMB-denominated settlement, not only merchandise trade.
21. Massive Trade Surpluses No Longer Translate Linearly Into Reserves; Capital Is Staying Offshore
China continues to accumulate large trade surpluses, but its foreign-exchange reserves have not risen proportionally. 李丰 acknowledges that he cannot provide a precise decomposition, but identifies 2 main explanations: the greater use of direct RMB pricing and settlement, and the fact that since mandatory foreign-exchange conversion was abolished in 2012, companies and individuals can leave their foreign-exchange earnings offshore.
Surpluses retained offshore ultimately have roughly 2 uses: continued investment in factories, capacity and operations, or purchases of financial assets. Flat reserves do not mean wealth has disappeared; they may indicate that Chinese companies and households are accumulating more physical and financial assets overseas. He sees this as one possible source of net funds settling in Hong Kong and other offshore markets.
Reducing mandatory conversion also has monetary-policy implications. In the past, rising reserves required corresponding RMB issuance, meaning the volume and pace of RMB issuance were not determined entirely by domestic conditions. Lower conversion pressure gives authorities more independent control over RMB supply, demand and issuance.
There are 2 end points to watch. On the trade side, China needs to gradually reduce its oversized surplus by buying more products from developing countries. On the financial side, overseas entities need an answer to 2 questions: what can they do with RMB, and can they raise RMB? Dim sum bonds, CIPS and offshore RMB assets are all building blocks for internationalization while the capital account remains not freely convertible.