At a recent weekly closed-door macro strategy session, Morgan Stanley's chief China economist and chief China equity strategist offered their latest perspectives on US Treasuries, US equities, A-shares, and Hong Kong-listed stocks. The key takeaways from the high-level discussion are detailed below.
First, the strategists have maintained a consistent view since more than a week ago: investment positioning should be tilted further back toward A-shares. This is closely tied to the shifting international landscape, which was a major theme of the conversation. A distinguishing feature of this cycle is that A-shares are now more correlated with the global cycle than at any time in the past. Therefore, when assessing A-shares, Hong Kong equities, or the broader China market, one must not overlook the evolving dynamics of global equities and the overall cross-asset portfolio ecosystem.
Second, the recent breach of the $40 trillion milestone by US public debt should not be simplistically interpreted as a drastic decline in overall solvency or a severe deterioration of America's fiscal health, nor should it be seen as a precursor to the collapse of US economic and capital market pricing mechanisms. The overall environment is assessed to remain relatively sound. After the debt hit a new record, the key question is whether corporate spending and consumer spending—the two sectors most sensitive to the economy—will face significant repercussions. The conclusion is that they won't. Neither households nor corporations currently show signs of accumulating or triggering major risk points. For asset allocation, there is also no current evidence that high-risk assets like equities will suffer a major negative impact.
Third, the pricing pressures from high US debt are also manifesting in commodities, including gold and copper prices. The forecast is that gold prices have a high probability of breaking above $5,000 per ounce again by 2027.
Fourth, the team is highly bullish on the overall asset allocation space in Korea. They project that the KOSPI index has roughly 30% upside potential by the middle of next year, advising qualified investors to consider reallocating a portion of their funds to Korea and Japan on a phased basis.
Fifth, short-term liquidity in the Hong Kong market may be relatively weak. However, a window of opportunity could emerge around mid-to-late September, potentially driving Hong Kong equities higher once again.
One strategist noted that AI has evolved from a commercial cycle into a geopolitical race. Sovereign AI-driven infrastructure redundancy on a global scale will lengthen the investment cycle. However, this process involves massive financing draws that, combined with US fiscal imbalance, push up long-end interest rates. The Treasury's intervention offers only temporary relief. Two scenarios are likely: the US Treasury yield curve may continue to steepen, and the US dollar may continue to weaken. The strategist also believes China can forge a unique path in AI development.
The recent impact on global capital markets has been driven by a shock to US Treasury liquidity, coinciding with the earnings season in both the US and China. Market participants are intensely focused on the rotation of positions among US stocks, Hong Kong equities, and A-shares. Three key topics were addressed: the shock to the global bond market, particularly US Treasuries; whether AI financing momentum can persist amid liquidity tightening and geopolitical competition; and the specific rotation strategy for A-shares and Hong Kong stocks.
The global bond market is experiencing a shock from high long-end interest rates. The question is whether the US Treasury's announcement of long-duration bond buybacks can resolve these issues. The core concern is whether tightening liquidity will hit a wall—whether such massive financing demand will encounter real liquidity constraints. It must be acknowledged that AI does require enormous financing, which will create structural differentiation in credit market liquidity and stress-test the weakest points of macro liquidity. Over the past few months, the financing demand related to AI in the US over the next two years has been repeatedly calculated to be immense. Investment-grade corporate bond issuance alone reached $270 billion in the first seven months of the year, surpassing the total for all of last year. Global AI-related credit issuance this year has reached $450 billion and is expected to exceed $550 billion by year-end. This scale of financing reflects the enormous capital expenditure needs of overseas tech giants and cloud providers, which has deteriorated their free cash flow. It is predicted that after capital expenditures by several major US cloud and tech giants reach between $1.3 trillion and $1.4 trillion by next year, there will be a lag in converting computing power investment into commercial revenue. By 2027, the free cash flow of US internet giants is expected to turn negative to -$140 billion, widening the financing gap significantly.
Liquidity differentiation is already appearing in the bond market with widening spreads. Credit spreads on unsecured bonds issued by high-grade cloud providers and internet companies have widened by about 30 basis points this year, showing clear structural differentiation. Supply shocks are primarily affecting unsecured corporate bonds. Conversely, asset-backed securities (ABS or CMBS) for data centers, which are supported by tangible physical assets, are performing relatively well. The market is more willing to pay for securitized products backed by completed, energized assets with stable long-term leases. On the other hand, unsecured corporate bonds from internet and cloud giants face more pronounced liquidity supply shocks. The question of whether Asia will be affected also arises. Asian tech and cloud companies looking to emulate US investments in large-scale computing centers may face liquidity shocks. Overall, the Asian financing model is more resilient than that of the US. The US relies heavily on the public bond market, with bond financing accounting for over 55% of cloud provider financing. In contrast, Asia's AI capital expenditure mainly relies on internal cash flow and bank loans, giving it natural immunity to fluctuations in the global bond market. The most noticeable event last week was the sudden and persistent rise in long-end US Treasury yields. Investors are puzzled: US headline inflation has not accelerated—July data even showed some softening and a downtrend—so why did long-end yields and term premiums suddenly rise? This is attributed to two major structural forces resonating. On the corporate side, AI computing infrastructure is exerting an unprecedented draining effect on the market—trillion-dollar financing needs are absorbing funds from the fixed-income market at a record pace, directly competing with government sovereign debt for limited long-term capital. On the government side, US economic growth and employment are relatively stable, but the fiscal deficit has remained at extreme levels near 6% of GDP for years. Logically, sound economic and employment conditions should prompt fiscal consolidation and deficit reduction, but with deep partisan divisions in Washington, there is no clear roadmap for fiscal rectification. The market will continue to see a flood of new US Treasuries. Combined with geopolitical instability in the Middle East, upside uncertainty for oil prices, long-term inflation uncertainty, and policy uncertainty surrounding the incoming Fed chair, investors are demanding higher term premiums on long-term treasuries, leading to the recent rise in long-end yields in the US and globally.
The Treasury's intervention to support the bond market is a temporary fix that fails to address the root cause. One possible scenario is a continued weakening of the US dollar. In an attempt to ease tightening financial conditions, the US Treasury broke with convention last week by announcing that, starting September 9th, it will double the size of its liquidity repurchases of long-dated nominal bonds to at least $4 billion from $2 billion to curb the upward move in yields. However, this intervention is only a temporary fix. While buybacks can suppress long-term interest rates in the short term, they cannot change market concerns about term premiums due to long-term inflation and Fed policy uncertainty, nor can they alter the massive debt stock and deficit trajectory. Unless the US government suddenly unveils a clear fiscal consolidation plan—unlikely in the current political environment—such interventions could create path dependency, leading investors to bet that the Treasury must step in whenever long-end yields rise, effectively creating an implicit "fiscal put option." It could also create potential tension between the Treasury and the Fed. New Fed Chair Kevin Warsh recently affirmed the rise in long-term yields, calling it a necessary measure to tighten financial conditions and curb inflation. However, the Treasury's insistence on suppressing long-end yields through buybacks effectively loosens financial conditions, creating inconsistency in policy coordination with Warsh's stance. If the Fed becomes hesitant to tighten further or is compelled to maintain low rates due to concerns about debt service burdens, it forms a classic pattern of "fiscal dominance over monetary policy." This would weaken the macro credibility of Treasuries and continue to suppress the dollar, pushing the dollar index lower. In other words, one cannot have it both ways. If Treasury intervention is used to lower long-end rates without addressing longer-term inflation, oil price uncertainty, and concerns about US debt, it will fail to address the root cause and may further pressure the dollar. At the upcoming global central bank symposium in Jackson Hole, despite the theme focusing on financial innovation, stablecoins, tokenization, and instant payments rather than short-term monetary policy, it is expected the Fed will not raise rates, and the dollar's outlook may remain bearish.
With the Fed on hold and some overseas central banks gradually tightening, the narrowing rate differential between the US and abroad, coupled with the implicit erosion of US fiscal credibility, will continue to drive dollar depreciation. This aligns with the previous assertion that global sovereign asset allocators will not put all their eggs in one basket, such as dollar assets, but will gradually shift away. The question of whether this will affect the innovative capacity and profitability of US companies, particularly leading firms, was also addressed by the equity strategist.
AI capital expenditure is unlikely to pause anytime soon because the world has entered the era of "sovereign AI," elevating AI to a matter of political economy. As discussed, a "one world, two systems" scenario is unfolding. Many countries pursue computing autonomy, data security, and supply chain control, inevitably leading to redundant construction and backup across regions between developed nations led by the US and other third-party countries. For example, US AI capital expenditure is expected to be around $860 billion this year, potentially reaching $1.3 trillion to $1.4 trillion next year. China's investment is also substantial. Europe plans to invest 750 to 800 billion euros in sovereign clouds by 2030. ASEAN's data center capacity is projected to explode from current 3.3 gigawatts (GW) to 19 GW by 2035. This means each country builds its own system, ensuring computing autonomy, data security, and supply chain control. While this sacrifices global allocation efficiency, it strengthens redundant investment and significantly extends and deepens the investment cycle for data centers, power generation and grid equipment, liquid cooling, and optical communications.
China faces both advantages and disadvantages in its external environment. The US may use a combination of measures, including restrictions on physical chips and potential administrative and legislative limits on models and services. However, China can forge a different path: shifting from pursuing frontier parameter scales to emphasizing inference efficiency, cost advantages, and system integration, leveraging low-cost open-source large models. Additionally, there is potential for China to provide subsidies for computing power if economic support is needed, alongside measures like expanding trade-in programs to support service consumption and providing interest subsidies for targeted mortgages. Furthermore, China is well-positioned to export digital infrastructure compatible with its technological architecture to the global South, carving out a global commercial market share that rivals the US.
The oil market will face sustained supply-demand pressure in the fourth quarter of this year and the first quarter of next year. Since the collapse of the US-Iran memorandum of understanding attempts in mid-June, the Strait of Hormuz crisis remains unresolved, with crude oil transit volume falling to about 6 million barrels per day, an extreme low reminiscent of the early days of the US-Iran war crisis. Diversion routes, such as Saudi Arabia's Red Sea route, are also suppressed to about 1.5 million barrels per day due to various threats and impacts. Meanwhile, the buffers that previously helped stabilize global oil prices are being depleted. US commercial and strategic petroleum reserves have fallen below 300 million barrels, the lowest since 1983, and physical facility limitations prevent some from being released quickly. Consequently, the US release plan for August and September is only about a few hundred thousand barrels per day. Starting in Q4, the cushioning effect of US official releases will largely disappear. While there was some floating storage at sea, from July to mid-August, global floating storage was drawn down by about 5 million barrels per day. The oil market will face sustained supply-demand pressure in Q4 and Q1. Neither Iran nor the US has found a solution to the crisis, Iran faces internal divisions, and the US is attempting to escalate economic sanctions, which could trigger further price shocks to the global supply chain.
In this context, the focus should be on fundamentals. Due to sovereign AI and geopolitical economics, infrastructure redundancy demand is keeping financing for data centers (especially those with high-quality underlying assets) relatively firm. However, caution is advised against many unsecured high-grade corporate bonds, whose spreads will continue to widen amidst supply shocks. Since US intervention in long-end yields is a temporary fix, while the yield curve may continue to steepen, the dollar index may continue to weaken due to artificially narrowed spreads between the US and abroad.
The recommendation to shift positions back toward A-shares was reinforced. Since more than a week ago, the call has been to rotate more of the China portfolio into A-shares. In early July, there was a view that Hong Kong stocks presented a good opportunity, and the rebound from July to mid-August was indeed impressive, led by internet giants and followed by other sectors. Investors who participated actively likely earned solid relative and absolute returns. The current view remains unchanged: allocations should shift more towards A-shares. The reason is closely tied to the international situation. The standout feature of A-shares this cycle is their exceptionally high correlation with the global cycle. A-shares serve as a real-time barometer for new productive forces as a growth engine, often preceding the macro economy in capital markets. The proportion of sectors representing new productive forces in the capital markets is increasing rapidly, whereas the macro economy is still dragged down by the property downturn, weak demand, and insufficient consumption. Given the resonance with the global supercycle, A-shares were also hurt during the global correction after late June. Additionally, the much-anticipated IPOs in July and August squeezed liquidity. After these short-term risk factors and high volatility factors are gradually digested, the A-share market ecosystem has improved significantly.
The impact of the US debt cycle cannot be ignored. The market has significant concerns about the implications of the $40 trillion debt for US solvency and changes in the yield curve, which also affects other asset classes like gold, silver, and cryptocurrencies due to their linkage to fiat currencies, especially the dollar. However, the view is that one should not interpret the record debt as a sign of a major solvency decline or a deterioration of US financial health. The overall environment remains relatively sound. The question is whether the debt record will significantly impact corporate and consumer spending. The conclusion is no. Corporate debt ratios are historically moderate. While large enterprises are borrowing more, and profitability and free cash flow are gradually affected, they remain in a relatively healthy position. Moreover, the corporate debt-to-GDP ratio in the US hasn't changed much in the past decade and has even decreased compared to the post-pandemic period. A recent report from the US technology team analyzed the commercialization path for internet giants and cloud providers facing free cash flow pressures and issuing more corporate debt. The conclusion is that the ROI on AI capex over the next few years is roughly between 25% and 50%, so solvency is not a major concern for now. On the consumer side, the debt-to-GDP ratio for the US household sector is currently 67%, lower than the 70% during the 2000 internet bubble and 74% in 2019, indicating a relatively healthy and manageable debt level. It's also important to understand that although treasury yields are relatively high, the bulk of household debt is locked into mortgages. Many homeowners locked in favorable long-term rates, meaning actual interest payment pressures are much lower than suggested by current interest rate levels. So far, there are no signs of major risk points accumulating or erupting on either the household or corporate side. For asset allocation, there's currently no sign of a major negative impact on high-risk assets. The risk of a large-scale capital exodus from US stocks to other assets is not yet significant. In an environment of high yields, the question is at what point risk-free rates become attractive enough to draw investors from risky equities. Currently, US equities remain firm, and overall risk premiums haven't risen dramatically. However, the real yield on 30-year US Treasuries is 300 basis points above inflation, offering a very rich yield, with nominal rates above 5%. Long-term investment-grade bond yields are above 6%. For large institutional allocators, this near-risk-free guaranteed return is highly attractive. Whether this triggers a migration from risk assets to safe assets hinges on US earnings growth. US earnings have consistently beaten expectations over the past few quarters. This year, S&P 500 earnings are projected to grow at an annualized rate of about 17%, with around 13% growth next year. Even with relatively stable valuations, earnings growth alone could drive stock prices or index levels up by over ten percentage points annually. As long as earnings growth remains significantly ahead of the risk-free rate at 5-6%, the risk of a massive capital withdrawal from US stocks to other assets is low. However, it's advisable to closely monitor each US earnings season. If initial signs of slowdown appear—especially if big tech firms' capex, earnings, or cash flow slow down due to delayed monetization, geopolitical risks, or weaker global demand, even just a slowdown in second derivatives—it could trigger a capital migration at the asset allocation level.
The pricing of high US debt is also reflected in commodities, including gold and copper prices. Gold prices are judged to have a high probability of breaking above $5,000 per ounce again by 2027. Global ETF flows have resumed strong buying of gold, with central banks globally (especially Poland and China) also actively accumulating. The weak dollar pressure mentioned earlier is supportive for gold. In terms of equity exposure, there are direct ways to gain exposure to gold and copper, such as the frequently mentioned long-term holding, Zijin Mining, which investors are advised to watch closely.
Regarding the relative shifts between A-shares and Hong Kong stocks, several major positives for Hong Kong from July to mid-August have largely been realized. The better-than-expected Q2 earnings results for internet and e-commerce sectors have been fully priced in, and some investors may be taking profits. The capital markets communication following the July listing of tech giants on the Hong Kong exchange, showcasing new AI models and features, has also been fully demonstrated and priced. Considering the ebb and flow of funds among Hong Kong, Korea, and Japan, the outlook for Korea's overall asset allocation space is very bullish. The KOSPI index is predicted to have roughly 30% upside by mid-next year, with suggestions for qualified investors to reallocate some funds to Korea and Japan on a phased basis. Short-term liquidity in Hong Kong may be relatively weak, but a window of opportunity may appear around mid-to-late September. Several timelines converge: whether both China and the US show goodwill on tech restrictions and trade; whether the late-September Politburo meeting reviews the economic trajectory and makes targeted fiscal breakthroughs, which would be a clear positive for the market; and a new round of model releases and AI features in the tech sector concentrated after September. If these factors evolve in the expected direction and create a resonance effect, Hong Kong stocks could strengthen again in mid-to-late September.
Following the massive IPOs, A-share liquidity will gradually stabilize in the coming days or within one to two weeks. A-shares are showing resonance with the global cycle, and liquidity will improve after two super-large IPOs. As detailed in a recent report, after large IPOs, especially for high-tech stocks on the STAR or ChiNext boards, stock prices surge in the first one to two weeks post-listing. With unlimited price limits in the first five trading days, these stocks cause a significant siphoning effect on daily market liquidity. While overall A-share daily trading volume remains high, the siphoning effect from these single large stocks reduces trading activity and participation in other sectors. This effect typically subsides within one to two weeks. Following the last major IPO, A-share liquidity is expected to stabilize in the next few days or one to two weeks. Additionally, strong support from the national team, medium-to-long-term confidence in new productive forces, and relatively better-than-expected earnings season performance across sectors are key reasons for a superior A-share ecosystem. It's advised to maintain patience and agility to capture the maximum return opportunities each component of the Chinese stock market offers at different stages.
The team will continue to monitor market developments closely and provide timely updates.
Comments