China Securities Co., Ltd. released its 2027 US economic and asset outlook, stating that since 2022, the US economy and US assets have essentially revolved around three core contradictions: AI, fiscal policy, and inflation.
After experiencing the technology diffusion and energy shocks of 2026, US economic growth momentum is generally weakening in 2027 — investment is cooling and consumption is being suppressed. After sufficient pricing, US equities will enter an AI return verification period in 2027 — cloud players look at cash flow, model companies look at first-mover advantage, and application players look at pricing power. After five consecutive years of being "overbought," US Treasuries will in 2027 await a final round of liquidity shock in a prolonged bear market — non-US capital returning to home countries and withdrawing from the US market. This is the end of this round of the dollar tide, and also the full stop of this round of the dollar credit cycle. The direction is actually fairly clear; what is left for the market is to wait for the rhythm and observe the asset waves stirred up by this round of dollar ebb.
The main views of China Securities Co., Ltd. are as follows: What have been the main threads of the US macroeconomy and assets over the past five years, and what turning points occurred in 2026? Over the past five years, the US economy and US assets have been driven by two main threads: technology and energy. Growth has always been resilient, inflation has always been hard to bring down, and K-shaped divergence has continued to deepen; US financial markets have always been attractive, US equities have been strong, and US Treasury yields have consistently been unable to fall. Technology has propped up growth and also propped up US assets. Since 2025, AI-related investment has contributed about 40% of growth; profit and share price gains in the AI chain have created a wealth effect, supporting high-income household consumption and attracting overseas capital into US equities. AI debt issuance and Treasuries compete for long-term funds, and the term premium has risen accordingly. Energy determines the direction of inflation and interest rates. At the end of February, conflict broke out in the Middle East, oil prices surged, the Federal Reserve restarted rate hikes in September, and the 10-year Treasury yield rose to 5.31% in October. High oil prices eroded the real incomes of low- and middle-income households and, together with the wealth effect brought by technology, deepened K-shaped divergence. In 2026, both aggregate economic resilience and structural divergence widened simultaneously. Consumption driven by investment and the wealth effect propped up the aggregate, but potential consumption power was weakening, and K-shaped divergence intensified. Changes in US assets in 2026: technology sectors led gains early on while other sectors performed relatively modestly; after October, the market began to worry about AI input-output returns; after the midterm elections in early November, technology sectors fluctuated sharply; US Treasury yields remained at high levels, with volatility increasing after geopolitical shocks.
How should we view the US economy in 2027? Under inflation and K-shaped divergence, growth gradually weakens. The US economy is set to weaken in 2027. The two pillars supporting 2026 are loosening at the same time: AI capital expenditure is slowing, and investment's contribution to growth is declining; the savings rate is only 3.8%, the wealth effect fades after US equity adjustments, and low- and middle-income households are squeezed by high oil prices and K-shaped divergence. It is expected that in 2027 both AI investment and the wealth effect will retreat simultaneously, and the US growth advantage will narrow. AI is widening K-shaped divergence, with gains flowing to capital while labor income slows. Labor productivity is growing more than 2% year over year, while real hourly wages have already turned negative year over year. On the employment side, entry-level positions are under pressure first, and the unemployment rate for 20-24 year-old college graduates is close to 7%. Republicans lost the midterm elections because of the cost of living, and young voters shifted toward the Democratic Party. Nearly 90% of voters regard the cost of living as a key factor affecting their vote. The share of voters aged 18 to 29 voting Democratic rose from 54% in 2024 to 72%; the Democratic Party won both chambers, and the United States entered a divided government. At the same time, fiscal hedging capacity has declined. The deficit ratio is already above 6% in non-recession periods, and interest expenses have risen to 4% of GDP. When growth slows, further fiscal expansion will push the term premium even higher. Under a divided government, both large deficit reduction and large-scale stimulus are difficult to achieve. The stimulus from last year's tax cut and spending bill fades, while Democratic efforts to restore health insurance subsidies and other spending expand modestly; the two roughly offset each other, and the fiscal impulse is broadly neutral. The number of game-theory flashpoints is increasing: the funding deadline of December 11 will likely be extended to February-March 2027; federal debt will hit its ceiling in the first quarter of 2027, raising the risk of a government shutdown.
US Treasuries in 2027: yields may spike in the first half due to non-US capital withdrawal, while yields may fall in the second half due to a Federal Reserve pivot. The US Treasury bear market has been discussed for five years, but in 2027 what is more noteworthy is whether non-US capital that is overweight Treasuries will reverse. The institution judges that yields may spike in the first half due to non-US capital withdrawal; in the second half, yields may fall due to a Federal Reserve pivot. The long-bond bear market is not over. The 10-year term premium has risen to its highest level since 2014, and there is still room relative to the 1990-2007 average. The market has already made its judgment: the persistently high US fiscal deficit may require a debt shock to reverse. In 2027, the supply and rollover pressure on US Treasuries has not diminished, geopolitical conflict is pushing the inflation center higher, and the medium- to long-term bear market trend in US Treasuries has not yet ended. The buyer structure has changed, making US Treasuries more sensitive to capital flows. Over the past five years, price-insensitive buyers have exited, and price-sensitive funds have taken over: the Federal Reserve's share of US Treasury holdings fell from 26% at the end of 2021 to 14%; since the end of 2021, overseas official holdings have reduced by about $360 billion. Private funds have short holding periods and are more likely to withdraw when interest rates fluctuate. The fragility of global dollar liquidity has already emerged. Apart from China, official institutions represented by Japan are also reducing Treasury holdings, with Japan's holdings down about 14% from the February high. This fragility will intensify with the transmission of global inflation. If in the first half of 2027 the pass-through of higher oil price levels keeps US inflation stably at a higher plateau, US Treasuries may continue to come under pressure. The ultimate buyer may be the Federal Reserve. The Fed maintains high rates until the end of 2027 and begins cutting rates at year-end to hedge the above pressures. When the ebb of dollar liquidity hits the long end, the Federal Reserve may, in addition to cutting rates, expand medium- and long-term Treasury purchases to stabilize the market; later in 2027, there is a possibility that US Treasury yields fall. In 2026, the credibility of US policy declined, and the risk premium on dollar assets rose accordingly. Treasury intervention in the bond market failed to reverse the rise in yields: after announcing an expansion of long-bond buybacks in mid-August, the 10-year Treasury yield briefly rose by about another 65 basis points. The White House pressured the Federal Reserve, and the Fed's independence was questioned. From a price perspective, signals of ebbing dollar confidence have already appeared. The relationship between gold and real interest rates has reversed: many central banks are increasing gold holdings, and reserve assets are diversifying from dollars to gold. The dollar itself also requires higher risk compensation. In 2025, the dollar index fell 9.4%; in 2026, safe-haven demand combined with Federal Reserve rate hikes brought only about a 3.7% rebound.
For US equities in 2027, the numerator side depends on whether AI can pass verification, and the denominator side depends on whether non-US capital ebbs. Both point to the same question: can the technology premium supporting US assets continue? The key to understanding US assets in 2027 lies in one question: can AI deliver returns? AI determines dollar confidence. Over the past decade or so, dollar confidence has been built on growth leadership, fiscal soundness, and institutional credibility. In 2026, two pillars — fiscal and institutional — weakened at the same time: the federal deficit is close to $2 trillion, debt has exceeded $40 trillion, the Treasury intervened in long bonds, and the White House pressured the Fed to hike. Dollar confidence increasingly depends on growth and capital returns, and growth and capital returns increasingly depend on AI. The strength of the United States over the past five years has essentially been the strength of technology. The resilience of the US economy, the strength of the dollar, and even the large overweight in US Treasuries by non-US capital are all built on technological leadership. The more fully priced it is, the greater the volatility. Technological strength has already translated into historically high US equity prices and valuations; the US equity risk premium has fallen to negative territory, and volatility in US equities and even global technology beneficiary stocks widened in 2026. Looking ahead to 2027, US equities will face two tests: the first test is that AI enters a return verification period. The observation framework is that cloud players look at cash flow, model companies look at first-mover advantage, and applications look at pricing power. Specifically: first, cloud vendors bear capital expenditure, and returns depend on cash flow over the full life of assets; second, model companies bear training and computing contracts, and returns depend on whether their lead can retain customers; third, the application layer bears distribution and customer acquisition, and returns depend on whether efficiency gains go to enterprises or users. In 2027, the consensus expectation for free cash flow of cloud giants is about -$150 billion; whether evidence of monetization can be produced before 2028 will determine the direction of valuations. The second test is whether the ebb of non-US capital will reshape valuations. Over the past five years, non-US capital has consistently overweighted US equities; foreign holdings of US equities rose from $13.7 trillion at the end of 2021 to $22.2 trillion, accounting for about 20% of US equity market capitalization, more than double the level in 2000 and about 2.4 times foreign holdings of US Treasuries. With high valuations and concentrated positioning, overseas funds need only slightly reduce their overweight to create sustained selling pressure. Circumstances that would change the judgment: AI revenue and profit margins are realized earlier and capital flows back to the United States; Middle East tensions ease markedly, oil prices fall sharply, and rate cuts come earlier; a sudden contraction in AI investment triggers a deep recession, and US Treasuries turn to gains.
Comments