Chaos Path: Stability and Security Become the Main Line of Allocation | Guotai Haitong Major Asset Allocation Monthly Plan 202604

(Source: Yi Guan the Big Picture)

Author: Fang Yi / Li Jian / Wang Zi-yu / Wang He / Guo Jiao-jiao

Core viewpoint: Based on Guotai Haitong’s large-class asset allocation framework, we believe that in the context of an unclear outlook for Middle East geopolitical developments, safe and stable assets are the main theme of large-class asset allocation. In April, we focus on the advantage of China assets—overweight equities, A/H shares—and also gold and crude oil, with industrial metals as a standard allocation.

Summary

▶ We have built Guotai Haitong’s large-class asset allocation framework, composed of “Strategic Asset Allocation (SAA) — Tactical Asset Allocation (TAA) — Major Event Review and Adjustment,” to serve as an overall guideline for investment decision-making. Based on this framework, we believe that, against the backdrop of the accelerating reshaping of the global order and the worsening trend of geopolitical conditions, safety once again becomes the scarcest resource, while gold is a tangible expression of how to hedge against this uncertainty. We recommend overweighting A shares, gold, and crude oil in April.

▶ It is recommended that in April 2026 the equity allocation weight is 40.00%: overweight A shares (10.00%), standard allocation to Hong Kong shares (7.50%), standard allocation to U.S. shares (12.50%), standard allocation to European shares (5.00%), and standard allocation to Japanese shares (5.00%). The Chinese stock market has strong resilience; we recommend overweighting A shares. Stability is scarce; the Chinese market has a lower risk premium. Micro-level trading shocks will not last long. At the current level, it is not advisable to blindly sell off. The Chinese stock market is expected to form an important bottom and a “hitting zone.” China’s supportive easing stance and diversified reserves/diversified growth can help break the risk narrative faster.

▶ It is recommended that in April 2026 the bond allocation weight is 40.00%: long-duration government bonds (10.00%), short-duration government bonds (10.00%), long-duration U.S. Treasuries (10.00%), and short-duration U.S. Treasuries (10.00%). Within bond assets, (1) reinforced or suppressed inflation expectations will affect the performance of long-duration bonds. Financing demand and credit supply imbalance remain objective realities. However, the central trend in risk appetite is moving upward; resident enterprises may conduct asset allocation rebalancing. Monetary policy will take effect relatively prudently and in a controlled manner. Against the backdrop of geopolitics pushing global energy prices higher and endogenous inflation rising beyond expectations, the allocation value-for-money of medium- and short-duration bonds is better than that of ultra-long-duration bonds. (2) The marginal convergence of the U.S. economy, with inflation expectations strengthening, will suppress the performance of long-duration U.S. Treasuries. Fed Chair Waller (as named by Trump) supports balance-sheet reduction and also advocates a moderate rate cut for monetary policy. Going forward, U.S. Treasury yields may move downward in a moderate way. Policies implemented by the Trump administration have significantly weakened U.S. sovereign credit. Global central banks and large asset management institutions are trend-selling U.S. Treasuries. Under geopolitical risk shocks, safe-haven funds may take defensive allocations, but they are constrained by the re-inflation trade.

▶ It is recommended that in April 2026 the commodity allocation weight is 20.00%: overweight gold (10.00%), overweight crude oil (6.25%), and standard allocation to industrial metals (3.75%). In commodities, (1) inflation expectations and oil prices have upside momentum; gold prices may see heightened volatility or increased in certain phases. In the long run, the strategic allocation value of gold remains intact: a series of policies implemented by the Trump administration will gradually dismantle the global order after World War II. Against the backdrop of the accelerating reshaping of the global order and the worsening trend of geopolitical conditions, safety once again becomes the scarcest resource, while gold is a tangible expression of hedging against this uncertainty. However, as speculative capital withdraws and the re-inflation trade continues to play out—or greatly intensifies—short-term volatility may be amplified. (2) Geopolitical conditions in the Middle East continue to worsen; we recommend overweighting crude oil. Global crude oil demand is relatively weak, and OPEC+ production-policy decisions are changeable. In recent times, Middle East geopolitical conditions have deteriorated sharply, and there is a trend toward further expansion. With the Strait of Hormuz being continuously blocked and crude oil inventories across major economies gradually declining, crude oil prices may still have upside momentum, but the uncertainty in geopolitical developments will further increase volatility in the energy market. (3) The re-inflation trade evolves into a stagflation-inflation trade, or it suppresses demand for industrial metals. In recent years, the rapid development of power-related construction equipment and transportation tools, as well as the expansion of AI computing power and updates to military facilities, have brought additional demand for industrial commodities. Industrial metals represented by copper may be in a temporary state of supply-demand imbalance. But today, as global macro re-inflation trade evolves into a stagflation trade—or suppresses demand for industrial metals—it also increases price volatility.

▶ Risk warning: There are limitations in the analysis dimensions; the model design involves subjectivity; historical and expected data may deviate; market consensus expectations may adjust; and the quantitative model has limitations.

Table of contents

01

Review of asset performance and macro tracking

We review, in the form of a monthly report, the events and key data that attracted high market attention in March 2026 and that had a relatively large impact on major asset classes, and we conduct necessary review and commentary. Meanwhile, based on Guotai Haitong’s large-class asset allocation framework, we analyze how marginal changes in the macroeconomy affect projections for major asset classes and the impact on TAA (Tactical Asset Allocation). Tactical allocation viewpoints reflect our expectations for the risk-return ratio of a particular asset over the next 1–3 months relative to other assets.

1.1. Guotai Haitong’s asset allocation framework

Guotai Haitong’s strategy team’s asset allocation group focuses on research into major asset allocation that combines both active and passive approaches, actively using macro analysis and quantitative model strategies. We organically combine the strengths of both. We build a strategic asset allocation framework based on macro factor risk parity and a tactical asset allocation framework driven by factor expectations. At the same time, we actively utilize experience from scenario-driven research on market trends to review key events that affect capital markets and capture key investment opportunities.

1.2. Tracking marginal changes in macro-consensus expectations

Changes in macro-consensus expectations affect asset pricing and valuation. We believe that when using macro analysis to guide asset allocation, we should attach importance to changes in market-consensus expectations. Important macro conditions—such as expectations for growth and inflation rising or falling rapidly—will affect asset prices through changes in valuation, and thereby affect asset returns.

02

Strategic allocation: Diversify risk with a macro factor risk parity model

Guotai Haitong’s strategy team developed a macro factor risk parity model for the SAA (strategic asset allocation) stage. The model can make good use of the advantages of factor allocation while avoiding difficulties in constructing and applying macro factor components.

Unlike the macro factor models used in some research, Guotai Haitong’s macro factor risk parity model places greater emphasis on controlling macro risks—namely, risks arising from macro real data coming in above or below expectations. Asset prices mainly reflect expectations about future information. Only unexpected shocks beyond expectations will cause price fluctuations. In multi-factor macro models, returns come from risk premia for macroeconomy. Eugene F. Fama, the 2013 Nobel Prize in Economics laureate and the proposer of the Fama-French three-factor model, published “The Efficient Capital Markets: A Survey of Theory and Empirical Work” in 1970, proposing the Efficient Market Hypothesis (EMH), which applies rational expectations directly to asset pricing. This theory holds that in efficient markets, asset prices instantaneously reflect all publicly available information. Investors can only obtain normal returns that match their risk and cannot outperform the market over the long run. This also implies that macro expectations consistent with the market are difficult to generate returns beyond the market.

In the factor selection stage, we use real macro indicators to construct raw macro factors. This approach avoids interference from trading factors and other non-macro asset price drivers affecting macro factors. In the SAA stage, the model’s objective is to seek diversified risk and relative stability in allocation positions. This avoids issues related to the frequency of macro indicator factors. The adjustment frequency of SAA itself is far lower than the release frequency of indicators, and data lags are almost negligible for long-term investment. Although macro indicator factors cannot be directly invested in, in the SAA stage, the target is to use the benchmark proportions for each major asset class in the asset pool as the goal. This does not conflict with investment methods that directly invest in factors or assets, and it can be effectively connected with investment methodologies such as macro analysis.

Therefore, we preprocess raw macro data: we take the predicted values obtained after STL seasonal adjustment as market expectations, use the difference between the actual value and the predicted value as the macro risk factor, and then standardize each factor.

When selecting raw macro factors, the Guotai Haitong macro factor risk parity model chooses domestic economic factors and overseas asset risk premium factors. The domestic macroeconomic factors controlling for risk exposure include growth, inflation, interest rates, credit, exchange rates, and liquidity factors. Overseas asset factors include factors for the United States, Europe, Japan, and India. In the model estimation process, we use an approach similar to the Barra model. The advantage of this explicit factor model is that constructing factors from asset characteristics is more consistent with economic meaning, and the direction of the factor risk parity model is also clearer—namely, avoiding overly concentrated risk exposures in the unexpected parts of macro factors.

Before regressing factor exposures, we first use subjective prior information to specify which macro factors each major asset class may relate to. For example, this article holds that the credit factor only affects the prices of credit bonds and corporate bond assets. Therefore, other assets do not include credit factors in the multivariate regression calculation for factor exposures. When calculating the factor exposure matrix at each month-end, we use multivariate linear regression to determine regression coefficients, with a rolling past-5-year regression window. The half-life of regression weights is 1 year.

The backtest results of the strategy validate the effectiveness of the macro factor risk parity model. We use seven representative major asset classes—CSI 800, Hang Seng Index, S&P 500, interest rate bonds, corporate bonds, CCMA commodities, and international gold—to conduct monthly rebalancing backtests, and we also use the risk parity model based on the same seven asset classes as a benchmark reference.

The strategy backtest results show that, compared with the factor risk parity model, the macro factor risk parity model significantly improves returns, though the Sharpe ratio weakens. Considering that the model aims to diversify macro risks at the SAA stage, the results are satisfactory.

From the positioning perspective, assuming no position limits are imposed, although the macro factor risk parity strategy shows slightly stronger volatility in position changes compared with the factor risk parity model, the allocation proportions across major asset classes remain relatively stable and balanced. The proportion of bond-type assets is located in the 40–50% range, equity assets are about 50%, and commodity asset allocation is less than 10%, which can basically meet mainstream investment needs.

Based on the above model and the comprehensive analysis by Guotai Haitong’s strategy and asset allocation team of the macro environment and asset allocation, we use the asset proportions calculated by the macro factor risk parity SAA model as a reference. We set the benchmark proportions for equities, bonds, and commodities as 45%, 45%, and 10%, respectively, and the maximum deviation limit is set at 10%.

03

Tactical allocation: BL strategy blends active and passive viewpoints to enhance returns

3.1. From single-asset macro factor modeling to multi-asset rotation

The TAA methodology in the Guotai Haitong asset allocation framework is based on an understanding of the investment clock and nested cycle models, while the BL rotation strategy relies on the cycle methodology. We build quantitative modeling for a single asset: that is, we quantify the degree of environmental pressure of a certain type of cycle in a given economic entity, thereby forming several underlying macro factors with distinct cycle characteristics. Then we apply an inverse-order quantile treatment to these underlying macro factors, producing a macro cycle scoring indicator for a specific cycle type. Finally, we combine multiple macro cycle scoring indicators related to cycles. Based on correlation and economic logic, we synthesize a macro composite scoring indicator for a certain asset or style. In short, we tailor a set of proprietary macro fundamentals quantitative indicators for each asset class, which becomes an important basis for understanding asset price performance and forming our subjective viewpoint matrix.

In the special report “From Macro-Friendly Scoring to the BL Model Viewpoint Matrix—A New Idea for Large-Class Asset Allocation Combining Active and Passive Approaches,” published on March 11, 2024, we create subjective viewpoint matrices by simply processing the asset macro composite scoring indicator, and combine them with the Black-Litterman model in the quantitative allocation model. The process begins with transforming consistent macroeconomic expectations into expected values for asset macro composite scoring indicators, then converting those into expected returns for each major asset class, and finally embedding them into the BL viewpoint matrix. With the available assets including equities (AH U.S., Japan, India, and other combinations), bonds (China and the U.S.), commodities, the U.S. dollar, and gold, the global large-class asset allocation BL strategy (i.e., a BL model strategy that introduces subjective macro composite viewpoints for assets including exchange rates) can achieve annualized returns of 23.1% over the five-year backtest period (2019/01/2–2024/02/29). This performance is significantly better than other comparison strategies, demonstrating the effectiveness of combining subjective and quantitative research.

In the out-of-sample performance since March 2024, the global large-class asset allocation BL strategy has also remained robust, even more striking. The strategy’s return in 2024 reached 24%, in 2025 it reached 52%, and the Sharpe ratio across the entire backtest period is as high as 1.76. Based on positions across each period, we can see that the strategy successfully captured the historical bull market in gold in the first quarter of 2025, as well as the A-share repair and a new high breakout in the second quarter. As a monthly rebalanced strategy guided by macroeconomic logic, the performance can be considered outstanding. In the new framework, we use the global large-class asset allocation BL strategy as a method for enhancing returns in the TAA stage. Although this strategy has relatively high position concentration, once the macro factor risk parity model is used in the SAA stage to establish benchmark proportions for major asset classes, the strategy only needs to make limited deviations from the benchmark positions in the TAA stage, which can effectively resolve this issue.

The TAA model, i.e., the global large-class asset allocation BL strategy, achieved a 43.3% return in 2025 out-of-sample, and a 5.2% return in 2026 YTD.

3.2. A scheme that combines the SAA allocation “center” and TAA to thicken returns performs exceptionally well

We use as the standard quantitative workflow an asset allocation plan that takes the SAA macro factor risk parity model-determined allocation center plus TAA asset rotation to enhance returns. We set the strategic asset allocation calculated from the macro factor risk parity model as the portfolio benchmark, with major asset weights of equities at 45.00% (A shares 7.50%, Hong Kong shares 7.50%, U.S. shares 15.00%, European shares 5.00%, Japanese shares 5.00%, and Indian shares 5.00%), bonds at 45.00% (including Chinese government bonds 22.50% and U.S. government bonds 22.50%), and commodities at 10.00% (including gold 5.00%, crude oil 2.50%, and CCMA commodity index 2.50%).

We set the upper and lower bounds for deviations in major asset classes at ±10%. For internal sub-asset weights, we mainly reference the operating results of the Black-Litterman tactical asset allocation model, and also set them based on subjective analysis. This strategy combines the SAA allocation center and TAA to add return, enabling effective control of volatility and maximum drawdown, and enhancing the returns of the original strategic asset allocation strategy. In 2025, this model achieved an annualized return of 21.7%, a Sharpe ratio of 2.29, a Calmar ratio of 3.84, and a maximum drawdown of 5.6% within the year.

04

Review of major macro events and allocation plan

4.1. Review of recent major macro events

Guotai Haitong’s strategy and asset allocation research team strives to effectively combine quantitative models with subjective analysis. In the actual investment process, major event reviews and adjustments are also key steps. Although they appear “subjective” within a highly quantitative asset allocation framework, in real application—when combined with quantitative strategies—they can effectively improve the accuracy and adaptability of investment decisions. Subjective assessments and adjustments also serve as a “safety net” for quantitative investment, and are especially indispensable in extreme events, policy changes, and model failures. In this section, when conducting event reviews, we mainly analyze major events that may affect monthly market trends.

4.2. Tactical large-class asset allocation plan for February 2026

For equities: The Chinese stock market has strong resilience; we recommend overweighting A shares. Stability is scarce; the Chinese market offers a lower risk premium. Micro-level trading shocks will not last long. At the current level, it is not advisable to blindly sell off. The Chinese stock market is expected to show an important bottom and a “hitting zone.” China’s supportive easing stance and diversified reserves/diversified growth can help break the risk narrative faster.

For bonds: (1) Reinforced or suppressed inflation expectations will affect the performance of long-duration bonds. The imbalance between financing demand and credit supply remains an objective reality, but the central trend in risk appetite is moving upward, and resident enterprises may conduct asset allocation rebalancing. Monetary policy is relatively cautious and restrained. Against the backdrop of geopolitics pushing global energy prices higher and endogenous inflation rising beyond expectations, the risk-return attractiveness of medium- and short-duration bond allocation is better than that of ultra-long-duration bonds. (2) The marginal convergence in the U.S. economy, with inflation expectations strengthening and thereby suppressing the performance of long-duration U.S. Treasuries. The Fed Chair Waller (as named by Trump) supports balance-sheet reduction and proposes a moderate cut in monetary policy rates. Going forward, U.S. Treasury yields are expected to move downward moderately. Policies implemented by the Trump administration have greatly weakened U.S. sovereign credit, and global central banks and large asset management institutions have been trend-selling U.S. Treasuries. Under geopolitical risk shocks, safe-haven funds may make defensive allocations, but they are constrained by the re-inflation trade.

For commodities: (1) Inflation expectations and oil prices have upside momentum; gold prices’ volatility may increase or intensify in certain phases. In the long run, gold’s strategic allocation value remains: a series of policies implemented by the Trump administration cause the global order to gradually fall apart after World War II. In the context of the global order accelerating its reshaping and the trend of geopolitical conditions worsening, safety once again becomes the scarcest resource, while gold is a tangible representation for hedging this uncertainty. But as speculative funds withdraw and the re-inflation trade continues to play out—or greatly intensifies—short-term volatility may rise sharply. (2) Geopolitical conditions in the Middle East continue to worsen; we recommend overweighting crude oil. Global crude oil demand is relatively weak, and OPEC+ production policy is changeable. In recent times, the Middle East geopolitical situation has deteriorated sharply and has a further expansion trend. With the Strait of Hormuz being continuously blocked and crude oil inventories across major economies gradually declining, crude oil prices may still have upside momentum. However, uncertainty in geopolitical developments will further increase volatility in the energy market. (3) The re-inflation trade evolves into a stagflation trade, or it suppresses industrial metal demand. In recent years, the rapid development of power-related construction equipment and transportation tools, as well as AI computing power expansion and updates to military facilities, have brought new demand for industrial commodities. Industrial metals represented by copper may be in a temporary imbalance of supply and demand. But now, as global macro re-inflation trade evolves into a stagflation trade—or suppresses industrial metal demand—it also increases price volatility.

Based on the research framework described in “A Brief Research Introduction to the Active Large-Class Asset Allocation Research System,” published in March 2025, and according to the TAA model calculations in Chapter 3 and the event review conclusions in Section 4.1 of Chapter 4, we update the February 2026 tactical large-class asset allocation plan as follows:

Equity allocation weight is 40.00%: overweight A shares (10.00%), standard allocation to Hong Kong shares (7.50%), standard allocation to U.S. shares (12.50%), standard allocation to European shares (5.00%), and standard allocation to Japanese shares (5.00%).

Bond allocation weight is 40.00%: long-duration government bonds (10.00%), short-duration government bonds (10.00%), long-duration U.S. Treasuries (10.00%), and short-duration U.S. Treasuries (10.00%).

Commodity allocation weight is 20.00%: overweight gold (10.00%), overweight crude oil (6.25%), and standard allocation to industrial metals (3.75%).

05

Risk warning

There are limitations in the analysis dimensions: the research framework is based on analysts’ viewpoints, and the analysis dimensions may not fully reflect market pricing factors.

There is subjectivity in the model design: the selection of factors and weights in the macro factor model is based on a combination of objective and subjective considerations; the objective part comes from quantitative backtesting, while the subjective part comes from experience and judgment, which may lead to some deviation.

There are discrepancies between historical and expected data: the historical data and expected data used in the report may not precisely represent actual market expectations.

Market-consensus expectation adjustments: all conclusions in the report are based on a neutral assumption for market-consensus expectations. If an event occurs that is beyond expectations, causing market expectations and the corresponding macro factor(s) to adjust, it may also lead to changes in the model conclusions.

Limitations of the quantitative model: this conclusion is derived only from the quantitative model and does not overlap with the viewpoints of other research teams at the research institute. For the research institute’s other research teams’ views on the above industries, please refer to the relevant published research reports.

Disclaimer

Guotai Haitong Strategy Team

Massive information, precise interpretation—available on the Sina Finance app