In the month of May 2026, major asset performance was ranked as follows: South Korea's KOSPI > Nikkei 225 > STAR 50 > ChiNext Index > Nasdaq > S&P 500 > CSI 300 > European stocks > US Dollar > Chinese bonds > Global bonds > Hang Seng Tech Index > 0 > Gold > Long VIX > Brent crude.

Major assets exhibited a nuanced phenomenon of "tech concentration alongside high rotation in non-tech assets." On one hand, global equity styles were unidirectional, with a deepening concentration in technology stocks, as evidenced by the Philadelphia Semiconductor Index surging 21.1% for the month. On the other hand, non-tech assets failed to establish clear trends, with capital still pricing according to a "Bayesian mode." The NACHO narrative first rose then fell, and the speed of rotation increased during the month due to several tentative shifts.

Key Observations for May

Specifically: (1) Global stocks, bonds, and commodities shifted from their synchronized recovery in April to a pattern of "strong stocks, volatile bonds, weak commodities." (2) Commodities showed deep divergence: "oil down, copper up; steel weak, coal strong; gold oscillating, silver closing higher." (3) The technology concentration in global equity markets, driven by AI profitability, deepened, with Japan and South Korea leading gains. (4) G7 government bond yields peaked simultaneously mid-month and collectively retreated by month-end. US yields experienced a bear flattening, Japanese yields a bear steepening, UK yields an extreme rollercoaster, with the US 30-year yield touching 5.195% intraday. (5) The US dollar rebounded, non-US currencies broadly declined, the Renminbi strengthened against the trend, and the effects of yen intervention were eroded. (6) Domestically, the "triple bull market in stocks, bonds, and currency" continued, but equity elasticity weakened significantly; the bond market diverged sharply from overseas markets, with the yield curve experiencing a bull steepening. (7) The Wind All-A Index closed slightly up 0.6% for the month. AI hardware continued its one-way trend, with optical modules and semiconductors showing greater赚钱效应 and concentration depth than in April. Market breadth fell to a rolling three-year -1.73 standard deviations. Since the start of the year, the performance gap between the strongest and weakest Shenwan industries has widened to 60%. (8) The property market continued its recovery, but the divergence between new and second-hand homes widened. New home sales in 30 cities saw their year-on-year growth slightly drop to near zero. The risk premium between the rental yield of the 100-city index and the 30-year government bond yield remained in the 18-20 basis points range.

Underlying Macro Trading Themes

Looking at the macro trading themes behind asset performance, three core threads have been at play since May 2026. First, the deepening of the global AI computing power theme. Despite the 10-year US Treasury yield rising 6 basis points, capital has been highly concentrated in segments with the most certain earnings realization, compressing the "duration risk" of US tech stocks. Japanese and South Korean equities have become the most elastic representatives within the global AI industrial chain. The forward P/E of the Philadelphia Semiconductor Index is currently around 26-28x (above its historical median but below its 2021 peak of 40x). Second, the NACHO trade first rose then fell. Oil was priced alternately between ceasefire expectations and the reality of physical blockades, volatility intensified in both the global energy sector and global bond markets, and gold and oil exhibited a seesaw relationship in May. A Walsh premium emerged, and the 10-year US Treasury yield broke above 4.6% during the month amid multiple Federal Reserve policy uncertainties. Third, Chinese assets followed a relatively independent logic. Chinese bond yields trended lower overall, with the "China bond-US bond" divergence reaching a historical extreme. The Renminbi appreciated steadily amidst a broad decline in non-US currencies and a deepening inversion of the China-US yield spread. The stock market showed clearer resonance with global styles, characterized by narrowing breadth and leadership from the AI sector. However, research by Kritzman & Turkington (2026) indicates that timing models based solely on rising concentration have historically not generated alpha in the US stock market over the long term. Similarly, our research in "Macro Conditions and Timing Implementation for High-Growth Narratives" finds that for A-shares, neither concentration nor market breadth alone serves as an effective timing signal.

Current Macroeconomic Positioning

What is the current position of the macroeconomic landscape? A comparison of our constructed "hard data" and "soft data" indices shows that the rise in Japanese stocks in May was somewhat validated by both hard and soft fundamentals. US stocks continued to see expectations of recovery precede reality, with hard data showing resilience but soft data weakening. The modest gains in European stocks matched their weak fundamentals. In China, soft data slightly outperformed hard data. From a perspective of expectation gaps, the Citigroup Economic Surprise Index ranking flipped from China > US > Japan > Europe at the start of the month to Japan > US > China > Europe by month-end. Adjusted for official statistics bureau figures, we estimate May's industrial added value growth at 4.50% YoY, retail sales at 0.16% YoY, CPI at 1.42% YoY, PPI at 3.47% YoY (with the price scissors gap widening to 1.95 percentage points), the real GDP index at approximately 4.13%, nominal GDP rising to 6.08%, and the GDP deflator near 2.0%. Within the economic mix of quantity and price, the contribution of prices increased while real growth momentum slowed.

Potential Catalysts for Equity Assets

Where might the next changes for equity assets come from? We believe there are three sources. First, "overseas geopolitical and monetary policy switches." A window for change is approaching. If a US-Iran agreement materializes and the Strait of Hormuz reopens, cost pressures for mid-to-downstream industries could ease, potentially leading to rebounds in some previously suppressed assets. Additionally, the Federal Reserve's June interest rate decision is unlikely to hold major surprises, but the signals it releases will be crucial. If the market interprets the new framework as leaning "hawkish," it could trigger adjustments in some currently high-valuation assets. Second, the implementation of domestic "Six Networks" policies. May's BCI corporate investment forward-looking index hit a yearly high, and the construction PMI bottomed and rebounded, both indicating that the impact of ultra-long-term special treasury bonds and policy-based financial tools is beginning to materialize, with concentrated support for "Six Networks" sectors starting to be transmitted. If fixed asset investment shows improvement in June, pro-cyclical assets, including cyclical and consumer sectors, could be lifted from low levels. Third, the domestic "inflation-growth" scissors gap may have preliminarily peaked. If a turning point is confirmed in June-July, the economy would preliminarily possess broad-based pricing power. While concentration may not completely unravel, it is likely to converge somewhat. The rotation speed among Shenwan secondary industries in A-shares had already shown a marginal rebound by late May. In summary, following a globally quite unidirectional May, June presents a series of "potential clues" that could, to some extent, promote rebalancing.

Major Asset Narrative Factor

Looking at the situation regarding major asset narrative factors. In May, the "herding effect" in major assets strengthened marginally, but cross-asset correlations remained at low-to-medium levels. The CSAD (cross-sectional absolute deviation) fell from the 88.8th percentile in April to within the 80th percentile, fundamentally different from the period of unidirectional narratives from May to September 2025 (30-50th percentile). Major assets are still in a process of "Bayesian updating," with divergent interpretations of the macro picture. Consequently, despite equity market concentration, we observe US stocks relatively strong, A-shares flat, divergence between Chinese and US bond markets, oil down/copper up, steel weak/coal strong, and gold oscillating. As a contrarian indicator, the CSAD has currently retreated from extreme levels to a normal range. The US equity risk premium (ERP) fell to -1.03, at the 93rd percentile since 2000 (74th percentile for the full sample), having touched -1.21 (95th percentile) mid-month. This prices in extreme market optimism towards AI profitability, but index-level value has become extremely mediocre.

M1-BCI-PPI Timing System

Regarding the "M1-BCI-PPI timing system," under the combination of BCI continuing to dip in May 2026, low-base support for M1, and PPI continuing its upward trend, the total cyclical component score for these three factors remained neutral in May (0.60→0.59→0.59). The growth-side BCI declined for two consecutive months, becoming the primary drag among the three-dimensional factors. The May financial data to be released in mid-June is key for the marginal direction. If M1 breaks above 5.4%, the signal might shift from neutral to slightly positive. After additionally incorporating the "herding factor" CSAD, representing narrative trading intensity, the next period's "M1-BCI-PPI-CSAD" total signal also remained neutral (0.50→0.42→0.42). If the benchmark is the CSI 300, this strategy has a monthly win rate of 78.47%, a cumulative return since 2024 of 49.31%, and an alpha of 4.72%. If the benchmark is the CSI 1000, it has achieved an annualized return since 2009 of 16.84%, a Sharpe ratio of 0.73, and a monthly win rate of 79.43%; the cumulative return since 2024 is 41.90%; the YTD return for 2026 is 16.32%, with an alpha of 5.06%.

Equity-Bond Relative Value Timing Strategy

Regarding the equity-bond relative value timing strategy. In May, the 10-year Chinese government bond yield fell 3.8 basis points to 1.70%, a new low for the year, while the All-A dividend yield also declined. The relative value of domestic bonds versus equities did not weaken substantially. In terms of timing signals, as of early June 2026, the centrally-adjusted equity-bond yield spread ("10-year government bond yield minus dividend yield") stood at the 58.8th percentile since 2015, at +0.79 standard deviations on a rolling three-year basis, having fluctuated narrowly near the 60th percentile for two consecutive months. Combined with the neutral signal from the three-dimensional M1-BCI-PPI-CSAD win rate, and domestic nominal growth showing "quantity contraction, price rise" (PPI up to 3.47% but real GDP slowing to 4.13%), the relative signal tends towards neutral. The "M1-BCI-PPI" + equity-bond relative value strategy has achieved a return of 53.9% since the beginning of 2024, with an alpha of 21.2%; the "M1-BCI-PPI-CSAD" + equity-bond relative value strategy has achieved a return of 51.72% since the beginning of 2024, with an alpha of 19.43%.

Valuation-Macro Deviation Framework

Regarding the "valuation-macro deviation" framework, a reading of +1.74 standard deviations suggests that the valuation safety margin for domestic stocks has been partially realized. We measure market pricing position using the All-A "P/E minus nominal GDP growth" – historically, +2 standard deviations is an extreme level, and breaching the warning level indicates that "low risk premium" will exert pricing power. The periods of September 2025 (+2.56σ) and January 2026 (+2.01σ) both corresponded to subsequent market adjustments. As of June 2, the Wind All-A P/E closed at 23.95x, with the deviation slightly rising to +1.74σ (previous month +1.73σ). Observing US stocks in the same manner, the S&P 500 CAPE deviation jumped from +1.06σ to +1.53σ, indicating more pronounced forward pricing in US stocks. Using a benchmark of 50% Wind All-A + 50% cash, this strategy has achieved a return of 54.4% since the beginning of 2024, with an alpha of 22.6%; the 2026 YTD return is 9.97%, with an alpha of 4.67%. After integrating the three-dimensional timing signal, the return since 2024 is 45.5%, with an alpha of 15.5%. The current deviation, combined with the three-dimensional signal, suggests maintaining a neutral equity allocation for June.

High-Growth Timing Model

Regarding the high-growth timing model, the latest iteration (indicative for June 2026) of the "5+1" timing model suggests maintaining a neutral allocation to the technology sector relative to the All-A Index, with the timing score unchanged from the previous value. On the macro risk side, the US Treasury MOVE volatility index oscillated higher after the synchronized shock to global bond markets in May (once rising above 86 mid-month), offsetting the continued convergence of nominal GDP downside volatility. The macro uncertainty facing tech assets shifted from "risk clearing" in April to "volatility rebound." Domestic nominal growth consolidated its upward trend but remains at low levels. The external liquidity score declined, but the relatively loose domestic liquidity landscape provided a hedge, making duration risk neutral. Industrial narrative pricing improved slightly (P/S-P/E spread convergence), but the risk premium disadvantage for tech assets deepened. Since the beginning of 2024, the strategy has achieved a return of 83.96%, with an alpha of 16.11%.

Dividend Asset Timing Model

Regarding the dividend asset timing model, the latest iteration (indicative for June 2026) of the six-dimensional timing factors suggests the dividend score continued to rise slightly, with the allocation relative to the Wind All-A Index adjusted to 45%. There were three changes in win rate. First, the US Treasury yield once broke above 4.67% mid-month, and MOVE surged above 86; rising external interest rates favor short-duration assets with stable cash flows. Second, the Chinese bond yield fell to 1.70%, a new low for the year; the "asset scarcity" logic highlights the coupon advantage of dividends (dividend yield 5.1% vs. government bond 1.70%). Third, risk appetite recovered, and the All-A equity risk premium level declined, which is unfavorable for dividends. However, inflows into coal, utilities, and banking following the late-May tech crowding squeeze provided marginal support for dividends through crowding relief. In terms of value, the CSI Dividend Index dividend yield has broken through the 5.0% value threshold, and its dividend yield advantage relative to the Wind All-A has risen to the 62nd historical percentile. The crowding of dividend assets further decreased from -1.2σ to -1.5σ, with the extreme gap in crowding versus tech assets reaching its widest this year. This strategy has achieved a return of 69.61% since 2024, with an alpha of 22.92% compared to a "50% Wind All-A + 50% CSI Dividend" benchmark.

Cross-Asset Win Rate and Value Comparison Framework

Regarding the cross-asset win rate and value comparison framework. On win rate: The May PMI fell to the 50.0% boom-bust line critical level. The EPMI and BCI historical percentiles fell to 67% and 39.2% respectively, indicating no improvement in the momentum for broad-based recovery. The K-shaped gap between high-tech manufacturing (PMI 52.9) and traditional industries (consumer goods PMI 49.7, raw materials PMI 47.1) widened, and construction sector sentiment preliminarily bottomed. The upward price trend consolidated in May, but the pattern of upstream recovery outpacing downstream strengthened further. Narrow liquidity remained extremely loose, while broad liquidity is not a catalyst for equity valuations. Q2 fixed asset investment and social financing data are key nodes for verifying whether the "domestic demand switch" has been turned on. On value: In May, US Treasury yields continued to maintain their position at the top of the value spectrum among major assets. The China-US yield spread inversion deepened further to -277 basis points. A-shares rose slightly while US stocks accelerated upwards; the "A-share minus US stock" dividend yield differential rebounded to 0.70% (-0.55σ), indicating a renewed expansion of A-shares' value advantage relative to US stocks. The "US Treasury minus A-share" dividend yield spread ratio rose to +0.94σ, approaching the threshold where US Treasuries are favored over A-shares (+1.0σ). The "China bond minus A-share" perspective from a dividend yield standpoint is at the 59th percentile (+0.79σ), a neutral signal. The copper-gold ratio rebounded to -0.22σ on a rolling three-year basis, while the copper-oil and gold-oil ratios recovered upwards to +0.3σ and +0.4σ. In summary, after A-shares' modest gains and US stocks' accelerated rise in May, A-shares' value advantage relative to US stocks re-expanded; domestic long-term bonds have realized some value but still have room; US Treasury value rose comprehensively compared to both A-shares and US stocks. Commodity ratios reflect that overseas risk-on sentiment is being driven by expectations of a macro recovery.

Equity-Bond Asset Scarcity Framework

Regarding the equity-bond asset scarcity framework. The latest iteration (for mid-June to mid-July 2026) shows the intensity of bond market asset scarcity has recovered from the previous period's "underweight" to "neutral." Credit spreads have entered a balancing period after reaching historical extremes. Equity asset scarcity divergence intensified, with a neutral composite score. The end-of-month value for micro trading concentration breached the upper bound of the three-year ±1 standard deviation band, but the monthly average remained in the neutral range. The premium for certainty remained above the threshold but its decline accelerated (1.58→1.08), indicating marginal visibility of concentration fragility. The market's pursuit of certainty remains high, coexisting with scarce high-growth sectors and concentration risk.

Gold Pricing Model

Looking at the gold pricing model. Since May, spot London gold, COMEX gold futures, and SHFE gold have all recorded their third consecutive monthly decline. The 10-year US TIPS real yield rose from 1.94% to 2.07% (once touching 2.18% mid-month). The duration parameter in the "duration-convexity" model slightly decreased to 10.05 but remains far above the 4.6 level at the end of 2025. Although the pricing power of the "interest rate anchor" has loosened from extreme levels, it maintains strong pricing power, with no clear medium-to-long-term catalysts for gold prices yet evident. Under cautious and optimistic rate cut scenarios, the model's central prices are $4442 and $4766 per ounce respectively, with the current gold price lying between them. On the trading front, sentiment from both domestic and international ETF flows was consistently weak, but positioning意愿 from trading-oriented institutions showed initial signs of stabilization. SPDR holdings decreased by 7.77 tonnes; domestic gold ETFs saw a net outflow of 9.403 billion yuan, with outflows for three consecutive weeks. However, the COMEX net long percentage recovered to 32.9% (85.8th percentile), and the RSI closed at 43.61 in the neutral zone. The gold-silver ratio remained in the normal 60-63 range. Gold implied volatility held around 25% for the month, with a slight rise at month-end roughly synchronized with the back-and-forth in US-Iran tensions. Geopolitical tail risks continue to disturb gold's risk-reward profile.

Near-Term Key Uncertainties

Short-term major uncertainties are detailed in a prior report titled "Four Switches Behind Assets." Currently, all four switches are at critical junctures for directional decisions: Fluctuating Federal Reserve policy expectations may impact equity assets that have already undergone valuation expansion, especially as the value buffer for US stocks has been rapidly realized, with the US VIX index operating in the historically low 15-17 range throughout the month. Geopolitical tensions have intensified since June, and the global physical crude oil inventory shortage is nearing a critical point, posing a potential tail risk. US employment data for April-May showed divergent characteristics, and the impact on AI-exposed industries is entering an observation period. The trajectory of domestic fixed asset investment data in May-June is also key for determining whether a pro-cyclical fundamental recovery can materialize.