Quantitative strategies promote the K-shaped recovery? CITIC SEC: After the cooling down of the game, it is necessary to pay attention to the repair of some non-AI sectors.

CITIC SEC released a research report stating that three narrative logics correspond to three levels of K-shaped differentiation, among which there is a reasonable differentiation based on the difference in economic trends, which is common in global markets, but there are also additional impacts caused by emotions and fundaments. Since June, domestic AI has been independently strengthening apart from overseas, making it the strongest K-type currently. This includes both industrial narratives and potential impacts from quantitative strategy rebalancing. The core is that some quantitative stock selection and index enhancement strategies need rebalancing under the pressure of excess returns and fund pressure, by increasing exposure to factors similar to "double innovation" to mitigate negative excess returns. The subjective long positions on core stocks of the Growth Enterprise Market are significantly stronger in terms of allocation and pricing power than those of the Science and Technology Innovation Board led by ETFs. This difference in ownership structure means that quantitative strategies have a more significant impact on the Science and Technology Innovation Board in the process of increasing exposure to the "double innovation" factor, with a more prominent driving force. Observing the trend since June, more and more mutual fund index-enhancement products have transitioned from continuous negative excess returns to positive excess returns, proving that this rebalancing is gradually progressing. Overall, the K-shaped differentiation in the domestic market is mixed with too many short-term narratives and fund impacts. After the cooling of fund games, it is necessary to pay attention to the recovery of some non-AI sectors. The main points of CITIC SEC are as follows: Three narrative logics correspond to three levels of K-shaped differentiation In the past few weeks, it has been repeatedly emphasized that the global K-shaped differentiation has been strengthened by narratives, which can be broken down into three clues. The first is the differentiation between AI and non-AI globally, with a strong U.S. dollar and rate hike expectations acting as macro amplifiers, and also the most easily loosened of the three, this differentiation stems from the difference in economic trends between AI and non-AI (resulting in different sensitivity to interest rates), and has its rationale, being relatively common globally. The second is the unique narrative of "carbon-based" versus "silicon-based" in the A-share market, causing significant underperformance of non-AI sectors in A-shares compared to similar sectors overseas, even if they operate the same global exposure business. This differentiation stems from the weak domestic demand data in China affecting fund sentiment, with domestic investors scrutinizing non-AI stocks more rigorously. The third is the outperformance of domestic AI since June relative to global AI. This is not simply an extension or mapping of the prosperous North American AI chain, but includes factors such as the autonomous and controllable valuation premium of the domestic chain, the recovery of parts of the North American chain that lagged since last year, and the impact of rebalancing from quantitative/index-enhancement strategies. Strong performance of domestic AI separately from overseas, driven by both industrial narratives and quantitative strategy rebalancing 1) Quantitative stock selection and index enhancement strategies are facing pressure on returns and funds, requiring rebalancing. According to the Ppww data of private funds, as of June 30, there were 1236 index-enhancement products with performance shown, with an average excess return in the first half of the year of only 3.1%, significantly lower than the 14.2% of the same period last year. The sample of private funds tracking the price index of index-enhancement products shows that the average excess returns of the Shanghai and Shenzhen 300 index-enhancement, CSI 500 index-enhancement, and CSI 1000 index-enhancement products in the first half of this year were 5.02%, 2.84%, and 2.94%, respectively, which were significantly lower than the 7.60%, 10.86%, and 14.50% of the same period last year, showing a noticeable reduction, especially for the CSI 1000 index enhancement products. The slide in excess returns will bring redemption pressure and may force some managers to adjust their strategies and rebalance their portfolios. 2) To accurately capture the excess returns in this round of market trends, it is necessary to introduce the "double innovation factor". The continued downturn in micro-cap stocks in recent period is one of the reasons for the sliding excess returns of quantitative stock selection and index-enhancement products. The equally weighted index of WIND Micro-Cap Stocks has fallen by 21.4% from the high point in mid-May to July 10, while the Science and Innovation 50 index has risen by 10% during the same period, posing a significant challenge to strategies with exposure to micro-cap stocks. The relative cumulative excess return of the micro-cap index to the Science and Innovation 50 index (since 2025) has fallen from a peak of 51.0% to -25.8%. As of July 10, only 7 out of the 30 primary industries have positive yields (electronics 65.6%, communications 52.3%, building materials 25.2%, machinery 9.7%, coal 7.7%, basic chemicals 5.7%, power and utilities 0.4%), with the vast majority of industries turning negative. The division between sectors and industries is so significant that as long as traditional quantitative factors have wide coverage, they may drag down returns. 3) Differences in ownership structure make the rebalancing of the "double innovation factor" have a more significant effect on the Science and Innovation Board than on the Growth Enterprise Market. By the end of the 2025 annual report (calculated based on the top ten shareholders), institutional holdings of the Growth Enterprise Market Index accounted for 67%, while institutional holdings of the Science and Innovation 50 accounted for only 42%. In addition, from an industry and thematic perspective, institutions with significant exposure to the technology track ETFs on the Science and Innovation Board account for only 25%. The institutional concentration of holdings in the constituent stocks of the Growth Enterprise Market is higher, and the core targets are held long-term by stable institutions such as public funds, making the chips more "sticky". Individual investors have a higher proportion of holdings in the constituent stocks of the Science and Innovation Board and ETFs, with fewer stable institutions absorbing the free float, resulting in greater price elasticity corresponding to marginal inflows of funds, and usually leading to panic and trading based on narrative changes, with emotions having a stronger secondary amplification effect on prices. In terms of trading turnover, the Science and Innovation Board has significantly higher turnover and volatility than the Growth Enterprise Market, and the lower the proportion of institutional holdings, the more significant the impact of quantitative stock selection and index enhancement funds on the Science and Innovation Board. The sensitivity of the Science and Innovation 50 and the Growth Enterprise Market 50 to micro-cap stocks increased simultaneously from January to June this year, reflecting a stronger linkage with small-cap/quantitative funds; however, since June, the Science and Innovation Board has clearly decoupled from micro-cap stocks and the correlation rapidly declined, during which the Science and Innovation Board has seen an independent trend with higher volatility, represented by semiconductors. This decoupling coincides with the weakening performance of micro-cap stock strategies this year, indicating that quantitative stock selection/index-enhancement funds may be shifting their focus from micro-cap/small-cap exposure towards the "double innovation" direction. The Science and Innovation Board is stronger than the Growth Enterprise Market, and significantly stronger than the markets in South Korea and the United States, with the expression in terms of industrial narratives being that the domestic AI chain is considerably stronger than the North American AI chain. The proportion of positive excess returns of index-enhancement products is increasing, and quantitative strategy rebalancing is gradually progressing Using a sample of 221 broad-based mutual fund index enhancement initial funds covering the Shanghai and Shenzhen 300, the CSI 500, CSI 800, and CSI 1000, the daily cumulative excess returns of each fund relative to its benchmark were calculated. Since the beginning of this year, the pace of excess returns accumulation of mutual fund index enhancement products was similar to that of last year until mid-April, but started significantly underperforming last year from the end of April, continuing until mid-May, exactly coinciding with the intensification of the K-shaped differentiation. The more extreme differentiation between AI and non-AI reduces the effectiveness of traditional quantitative factors and affects portfolio returns, forcing quantitative products to rebalance. However, since mid-May, the average cumulative excess return rate of mutual fund index enhancements has gradually begun to rise, reaching 1.76% on July 8 (compared to 3.42% during the same period last year), even surpassing the previous high of 1.43% on April 3, indicating that many products have indeed adjusted their portfolios under the pressure of K-shaped differentiation. Looking at monthly average excess returns, the samples were -1.81%, -1.22%, and -1.02% for April to June, and +0.62% since July; similarly, the proportion of products with positive excess returns in April to June was 41.18%, 51.58%, and 51.13%, and since July has been 71.49%. Overall, the absolute level of negative excess returns and the proportion of products with negative excess returns is decreasing month by month, reflecting that many mutual fund index enhancement products are actively adjusting their holdings, which is further intensifying the K-shaped differentiation. The K-shaped differentiation in the domestic market is mixed with too many short-term narratives and fund impacts, and after the cooling of fund games, it is necessary to focus on the recovery of some non-AI sectors Since March of this year, three narrative logics have driven the differentiation of the three K-shapes: the differentiation between AI and non-AI globally, the unique narrative of "carbon-based" versus "silicon-based" in the A-share market, and the outperformance of domestic AI relative to global AI since June. There is a reasonable differentiation based on the difference in economic trends, relatively common globally, as well as additional impacts caused by emotions and funds. As more and more products complete passive rebalancing, the impacts of emotions and funds will diminish, and after the cooling of fund games, it is necessary to focus on the recovery of some non-AI sectors. Specifically, in terms of allocation, for the AI side, there is a preference for categories with strong supply constraints, undervalued valuations, and focuses on midstream and downstream, such as cloud service providers, storage, servers, gas turbines, and diesel generators; for the non-AI side, the focus is on certain basic metals (copper, tin), new energy (waiting for the negative industry sentiment to dissipate and return to objective performance), and chemicals (refrigerants, phosphorus chemicals, spandex, dyes, large refining, etc.). In addition, the continued bullish outlook on the overseas logic continuing and relatively thorough clearance of innovative drugs, with the recently released new version of the basic drug catalog being the first adjustment since 2018, which may have a significant positive impact on the profit forecasts of certain companies. Still positive on undervalued securities firms, with the suppression of liquidity and other defects expected to gradually fade in the second half of the year, where short-term logic can focus on investment income and interim report performance, while the long-term logic is the broader trend of Chinese companies driving overseas investment and financing, leading to the globalization of securities business. Risk factors Escalation of friction in the technology, trade, and financial sectors between China and the U.S.; Potential weakening of the effectiveness of domestic policies, implementation, or economic recovery falling short of expectations; Unexpected tightening of macro liquidity at home and abroad; Further escalation of conflicts in regions such as Russia, Ukraine, and the Middle East; Failure to meet expectations in the digestion of real estate inventories in China.