Zhongtai: The market shows the characteristics of large funds rushing to grab small funds. Institutional willingness to go long remains strong.

Zhongtai releases research report stating that since the easing of tensions between the US and Iran at the end of March, the market has shown characteristics of large funds rushing in and small funds, indicating that institutional bullish sentiment remains strong. In terms of sector allocation, in the past two weeks, institutional funds have mainly flowed into the TMT sector. Investment recommendations include: first, monitor the price increases in AI and the disclosure progress of sectors related to the first quarter report. Second, non-banking financial institutions have high 25-year performance growth rates and are currently undervalued, with recent increased attention from large funds, suggesting potential strong valuation recovery space in the future. Key points from Zhongtai: 1. Methodology: Describing fund behavior based on order splitting The differences in behavior of different investors may lead to differentiated long-term investment returns. By tracking the behavior of "smart funds" within the investor group, one can to a certain extent "follow" and increase returns. This article attempts to separate the behaviors of investors with different fund sizes based on order splitting, and determine whether these behaviors serve as guides for investment. Specifically, orders are divided into small orders, medium orders, large orders, and super large orders, corresponding to four categories: less than 40,000 RMB, 40,000-200,000 RMB, 200,000-1,000,000 RMB, and over 1,000,000 RMB, to describe the behaviors of retail investors, medium investors, large investors, and institutional investors. The daily flow of funds has randomness, therefore the cumulative net inflow characteristics of funds are mainly observed. From actual data observations, different fund sizes show distinct characteristics in terms of cumulative net inflows. Small fund orders tend to show continuous net inflows, while large and super large fund orders tend to show long-term net outflows. This phenomenon may be due to two reasons: structural biases in trading behavior, where investors tend to use small amounts for buying in the accumulation phase, and larger amounts for selling during reduction or liquidation; and differences in long-term returns among different investors. If some funds are experiencing long-term losses or low profits, they will continue to show relatively large purchases and sell for amounts lower than the purchase amount after incurring losses. This article uses the HP filtering method to remove trend components from the cumulative net inflow sequence, thereby separating the four types of marginal changes in funds. After this processing, the marginal changes of funds of different sizes at specific stages can be more clearly described. 2. Historical review: How does fund structure changes guide bull and bear cycles? Overall, the fund behavior described based on order splitting shows clear leading characteristics in multiple bull and bear cycles. Institutional fund net inflows often correspond to market bottom areas and the beginning stages of primary uptrends, while their net outflows are more likely to occur in the later stages of the primary uptrend and during market consolidation and volatility periods. Compared to other types of funds, super large orders (institutional) often exhibit stronger foresight and better timing abilities during market operations, to a certain extent possessing characteristics of "smart funds." The firm has also constructed a Wind A-share Index timing model to validate the effectiveness of "smart funds". The core logic is to be fully invested after identifying the "institutional buying, retail selling" characteristic in the market, and to be out of the market after a reversal. Based on backtesting results, this strategy can effectively capture the major upward trends in the market, and maintain lower positions during periods of consolidation and downtrends, significantly improving the risk-return ratio. 3. Industry review: The guiding ability of "smart funds" in structural market trends From an industry perspective, the behavior of large funds shows structural differences. In cyclical sectors, institutional funds are more inclined to continue positioning during downturns and low valuation levels, and gradually realize profits during the process of valuation recovery and uptrends; in the consumer sector, large funds overall are highly synchronized with market trends, more exhibiting marginal adjustments in trend following, with relatively limited leading capabilities. The behavior of funds in the growth sector is more complicated, and although there may be certain "misjudgments", they can still capture most of the primary uptrend. 4. Industry allocation: Horizontal comparison strategy based on "smart funds" In addition to index timing, this article further expands the "smart funds" approach to the industry level, constructing a horizontal comparison and optimal allocation strategy based on fund behavior. Firstly, the marginal changes in cumulative net inflows of institutional funds within primary industries are calculated after trend removal, and then subjected to percentile processing for comparability across different industries. Based on this, industries are ranked at each rebalancing point, favoring industries with leading net inflows for allocation, thereby illustrating how changes in market structure guide market trends. Backtesting results show that using the past 10-day net inflows as the basis for ranking, combined with weekly rebalancing, the portfolio can achieve a stable excess return. The highest portfolio annualized return reaches 9.15%, while the lowest is -1.39%, with significant and stable differences between long and short positions. Strategies that do not undergo percentile standardization show more prominent performance in terms of returns, with an annualized return of 13% from 2016 to 2025, but display relatively weaker stability in long and short portfolio returns. From the results, percentile processing sacrifices some return elasticity, but enhances the robustness and comparability of the strategy. Risk warnings: Changes in macroeconomic and policy environments, market liquidity fluctuations, changes in market style characteristics, model failure risks, and historical performance does not guarantee future results.