Tianfeng: Focus on three annual themes in 2025: undervalued dividend opportunities, AI catalysis, and consumer recovery.

Tianfeng released a research report stating that the market may have already "pre-voted" and "reduced the dimension" of the main line for 25 years to three scenarios: the undervalued dividend continues to rise; technology under the catalyst of Byte AI; valuation repair of consumer stocks and gradual recovery of consumer stratification. Dividend withdrawals often occur when there is a strong industry trend, so the height of the undervalued dividend depends on the progress of the AI industry trend, and the progress of the AI industry trend depends on the breakthroughs in the AI application end and the consumer end. Depending solely on the AI hardware end, the industry trend may be limited in long-term valuation: the core contradiction of the manufacturing sector giving high long-term valuations is that its growth cannot be linearly extrapolated, and high growth followed by high CAPEX leads to industry deteriorations + technological iterations, which may lead to the disappearance of the industry trend two years later; while the growth of consumer leaders can often be "linearly extrapolated", which is also a major logic for the valuation repair of the consumer sector in 25 years. Key points from Tianfeng are as follows: After the Central Economic Conference at the end of 24 years, the overall A-share market was weak, and the top 5 in terms of relative excess returns to the overall A-share market were banks (undervalued dividends), communications (AI technology), home appliances, commercial retail, and social services (some consumer sectors), corresponding to three scenarios: the undervalued dividend continues to rise; technology under the catalyst of Byte AI; valuation repair of consumer stocks and gradual recovery of consumer stratification. The correlation between banks in 2024 and government bond yields has changed, reflecting a "grading up" logic for sectors like banks with undervalued dividends: 1) From 2020 to 2023, the trading logic of the banking sector was more of "large-cap value cyclical stocks," reflecting expectations for the economic fundamentals, with bank stock movements roughly in sync with the direction of the ten-year government bond yield. 2) The difference in 2024 is that the banking sector and the ten-year government bond yield suddenly turned negative correlated, with an R of 61%, which can roughly be understood as the government bond yield explaining 61% of the fluctuations in bank stocks; this phenomenon reflects a "grading up" in the logic of the banking sector: from "large-cap value cyclical stocks" to "income assets that have to be overweight under the downward trend of government bond yields," which becomes particularly significant when insurance funds and other absolute return funds are the main marginal participants. Most of the time, the undervalued dividend style can provide absolute returns, but there are two situations where there may be absolute withdrawals: One is when the growth trend of the industry is strong and competing for funds, leading to withdrawals; for example, from June 2012 to December 2012, from March 2013 to December 2013, from March 2019 to March 2020. The second is when the previous gains in the CSI Dividend Index are large + the overall market declines, causing resonances, although dividends are theoretically a safe-haven sector, their essence still lies in stocks, influenced by market beta; therefore, such situations may cause significant withdrawals; for example, from June 2015 to January 2016, from January 2018 to October 2018, from June 2024 to September 2024. The height of the undervalued dividend in 25 years needs to focus on the AI industry trend, which has been increasingly evident since the second half of 24: core dividend varieties (Industrial and Commercial Bank of China and China Yangtze Power) gradually become "symbolized," with their rise and fall relatively decoupled from their dividend logic, exhibiting characteristics of "passive rise" during theme market rallies and weak industry trends, and "passive sideways trend" during strong industry trend rallies. The performance of the AI industry trend in 25 years depends on the progress of the AI application end. We have selected manufacturing and consumer leaders from the CSI 800 index, and statistically analyzed the growth rates of non-recurring net profits for all companies in 2021 and 2023. We found that the high growth of manufacturing companies is statistically unsustainable, as manufacturing companies with high growth rates in 2021 did not show higher growth rates in 2023. However, when comparing consumer leaders, we can see that the growth of consumer leaders can be "linearly extrapolated." Therefore, during the interpretation of the AI sector's major market rally (such as March to May 2023 and September to October 2024), there are characteristics of soft logic (media representing applications and consumption scenarios); and the height of the AI industry trend in 25 years also needs to focus on whether there can be leading companies in AI applications and consumption scenes, grasp more AI applications and consumption scenarios, with moats similar to Tencent, Maotai, Yili, and other leading companies, making it difficult for other competitors to quickly replace them through capital expenditures. Industry allocation suggestions: Competitiveness 2.0 is still in the second stage of tug-of-war, return after Spring Festival, waiting for the "bullish period" of consumption. In a market with wide fluctuations and volatile theme market movements, it is difficult for institutions to participate, so it is important to avoid buying high and selling low until the improvement in fundamentals is confirmed after the Spring Festival. In the second stage of "strong expectations, weak reality," policy drivers and thematic rotations will lead to a barbell trend in technology themes + low volatility dividends. Being overly pessimistic about consumption based on macro narratives is a risk, as the core factor for investing in the consumer sector is valuation. In the current environment of low valuations in the consumer sector, declining interest rates, and the recovery phase under policy stimulation (even with a weak slope), with all three core conditions suitable, we are not pessimistic about consumer being a major investment theme in 2025, emphasizing the Hong Kong Internet sector. Risk warning: 1) Past historical experience is for reference only; 2) Style classification is for reference only; 3) Policies may have uncertainties during implementation.