Group 1: Construction of High-Growth Earnings Portfolio - The selection of high-growth earnings stocks is based on the top 80% of total market capitalization and average daily trading volume within the CSI All Share Index, excluding those with negative net profit from the previous year, and selecting the top 50% based on analysts' consensus earnings growth expectations [7][9][13]. - The average number of stocks in the high-growth earnings portfolio during the backtesting period from August 31, 2011, to October 31, 2025, is 571 [9]. - The portfolio is rebalanced at the end of April, August, and October, and stocks are selected based on the expected earnings growth factor [15][21]. Group 2: Achievement of High Earnings Growth - From April 27, 2012, to October 31, 2024, the median earnings growth rate of the high-growth earnings portfolio is 99%, significantly higher than the median earnings growth rate of 36.86% for the overall stock pool [24][28]. - The high-growth earnings portfolio consistently ranks in the top two deciles of the overall market earnings growth distribution during the backtesting period [28][30]. Group 3: Predictive Factors for Earnings Growth - The analysis identifies several factors that effectively predict annual earnings growth, including analyst ratings, growth, profitability, and valuation [35][37]. - The RankIC (Rank Information Coefficient) values indicate that analyst and growth factors have the most significant predictive power for identifying stocks with higher earnings growth [37]. Group 4: Differences Between Progressive and Parallel Stock Selection - The high-growth earnings portfolio employs a progressive stock selection method, first filtering stocks based on expected earnings growth and then refining the selection using changes in analyst earnings forecasts [43][46]. - In contrast, parallel stock selection methods, which use both factors simultaneously, have shown to be less effective in terms of annual returns over the backtesting period [47][49].

Group 1: High Growth Portfolio Construction - The high growth profit stock pool is selected from the top 80% of stocks by market capitalization and average daily trading volume, excluding those with negative net profit from the previous year, and then selecting the top 50% based on analyst consensus profit growth expectations[6] - During the backtesting period from August 31, 2011, to October 31, 2025, the average number of stocks in the high growth profit stock pool was 571[11] - The high growth profit portfolio achieved an annualized return of 13.32% compared to the CSI 500 total return index's 7.31% during the same period, with a Sharpe ratio of 0.50[13] Group 2: Performance Attribution and Comparison - The median profit growth rate of the high growth profit portfolio was 99% during the period from April 27, 2012, to October 31, 2024, significantly higher than the median profit growth rate of 36.86% for the stock pool[29] - The high growth profit portfolio consistently ranked in the top two deciles of market profit growth during the same period, indicating strong performance relative to the market[32] - The portfolio's performance was primarily driven by growth and analyst factors, aligning with the portfolio construction logic[23] Group 3: Factor Analysis and Selection Methodology - The analysis identified that the analyst factor, growth factor, and profitability factor had significant predictive power for annual profit growth, with average RankIC values of 37.81%, 30.22%, and 33.46% respectively[41] - The sequential selection of factors in constructing the portfolio outperformed parallel selection methods, with the sequential method winning in 9 out of 14 years from 2012 to 2025[59] - The high growth profit portfolio's average market capitalization during the backtesting period was 15.8 billion yuan, indicating it is a mid-cap growth portfolio[20]

Group 1: Report's Core View - The report focuses on constructing high-growth and stable portfolios based on predicted performance. It analyzes the performance of stock portfolios with known annual profit growth rates, screens high-growth stock pools by domain, and constructs portfolios using specific factors [1][4][16] Group 2: Future Perspective - Performance of Stock Portfolios with Known Annual Profit Growth Rates - In the CSI All-Share sample space, from 2011/12/30 to 2025/4/30, stocks were screened at the end of April, August, and October each year. After certain filtering steps, stocks were grouped into deciles by annual profit growth rate, and equal-weighted portfolios were constructed. The table shows the annual and annualized returns, volatility, and Sharpe ratios of the top four groups [6][10] Group 3: Domain-based Screening of High-Expected Profit Growth Stock Pools - In the CSI All-Share sample space, at the end of April, August, and October each year, samples were screened by market capitalization and trading volume and divided into two sub-samples based on analyst coverage. Two high-expected profit growth stock pools (Domain 1 and Domain 2) were obtained. During the backtesting period from 2011/12/30 to 2025/7/31, the long-term annualized returns of the equal-weighted portfolios of the two domains were similar, but Domain 2 was more affected by small and micro-cap stocks [18][20][23] Group 4: Domain-based Construction of High-Growth and Stable Portfolios High-Growth Portfolio Construction - In Domain 1, the top 50% of samples in terms of consensus expected performance growth were selected. High-growth portfolios were constructed by selecting stocks based on the consensus expected performance change factor. During the backtesting period from 2011/12/30 to 2025/7/31, the portfolio was updated at the end of April, August, and October each year. As the number of holdings increased, the annualized return decreased slightly, and the median market capitalization and turnover rate varied little [42][44][47] Stable Portfolio Construction - In Domain 2, the top 50% of samples in terms of performance acceleration were selected. Stable portfolios were constructed by selecting stocks based on the volatility factor. During the backtesting period from 2011/12/30 to 2025/7/31, the portfolio was updated at the end of April, August, and October each year. As the number of holdings increased, the annualized return decreased slightly, and the median market capitalization and turnover rate varied little [42][60][61] Group 5: Appendix - Improvement of Analyst Factors - A new method is proposed to calculate analyst factors based on the adjustment of earnings forecasts after performance disclosure. Two sub-factors (consensus expected net profit change and analyst earnings upgrade ratio) were reconstructed, and the stock selection effects of the improved factors were tested against the initial factors [68][71][82]

Report Industry Investment Rating No relevant content provided. Core Viewpoints of the Report - The China Securities Regulatory Commission has strengthened the binding effect of performance comparison benchmarks, which will impact fund company evaluation and fund manager assessments [8]. - A new dividend - growth portfolio is constructed to reduce the annualized tracking error compared to the original portfolio and achieve better returns [8][19]. - A preliminary exploration of the Hong Kong stock dividend - growth strategy shows different performance results under different construction methods [33][37]. - A new method is proposed to calculate analyst - related factors, and the improved factors are tested for their stock - selection effectiveness [42]. Summary by Relevant Catalogs 1. Adjusting the Dividend - Growth Portfolio under Benchmark Constraints - The China Securities Regulatory Commission issued the "Action Plan for Promoting the High - Quality Development of Public Funds", emphasizing the binding effect of performance comparison benchmarks on fund company and manager evaluations [8]. - From 2016 to 2024, the annualized tracking error of the original dividend - growth portfolio relative to the CSI Dividend Index was 9.08%, while that of the new portfolio was 5.82% [8]. 2. Construction of the New Dividend - Growth Portfolio - The new portfolio is constructed from two dimensions: past three - year stable dividend ratio and expected profit growth, and past two - year consecutive dividend increase and expected profit growth [12]. - The stocks screened from the two dimensions are combined to form a stock pool with an average of 572 stocks since 2016 [13]. - Further stock - selection is carried out in the stock pool: first select the top 100 stocks with the highest average dividend yield in the past three years, then select the top 50 stocks with high growth factors, and weight the stocks in the portfolio by the latest annual dividend yield [17][18]. - From 2016/1/1 to 2025/5/31, the new dividend - growth portfolio had an annualized return of 12.08%, compared to 6.84% of the CSI Dividend Total Return Index, with an annualized excess return of 5.24% [19]. - In terms of dividend yield, the new portfolio is similar to the CSI Dividend Index historically and slightly higher in the next - year weighted dividend yield [23]. - In industry distribution, the new portfolio is under - weighted in transportation, power and utilities, steel, and petroleum and petrochemical industries, and over - weighted in banking, construction, and non - banking finance industries compared to the CSI Dividend Index [25]. - In terms of market capitalization, the new portfolio has a smaller market capitalization than the CSI Dividend Index but is still large - cap in absolute terms [28]. 3. Preliminary Exploration of the Hong Kong Stock Dividend - Growth Strategy - From a posteriori perspective, in the Hong Kong Stock Connect sample pool, a known dividend - growth portfolio and a benchmark portfolio are constructed. From 2014/12/31 to 2025/4/30, the annualized return of the known dividend - growth portfolio was 14.21%, compared to 8.28% of the benchmark portfolio, with an annualized excess return of 5.92% [33][34]. - Using a simple momentum strategy to construct a Hong Kong stock dividend - growth portfolio did not achieve significant excess returns, indicating its ineffectiveness [37]. 4. Appendix: Improvement of Analyst Factors - A new method is proposed to calculate analyst - related factors, considering the timeliness of analyst earnings forecasts after performance announcements [42]. - Taking the "consensus expected net profit change" and "analyst earnings upgrade ratio" as examples, the improved factors are reconstructed and their stock - selection effectiveness is tested [45]. - Back - testing shows that the improved factors generally have better stock - selection effectiveness than the initial factors [47][48][57].