高维宏观周期驱动风格、行业月报(2025/11)：经济景气下行、通胀细分项下行看好大盘价值风格-20251208

• The report constructs a high-dimensional macro cycle recognition framework based on five dimensions: economic prosperity, inflation, interest rates, inventory, and credit. This approach addresses the instability of single-dimension indicators and improves the identification of macroeconomic cycles[9][8][3] - Macro factor variables are constructed by regressing macro indices against broad-based indices and proxy macro variables, selecting significant variables based on t-values. The inverse of the past year's standard deviation is used for weighting. Additionally, a one-sided HP filter is applied to macroeconomic data to eliminate short-term fluctuations and focus on long-term trends. Macro trends are categorized using factor momentum (upward or downward) and macro states are divided into high, medium, and low positions using time-series percentiles[2][8][9] - A broad-based timing strategy for CSI All Index is constructed using macro variable combinations. The strategy predicts future returns based on liquidity and inventory sub-strategies. If either prediction exceeds a threshold (0.6), the index is bought; otherwise, it is sold. From January 2012 to November 2025, the strategy achieved an annualized return of 14.71%, with an excess return of 10.5% relative to the CSI All Index[29][30][33] - A dividend index timing strategy is constructed using macro variable combinations, specifically inflation+inventory and inventory+credit sub-strategies. The average prediction value determines whether to buy or sell the index. From January 2012 to November 2025, the strategy achieved an annualized return of 10.64%, with an excess return of 8.41% relative to the dividend index[35][39][38] - A style rotation strategy is constructed using effective macro factor combinations, specifically inflation+inventory and inflation+credit. Monthly predictions of future returns for six style indices are ranked, and the top two indices are equally weighted for allocation. From September 2014 to November 2025, the strategy achieved an annualized return of 12.64%, with an excess return of 5.49% relative to equal-weighted style indices[43][47][50]