金融工程定期：资产配置月报（2025年1月）

Quantitative Models and Construction Methods 1. Model Name: High-Frequency Macroeconomic Factors - Construction Idea: Utilize asset portfolio simulations to construct a high-frequency macroeconomic factor system, capturing market expectations for macroeconomic changes[10] - Construction Process: 1. Combine real macroeconomic indicators to synthesize low-frequency macroeconomic factors 2. Select assets leading low-frequency macroeconomic factors 3. Use rolling multivariate leading regression with asset year-over-year returns as independent variables and low-frequency macroeconomic factors as dependent variables to determine asset weights and simulate macroeconomic factor trends[10] - Formula: Rolling regression fitting for specific factors: - High-frequency economic growth: Industrial added value YoY, PMI YoY, social retail sales YoY → Hang Seng Index, CRB Metal Spot Index[11] - High-frequency inflation (consumer): CPI YoY → Edible Agricultural Product Price Index, Pork Price Index[11] - High-frequency inflation (producer): PPI YoY → Producer Material Price Index, CRB Industrial Spot Index, CRB Composite Spot Index[11] - High-frequency interest rate: Treasury yield → Short Treasury Index (1-3 years)[11] - High-frequency credit: Credit bond yield - Treasury yield → Corporate Bond AAA Index (long), Treasury Index (short)[11] - High-frequency exchange rate: RMB/USD midpoint → Dollar Index (long)[11] - High-frequency term spread: Long-term Treasury yield - Short-term Treasury yield → Short-term Bond Index (long), Long-term Bond Index (short)[11] 2. Model Name: Duration Timing Model - Construction Idea: Predict yield curve movements using an improved Diebold2006 model and map expected returns for bonds of different durations[19] - Construction Process: 1. Predict level, slope, and curvature factors 2. Level factor prediction: Based on macro variables and policy rate tracking 3. Slope and curvature factors: Predicted using AR(1) model[19] - Recommendation: Hold 1-year short-duration bonds for the next three months based on predictions as of December 31, 2024[19] 3. Model Name: Convertible Bond Valuation and Rotation - Construction Idea: Compare convertible bonds with equities and credit bonds, and implement style rotation within convertible bonds[25] - Construction Process: 1. Relative valuation with equities: Construct "100-yuan conversion premium rate" and calculate rolling historical percentiles to assess relative value[25] 2. Relative valuation with credit bonds: Use "Adjusted YTM - Credit Bond YTM" to measure relative value, stripping out the impact of conversion terms[25] 3. Style rotation: Use 20-day momentum and volatility deviation as market sentiment indicators, rebalancing bi-weekly[29] 4. Model Name: Gold Expected Return Model - Construction Idea: Link gold's forward real return with U.S. TIPS (Treasury Inflation-Protected Securities) real return[30] - Construction Process: - Formula: $E[Real_Return^{gold}]=k\times E[Real_Return^{TIPS}]$ $E[R^{gold}]=\pi^{e}+k\times E[Real_Return^{TIPS}]$ - Parameter k is estimated using extended OLS, with the Fed's long-term inflation target of 2% as a proxy for expected inflation[30] 5. Model Name: Active Risk Budget Model - Construction Idea: Combine risk parity with active signals to dynamically adjust asset weights based on stock-bond relative valuation, stock valuation levels, and market liquidity[33][36] - Construction Process: 1. Stock-bond relative valuation: Use the Fed model to define equity risk premium (ERP): $ERP={\frac{1}{PE_{ttm}}}-YTM_{TB}^{10Y}$[37] - Overweight equities when ERP > 5%, underweight when ERP < 2%[38] 2. Stock valuation: Calculate rolling 5-year historical percentiles of stock valuations; overweight equities when below 25%, underweight when above 75%[39] 3. Market liquidity: Use M2-M1 spread as a proxy; overweight equities when M2-M1 ≥ 5%, underweight when M2-M1 ≤ -5%[41] 4. Aggregate signals into a composite score and convert to equity risk budget weights using the softmax function: $softmax(x)={\frac{\exp(\lambda x)}{\exp(\lambda x)+\exp(-\lambda x)}}$[44] 6. Model Name: Industry Rotation Model 3.0 - Construction Idea: Construct sub-models from six dimensions (trading behavior, prosperity, capital flow, chip structure, macro drivers, technical analysis) and dynamically synthesize signals for bi-weekly industry selection[51][54] - Construction Process: 1. Trading behavior: Capture intraday momentum + overnight reversal effects 2. Prosperity: Capture earnings momentum effects 3. Capital flow: Capture active accumulation + passive distribution behaviors 4. Chip structure: Capture holding returns + resistance-support effects 5. Macro drivers: Map high-frequency macro expectations to industries 6. Technical analysis: Capture trading signals from trend, oscillation, and volume indicators[54] --- Model Backtest Results 1. High-Frequency Macroeconomic Factors - High-frequency economic growth YoY: Upward trend as of December 31, 2024[11] - High-frequency inflation (consumer): Downward trend as of December 31, 2024[16] - High-frequency inflation (producer): Upward trend as of December 31, 2024[16] 2. Duration Timing Model - December 2024 return: 37.8bp; benchmark return: 176.4bp; excess return: -138.6bp[22] - 1-year return: 3.10%; benchmark return: 8.25%; excess return: -5.15%[22] 3. Convertible Bond Valuation and Rotation - "100-yuan conversion premium rate": 23.22% as of December 27, 2024[25] - "Adjusted YTM - Credit Bond YTM": 0.30% as of December 27, 2024[25] - Style rotation (2018/2/14-2024/12/27): Annualized return: 22.38%; max drawdown: 15.54%; IR: 1.39; monthly win rate: 64.63%[29] 4. Gold Expected Return Model - Predicted 1-year return: 22.0% as of December 31, 2024[30] - Past 1-year absolute return: 27.23%[32] 5. Active Risk Budget Model - December 2024 portfolio return: 1.63%; equity weight: 18.50%; bond weight: 81.50%[46] - Full sample (2008/1/2-2024/12/31): Annualized return: 6.67%; max drawdown: 4.89%; return volatility ratio: 1.67; return drawdown ratio: 1.36[50] 6. Industry Rotation Model 3.0 - December 2024 multi-long portfolio return: -2.23%; multi-short return: 2.13%; benchmark return: -2.25%; multi-long excess: 0.02%; multi-short excess: 2.11%[60] - ETF rotation (December 2024): Portfolio return: -3.63%; benchmark return: -1.86%; excess return: -1.76%[68]