大类资产配置模型周报第31期：商品指数与国债指数收涨，BL策略上周收益领先-20250624

Quantitative Models and Construction Methods 1. Model Name: Black-Litterman (BL) Model - Model Construction Idea: The BL model is an improvement over the traditional Mean-Variance Optimization (MVO) model. It integrates subjective views with quantitative models using Bayesian theory to optimize asset allocation weights. This approach addresses MVO's sensitivity to expected returns and provides a more robust asset allocation framework[12][13]. - Model Construction Process: 1. Select underlying assets, such as equity indices, bond indices, commodities, and gold. 2. Incorporate subjective views on market returns into the model using Bayesian theory. 3. Optimize asset allocation weights based on the combined subjective and quantitative inputs. 4. Construct two domestic BL models and two global BL models using selected assets like CSI 300, Hang Seng Index, S&P 500, and others[13][14]. - Model Evaluation: The BL model effectively combines subjective and quantitative perspectives, offering higher fault tolerance and efficient asset allocation solutions compared to purely subjective or traditional MVO approaches[12]. 2. Model Name: Risk Parity Model - Model Construction Idea: The risk parity model aims to equalize the risk contribution of each asset (or factor) in the portfolio. It is an improvement over the traditional MVO model, focusing on risk distribution rather than return maximization[17]. - Model Construction Process: 1. Select appropriate underlying assets, such as equity indices, bond indices, commodities, and gold. 2. Calculate the risk contribution of each asset to the portfolio based on expected volatility and correlations. 3. Optimize the portfolio weights to ensure equal risk contribution from each asset. 4. Construct domestic and global risk parity models using selected assets like CSI 300, S&P 500, Hang Seng Index, and others[18]. - Model Evaluation: The model provides a stable performance across economic cycles by balancing risk contributions, making it suitable for all-weather strategies[16][17]. 3. Model Name: Macro Factor-Based Asset Allocation Model - Model Construction Idea: This model bridges macroeconomic research and asset allocation by constructing a framework based on six macroeconomic risks: growth, inflation, interest rates, credit, exchange rates, and liquidity[22][23]. - Model Construction Process: 1. Use the Factor Mimicking Portfolio method to construct high-frequency macroeconomic factors for the six risks. 2. Calculate factor exposures for selected assets at the end of each month. 3. Use a risk parity portfolio as the benchmark and adjust factor exposures based on subjective macroeconomic views. 4. Solve the optimization problem to determine asset weights for the next month[23][24]. - Model Evaluation: The model effectively translates macroeconomic views into actionable asset allocation strategies, providing a systematic approach to incorporating macroeconomic insights into portfolio construction[22][23]. --- Model Backtesting Results 1. Black-Litterman (BL) Models - Domestic BL Model 1: Weekly return 0.01%, June return 0.55%, YTD return 1.52%, annualized volatility 2.36%, max drawdown 1.31%[14][16] - Domestic BL Model 2: Weekly return 0.02%, June return 0.47%, YTD return 1.53%, annualized volatility 2.10%, max drawdown 1.06%[14][16] - Global BL Model 1: Weekly return 0.15%, June return 0.49%, YTD return 0.04%, annualized volatility 2.09%, max drawdown 1.64%[14][16] - Global BL Model 2: Weekly return 0.12%, June return 0.45%, YTD return 0.72%, annualized volatility 1.83%, max drawdown 1.28%[14][16] 2. Risk Parity Models - Domestic Risk Parity Model: Weekly return 0.10%, June return 0.50%, YTD return 2.24%, annualized volatility 1.62%, max drawdown 0.76%[21][22] - Global Risk Parity Model: Weekly return 0.10%, June return 0.43%, YTD return 1.86%, annualized volatility 1.86%, max drawdown 1.20%[21][22] 3. Macro Factor-Based Asset Allocation Model - Macro Factor Model: Weekly return 0.09%, June return 0.62%, YTD return 2.20%, annualized volatility 1.49%, max drawdown 0.64%[28][29]