海外量化季度观察：量化如何应对宏观不确定性冲击？

• AQR has started to embrace AI in its investment decisions, using more AI algorithms to potentially provide higher returns despite occasional difficulties in explaining drawdowns[11] - BlackRock's quantitative system aims to identify more granular risk factors and maintain neutrality to most risks, while seeking short-term reversal opportunities in dense market trading to outperform the market[1][15][16] - Bridgewater is focusing on the impact of "modern mercantilism" on asset prices, noting that U.S. assets still face significant uncertainty and highlighting the strong allocation value of gold[21][22] Quantitative Models and Construction Methods 1. Model Name: BlackRock's Safety Engineering System - Construction Idea: To handle high uncertainty by identifying more granular risk factors and maintaining neutrality to most risks - Construction Process: - The system has evolved to control risk exposure not only to conventional factors like market cap, momentum, and growth value but also to thousands of more granular risk factors such as Japan export factor and domestic demand stock factor - The system adjusts these factors based on macroeconomic changes and increases the frequency of monitoring event-related factors to hourly or minute levels - Evaluation: The system's performance during the pandemic demonstrated that broader data dimensions and more precise risk control are more important than complex models[15][16][17] 2. Model Name: AQR's High Volatility Factor Model - Construction Idea: To embrace high volatility factors for their long-term Sharpe ratio despite short-term drawdowns - Construction Process: - AQR uses the variance ratio to measure the volatility level of factors: $ \text{Variance Ratio} = \frac{\text{Annual Factor Return Variance}}{\text{Monthly Factor Return Variance} \times 12} $ - Factors with higher variance ratios are considered high volatility factors - AQR analyzed 13 major categories and 153 sub-factors for their variance ratios and Sharpe ratios - Evaluation: Long-term high volatility factors show a significant positive correlation with Sharpe ratios, suggesting that quantitative managers should embrace these factors and use diversification to reduce short-term volatility[23][24][25] Model Backtesting Results 1. BlackRock's Safety Engineering System - Information Ratio (IR): - Economic regime: 1.02 - Valuation: 0.77 - Sentiment: 0.43 - Growth timing: 1.06 - Aggregate signal: 1.83 - Max Drawdown: - Economic regime: -2.5% - Valuation: -3.4% - Sentiment: -4.2% - Growth timing: -2.7% - Aggregate signal: -1.9%[40] 2. AQR's High Volatility Factor Model - Variance Ratio: - Debt Issuance: 1.8 - Accruals: 1.6 - Profitability: 1.5 - Low Leverage: 1.4 - Investment: 1.4 - Profit Growth: 1.4 - Value: 1.4 - Core Stream Size: 1.2 - Quality: 1.2 - Seasonality: 1.1 - Low Risk: 1.0 - Momentum: 1.0 - Short-Term Reversal: 0.9 - Sharpe Ratio: - Debt Issuance: 0.7 - Accruals: 0.6 - Profitability: 0.3 - Low Leverage: 0.0 - Investment: 0.4 - Profit Growth: 0.4 - Value: 0.4 - Core Stream Size: 0.0 - Quality: 0.4 - Seasonality: 0.2 - Low Risk: 0.1 - Momentum: 0.3 - Short-Term Reversal: 0.1[24][25][27]