“学海拾珠”系列之二百四十七：分散化投资是否驱动大盘股需求？

Quantitative Models and Construction - Model Name: Active and Passive Rebalancing Metrics Construction Idea: Decompose quarterly portfolio weight changes into active discretionary decisions and passive return-driven changes to analyze fund manager behavior [38][40][42] Construction Process: - Formula: $W_{i,j,t}-W_{i,j,t-1}=\underbrace{W_{i,j,t}-\widehat{W}{i,j,t}}{\text{Active}{i,j,t}}+\underbrace{\widehat{W}{i,j,t}-W_{i,j,t-1}}{\text{Passive}{i,j,t}}$ $\widehat{w}{i,j,t}=\frac{\left(1+r{i,t}\right),w_{i,j,t-1}}{\sum\left(1+r_{i,t}\right),w_{i,j,t-1}}$ - Active: Residual weight changes after removing mechanical effects, capturing discretionary rebalancing [40][42] - Passive: Weight changes driven by market returns assuming no trading activity [40][41] Evaluation: Captures fund managers' preferences for managing portfolio concentration and distinguishes between minor adjustments and large-scale asset rotation [42][43] - Model Name: Threshold Demand Construction Idea: Focus on concentrated positions exceeding 2% of fund AUM to measure diversification-driven demand [82][83] Construction Process: Formula: $Threshold Demand_{i,t}=\frac{\sum_{j}\left(\widehat{w}{i,j,t}-w{i,j,t-1}\right)\cdot I(w_{i,j,t-1}>2%)\cdot\text{Shares}{i,j,t-1}}{\sum{j}\text{Shares}{i,j,t-1}}$ - Uses only concentrated positions (10% of fund holdings) where portfolio size and concentration matter [82][83] Evaluation: Effectively isolates positions where diversification constraints are most impactful [83] - Model Name: Fitted Demand Construction Idea: Use spline coefficients from weight ranges to construct demand metrics based on rebalancing intensity [83][84] Construction Process: Formula: $Fitted Demand{i,t}=\frac{\sum_{j}\left(\widehat{W}{ij,t}-W{ij,t-1}\right)\cdot\beta_{weight}\cdot\text{Shares}{i,j,t-1}}{\sum{j}\text{Shares}{i,j,t-1}}$ - $\beta{weight}$ represents rebalancing intensity coefficients for different weight ranges [83][84] Evaluation: Focuses on positions within 2%-6.5% of fund AUM, capturing nuanced rebalancing behavior [83][84] Model Backtesting Results - Active and Passive Metrics: - Contemporaneous Active adjustment for 1% Passive weight change: -0.234% [44][49] - Next-quarter Active adjustment for 1% Passive weight change: -0.171% [44][49] - Threshold Demand: - Standard deviation: 0.15% - Predicts equity fund sell probability increase by 1.28%-2.20% [85][86] - Fitted Demand: - Standard deviation: 0.03% - Predicts equity fund sell probability increase by 0.5%-0.67% [85][86] Quantitative Factors and Construction - Factor Name: Rebalancing Demand Construction Idea: Aggregate passive-driven portfolio changes to measure demand for large-cap stocks [81][82] Construction Process: Formula: $Rebalancing Demand_{i,t}=\frac{\sum_{j}\left(\widehat{w}{i,j,t}-w{i,t-1}\right)\cdot\text{Shares}{i,j,t-1}}{\sum{j}\text{Shares}_{i,j,t-1}}$ - Aggregates passive-driven changes across all observed mutual funds [81][82] Evaluation: Predicts short-term price pressure and subsequent reversals for large-cap stocks [82][88] Factor Backtesting Results - Rebalancing Demand: - Predicts short-term returns: -0.44% (t=-3.21) for first 35 trading days [88][89] - Predicts subsequent reversals: +0.27% (t=2.60) for remaining quarter [88][89] - Threshold Demand: - Predicts short-term returns: -0.348% (t=-3.719) for first 35 trading days [88][89] - Predicts subsequent reversals: +0.178% (t=2.508) for remaining quarter [88][89] - Fitted Demand: - Predicts short-term returns: -0.460% (t=-3.598) for first 35 trading days [88][89] - Predicts subsequent reversals: +0.253% (t=2.616) for remaining quarter [88][89] Additional Observations - Impact on Momentum Portfolios: - Adjusting for rebalancing demand improves momentum portfolio returns by 230% for large-cap stocks [114] - Suggests diversification-driven demand weakens traditional momentum strategies [114] - Price Pressure and Reversals: - Large-cap stocks experience V-shaped return patterns due to rebalancing demand [93][94] - Short-term price pressure followed by reversals aligns with non-fundamental demand effects [93][94]