Orient Securities(SH:600958) Orient Securities·2025-07-06 14:44Quantitative Models and Factor Construction Factor Names and Construction Details - Factor Name: Trend - Construction Idea: Captures the market's preference for trend-following strategies, using exponential weighted moving averages (EWMA) with different half-lives to measure price trends[11][16] - Construction Process: - Trend_120: $ \text{EWMA(halflife=20)}/\text{EWMA(halflife=120)} $ - Trend_240: $ \text{EWMA(halflife=20)}/\text{EWMA(halflife=240)} $[16] - Evaluation: Demonstrates strong performance in short-term market environments, reflecting increased preference for trend-following strategies[11] - Factor Name: Certainty - Construction Idea: Measures market confidence in stable and predictable investments, using metrics like institutional holdings and analyst coverage[16] - Construction Process: - Instholder Pct: Proportion of institutional holdings - Cov: Analyst coverage adjusted for market capitalization - Listdays: Number of days since listing[16] - Evaluation: Improved performance indicates restored market confidence in certainty-driven strategies[11] - Factor Name: Value - Construction Idea: Focuses on valuation metrics such as book-to-price (BP) and earnings yield (EP)[16] - Construction Process: - BP: $ \text{Net Assets}/\text{Market Value} $ - EP: $ \text{Earnings}/\text{Market Value} $[16] - Evaluation: Shows recovery in market preference for value-oriented investments[11] - Factor Name: Liquidity - Construction Idea: Assesses the impact of liquidity on asset pricing using turnover rates and liquidity betas[16] - Construction Process: - TO: Average logarithmic turnover over 243 days - Liquidity Beta: Regression of individual stock turnover against market turnover[16] - Evaluation: Underperformed significantly, reflecting reduced demand for high-liquidity assets[12] - Factor Name: Volatility - Construction Idea: Measures the impact of price volatility on asset returns using historical and idiosyncratic volatility metrics[16] - Construction Process: - Stdvol: Standard deviation of returns over 243 days - Ivff: Idiosyncratic volatility from Fama-French 3-factor model over 243 days[16] - Evaluation: Weak performance indicates declining interest in high-volatility assets[12] - Factor Name: Momentum - Construction Idea: Captures the continuation of price trends over different time horizons[16] - Construction Process: - UMR_1Y: Risk-adjusted momentum over 12 months - UMR_6M: Risk-adjusted momentum over 6 months[16] - Evaluation: Mixed results, with long-term momentum factors underperforming[12] Factor Backtesting Results - Trend Factor - Weekly return: 2.26%[11] - Monthly return: 2.98%[13] - YTD return: -3.81%[13] - 1-year return: 24.24%[13] - Historical annualized return: 14.10%[13] - Certainty Factor - Weekly return: 1.36%[11] - Monthly return: -2.87%[13] - YTD return: -11.74%[13] - 1-year return: -20.09%[13] - Historical annualized return: 2.63%[13] - Value Factor - Weekly return: 0.78%[11] - Monthly return: -2.14%[13] - YTD return: -10.78%[13] - 1-year return: -27.42%[13] - Historical annualized return: 7.14%[13] - Liquidity Factor - Weekly return: -3.85%[12] - Monthly return: 0.07%[13] - YTD return: 15.79%[13] - 1-year return: 29.31%[13] - Historical annualized return: -3.52%[13] - Volatility Factor - Weekly return: -2.83%[12] - Monthly return: -1.05%[13] - YTD return: 5.31%[13] - 1-year return: 28.00%[13] - Historical annualized return: -13.15%[13] - Momentum Factor - Weekly return (1-year UMR): 0.15%[24] - Monthly return (1-year UMR): 0.21%[24] - YTD return (1-year UMR): 1.69%[24] - 1-year return (1-year UMR): 1.22%[24] - Historical annualized return (1-year UMR): 3.87%[24] MFE Portfolio Construction - Construction Process: - Objective: Maximize single-factor exposure while controlling for industry, style, and stock-specific constraints - Optimization Model: $ \begin{array}{ll} max & f^{T}w \ s.t. & s_{l} \leq X(w-w_{b}) \leq s_{h} \ & h_{l} \leq H(w-w_{b}) \leq h_{h} \ & w_{l} \leq w-w_{b} \leq w_{h} \ & b_{l} \leq B_{b}w \leq b_{h} \ & 0 \leq w \leq l \ & 1^{T}w=1 \ & \Sigma|w-w_{0}| \leq to_{h} \end{array} $[61][62] - Constraints: - Style and industry exposure limits - Stock weight deviation limits - Turnover rate limits[64][65] - Backtesting: Monthly rebalancing, transaction cost of 0.3% applied, and performance evaluated against benchmarks[66]