20230816-国盛证券-_薪火_量化分析系列研究（三）_红利低波的增强方案_16页_1mb

Summary of Dividend Low Volatility Index Enhancement

Introduction and Overview

The report analyzes the Dividend Low Volatility Index (H30269CSI), which demonstrates strong long-term performance with a 1326% annualized return since 2014, outperforming the CSI 300 Index by approximately 5% annually in terms of excess returns while exhibiting lower volatility and drawdowns. The index is effective in market downturns due to its dividend characteristics.

Key limitations identified in the current construction are: a lengthy adjustment frequency and instability in the low volatility factor selection. These issues motivate three enhancement schemes aimed at improving performance and stability.

Dividend Low Volatility Index Performance

• Long-term data (2014–present) shows the Dividend Low Volatility All-in-one index has a high annualized return of 1326%, low annualized volatility of 2108%, a low maximum drawdown of 4249%, and a positive information ratio.
• Compared to the CSI 300, it provides consistent excess returns and better risk-adjusted metrics, making it a defensive strategy in volatile markets.
• Industry distribution includes sectors like banking and publishing, contributing to its stability across "middle-tech" and AI-related trends.

Limitations of Current Index Construction

• Annual adjustment frequency leads to delays in capturing market changes, potentially allowing riskier stocks to enter the portfolio.
• The use of traditional volatility factor (standard deviation) results in suboptimal stock selection, with backtests showing moderate success in the Dividend Stock Pool.

Improvement Scheme 1: Monthly Dividend Low Volatility Combination

• Aim: Address the annual adjustment by switching to monthly frequency to capture real-time changes.
• Methods: Sample space unchanged, selection based on dividend yield and monthly high-frequency volatility factor for better stability.
• Performance: Achieves a 1668% annualized return, with metrics improving over time, though some years (2016–2018) show underperformance due to the traditional volatility factor's inconsistencies.
• Benefit: Shorter adjustment period allows faster response to market shifts, enhancing overall reliability.

Improvement Scheme 2: High-Frequency Volatility Factor Integration

• Enhance Scheme 1 by using a more stable high-frequency volatility factor, derived from minute return standard deviation over 20 trading days.
• Performance: Yield increases to 2092% annualized, with improved information ratio and reduced drawdown (e.g., 2905% max drawdown), outperforming both the traditional month-based combinations and the original index in multiple performance metrics.
• Conclusion: This scheme offers superior risk-adjusted returns and adaptability, driven by better volatility selection.

Improvement Scheme 3: Valuation Difference-Based Timer

• Utilize the valuation difference between dividend index components and the broader market to time entries/ exits.
• Method: EMA calculates BP spread based on trimmed mean values; timers detect when valuation is overextended and signal avoid/short alternatives.
• Performance: Timer signals achieve a 69.23% win rate, offering 2320% annualized gain with lower volatility and drawdowns compared to the index, enhancing defensive capabilities during drawdowns.
• Effectiveness: Complements other schemes by minimizing losses through strategic withdrawal recommendations based on indicators.

Overall Summary and Conclusion

• The report's enhancement schemes significantly boost performance by addressing the original index's weaknesses:

  • Monthly combinations improve returns.
  • High-frequency volatility factors provide stability and high gains.
  • Valuation-timing strategies offer superior risk mitigation with positive returns.

• However, backtests rely on historical data, and model failure is possible under market structural changes. No actual investment advice is given.

• Recommendation: Enhanced combinations show promise for stabilizing Smart Beta strategies in a volatile environment.

• Risk Acknowledgment: Potential for model inaccuracy if market conditions shift.