20240705-国联证券-金融工程专题_融合股指贴水的四因子择时策略_23页_2mb

报告摘要

Summary in English

Analysis of the Equity Asset Allocation Strategy

This report introduces a four-factor timing strategy designed to optimize equity asset allocation in the Chinese market. It builds upon a framework integrating macroscopic, mesoscopic, and microscopic dimensions.

Key Elements:

Three-Dimensional Signals:
- Constructs signals from macroeconomic sentiment (using Logit models to gauge future 60-day performance), mesoscopic leading indicators (evaluating industry profitability trends and predicting earnings cycles), and microscopic risk factors (tracking valuation, liquidity, and volatility extremes).
- The non-linear portfolio allocation models these dimensions through eight distinct market states, improving dynamic asset allocation decision-making.
Stock Index Futures Signal:
- Incorporates futures backwardation (price discrepancy between futures and spot indices) as an indicator to reflect market sentiment and risk aversion.
- A high-frequency correlation analysis between price movements in the index and futures helps identify timely bearish or bullish signals.
Signal Synthesis:
- The four-factor composite signal integrates the three-dimensional and index futures signals. This combination mitigates the limitations of the traditional model in predicting market turning points.

Backtesting Results:
The strategy was tested from 2017 to 2024, contributing substantial excess returns across major indices (e.g., 986% annualized excess for the CSI 300, 936% for CSI 500, and 870% for CSI All-Share). Backtesting results show consistent performance improvements compared to traditional approaches, particularly during market shifts.

Risk Insight:
The strategy uses a quantitative approach with inherent risks associated with timing-model dependencies on historical data. While past performance is indicative, future results are subject to market dynamics and model adjustments.

摘要（中文总结）

本报告提出了一种基于多维度因子合成的选股策略，旨在优化权益资产配置。该策略结合宏观、中观和微观三个维度，形成一个四因子择时模型。

核心要素：

三维度构建：
- 宏观层：通过Logit模型预测60日市场情绪，判断当前宏观环境利好或利空程度。
- 中观层：跟踪行业盈利趋势和领先指标，预测景气周期变化。
- 微观层：衡量估值水平、市场风险溢价及流动性，作为配置反转预警指标。
- 非线性模型：将市场划分为八个维度状态，动态调整资产配置比例。
股指期货贴水信号：
- 通过股指期货与现货价格基差关联性，每日高频指标反应市场情绪，识别交易者避险行为。
- 当期合约基差成为判断恐慌/乐观情绪的有效补充。
信号合成：
- 三维度择时信号与基差信号融合，形成四因子定时报时决策模型，大幅改进拐点预见时效。