Quantitative Models and Construction Methods 1. Model Name: CRR Pricing Model for Convertible Bonds - **Model Construction Idea**: The CRR pricing model is based on the binomial tree framework, which incorporates embedded options, credit spreads, and other factors to calculate the theoretical value of convertible bonds. This model aims to identify pricing deviations in the convertible bond market and exploit the mean-reversion of these deviations[44][45]. - **Model Construction Process**: 1. Use the CRR binomial tree model to calculate the theoretical value of convertible bonds, considering embedded clauses and credit spreads. 2. Define the "pricing deviation" as the difference between the CRR theoretical price and the market price. 3. Select convertible bonds with the highest CRR price-to-market price ratio within each style (equity-like, balanced, and debt-like) for portfolio inclusion. 4. Portfolio construction: - Initial screening criteria: - Bond rating of AA- or above - Outstanding balance of at least 200 million yuan - Exclude stocks with ST or other special treatments - No historical rating downgrades or negative outlooks - Major shareholder pledge ratio below 90% - Active trading in the past 10 days - Redemption progress less than 5 days - Not below the bond floor - Style classification: - Equity-like: Parity > 110 - Balanced: Parity between 90 and 110 - Debt-like: Parity < 90 - Select the top 10 bonds with the highest CRR price-to-market price ratio in each style, totaling 30 bonds[45][51]. 5. Weighting and rebalancing: Equal weighting with monthly rebalancing[44][45]. - **Model Evaluation**: The CRR pricing model is more accurate than traditional methods like BSM due to its consideration of embedded clauses and credit spreads, making it suitable for identifying pricing inefficiencies in the convertible bond market[15][44]. 2. Model Name: Low Valuation Momentum Strategy for Convertible Bonds - **Model Construction Idea**: This strategy combines low valuation and momentum factors to identify convertible bonds with low conversion premium levels and positive momentum in their underlying stocks[48][49]. - **Model Construction Process**: 1. Initial screening criteria: Same as the CRR pricing model[51]. 2. Style classification: Same as the CRR pricing model[51]. 3. Selection criteria: - Evaluate the relative valuation of convertible bonds (absolute and relative pricing) - Assess the short-term momentum of the underlying stocks - Combine these factors into a comprehensive score and select the top 10 bonds in each style, totaling 30 bonds[51]. 4. Weighting and rebalancing: Equal weighting with monthly rebalancing[51]. - **Model Evaluation**: The low valuation momentum strategy effectively combines valuation and momentum factors, making it a robust multi-factor strategy for convertible bond selection[49][51]. --- Backtesting Results of Models 1. CRR Pricing Model - **Monthly Return**: -0.51%[48] - **Annualized Return**: 15.17%[48] - **Annualized Volatility**: 12.82%[48] - **Maximum Drawdown**: 12.08%[48] - **Return-to-Volatility Ratio**: 1.18[48] - **Return-to-Drawdown Ratio**: 1.26[48] - **Monthly Win Rate**: 61.36%[48] 2. Low Valuation Momentum Strategy - **Monthly Return**: -0.28%[55] - **Annualized Return**: 15.16%[55] - **Annualized Volatility**: 12.84%[55] - **Maximum Drawdown**: 11.26%[55] - **Return-to-Volatility Ratio**: 1.18[55] - **Return-to-Drawdown Ratio**: 1.35[55] - **Monthly Win Rate**: 64.77%[55]

证券研究报告 | 金融工程 gszqdatemark 2025 03 15 年 月 日 量化分析报告 择时雷达六面图：估值面略有弱化 择时雷达六面图：基于多维视角的择时框架。权益市场的表现受到多维度 指标因素的共同影响，我们尝试从流动性、经济面、估值面、资金面、技 术面、拥挤度选取二十一个指标对市场进行刻画，并将其概括为"估值性 价比"、"宏观基本面"、"资金&趋势"、"拥挤度&反转"四大类，从而生成 [-1,1]之间的综合择时分数。 本周综合打分。本周市场的估值性价比有所下降，宏观基本面无显著变化， 市场的资金&趋势仍然维持较高水平，而拥挤度&反转指标仍维持看空信 号，综合打分位于[-1,1]之间，当前的综合打分为-0.21 分，整体为中性偏 空观点。当前六面图各个维度的观点如下： 流动性。本周货币方向、货币强度、信用方向、信用强度均发出看空信 号，当前流动性得分为-1.00 分，综合来看发出显著看空信号。 估值面。由于市场回升，本周席勒 ERP、PB 与 AIAE 指标的 zscore 均 下降，当前市场的估值面得分为-0.17 分，信号较为中性。 资金面。本周内资的两融增量与成交额继续维持上行趋势，均发 ...