Quantitative Models and Construction Methods Model 1: Convertible Bond Valuation Model - **Model Name**: Convertible Bond Valuation Model - **Model Construction Idea**: The model aims to compare the valuation of convertible bonds with their underlying stocks using a time-series comparable valuation metric called "100 Yuan Conversion Premium Rate" and the median of "Adjusted YTM - Credit Bond YTM" to measure the relative allocation value between debt-biased convertible bonds and credit bonds[4][5][15] - **Model Construction Process**: - **100 Yuan Conversion Premium Rate**: Fit the relationship curve between the conversion premium rate and conversion value in the cross-sectional space at each time point, and substitute the conversion value = 100 into the fitting formula to obtain the "100 Yuan Conversion Premium Rate" - Formula: $$ y_{i}=\alpha_{0}+\,\alpha_{1}\cdot\,{\frac{1}{x_{i}}}+\epsilon_{i} $$ where \( y_{i} \) is the conversion premium rate of the i-th convertible bond, and \( x_{i} \) is the conversion value of the i-th convertible bond[43] - **Adjusted YTM - Credit Bond YTM**: Adjust the YTM of debt-biased convertible bonds by stripping out the impact of conversion terms - Formula: $$ \text{Adjusted YTM} = \text{Convertible Bond YTM} \times (1 - \text{Conversion Probability}) + \text{Expected Conversion Annualized Yield} \times \text{Conversion Probability} $$ The conversion probability is calculated using the BS model, substituting the closing price of the underlying stock, option exercise price, stock volatility, remaining term, and discount rate to calculate the conversion probability \( N(d2) \)[44] - **Model Evaluation**: The model provides a systematic approach to evaluate the relative allocation value of convertible bonds compared to their underlying stocks and credit bonds[15] Model 2: Convertible Bond Comprehensive Valuation Factor - **Model Name**: Convertible Bond Comprehensive Valuation Factor - **Model Construction Idea**: The model combines the deviation of the conversion premium rate and the theoretical value deviation (Monte Carlo model) to construct a comprehensive valuation factor for convertible bonds[6][19] - **Model Construction Process**: - **Conversion Premium Rate Deviation**: - Formula: $$ \text{Conversion Premium Rate Deviation} = \text{Conversion Premium Rate} - \text{Fitted Conversion Premium Rate} $$ - **Theoretical Value Deviation (Monte Carlo Model)**: - Formula: $$ \text{Theoretical Value Deviation} = \frac{\text{Convertible Bond Closing Price}}{\text{Theoretical Value}} - 1 $$ The Monte Carlo model fully considers the conversion, redemption, downward revision, and repurchase terms of convertible bonds, simulating 10,000 paths at each time point and using the same credit term interest rate as the discount rate to calculate the theoretical value of the convertible bond[20] - **Comprehensive Valuation Factor**: - Formula: $$ \text{Convertible Bond Comprehensive Valuation Factor} = \text{Rank}(\text{Conversion Premium Rate Deviation}) + \text{Rank}(\text{Theoretical Value Deviation (Monte Carlo Model)}) $$ - **Model Evaluation**: The comprehensive valuation factor performs well in the overall, balanced, and debt-biased convertible bonds, while the theoretical value deviation (Monte Carlo model) performs better in equity-biased convertible bonds[19][20] Model 3: Convertible Bond Style Rotation Model - **Model Name**: Convertible Bond Style Rotation Model - **Model Construction Idea**: The model uses convertible bond momentum and volatility deviation as market sentiment capture indicators to construct a convertible bond style rotation portfolio, with bi-weekly rebalancing[7][26] - **Model Construction Process**: - **Market Sentiment Capture Indicators**: - Formula: $$ \text{Convertible Bond Style Market Sentiment Capture Indicator} = \text{Rank}(\text{Convertible Bond 20-Day Momentum}) + \text{Rank}(\text{Volatility Deviation}) $$ - **Style Rotation Position Calculation**: - Example Calculation: | | Convertible Bond Equity-Biased Low Valuation | Convertible Bond Balanced Low Valuation | Convertible Bond Debt-Biased Low Valuation | | --- | --- | --- | --- | | Equal Weight Index | 1 | 2 | 3 | | Volatility Deviation Ranking | 2 | 1 | 3 | | Market Sentiment Capture Indicator | 3 | 3 | 6 | | Style Rotation Position | 50% | 50% | 0% | - **Model Evaluation**: The style rotation model effectively captures market sentiment and allocates positions accordingly, showing superior performance compared to the equal-weight index[26][27][28] Model Backtesting Results Convertible Bond Valuation Model - **100 Yuan Conversion Premium Rate**: Rolling three-year percentile at 98.70%, rolling five-year percentile at 94.90%[4][15] - **Adjusted YTM - Credit Bond YTM**: Current median at -2.36%[5][15] Convertible Bond Comprehensive Valuation Factor - **Equity-Biased Convertible Bond Low Valuation Index**: - Annualized Return: 26.10% - Annualized Volatility: 20.55% - Maximum Drawdown: -22.94% - IR: 1.27 - Calmar Ratio: 1.14 - Monthly Win Rate: 62.22%[23] - **Balanced Convertible Bond Low Valuation Index**: - Annualized Return: 14.80% - Annualized Volatility: 11.82% - Maximum Drawdown: -15.95% - IR: 1.25 - Calmar Ratio: 0.93 - Monthly Win Rate: 62.22%[23] - **Debt-Biased Convertible Bond Low Valuation Index**: - Annualized Return: 13.37% - Annualized Volatility: 9.43% - Maximum Drawdown: -17.78% - IR: 1.42 - Calmar Ratio: 0.75 - Monthly Win Rate: 57.78%[23] Convertible Bond Style Rotation Model - **Convertible Bond Style Rotation**: - Annualized Return: 25.27% - Annualized Volatility: 16.68% - Maximum Drawdown: -15.89% - IR: 1.51 - Calmar Ratio: 1.59 - Monthly Win Rate: 65.56%[32] - **Convertible Bond Low Valuation Equal Weight Index**: - Annualized Return: 14.71% - Annualized Volatility: 10.97% - Maximum Drawdown: -15.48% - IR: 1.34 - Calmar Ratio: 0.95 - Monthly Win Rate: 61.11%[32] - **Convertible Bond Equal Weight Index**: - Annualized Return: 9.75% - Annualized Volatility: 11.66% - Maximum Drawdown: -20.60% - IR: 0.84 - Calmar Ratio: 0.47 - Monthly Win Rate: 60.00%[32]

Quantitative Models and Construction Methods - **Model Name**: "百元转股溢价率" (Premium Rate per 100 Yuan Conversion) **Model Construction Idea**: Compare convertible bond valuation with equity valuation using historical percentile metrics to assess relative allocation value [4][15] **Model Construction Process**: Fit a cross-sectional curve of conversion premium rate and conversion value at each time point. Substitute conversion value = 100 into the fitted formula to derive "百元转股溢价率". Formula: $$ y_{i}=\alpha_{0}+\,\alpha_{1}\cdot\,{\frac{1}{x_{i}}}+\epsilon_{i} $$ Here, \( y_{i} \) represents the conversion premium rate of the \( i \)-th bond, and \( x_{i} \) represents the conversion value of the \( i \)-th bond [44] **Model Evaluation**: Provides a relative valuation perspective for convertible bonds versus equities [15] - **Model Name**: "修正 YTM – 信用债 YTM" (Adjusted YTM Minus Credit Bond YTM) **Model Construction Idea**: Adjust convertible bond yield-to-maturity (YTM) by removing the impact of conversion clauses to compare with credit bond YTM [4][15] **Model Construction Process**: $$ \text{Adjusted YTM} = \text{Convertible Bond YTM} \times (1 - \text{Conversion Probability}) + \text{Expected Conversion Annualized Return} \times \text{Conversion Probability} $$ Conversion probability is calculated using the Black-Scholes model, incorporating stock price, strike price, stock volatility, remaining term, and discount rate. The median of the differences between adjusted YTM and credit bond YTM is then computed: $$ \text{"修正 YTM – 信用债 YTM" Median} = \text{median}\{X_1, X_2, ..., X_n\} $$ Here, \( X_i \) represents the difference between adjusted YTM and credit bond YTM for the \( i \)-th bond [45][46] **Model Evaluation**: Suitable for assessing relative allocation value between debt-heavy convertible bonds and credit bonds [15] Quantitative Factors and Construction Methods - **Factor Name**: 转股溢价率偏离度 (Conversion Premium Rate Deviation) **Factor Construction Idea**: Measure deviation of conversion premium rate from fitted values to assess valuation differences [21] **Factor Construction Process**: $$ \text{Conversion Premium Rate Deviation} = \text{Conversion Premium Rate} - \text{Fitted Conversion Premium Rate} $$ Fitted values are determined by the cross-sectional curve fitting process [21] **Factor Evaluation**: Effective in comparing valuation across different convertible bonds [21] - **Factor Name**: 理论价值偏离度 (Theoretical Value Deviation) **Factor Construction Idea**: Assess price expectation differences using Monte Carlo simulation [21] **Factor Construction Process**: $$ \text{Theoretical Value Deviation} = \frac{\text{Convertible Bond Closing Price}}{\text{Theoretical Value}} - 1 $$ Monte Carlo simulation considers conversion, redemption, downward revision, and repurchase clauses, simulating 10,000 paths at each time point using the same credit term limit rate as the discount rate [21] **Factor Evaluation**: Provides a comprehensive valuation perspective by incorporating multiple convertible bond clauses [21] - **Composite Factor Name**: 转债综合估值因子 (Convertible Bond Comprehensive Valuation Factor) **Factor Construction Idea**: Combine conversion premium rate deviation and theoretical value deviation for enhanced valuation analysis [21] **Factor Construction Process**: $$ \text{Convertible Bond Comprehensive Valuation Factor} = \text{Rank(Conversion Premium Rate Deviation)} + \text{Rank(Theoretical Value Deviation)} $$ **Factor Evaluation**: Demonstrates superior performance across various convertible bond categories [21] - **Factor Name**: 转债市场情绪捕捉指标 (Convertible Bond Market Sentiment Capture Indicator) **Factor Construction Idea**: Use momentum and volatility deviation to identify market sentiment [29] **Factor Construction Process**: $$ \text{Market Sentiment Capture Indicator} = \text{Rank(20-day Momentum)} + \text{Rank(Volatility Deviation)} $$ **Factor Evaluation**: Effective in guiding convertible bond style rotation strategies [29] Model Backtesting Results - **"百元转股溢价率" Model**: Rolling three-year percentile at 47.4%, rolling five-year percentile at 50.9% [4][15][18] - **"修正 YTM – 信用债 YTM" Model**: Current median value at -0.03% [4][15][18] Factor Backtesting Results - **转股溢价率偏离度 Factor**: Enhanced excess returns in the past four weeks for偏股,平衡,偏债 convertible bonds at 1.33%, 0.27%, and 0.04%, respectively [5][23] - **理论价值偏离度 Factor**: Demonstrates superior performance in偏股 convertible bonds [20][21] - **转债综合估值因子 Factor**: - 偏股转债低估指数: IR = 1.22, annualized return = 24.91%, annualized volatility = 20.39%, max drawdown = -22.83%, Calmar ratio = 1.09, monthly win rate = 63.64% [24] - 平衡转债低估指数: IR = 1.16, annualized return = 13.77%, annualized volatility = 11.87%, max drawdown = -16.04%, Calmar ratio = 0.86, monthly win rate = 60.23% [24] - 偏债转债低估指数: IR = 1.29, annualized return = 12.21%, annualized volatility = 9.45%, max drawdown = -17.59%, Calmar ratio = 0.69, monthly win rate = 56.82% [24] Style Rotation Backtesting Results - **转债风格轮动 Model**: - IR = 1.47, annualized return = 24.23%, annualized volatility = 16.54%, max drawdown = -15.54%, Calmar ratio = 1.56, monthly win rate = 65.91% [35] - Recent four-week return = 2.24%, year-to-date return = 26.75% [31][32]

Quantitative Models and Construction Methods Model Name: Convertible Bond Valuation Model - **Construction Idea**: The model aims to compare the valuation of convertible bonds with their underlying stocks and other credit bonds to determine relative investment value[4][13] - **Construction Process**: - **Convertible Bond and Stock Valuation**: Construct the "Hundred Yuan Conversion Premium Rate" to compare the valuation of convertible bonds and stocks over time. Calculate the rolling historical percentile to measure the relative configuration value[4][13] - Formula: $$ y_{i}=\alpha_{0}+\,\alpha_{1}\cdot\,{\frac{1}{x_{i}}}+\epsilon_{i} $$ where \( y_{i} \) is the conversion premium rate of the i-th bond, \( x_{i} \) is the conversion value of the i-th bond[46] - **Convertible Bond and Credit Bond Valuation**: Focus on the impact of conversion terms on the yield to maturity (YTM) of convertible bonds, and calculate the "Adjusted YTM - Credit Bond YTM" median to measure the relative configuration value between convertible bonds and credit bonds[4][13] - Formula: $$ \text{Adjusted YTM} = \text{Convertible Bond YTM} \times \text{Maturity Probability} + \text{Expected Conversion Yield} \times \text{Conversion Probability} $$ $$ = \text{Convertible Bond YTM} \times (1 - \text{Conversion Probability}) + \text{Expected Conversion Yield} \times \text{Conversion Probability} $$ $$ \text{Adjusted YTM - Credit Bond YTM Median} = \text{median}\{X_1, X_2, ..., X_n\} $$ where \( X_i \) represents the difference between the adjusted YTM of the i-th convertible bond and the YTM of a credit bond of the same grade and maturity[47][48] - **Evaluation**: The model provides a systematic approach to evaluate the relative investment value of convertible bonds compared to their underlying stocks and other credit bonds[4][13] Model Name: Convertible Bond Style Rotation Model - **Construction Idea**: The model captures market sentiment using momentum and volatility deviation indicators to construct a convertible bond style rotation portfolio[5][23] - **Construction Process**: - **Market Sentiment Indicators**: Use convertible bond 20-day momentum and volatility deviation as market sentiment capture indicators[5][23] - Formula: $$ \text{Convertible Bond Style Market Sentiment Capture Indicator} = \text{Rank}(\text{Convertible Bond 20-day Momentum}) + \text{Rank}(\text{Volatility Deviation}) $$ - **Portfolio Construction**: Rank the convertible bond style indices based on the sentiment indicators, and allocate the portfolio based on the rankings. If all three styles are selected, invest 100% in the balanced low valuation style[5][23][32] - **Evaluation**: The model effectively captures market sentiment and adjusts the portfolio allocation to optimize returns[5][23] Model Backtest Results Convertible Bond Valuation Model - **Hundred Yuan Conversion Premium Rate**: Rolling three-year percentile at 43.5%, rolling five-year percentile at 49.8%[4][13][16] - **Adjusted YTM - Credit Bond YTM Median**: Current median at 0.11%[4][13][16] Convertible Bond Style Rotation Model - **Recent 4-week Returns**: Convertible bond style rotation return at 8.58%, year-to-date return at 23.98%[5][33][35] - **Information Ratio**: Convertible bond style rotation IR at 1.46, convertible bond low valuation equal-weight index IR at 1.22, convertible bond equal-weight index IR at 0.71[38] Quantitative Factors and Construction Methods Factor Name: Convertible Bond Comprehensive Valuation Factor - **Construction Idea**: Combine the deviation of conversion premium rate and theoretical value deviation (Monte Carlo model) to construct a comprehensive valuation factor[5][24] - **Construction Process**: - **Conversion Premium Rate Deviation**: - Formula: $$ \text{Conversion Premium Rate Deviation} = \text{Conversion Premium Rate} - \text{Fitted Conversion Premium Rate} $$ - **Theoretical Value Deviation (Monte Carlo Model)**: - Formula: $$ \text{Theoretical Value Deviation} = \frac{\text{Convertible Bond Closing Price}}{\text{Theoretical Value}} - 1 $$ Monte Carlo model simulates 10,000 paths at each time point, using the same credit term limit rate as the discount rate to calculate the theoretical value of the convertible bond[25] - **Comprehensive Valuation Factor**: - Formula: $$ \text{Convertible Bond Comprehensive Valuation Factor} = \text{Rank}(\text{Conversion Premium Rate Deviation}) + \text{Rank}(\text{Theoretical Value Deviation (Monte Carlo Model)}) $$ - **Evaluation**: The factor provides a robust method to evaluate the valuation of convertible bonds comprehensively[5][24] Factor Backtest Results Convertible Bond Comprehensive Valuation Factor - **Recent 4-week Returns**: Low valuation factor enhanced excess returns in convertible bonds: 1.56% for equity-biased, 0.10% for balanced, 0.18% for debt-biased[27] - **Information Ratio**: Equity-biased convertible bond low valuation index IR at 1.22, balanced convertible bond low valuation index IR at 1.16, debt-biased convertible bond low valuation index IR at 1.29[28] Convertible Bond Low Valuation Index Components Equity-biased Convertible Bond Low Valuation Index Components - **Components**: Guangda Convertible Bond, Jindan Convertible Bond, Jingdang Convertible Bond, etc.[6][43] Balanced Convertible Bond Low Valuation Index Components - **Components**: Liqun Convertible Bond, Hebang Convertible Bond, Ying 19 Convertible Bond, etc.[6][44] Debt-biased Convertible Bond Low Valuation Index Components - **Components**: Dongnan Convertible Bond, Shunbo Convertible Bond, Huitong Convertible Bond, etc.[6][45]

Quantitative Models and Construction Methods 1. Model Name: "百元转股溢价率" (Premium Rate per 100 Yuan Conversion) - **Model Construction Idea**: This model aims to compare the valuation of convertible bonds and their underlying stocks by calculating a time-series comparable valuation indicator, "百元转股溢价率" (Premium Rate per 100 Yuan Conversion), and using rolling historical percentiles to measure the relative allocation value[4][15]. - **Model Construction Process**: - Fit the relationship curve between the conversion premium rate and conversion value in the cross-sectional space at each time point - Substitute a conversion value of 100 into the fitted formula to obtain the "百元转股溢价率" - Formula: $$y_{i}=\alpha_{0}+\,\alpha_{1}\cdot\,{\frac{1}{x_{i}}}+\epsilon_{i}$$ where \(y_{i}\) represents the conversion premium rate of the \(i\)-th bond, and \(x_{i}\) represents the conversion value of the \(i\)-th bond[44] - **Model Evaluation**: Provides a relative valuation perspective for comparing convertible bonds and their underlying stocks[15] 2. Model Name: "修正 YTM – 信用债 YTM" (Adjusted YTM - Credit Bond YTM) - **Model Construction Idea**: This model isolates the impact of conversion terms on the yield-to-maturity (YTM) of convertible bonds to assess the relative allocation value between debt-heavy convertible bonds and credit bonds[5][15]. - **Model Construction Process**: - Adjust the YTM of debt-heavy convertible bonds by considering the probability of conversion and maturity - Formula: $$\text{Adjusted YTM} = \text{Convertible Bond YTM} \times (1 - \text{Conversion Probability}) + \text{Expected Annualized Return of Conversion} \times \text{Conversion Probability}$$ - Conversion probability is calculated using the Black-Scholes (BS) model, incorporating stock price, strike price, stock volatility, remaining maturity, and discount rate - Calculate the median difference between the adjusted YTM of convertible bonds and the YTM of credit bonds of the same rating and maturity: $$\text{"Adjusted YTM - Credit Bond YTM Median"} = \text{median}\{X_1, X_2, ..., X_n\}$$ where \(X_i\) represents the difference for the \(i\)-th convertible bond[45][46] - **Model Evaluation**: Highlights the cost-effectiveness of debt-heavy convertible bonds compared to credit bonds[5] 3. Model Name: Convertible Bond Style Rotation Model - **Model Construction Idea**: This model captures market sentiment using momentum and volatility deviation indicators to construct a convertible bond style rotation portfolio, with bi-weekly rebalancing[6][27]. - **Model Construction Process**: - Calculate the following sentiment indicators for each convertible bond: - 20-day momentum - Volatility deviation - Rank the indicators in reverse order and sum the rankings to determine the market sentiment capture indicator for each style index: $$\text{Market Sentiment Capture Indicator} = \text{Rank (20-day Momentum)} + \text{Rank (Volatility Deviation)}$$ - Allocate portfolio weights based on the rankings, with a preference for the style index with the lowest indicator value. If rankings are equal, allocate weights equally. If all three styles are selected, allocate 100% to the balanced low-valuation style[28] - **Model Evaluation**: Demonstrates superior performance compared to the equal-weighted convertible bond index, with a focus on balanced low-valuation styles[27][33] --- Quantitative Factors and Construction Methods 1. Factor Name: 转股溢价率偏离度 (Conversion Premium Deviation) - **Factor Construction Idea**: Measures the deviation of the conversion premium rate from its fitted value, enabling comparability across different parities[19][20]. - **Factor Construction Process**: $$\text{Conversion Premium Deviation} = \text{Conversion Premium Rate} - \text{Fitted Conversion Premium Rate}$$ - The number of convertible bonds determines the fitting quality[20] - **Factor Evaluation**: Provides a robust measure for identifying valuation discrepancies in convertible bonds[20] 2. Factor Name: 理论价值偏离度 (Theoretical Value Deviation, Monte Carlo Model) - **Factor Construction Idea**: Quantifies the price expectation gap by considering various convertible bond terms (e.g., conversion, redemption, downward revision, put options) through Monte Carlo simulation[19][20]. - **Factor Construction Process**: $$\text{Theoretical Value Deviation} = \frac{\text{Convertible Bond Closing Price}}{\text{Theoretical Value}} - 1$$ - Simulate 10,000 paths at each time point using the Monte Carlo model, with the same credit term interest rate as the discount rate[20] - **Factor Evaluation**: Effectively captures valuation discrepancies, particularly for equity-heavy convertible bonds[19][20] 3. Factor Name: 转债综合估值因子 (Comprehensive Convertible Bond Valuation Factor) - **Factor Construction Idea**: Combines the rankings of the above two factors to enhance valuation analysis across all domains (equity-heavy, balanced, debt-heavy)[19][20]. - **Factor Construction Process**: $$\text{Comprehensive Convertible Bond Valuation Factor} = \text{Rank (Conversion Premium Deviation)} + \text{Rank (Theoretical Value Deviation)}$$ - **Factor Evaluation**: Demonstrates superior performance in identifying undervalued convertible bonds across different styles[19][20] --- Backtesting Results of Models 1. Convertible Bond Style Rotation Model - **Annualized Return**: 23.38% - **Annualized Volatility**: 16.48% - **Maximum Drawdown**: -15.54% - **IR**: 1.42 - **Calmar Ratio**: 1.50 - **Monthly Win Rate**: 65.12%[33] --- Backtesting Results of Factors 1. 转股溢价率偏离度 Factor - **Equity-Heavy Convertible Bonds**: Enhanced excess return of 0.9% over the past 4 weeks[22] - **Balanced Convertible Bonds**: Enhanced excess return of 1.2% over the past 4 weeks[22] - **Debt-Heavy Convertible Bonds**: Enhanced excess return of -0.3% over the past 4 weeks[22] 2. 理论价值偏离度 Factor - **Equity-Heavy Convertible Bonds**: Enhanced excess return of 0.9% over the past 4 weeks[22] - **Balanced Convertible Bonds**: Enhanced excess return of 1.2% over the past 4 weeks[22] - **Debt-Heavy Convertible Bonds**: Enhanced excess return of -0.3% over the past 4 weeks[22] 3. 转债综合估值因子 Factor - **Equity-Heavy Convertible Bonds**: Enhanced excess return of 0.9% over the past 4 weeks[22] - **Balanced Convertible Bonds**: Enhanced excess return of 1.2% over the past 4 weeks[22] - **Debt-Heavy Convertible Bonds**: Enhanced excess return of -0.3% over the past 4 weeks[22]