GF SECURITIES(SZ:000776) GF SECURITIES·2026-03-01 06:26Quantitative Models and Construction Methods 1. Model Name: Macro Indicator Trend Model - Model Construction Idea: This model establishes the relationship between macroeconomic indicators and the performance of major asset classes. It evaluates whether the trend of macro indicators (upward or downward) significantly impacts the monthly returns of asset classes[17][18] - Model Construction Process: 1. Use the monthly moving average of macro indicators to determine their trends (upward or downward) 2. Apply a T-test to assess whether the distribution of monthly returns for an asset class differs significantly under upward and downward trends of a macro indicator 3. The T-test formula is: - $\overline{R_{1}}$ and $\overline{R_{2}}$ represent the average monthly returns of an asset class under upward and downward trends, respectively - $S_{1}$ and $S_{2}$ are the standard deviations of monthly returns under upward and downward trends, respectively - $n_{1}$ and $n_{2}$ are the number of months under upward and downward trends, respectively[17] 4. Select macro indicators that significantly influence asset performance and assign monthly quantitative scores to each asset class based on these indicators[18] 2. Model Name: Technical Indicator Model - Model Construction Idea: This model evaluates asset trends, valuation, and fund flows to determine the technical outlook for major asset classes[22][23] - Model Construction Process: 1. Trend: - Use closing prices or LLT indicators to construct trend indicators for each asset class - Assign +1 if the trend indicator is positive (upward trend) and -1 if negative (downward trend)[22] 2. Valuation: - Calculate the equity risk premium (ERP) as the inverse of the PE(TTM) of the CSI 800 Index minus the 10-year government bond yield - Define the 5-year historical percentile of ERP as: - Assign scores based on ERP percentiles: - >90%: +2 - 70%-90%: +1 - 30%-70%: 0 - 10%-30%: -1 - <10%: -2[23][25] 3. Fund Flows: - Calculate the monthly net inflow of funds for individual stocks and aggregate them to obtain the index-level net inflow - Assess the marginal change in monthly net inflows to measure overall fund flow conditions - Assign +1 for positive fund flows (inflows) and -1 for negative fund flows (outflows)[26] 3. Model Name: Fixed Proportion + Macro Indicators + Technical Indicators Combination Model - Model Construction Idea: This model combines fixed proportion weights with macro and technical indicators to adjust asset allocation dynamically[35][36] - Model Construction Process: 1. Set fixed proportion weights for five asset classes: equity, bonds, commodities, industrial products, and cash 2. Adjust weights based on the latest monthly signals from macro and technical indicators 3. Increase or decrease the allocation to non-cash assets accordingly, while adjusting the cash allocation proportionally[36] 4. Model Name: Classic Allocation Model + Macro Indicators + Technical Indicators Combination Model - Model Construction Idea: This model incorporates macro and technical indicators into classic allocation strategies, such as risk parity or volatility control, to optimize asset allocation[46] - Model Construction Process: 1. Use risk parity or volatility control (e.g., annualized volatility ≤6%) as the baseline allocation strategy 2. Adjust weights dynamically based on the latest monthly signals from macro and technical indicators 3. Reallocate weights among non-cash assets and adjust cash allocation proportionally[46] --- Model Backtesting Results 1. Macro Indicator Trend Model - No specific backtesting results provided for this model 2. Technical Indicator Model - No specific backtesting results provided for this model 3. Fixed Proportion + Macro Indicators + Technical Indicators Combination Model - Annualized Return: 10.18% - Maximum Drawdown: 9.27% - Annualized Volatility: 6.17%[40] 4. Classic Allocation Model + Macro Indicators + Technical Indicators Combination Model - Volatility Control (6%) + Macro + Technical Indicators: - Annualized Return: 10.44% - Maximum Drawdown: 7.37% - Annualized Volatility: 5.57% - Risk Parity + Macro + Technical Indicators: - Annualized Return: 8.28% - Maximum Drawdown: 4.58% - Annualized Volatility: 3.40%[50] --- Quantitative Factors and Construction Methods 1. Factor Name: Equity Style Rotation Factors (Large/Small Cap, Growth/Value) - Factor Construction Idea: These factors assess the relative performance of equity styles (e.g., large vs. small cap, growth vs. value) based on macro and technical indicators[52][54] - Factor Construction Process: 1. Macro Indicators: - Evaluate the impact of macro indicators (e.g., M2 growth, US 10-year bond yield) on equity styles - Assign scores based on the direction and significance of these indicators[54] 2. Technical Indicators: - Use relative performance metrics (e.g., 1-month or 6-month return differences) to assess trends - Evaluate fund flows (e.g., net inflow differences) to measure capital allocation between styles - Assign scores based on the direction of trends and fund flows[55] --- Factor Backtesting Results 1. Equity Style Rotation Factors - Large/Small Cap Rotation: - Annualized Return: 14.42% - Maximum Drawdown: 49.10% - Annualized Volatility: 22.30%[59] - Growth/Value Rotation: - Annualized Return: 14.47% - Maximum Drawdown: 45.18% - Annualized Volatility: 21.56%[66]