机构境内资产配置指南：宏观胜率和微观赔率视角下的定价研究

Quantitative Models and Construction Methods 1. Model Name: Pring Cycle - Model Construction Idea: The Pring Cycle is an upgraded version of the Merrill Clock, incorporating financial data to enhance the predictive accuracy of asset allocation recommendations. It defines financial indicators as leading indicators, real economy indicators as coincident indicators, and price indicators as lagging indicators. These three groups of indicators form six economic states, each corresponding to specific asset allocation strategies [9][10][11] - Model Construction Process: 1. Leading indicators include M2 growth and new social financing, filtered for cyclical factors to observe trend components [14] 2. Coincident indicators include real estate investment and export growth, also filtered for cyclical factors [14][22] 3. Lagging indicators include CPI and PPI growth, filtered similarly [14][28] 4. The model identifies the current economic state based on the trends of these indicators. For example, the "Recovery" state is characterized by rising leading indicators, stable coincident indicators, and declining lagging indicators [10][11] - Model Evaluation: The model improves upon the Merrill Clock by incorporating financial data, but it cannot fully capture real economic states during extraordinary events [13] 2. A-Share Pricing Framework - Model Construction Idea: The pricing framework integrates short-term fundamentals, long-term confidence, and required return rates to determine the reasonable valuation range of A-shares. It emphasizes the PB-ROE relationship for valuation assessment [52][56] - Model Construction Process: 1. Decompose stock returns into components: dividend yield, net asset growth, and valuation changes [56] 2. Use a two-stage DDM model to calculate reasonable PB values: $PB_{current} = ROE_1 \times d_1 \times \sum_{t=1}^{T} \frac{(1+g_1)^{t-1}}{(1+R_f+R_p)^t}$ $PB_{stable} = \frac{ROE_2 \times d_2}{1+R_f+R_p-g_2} \times \frac{(1+g_1)^T}{(1+R_f+R_p)^T}$ where $ROE_1$ and $ROE_2$ are the return on equity for the current and stable growth phases, $d_1$ and $d_2$ are dividend payout ratios, $g_1$ and $g_2$ are growth rates, and $R_f+R_p$ is the required return rate [56][57] 3. Historical calibration suggests an 11-year duration for the first growth phase, with ROE assumptions adjusted for optimistic, neutral, and pessimistic scenarios [58] - Model Evaluation: The framework effectively captures valuation dynamics, but short-term ROE fluctuations introduce uncertainty [56][58] 3. Interest Rate Pricing Framework - Model Construction Idea: A three-factor model for long-term government bond yields, incorporating policy rates, inflation expectations, and growth expectations [59] - Model Construction Process: 1. Represent policy rates using one-year interbank CD rates, inflation expectations using CPI growth, and growth expectations using PMI levels [59] 2. Regression analysis reveals the relative importance of these factors: monetary policy > inflation expectations > growth expectations [62][63] 3. Additional analysis links bond yields to housing prices, reflecting cyclical economic drivers [67][69] - Model Evaluation: The model highlights the dominant role of monetary policy but acknowledges limitations in capturing short-term market dynamics [63][67] 4. Gold Pricing Framework - Model Construction Idea: Gold pricing is influenced by its commodity, financial, and monetary attributes, with monetary factors being the most consistent driver [74][75] - Model Construction Process: 1. Historical analysis identifies three gold bull markets driven by different factors: inflation and oil prices (1971-1980), financial crises and low real rates (2001-2011), and de-globalization and central bank purchases (2019-present) [74] 2. Introduce a valuation metric: - Pre-2022: Global gold reserves × gold price / US M2 - Post-2023: Global gold reserves × gold price / weighted M2 of reserve currency countries (USD, EUR, GBP, JPY, RMB) [80] - Model Evaluation: The framework effectively captures long-term trends but faces challenges in predicting short-term price movements [78][80] --- Model Backtesting Results 1. Pring Cycle - Current state: Recovery phase, favoring equity assets [35] 2. A-Share Pricing Framework - Reasonable PB range for CSI 800: 1.36-1.55 - Expected annual return: 5%-9% [58][59] 3. Interest Rate Pricing Framework - Predicted 10-year government bond yield: 1.36%-1.51% - Yield corridor: ±1.5 standard deviations around the central estimate [72][73] 4. Gold Pricing Framework - Current valuation percentile: 39% (10% below the median) - Long-term upward potential remains [80][82] --- Quantitative Factors and Construction Methods 1. Style Factors - Value Style: - ROE: 9.14% - PB range: 0.9-0.95 - Expected return: 6%-8% [96][100] - Growth Style: - ROE: 12.48% - PB range: 1.69-3.24 - Expected return: -7%-13% [103][108] - Small-Cap Style: - ROE: 5.99% - PB range: 0.65-1.82 - Expected return: Low [111][116] - Large-Cap Style: - ROE: 10.21% - PB range: 0.92-1 - Expected return: 0%-2% [119][123] - Quality Style: - ROE: 14.23% - PB range: 2.34-5.35 - Expected return: 8%-39% [128][131] - Dividend Style: - ROE: 9.09% - PB range: 0.83-0.87 - Expected return: 11%-12% [134][138] --- Factor Backtesting Results 1. Value Style - Historical PB-ROE alignment indicates moderate valuation [100][102] 2. Growth Style - High ROE volatility leads to wide valuation ranges [108][110] 3. Small-Cap Style - Valuation driven more by liquidity than fundamentals [113][117] 4. Large-Cap Style - Valuation closely tied to fundamentals, with limited upside [123][127] 5. Quality Style - Significant valuation recovery potential [131][133] 6. Dividend Style - Stable valuation with moderate upside [138][140]