股指分红点位监控周报：大盘风格持续领航，IH合约升水，IM合约深贴水-20250527

Quantitative Models and Construction Methods 1. Model Name: Dividend Points Estimation Model - Model Construction Idea: This model aims to estimate the dividend points of stock indices, which is critical for accurately calculating the basis and premium/discount levels of stock index futures. It incorporates the impact of constituent stock dividends on index points[12][36][42] - Model Construction Process: 1. Formula: $\text{Dividend Points} = \sum_{n=1}^{N} \left( \frac{\text{Dividend Amount of Stock}_n}{\text{Total Market Value of Stock}_n} \times \text{Weight of Stock}_n \times \text{Index Closing Price} \right)$ - $N$: Number of constituent stocks - Dividend amount and weight are adjusted based on the stock's ex-dividend date, which must fall between the current date ($t$) and the futures contract expiration date ($T$)[36] 2. Key Steps: - Dividend Amount Estimation: If not disclosed, it is calculated as: $\text{Dividend Amount} = \text{Net Profit} \times \text{Dividend Payout Ratio}$ - Net profit is predicted using historical profit distribution patterns, distinguishing between stable and unstable profit distributions[45][48] - Dividend payout ratio is estimated using historical averages, with adjustments for non-dividend-paying companies or extreme values[49][51] - Ex-Dividend Date Prediction: If not disclosed, it is predicted using historical intervals between announcement and ex-dividend dates, with adjustments for outliers or default dates[54] - Constituent Stock Weight Adjustment: Daily weights are calculated as: $W_{n,t} = \frac{w_{n0} \times (1 + r_n)}{\sum_{i=1}^{N} w_{i0} \times (1 + r_i)}$ - $w_{n0}$: Weight at the last disclosed date - $r_n$: Non-adjusted return of stock $n$ since the last disclosed date[43] - Model Evaluation: The model demonstrates high accuracy for indices like the SSE 50 and CSI 300, with prediction errors around 5 points. For the CSI 500, the error is slightly larger but remains stable at approximately 10 points[59] --- Model Backtesting Results 1. Dividend Points Estimation Model - SSE 50 Index: Prediction error ~5 points[59] - CSI 300 Index: Prediction error ~5 points[59] - CSI 500 Index: Prediction error ~10 points[59]