股指分红点位监控周报：H及IF主力合约升水，IC及IM合约均深贴水-20250709

Quantitative Models and Construction Methods - Model Name: Index Dividend Points Estimation Model Model Construction Idea: This model estimates the dividend points of index constituents to account for the natural drop in index levels caused by dividend ex-dates, which is critical for accurately calculating the basis and premium/discount levels of stock index futures[12][38][44] Model Construction Process: 1. Identify the index constituents and their weights. If daily weights are unavailable, adjust monthly weights using the formula: $W_{n,t} = \frac{w_{n0} \times (1 + r_{n})}{\sum_{i=1}^{N} w_{i0} \times (1 + r_{i})}$ where $w_{n0}$ is the weight of stock $n$ on the last disclosed date, and $r_{n}$ is the non-adjusted return of stock $n$ from the last disclosed date to the current date[45][46] 2. Estimate the dividend amount for each constituent: - If disclosed, use the reported dividend amount - If not disclosed, estimate using: $\text{Dividend Amount} = \text{Net Profit} \times \text{Dividend Payout Ratio}$ - Net profit is predicted using historical profit distribution patterns, distinguishing between stable and unstable profit distributions[47][50] - Dividend payout ratio is estimated using historical averages or prior-year values, with adjustments for outliers[51][53] 3. Predict the ex-dividend date using historical intervals and linear extrapolation, or default to specific dates if historical data is unavailable[55][56] 4. Calculate the dividend points for the index: $\text{Dividend Points} = \sum_{n=1}^{N} \left( \frac{\text{Dividend Amount}_n}{\text{Market Cap}_n} \times \text{Weight}_n \times \text{Index Closing Price} \right)$ where $n$ represents each constituent, and only constituents with ex-dividend dates between the current date and the futures contract expiration date are included[38][44] Model Evaluation: The model demonstrates high accuracy for indices like the SSE 50 and CSI 300, with prediction errors generally within 5 points. However, the error margin for the CSI 500 index is slightly larger, around 10 points[57][61] Model Backtesting Results - Index Dividend Points Estimation Model: - SSE 50 Index: Prediction error ~5 points[61] - CSI 300 Index: Prediction error ~5 points[61] - CSI 500 Index: Prediction error ~10 points[61] Quantitative Factors and Construction Methods - Factor Name: Historical Profit Distribution Factor Factor Construction Idea: This factor predicts net profit by analyzing historical profit distribution patterns, distinguishing between stable and unstable distributions[50] Factor Construction Process: 1. Classify companies into stable or unstable profit distribution categories based on historical quarterly profit data 2. For stable distributions, use historical patterns to predict future profits 3. For unstable distributions, use the previous year's profit as the prediction[50] Factor Evaluation: Effective for companies with consistent profit patterns but less reliable for those with volatile earnings[50] - Factor Name: Historical Dividend Payout Ratio Factor Factor Construction Idea: This factor estimates the dividend payout ratio using historical averages or prior-year values, with adjustments for extreme values[51] Factor Construction Process: 1. Use the prior year's payout ratio if the company paid dividends last year 2. Use the average payout ratio of the last three years if no dividends were paid last year 3. Assume no dividends if the company has never paid dividends 4. Apply truncation if the estimated payout ratio exceeds 100%[53] Factor Evaluation: Reliable for companies with stable dividend policies but may overestimate for companies with irregular payouts[51][53] - Factor Name: Ex-Dividend Date Prediction Factor Factor Construction Idea: This factor predicts ex-dividend dates using historical intervals and linear extrapolation[55] Factor Construction Process: 1. Use the disclosed ex-dividend date if available 2. If unavailable, estimate based on historical intervals between announcement and ex-dividend dates 3. Default to specific dates (e.g., July 31, August 31, or September 30) if historical data is insufficient[56] Factor Evaluation: Accurate for most companies, with 90% of predictions falling within expected timeframes[56] Factor Backtesting Results - Historical Profit Distribution Factor: Effective for stable profit companies, less so for volatile ones[50] - Historical Dividend Payout Ratio Factor: Reliable for stable dividend policies, prone to overestimation for irregular payouts[51][53] - Ex-Dividend Date Prediction Factor: 90% accuracy for companies with historical data, with most predictions aligning with expected timelines[56]