股指分红点位监控周报：市场交投活跃，IC合约贴水幅度大幅收窄-20260108

• The report introduces the dividend point estimation method for stock indices, emphasizing its importance in accurately calculating the premium or discount levels of stock index futures contracts, which track price indices rather than total return indices[43][12][44] - The dividend point estimation process involves key metrics such as component stock weights, dividend amounts, total market capitalization, and index closing prices. While some data like market capitalization and closing prices are directly accessible, component stock weights and dividend amounts require further estimation[48][43][44] - Component stock weights are refined from monthly data provided by mainstream data providers to daily weights using a formula that adjusts for non-reinvested price changes. The formula is: $W_{n,t}={\frac{w_{i0}\times(1+r_{n})}{\sum_{i=1}^{N}w_{i0}\times(1+r_{n})}}$ where $w_{i0}$ represents the weight of stock $n$ at the last disclosed date, and $r_{n}$ is the non-reinvested price change of stock $n$ during the period[49][50] - Dividend amounts are estimated using the formula: Dividend Amount = Net Profit × Dividend Payout Ratio Net profit is dynamically predicted based on historical profit distributions, categorizing companies into stable and unstable profit distribution groups. Stable companies follow historical patterns, while unstable ones use prior-year data as a reference[51][53][56] - Dividend payout ratios are predicted using historical averages. If a company paid dividends last year, the previous year's ratio is used; otherwise, a three-year average is applied. Companies with no dividend history are assumed not to pay dividends[54][57][55] - Ex-dividend dates are forecasted using historical intervals and linear extrapolation methods. If historical dates are deemed unreasonable, default dates are applied based on typical dividend announcement timelines[55][60][58] - The accuracy of the dividend point estimation model is validated by comparing predicted dividend points with actual points for major indices like the SSE 50, CSI 300, CSI 500, and CSI 1000. The model demonstrates high accuracy, with errors generally within 5 points for SSE 50 and CSI 300, and within 10 points for CSI 500 and CSI 1000[61][65][43] - The report highlights the predictive accuracy of the model for stock index futures contracts, showing minimal deviation between forecasted and actual dividend points for contracts across different indices[65][68][70]