金融工程周报：期指长周期因子下降-20251215

Group 1: Report Industry Investment Ratings - Stock Index: ☆☆☆ [1] - Treasury Bond: ☆☆☆ [1] Group 2: Core Viewpoints of the Report - As of the week ending December 12th, the A - share market showed a structured and volatile trend. The average daily trading volume of the whole market was 1.95 trillion yuan, an increase of nearly 260 billion yuan compared with the previous week. The three major indexes showed different trends, with the Shanghai Composite Index falling 0.34%. There was relatively limited information on short - term incremental policies and economic data, and market structural characteristics emerged [1]. - From the high - frequency macro - fundamental factor scores, for stock index futures, the inflation indicator scored 8 points, the liquidity indicator scored 9 points, the valuation indicator scored 11 points, and the market sentiment indicator scored 9 points. For bond futures, the inflation indicator scored 8 points, the liquidity indicator scored 10 points, and the market sentiment indicator scored 6 points [1]. - In terms of term structure, the weighted annualized basis rates (dividend - adjusted) of the ending positions of IH, IF, IC, and IM were 0.33%, - 2.32%, - 4.16%, and - 9.95% respectively, and the discount of far - month contracts widened again [1]. - The net value of the quantitative CTA strategy for financial derivatives did not change last week. In the long - term, although the social financing data slightly exceeded expectations, the credit data such as M1 and M2 showed sub - seasonal declines and were lower than expected. The credit factor put pressure on stock index futures, with a relatively large decline in IC. In the short - term, the high - frequency real estate and consumption sectors were still weak. The RMB continued to appreciate against the US dollar, and the capital situation remained relatively loose, but the short - term increase was relatively limited. In terms of positions, the risk appetite significantly recovered compared with the previous week. IF and IH remained relatively neutral, while IC and IM had relatively large declines. The overall comprehensive signal was in the neutral range. For bond futures, the capital situation remained loose. After a short - term rise, the positions of bond futures significantly declined. The stock - bond seesaw effect shrank, and the bond market was insensitive to fundamental feedback. The position factor of TF slightly declined, and institutional year - end allocation behavior was relatively cautious. The comprehensive signal was in a neutral and volatile state [1]. Group 3: Summary According to Relevant Catalogs 3.1 Macro - fundamental High - frequency Factor Scores - For economic kinetic energy indicators, including blast furnace开工率, PTA开工率, etc., different indicators showed different weekly changes, numerical values, historical quantiles, and correlations with stock and bond indexes. The scores for stock index futures and bond futures were both 8 points [2]. 3.2 Inflation Indicators - Various inflation - related indicators such as the vegetable basket product wholesale price index, coking coal, etc. had different weekly changes, numerical values, historical quantiles, and correlations with stock and bond indexes. The scores for stock index futures and bond futures were both 8 points [3]. 3.3 Liquidity Indicators - Liquidity - related indicators such as DR007, DR001, etc. had different weekly changes, numerical values, historical quantiles, and correlations with stock and bond indexes. The score for stock index futures was 9 points [4]. 3.4 Index Valuation - Index valuation indicators such as price - to - earnings ratio, price - to - sales ratio, etc. had different weekly changes, numerical values, historical quantiles, and correlations with the stock index. The score for stock index futures was 10 points [5]. 3.5 Market Sentiment: Stock Index - Stock - index - related market sentiment indicators such as margin trading balance, northbound capital inflow, etc. had different weekly changes, numerical values, historical quantiles, and correlations with the stock index. The score for bond futures was 9 points [6]. 3.6 Market Sentiment: Bond - Bond - related market sentiment indicators such as the yield of 10 - year government - developed bonds, the VIX index, etc. had different weekly changes, numerical values, historical quantiles, and correlations with the bond index. The score for bond futures was 6 points [7]. 3.7 Strategy Introduction (Quantitative CTA Strategy) - The product pool includes stock index futures and bond futures. The short - term model focuses on market style, external factors, and capital - related high - frequency financial data. The long - term model focuses on market expectations and macro - economic low - frequency indicators. The position data is synthesized considering institutional long and short positions [15]. - The comprehensive signal strength is weighted by the signals of three independent models (0 - 1). Contracts with the top 2 comprehensive signal strengths greater than or equal to 0.6 are considered for long positions, and those with the bottom 2 less than or equal to 0.4 are considered for short positions. Signals are shielded 7 days before the delivery date. The stop - loss point is set at a daily decline of more than 1%, with equal - weighted allocation of capital. Consecutive two - day same - direction signals are shielded [16][17]. 3.8 Last Week's Situation - The data of IF, IH, IC, IM, T, and TF main contracts from December 8th to 12th, 2025 were all 0 [18]. 3.9 Treasury Bond Futures Cross - variety Arbitrage Strategy - The cross - variety arbitrage strategy is based on the signal resonance of the fundamental three - factor model and the trend regression model. The fundamental factor uses the instantaneous forward - rate function proposed by Nelson and Siegel, which decomposes the interest - rate term structure into level, slope, and curvature. The three - factor model is constructed using PCA, factor rotation, and logistic regression, with signals divided into three types: '1', '0', and '- 1'. The trend regression model is used to filter signals, and trading is carried out when there is resonance. In actual operation, the 10 - 5Y spread is adjusted with a duration - neutral ratio of 1:1.8 [19]. 3.10 TF and T Main Contract Trading Signals - From December 8th to 12th, 2025, the N - S model and trend regression model signals of TF and T main contracts showed different situations [22].