中银量化绝对收益系列专题：宏观因子资产化框架下的国债期货择时策略

Core Insights - The report introduces a macro factor assetization framework for timing strategies in government bond futures, demonstrating robust return characteristics and strong risk resistance through backtesting [1][2]. Group 1: Macro Real-Time (PIT) Indicator Library Construction - The macro factor assetization strategy utilizes real-time macro data, contrasting with traditional models that lag by 1-2 months, by employing a precise macroeconomic calendar to obtain macro data disclosure dates and times [4][19]. - The PIT macro indicators are designed across four dimensions: economic growth, inflation, monetary credit policy, and central bank open market operations, creating a macro factor library [4][19]. Group 2: Strategy Construction and Backtesting - The strategy framework consists of three main steps: macro factor construction, macro trading logic net value realization, and dynamic factor selection and combination [4][29]. - The model achieved a post-fee Sharpe ratio of approximately 1.3 and a Calmar ratio of about 1.1, indicating strong performance despite challenges in capturing significant excess returns during the bull market from 2021 to 2024 [4][29]. - The model's performance is relatively insensitive to the lag parameter n, with optimal settings found between 10 to 30 minutes, leading to a standardized approach of a 10-minute lag for all signals [4][29]. Group 3: Factor Dynamic Selection and Combination - The macro factors are categorized into four types: economic growth, inflation, monetary credit, and open market operations, with each factor's performance analyzed for effective timing signals [4][29]. - The report emphasizes the importance of dynamic factor selection to enhance model performance, utilizing momentum factor selection methods to optimize the factor pool [4][29][56]. - The empirical results indicate that the combined signals from multiple factors significantly improve timing effectiveness compared to single-factor performance [4][29][74].