PACIFIC SECURITIES(SH:601099) Tai Ping Yang Zheng Quan·2025-12-22 11:45Quantitative Models and Construction Methods 1. Model Name: Industry Crowding Monitoring Model - Model Construction Idea: This model is designed to monitor the crowding levels of industries on a daily basis, focusing on the Shenwan First-Level Industry Index. It identifies industries with high or low crowding levels and tracks changes in crowding over time[3]. - Model Construction Process: The model calculates the crowding level of each industry based on specific metrics (not detailed in the report). It then ranks industries by their crowding levels and highlights those with significant changes in crowding. For example, the report notes that the military and retail industries had high crowding levels, while the computer industry had relatively low levels. Additionally, it tracks main fund flows into and out of industries over recent trading days[3]. - Model Evaluation: The model provides actionable insights into industry crowding trends, helping investors identify potential opportunities or risks in specific sectors[3]. 2. Model Name: Premium Rate Z-Score Model - Model Construction Idea: This model is used to screen ETF products for potential arbitrage opportunities by calculating the Z-score of their premium rates. It also serves as a warning signal for potential price corrections in ETFs[4]. - Model Construction Process: The model involves rolling calculations of the Z-score for the premium rates of various ETFs. The Z-score is calculated as: $ Z = \frac{(P - \mu)}{\sigma} $ where $ P $ is the current premium rate, $ \mu $ is the mean premium rate over a rolling window, and $ \sigma $ is the standard deviation of the premium rate over the same window. ETFs with extreme Z-scores are flagged as potential arbitrage opportunities or correction risks[4]. - Model Evaluation: The model is effective in identifying ETFs with significant deviations from their historical premium rates, providing opportunities for arbitrage or risk management[4]. Model Backtesting Results 1. Industry Crowding Monitoring Model - No specific numerical backtesting results are provided for this model in the report[3]. 2. Premium Rate Z-Score Model - No specific numerical backtesting results are provided for this model in the report[4]. Quantitative Factors and Construction Methods No specific quantitative factors are detailed in the report. Factor Backtesting Results No specific backtesting results for factors are detailed in the report.