金工ETF点评：跨境ETF单日净流入66.57亿元，医药拥挤持续满位，钢铁建材高位

• The report constructs an industry crowding monitoring model to monitor the crowding degree of Shenwan first-level industry indices on a daily basis[4] - The ETF product screening signal model is built using the premium rate Z-score model, which provides potential arbitrage opportunities through rolling calculations[5] Model Construction and Evaluation 1. Industry Crowding Monitoring Model - Construction Idea: Monitor the crowding degree of Shenwan first-level industry indices daily[4] - Construction Process: The model calculates the crowding degree of each industry index based on the daily trading data and ranks them accordingly[4] - Evaluation: The model effectively identifies industries with high and low crowding degrees, providing valuable insights for investment decisions[4] 2. Premium Rate Z-score Model - Construction Idea: Identify potential arbitrage opportunities in ETF products by calculating the Z-score of their premium rates[5] - Construction Process: - Calculate the premium rate of each ETF product - Compute the Z-score of the premium rate using the formula: $ Z = \frac{(X - \mu)}{\sigma} $ where $X$ is the premium rate, $\mu$ is the mean premium rate, and $\sigma$ is the standard deviation of the premium rate[5] - Evaluation: The model helps in identifying ETF products with significant deviations from their average premium rates, indicating potential arbitrage opportunities[5] Model Backtesting Results 1. Industry Crowding Monitoring Model - Top Crowded Industries: Pharmaceuticals, Steel, Building Materials[4] - Least Crowded Industries: Automobiles, Home Appliances[4] 2. Premium Rate Z-score Model - Top Potential Arbitrage Opportunities: Identified through rolling calculations, specific ETF products are not listed in the provided content[5]