公募基金市场观察系列：财富管理新范式，ETF投顾展现巨大潜力

• The report introduces five ETF quantitative strategies: macro timing strategy, momentum selection strategy, low-volatility diffusion industry rotation strategy, fund flow strategy, and quantile regression strategy[41][43][52] - Macro Timing Strategy: This strategy uses Gaussian distribution-based Black-Litterman (B-L) and Copula distribution-based B-L models to calculate ETF allocation weights. It incorporates economic cycle segmentation to constrain asset allocation weights across different ETF categories. Representative ETFs include stock ETFs (e.g., CSI 300 ETF), bond ETFs (e.g., government bond ETF), commodity ETFs (e.g., gold ETF), and currency ETFs (e.g., Silver Day Profit A). The strategy achieved an annualized return of 9.43%, Sharpe ratio of 0.66, Calmar ratio of 0.52, and maximum drawdown of -13.21% since 2020[43][45][46] - Momentum Selection Strategy: This strategy utilizes XGBoost to predict ETF upward probability as a momentum indicator and historical quantile of fund shares as a crowding indicator. It selects ETFs with high momentum and low crowding within sectors, adjusting allocation weights based on crowding levels. The strategy achieved an annualized return of 13.93%, Sharpe ratio of 1.33, Calmar ratio of 1.21, and maximum drawdown of -23.14% since 2020[53][54][55] - Low-Volatility Diffusion Industry Rotation Strategy: This strategy applies industry diffusion index factors combined with volatility adjustments to design a low-volatility industry rotation model. It matches ETFs to industry indices based on correlation and selects top ETFs for monthly rebalancing. The strategy achieved an annualized return of 12.22%, relative excess return of 3.21%, Sharpe ratio of 0.91, and maximum drawdown of -21.43% since 2020[61][62][63] - Fund Flow Strategy: This strategy uses weighted fund flow indicators and risk metrics to rank industries, selects ETFs based on turnover rate and premium/discount rate, and optimizes allocation weights using second-order stochastic dominance. The strategy achieved an annualized return of 11.24%, Sharpe ratio of 1.12, Calmar ratio of 1.03, and maximum drawdown of -19.21% since 2020[64][67][68] - Quantile Regression Strategy: This strategy employs quantile random forest (QRF) to predict future distribution characteristics of technology sector indices, selects ETFs based on liquidity and scale, and adjusts positions using MACD and volatility metrics. The strategy achieved an annualized return of 15.62%, Sharpe ratio of 1.21, Calmar ratio of 1.15, and maximum drawdown of -21.43% since 2020[71][72][73]