资产配置月报202602：如何衡量黄金的交易拥挤度？-20260206

• The report introduces a structured static factor model for predicting 10Y government bond yield movements, utilizing four macroeconomic factors: economic growth, inflation, debt leverage, and short-term interest rates[44][50][53] - The model has achieved a historical prediction accuracy of approximately 70% since 2006, with a sample-out accuracy of 68% since 2023[47][50] - For February 2026, the model forecasts a 6BP increase in the 10Y government bond yield to 1.88%, driven by upward trends in all four macroeconomic factors[50][53] - A structured static factor model is also applied to gold price movements, incorporating four key factors: US economy, US employment, US fiscal policy, and US external debt[54][57] - The gold model has demonstrated a historical prediction accuracy of 65% since 2008, with a sample-out accuracy of 78% since 2023[54][55] - The report highlights that fiscal and employment factors are currently supporting gold price increases, while economic and external debt factors show mixed signals[57][60] - A quantitative strategy for managing gold positions based on trading congestion is proposed, using two metrics: 40-day price deviation rate and SHFE gold implied volatility (IV)[19][21] - The strategy suggests reducing portfolio exposure to 40% when the 40-day price deviation rate exceeds 9% and SHFE gold IV surpasses 30%, achieving an excess return of 53.4% and improving the Sharpe ratio from 1.26 to 1.62 during backtesting from 2020 to February 2026[21][19] - The report recommends a multi-dimensional industry allocation strategy combining "win-rate and odds" and "clearance reversal" approaches, with industries such as non-ferrous metals, basic chemicals, and steel being highlighted[99][102][115]