Guotai Haitong: Recommend overweighting large-cap style in October, with equal-weight allocation to value and growth styles.

Guotai Haitong released a research report stating that, based on quantitative model signals at the end of September, it recommends overweighting large-cap style in October and allocating equally to value and growth styles. The market-cap factor valuation spread is 0.76, and small/micro-cap remains favored over the medium and long term; growth style has been dominant for two years, while value style is expected to outperform over the long term. Guotai Haitong's main views are as follows: Monthly large-cap/small-cap style rotation strategy. The quantitative model signal at the end of September was -1, pointing to large-cap; it is recommended to overweight large-cap style in October. From a medium-to-long-term perspective, large-cap and small-cap rotate once every 5-10 years. The current market-cap factor valuation spread is 0.76, still some distance from the historical top region of 1.7-2.6, so it is not crowded over the medium and long term, and small/micro-cap remains favored. As of the end of September, the model's year-to-date return was -9.64%, with excess return of -5.72% relative to the equal-weight benchmark (-3.93%). Reviewing the September overweight large-cap view, the view was wrong. Monthly value/growth style rotation strategy. The quantitative model signal at the end of September was 0, and it is recommended to allocate equally to value and growth styles in October. From a medium-to-long-term perspective, value and growth rotate once every 2-3 years. Growth style has already been dominant for two years, and there is a possibility of switching to a long-term period of value style dominance, so value style is more favored going forward. As of the end of September, the model return was -9.56%, with excess return of -6.55% relative to the equal-weight benchmark (-3.21%). Style factor performance tracking. Among the 8 major categories of factors, value and quality factors had relatively high positive returns this month; large-cap and momentum factors had relatively high negative returns. Year to date, volatility and momentum factors had relatively high positive returns; large-cap and liquidity factors had relatively high negative returns. Among the 20 style factors, book-to-market ratio, earnings quality, and earnings yield factors had relatively high positive returns this month; large-cap, residual volatility, and industry momentum factors had relatively high negative returns. Year to date, momentum, beta, and earnings yield factors had relatively high positive returns; industry momentum, short-term reversal, and residual volatility factors had relatively high negative returns. Factor covariance matrix update. Stock covariance matrix estimation is the core of stock portfolio risk forecasting. Using a multi-factor model, the stock covariance matrix can be decomposed into a combination of the factor covariance matrix and the stock-specific risk matrix, thereby achieving a relatively accurate estimate. Risk warning: Quantitative models are built based on historical data, and historical patterns are subject to the risk of failure.