The strategy trades mutual funds, predicting returns using lagged fund-level characteristics and a BRT model, longing top-decile funds and shorting bottom-decile funds, rebalancing value-weighted portfolios monthly.
ASSET CLASS: funds | REGION: United States | FREQUENCY:
Monthly | MARKET: equities | KEYWORD: Machine Learning, Mutual Fund
I. STRATEGY IN A NUTSHELL
Predict returns of actively managed mutual funds using machine learning (Boosted Regression Trees) on lagged, fund-level characteristics derived from 94 stock-specific factors. Each month, sort funds into deciles by predicted returns, going long the top decile and short the bottom, with value-weighted portfolios rebalanced monthly.
II. ECONOMIC RATIONALE
Mutual fund returns depend on complex, non-linear interactions among holdings, making traditional univariate methods ineffective. Machine learning captures these relationships, improving prediction accuracy. Boosted Regression Trees iteratively focus on poorly predicted funds, enhancing model performance while avoiding reliance on past fund returns.
III. SOURCE PAPER
Selecting Mutual Funds From the Stocks They Hold: a Machine Learning Approach [Click to Open PDF]
Bin Li, Alberto Rossi, Wuhan University, Georgetown University
<Abstract>
We select mutual funds in real time by combining individual fund holdings and a large number (94) of stock characteristics to compute fund-level characteristics on the basis of the stocks they hold. We show that, first, the majority of funds are largely exposed—both positively and negatively—to approximately 40-50 characteristics. Second, fund performance is non-linearly related to fund characteristics and there are significant degrees of interaction between different fund characteristics and fund performance. Third, when we predict fund performance, these non-linearities and interactions prove important as machine learning methods such as Boosted Regression Trees (BRT) outperform significantly standard linear frameworks and the BRT-generated forecasts encompass the ones generated by the predictors of mutual fund performance that have been proposed in the literature so far. Fourth, while in our setting BRT outperform the LASSO, elastic nets, random forests, and neural networks with 1 through 5 hidden layers, these other machine learning methods deliver good performance and they all outperform ordinary least squares models. Finally, while we detect signicant predictability using machine learning methods, the fund characteristics that matter the most in predicting fund returns and the functional relation between fund characteristics and fund performance are time-varying.
IV. BACKTEST PERFORMANCE
| Annualised Return | 6.32% |
| Volatility | 12.85% |
| Beta | N/A |
| Sharpe Ratio | 0.49 |
| Sortino Ratio | N/A |
| Maximum Drawdown | N/A |
| Win Rate | N/A |
