Abstract: To identify funds skilled in both stock picking and market timing, we develop a test for the zero product of these two skills to first single out funds with at least one zero skill. Our simulations confirm the test's accurate size and good power. We apply our test to active U.S. equity mutual funds to exclude zero-skill funds, and classify the remaining funds based on stock picking and market timing. We find that the 1% of funds with both skills are the only group with significant risk-adjusted performance. We also provide evidence for stock-picking and market-timing trade-offs along multiple dimensions.

Key words and phrases: ARMA-GARCH model, bootstrap, heteroscedasticity, hypothesis testing, weighted inference.