This study develops a style rotation model based on quarterly forecasts of style factor (SF) returns, across four style categories, generated using market and macroeconomic data. The prescriptions from this model are tested on a sample of US active equity mutual funds’ portfolio holdings. An annual buy-and-hold style timing strategy investing in the factor with the highest forecast return each quarter achieves an average annual excess return of 7.26%, significant at the 1% level during 1981–2011. However, a fund-of-fund (FoF) timing strategy investing in the funds with the greatest exposure (i.e. the preferred funds) to the style predicted to outperform over the following year does not generate statistically significant Daniel, Grinblatt, Titman and Wermers (DGTW)-adjusted performance. The lack of performance is primarily because the long-only funds are by nature unable to fully exploit the long-short SF returns. This highlights the issue of using long-short portfolio returns, particularly when evaluating fund performance.