We examine two forward-looking mutual fund ratings provided by Morningstar: the analyst rating and the quantitative rating based on the machine learning technique. The analyst rating identifies outperforming funds, while the quantitative rating fails to do so. The analyst rating outperforms the quantitative rating in predicting performance mostly due to the selection of analyst coverage. The tone in the analyst report contains incremental information. Retail investors do not follow analyst recommendations, but instead chase the quantitative rating. The overall evidence highlights the importance of mutual fund analysts in information production and implies a capital misallocation problem in mutual fund investment. 

Keywords: Analyst Rating; Quantitative Rating; Mutual Funds; Information Provision; Fund Flows; Machine Learning