Ratings and Asset Allocation: An Experimental Analysis

Investment ratings (e.g. by Morningstar) provide a simple ordinal measure to compare assets. Typically, ratings are assigned within categories — groupings in which assets share important characteristics. Comparing ratings across categories can be misleading. We study ratings using an asset allocation experiment focusing specifically on this problem. Subjects make repeated allocation decisions under complete information. Ratings that convey no new information and that are altered because of categorization affect subject investment choices and harm performance in the experiment.