The use of analyst forecasts as proxies for investors' earnings expectations is commonplace in empirical research. An implicit assumption behind their use is that they reflect analysts' private information in an unbiased manner. As demonstrated here, this assumption is not necessarily valid. There is shown to be a tendency for analysts to release forecasts closer to prior earnings expectations than is appropriate, given their information. Further, analysts exhibit herding behavior, whereby they release forecasts similar to those previously announced by other analysts, even when this is not justified by their information. These results are shown to have interesting empirical implications. Article published by Oxford University Press on behalf of the Society for Financial Studies in its journal, The Review of Financial Studies.