Investors face greater difficulty valuing loss-reporting than profitreporting firms: losses may be due to very different reasons (e.g., poor operating performance or investments in intangibles, and financial accounting information is of more limited use for valuing lossmaking firms than profit-making firms. Because of increased uncertainty about loss firms' future financial and business viability, we hypothesize that financial analysts will be more selective when choosing to follow loss firms than profit firms, with the result that "abnormal" analyst following will be more informative to investors regarding the future performance of loss firms than profit firms. Consistent with this prediction, we find that abnormal analyst coverage is useful for predicting firms' future prospects, and is more strongly associated with future performance (stock returns and ROA) for loss firms than for profit firms. The market, however, does not seem to use this useful information when pricing loss firms: for loss firms a portfolio investment strategy based upon abnormal analyst following can generate positive excess returns over 1-to 3-year holding periods. These results are stronger for persistent-loss firms than for occasional-loss firms. We conclude that abnormal analyst following contains useful information about firms' future prospects, and even more so for loss firms than for profit firms. K E Y W O R D Sanalyst following, future performance, loss firms, self-selection 140