dc.description.abstract | Models based on psychological biases can explain momentum and reversal in stock returns, but risk
overfitting of theory to data. We examine a central psychological bias, representativeness, which
underlies many behavioral-finance theories. According to this bias, individuals form predictions about
future outcomes based on how closely past outcomes fit certain categories. To produce out-of sample
tests, we use accounting performance to identify these categories and test the idea that investors
misclassify firms and thus make biased forecasts. We find evidence of short-term accounting
momentum, consistent with the idea that investors fail to immediately incorporate new information, but
find no support for long-term reversal related to accounting performance. Contrary to theory, we find
little evidence that the consistency of past accounting performance is related to future returns | en |