By Peter L. Bernstein

Abstract

Werner de Bondt and Richard Thaler have developed an elaborate set of statistical data to demonstrate what market practitioners have been insisting upon for a very long time indeed: equity valuation errors are systematic and therefore predictable. While I find the reasoning in their paper to be convincing and their test results to be most impressive, the authors fail to recognise some elements of the marketplace that would enrich their analysis and that - just maybe - might help investors to break their unfortunate habits.