WHAT ARE THE IMPLICATIONS OF THIS QUESTION ON ACCOUNTING? alue relevance research, broadly defined, seemed to be a default thesis topic for Ph.D. students in the 90s such as me, but I was lucky, or unlucky, to have Shyam Sunder as my adviser. Accordingly, I approached value relevance studies from a skeptical point of view. ost empirical research does not sufficiently contemplate the model-dependent nature of value elevance studies, and I want to emphasize that event studies are not free from this dependence ecause we cannot distinguish a reaction to news from a realization of risk premium. CAPM is no longer the asset pricing model except in introductory MBA courses, and its mpirical validity has been seriously questioned Campbell 2008; Cochrane 2007a . In particular, he constant equity risk premium assumption is empirically untenable. As Cochrane 2007b ointed out, the alpha and beta framework has been dead, and “there is no alpha and beta, there is ust beta you know about, and beta you don’t understand yet, and no clear separation between the wo.” That observation means that we can find a mean-variance efficient portfolio incorporating as any factors as we want, unless there is an arbitrage opportunity. So there are few, if any, estrictions for model construction. Because any value relevance study is model-dependent, and there is virtually no restriction on sset pricing models, we cannot get any objective policy implications from observed returns being bnormal or unexpected because we can always claim the return is not abnormal or unexpected ut is the result of the underlying asset pricing model’s being subnormal.