fama and macbeth revisited: A Critique

The main conclusion of the FM study relies on the fact that the average of the slopes of 402 regressions of the monthly returns on 20 portfolios on theirs beta coefficients is positive. Considering this set of 402 slopes as a random sample drawn from the same normally distributed population, FM performed a t-test on the mean and conclude that the true mean significantly differs from zero. Then they took this result as a proof in favour of the theory that there is in the real world a perfect linear relationship between the expected return and the true beta of securities and portfolios or, in other terms, in favour of the theory that the market portfolio is efficient. In this article, we present several tests and arguments that put some shadow on these conclusions. In fact, several tests lead us to the conclusion that the 402 random observations above mentioned are not drawn from a normal (or symmetric stable) distribution, neither are they independent or identically distributed. Indeed, the most disturbing fact is that those observations are likely not independent.