Robust Fundamental Analysis

According to market efficiency, it would not be possible to profit from mispricing in stock prices. In this scenario, in case investors are able to correctly price the outstanding information, there would not be mispricing signals to invest on. If instead, investors would not always value prices correctly, by estimating how, on average, investors reflect information into prices,it would be possible to detect firms where certain information have been over (under) weighted at a specific point in time. In order to estimate the weights that investors give to the information it is necessary to run a cross-sectional regression at company level with on the LHS the market cap and on the RHS the chosen explanatory variable. However, this analysis does not consider periods where the whole market has an irrational behavior (e.g. speculative bubble). I give evidence in this paper that it is essential to consider the time dimension to correctly estimate the full deviation of the market price from its fundamental. I achieve this by applying two different panel estimation techniques: Fama-MacBeth and Cointegration panel estimation. Moreover, I show what happens to the cross-sectional mispricing estimation when I add dummy variables to the model (Dummy for sectors, Dummy for Book to Market), in addition to the publicly available information, in order to control for differences in the price level for sub-groups of companies. I also give evidence that fundamental analysis actually works, by building a trading strategy that relies on price deviations from their fundamental values.