Dynamic adjustment of stock prices to the fundamental valueAn error correction approach

This thesis analyses the long term dynamic adjustment of stock prices to the fundamental value process. Whereas most research on relative mean reversion is based on crude measures of the fundamental value, such as dividends and earnings, we model various proxies of thefundamental value process which explicitly take into account expected future cash flows instead of realised cash flows. First, we consider the ability of the simple mean reversion model, the two step Error Correction Model (ECM) and the single equation ECM to account forthe dynamics of the mean reversion process. It is found that the single equation ECM, which takes into account both the short- and long term effect of changes in the fundamental value on the stock price, as well as changes in the co-integrating relation between these two variables over time, is the preferred model specification. When the single equation ECM is estimated at country level, we find significant evidence of mean reversion, unlike previous studies onrelative mean reversion. Secondly, we estimate the single equation ECM within 2 panel datasets: one long dataset from 1922 to 2009 including 3 countries and one broad dataset from 1973 to 2009 including 13 countries. In the first dataset we find a significant positive speed of mean reversion with a half-life ranging between 8.5 and 18.8 years, depending on the proxy included in the model. For our second dataset, we do not find significant estimates of the speed of mean reversion, which underlines the importance of obtaining sufficiently long time series to find evidence in favour of mean reversion. Using rolling window estimation, we find large fluctuations in the speed of mean reversion over time, and in a substantial number of periods no mean reversion is found at all.