Title: Multidimensional time-frequency analysis of the CAPM
Authors: Roman Mestre - University of Montpellier (France) [presenting]
Michel Terraza - lameta (France)
Abstract: The CAPM theory provides a measure of the sensitivity of an asset to the market called the systematic risk. The Beta is estimated by its market line. According to the OLS hypothesis, it is stable over time but this is not empirically verified. Many studies support this fact, and more particularly the Beta dispersion according to the frequencies which is related to the heterogeneous behavior of agents. Using the wavelets method, we can calculate the coherence and the phase between the stock's returns and those of the market over time. In order to confirm the correctness of the methodology, we use three French equities with different Betas (AXA, LVMH and Orange) from 2005 to 2015 period including the crisis. We show that the wavelets coherence, associated with the phase, improves our understanding about the Equity-Market relationship and then, the classification of equities according to their frequency characteristics (the contagion and interdependence phenomenon). The contagion effects (from the market to the stock) is principally located on the High Frequencies whereas the interdependence effect is located on the Low-Frequencies (Long-run investment). The link between beta and coherence-phase can help the investors to choose more efficiently the time they should invest.