Résumé
The OLS Estimation of the CAPM suffers a lot of statistical issues. We develop a TimeFrequency Multi-Betas Model with Gold and Oil as supplementary source of risk and with ARMA-EGARCH errors to take into consideration some of these statistical weaknesses. We use 30 french stocks listed on the CAC40 for the daily period from 2005 to 2015. The conjugaison of the wavelets decompositions and the parameters estimates constitutes an significant asset for managers choices according to their view (shor-medium-long term). The results represent a decision support by multiplying the interpretive possibilities, In short-run (''Noise Trader'' and ''High Frequency trader'') 1/3 of the equities are not affected by the Oil and Gold fluctuations, and the estimated Betas parameters related to the market movements are few different compare to the Model without wavelets. At the opposite, in long-run (fundamentalists investors), Oil and Gold affect all the stocks but their impact varies according to the Beta (sensitivity to the market). As example, we highlight that Oil prices negatively affect the stocks in long-run especially the equities with a high market beta (greater than 1) as Banking Stocks. We also observe significant differences between parameters estimated with and without wavelets.