Resumen
For many decades, asset-pricing theory was constructed on absence of frictions hypothesis. However, since pioneers? works of Chen et al. (1975) and Amihud and Mendelson (1986) in microstructure area, systematic violations of this assumption were observed and the standard CAPM generates biased estimates of returns. The main obstacle that we face when we use CAPM with frictions is the unavailability of the observable variables that directly measure illiquidity costs, transaction costs and information costs. Microstructure literature provides many models to estimate them. We use Amihud (2002) model to estimate illiquidity costs and Lin et al. (1995) model to estimate transaction and information costs. Before a study of a sample of 40 quoted securities in Tunisian financial market, on the period of 07/02/2011 to 31/01/2013, results appear conclusive. First, we show the existence of asset pricing bias compared to the standard CAPM. Furthermore, we find a strong relation between transaction costs, information costs and illiquidity estimators and returns; thus the use of the CAPM with frictions. This model appears very strong and the results prove the necessity to account for frictions in asset pricing. Keywords: CAPM; market microstructure; frictions; bid-ask spread components. JEL Classifications: G20; G21; G01; G31; G32