Asset Pricing Model under Costly Information Evidence from the Tunisian Stock Market

The purpose of this paper is to propose a new model for price estimation in financial markets. This model considers costly information; investors must buy information in order to reach an optimal decision. We use entropy statistics to estimate information cost. Asset's price is supposed to be a linear function of its: previous price; information cost; exchanged quantity, and the risk-free rate of interest. We find that this model proves a very significant aptitude to anticipate future asset's prices of the Tunisian Stock Market over the period extending from 2002 to 2008. The proposed model allows both institutional and particular investors to predict future's asset prices on the basis of knowledge of the previous price, the information, the exchanged quantity and the risk-free rate of interest.