Modelo CAPM condicional: uma melhor previsão para retornos de ativos brasileiros

Abstract

The Financial Market is a present theme in almost all the papers in the Finance literature, as it centralizes the most diverse negotiations around the world. With an interest in higher profits from more elaborate returns, investors need to allocate their resources in such a way that diversified combinations of assets are possible, aiming at decreasing their correlations and increasing the efficiency of their portfolio, as proposed by Markowitz (1952). Based on Portfolio Theory, other models have emerged to optimize asset return forecasts to ensure greater certainty of investment for market players. The percussive model of this condition is the Capital Asset Pricing Model, CAPM, proposed by Sharpe (1964) and Lintner (1965), which aims to relate the risk and return measures of assets in an Efficient Market scenario, with symmetrical information sharing as well as other simplifying assumptions. Globally known and respected, the CAPM model has become a reference when it comes to asset pricing and forecasting their returns, but because it has an empirical content that is far from reality, it has been criticized and reformulated. The main derivation of this model is called Conditional CAPM, where investors were not only concerned with non diversifiable risk or systematic market risk, but also encompassed in their choices the risk of each time traded asset. Tests by authors Jagannathan and Wang (1996) and Adrian and Franzoni (2009) have made room in the literature to validate it, as well as to offer closer predictions of the reality of tested stock returns in relation to the unconditional model. This paper aims to show the advantages of the Conditional CAPM model over the static model, through its application in daily data of the eighteen most traded Ibovespa index in the period 24/09/2013 to 09/21/2018, which returned forecasts more correct for most of these actions, confirming the hypothesis raised in this paper.