CAPM - MARKOV SWITCHING AND KALMAN FILTER: AN APPLICATION TO B3’S SUSTAINABILITY AND GOVERNANCE INDUSTRY INDICES

Abstract

This paper seeks to understand the behavior of the sectoral, governance
and sustainability indices of the Brazilian stock exchange. For this, a traditional methodology was used with the CAPM models in their versions: static, with regime change
(Markov Switching) and with betas varying at each point of time (Kalman Filter and
Smoother). The application of this methodology brought evidence that eight of the
nine indexes analyzed present structural change (alternating between two regimes). In
addition, it was noted that betas are unstable over time, ie there is a nonlinear relationship between risk and return. Overall, the results found indicate that the systemic risk
(beta) of the analyzed indices varies over time and depends on regimens. Finally, this
analysis allows the manager or investor to have access to a set of information relevant to his decision making regarding investments in sectors or a set of companies with good
practices, which make up the Brazilian stock exchange.