Interval-valued estimation for the five largest market capitalization stocks in the stock exchange of Thailand by Markov-Switching CAPM

Abstract

© 2018, Springer International Publishing AG. The paper aims to quantify risk using Capital Asset Pricing Model (CAPM) of the 5 top-traded stocks in Thailand’s Stock Exchange market as well as to investigate existence of structural change in CAPM. Thus, in this paper, we compare non-linear Markov-switching CAPM with linear CAPM. In addition, we use interval-valued data instead of conventional single-valued data because of its ability to capture the whole period rather than a point of time. This paper, therefore, introduces an approach to fit both Markov-switching and linear CAPM to interval-valued data. Interval value of each stock return is retrieved from its midpoint to fit in Markov-switching and linear regression estimations which apply the midpoint of interval value of market return. From empirical analysis, the results are satisfactory as AIC judged that our Markov-switching CAPM outperforms the linear benchmark four out of five stocks.