Asset pricing and volatility modeling : the case of Indonesia stock market

Abstract

In determining the rate of return on stocks, many models have been introduced to obtain optimal returns and able to minimize risk. Equilibrium model such as the CAPM, APT and multifactor models have been used in calculating the level of risk and returns through portfolio formation. Since the development initiated by Markowitz who invented portfolio theory, the empirical results of many researchers have produced different points of view relating to stock return and risk relationship. This study aims to look at what factors can be used as a basis to determine returns and at the same time can minimize the risk. As in previous research studies using the CAPM, APT and multifactor models, this study focused on determining the combination of the most significant variables that determine portfolio stock returns in Indonesia. In addition to using the standard in obtaining beta estimates, this study also uses an estimate of volatility models. In obtaining the best model, the first variable that were selected passed through several test models of equilibrium, so that the best model only includes several valid variables. The research was divided into three different economic conditions; full period and two sub periods indicating financial crisis (1998`s) and the global crisis (2008`s). The results showed that the CAPM is not valid and that market capitalization variable more able to explain changes in the portfolio yield. The model of the APT shows that macroeconomic and market risk premium are significant in explaining changes in portfolio returns, except for the production index. Several fundamental factors of the multifactor models are also found to be significant variables including rating, and that liquidity factor is still an investment benchmark in Indonesia. It is proven that the volume and frequency of trades consistently significant in all test models. Apart from that, the variables showed significant ratings that investors in Indonesia are still passive, traditional and avoid risk. The simulation results of this study indicate that beta is estimated using a standard similar to that estimated using ARCH beta (volatility modeling), and that both methods showed the same conclusion. As such, it can be said to be consistent in terms of portfolio formation. Also, the magnitude and direction of the regression coefficients were tested using several models. In addition, when the establishment of a portfolio simulation was made, it was found that there is an effect of market capitalization. Small-cap portfolios have higher returns than large-cap, and Value at Risk (VaR) value is similar relatively between the two methods of portfolio formation.