Pustaka Ilmiah Universitas Padjadjaran

Abstrak

Mean-MVaR Portfolio Optimization Under CAPM With Lagged, Non Constant Volatility and the Long Memory Effect

In this paper, we discus the methods of portfolio optimization based on the mean and the Modified Valueat-Risk(MVaR) under the Capital Asset Pricing Model (CAPM) with lagged, non-constant volatility and the long memory effect. In CAPM, the returns of individual stocks (or portfolios) are assumed influenced by the market return and risk-free return. Here, we estimate the stock return betas by extending the CAPM model with lagged market factors, where the market returns are assumed has non constant volatility, which will be estimated using GARCH models, and the long memory effect will be modeled using ARFIMA model. The risk is measured by MVaR that is calculated using normal distribution with a confidence level c. Mean and MVaR will be used for formulation of the portfolio optimization problems. The portfolio optimization is performed using the Lagrangean Multiplier and the solution is obtained by the Kuhn-Tucker theorems. We illustrate these methods using some stocks from the Indonesian capital market.