Forecasting Market Risk in ASEAN-5 Indices using Normal and Cornish-Fisher Value at Risk

Abstract

The purpose of this paper is to quantify and compare the risk in ASEAN-5 markets consist of Indonesia (IDX), Malaysia (MYX), Singapore (SGX), Thailand (SET), and Philippines (PSE) and design accurate and practical method to measure daily market risk using Value at Risk model. We compare two distribution model namely normal (Normal VaR) and Cornish-Fisher (CFVaR) to modify normal quantile to achieve most precise result. Besides, we employed ARMA-GARCH model to quantify one-day-ahead volatility. This study found that distribution of ARMA residual in all markets show asymetric characteristics with leptokurtic and negative skewness. The ARMA-GARCH (1,1) is powerful forecasting tool in emerging ASEAN but it is less effective in developed ASEAN due to the absence of randomness in ARMA residuals. The Normal VaR is best used in markets with less skewed distribution such as MYX and SGX. However, the CFVaR is rejected in PSE markets due to overestimation of risk. Meanwhile, the CFVaR is best used in IDX and SET markets which indicated by less failures produced by this model in both markets.

Keywords: ARMA-GARCH, normal distribution, modified quantile, value at risk