Comparative analysis of known time series models to forecast volatility in Philippine Stock and FOREX markets

Giovanni Moses Lim
Mariel I. Maranan

Abstract/Summary

The stock and FOREX markets are two of the known markets in the world of business, and in this study, a number of time series models to forecast volatility were compared. Furthermore, the research is aimed to study models for these two markets in the Philippine setting. Forecasting volatility is helpful in determining asset return distribution and in other finance applications such as investment, portfolio option pricing, hedging and risk management. Thus, it is important to examine and compare the time series models in forecasting volatility by computing the error statistics, pairwise comparison test, test for unbiasedness and the predictive power. Results show that the dominant model for FOREX is GARCH(1,1) and for stocks, Moving Average model of order 3. It can be noted that a number of forecasting models can be deemed accurate enough to forecast monthly volatility.