508 – Forecasting and Modeling Financial Volatility
Modeling and Forecasting Financial Volatility Using Composite CARR Models
Isuru Ratnayake
Missouri University of Science and Technology
V. A. Samaranayake
Missouri University of Science and Technology
Recent studies have shown that when it comes to forecasting realized volatility, conditional autoregressive range (CARR) models, that utilize the daily range of a commodity price, outperforms the traditional GARCH approach that models the daily returns. The CARR models, however, assume that the unconditional mean range is constant over time, which holds only if the unconditional volatility remains fixed over the duration of the study period. As several authors have pointed out, there is strong empirical evidence suggesting the feasibility of modeling a slow-varying change in the unconditional volatility over the study period using long term volatility component. In this paper we propose a new composite range based component model to analyze both long term and short term volatility components in a daily price range data. The proposed CCARR models long term volatility changes as a stochastic component which itself exhibits conditional volatility and the application of the proposed model is illustrated by using S&P500 and FTSE 100 stock indices.