A0771
Title: Modeling and forecasting range-based volatility using asymmetric CARR model
Authors: YouBeng Koh - Universiti Malaya (Malaysia)
Lei Chen - Universiti Malaya (Malaysia) [presenting]
Abstract: The aim is to propose an asymmetric conditional autoregressive range (AsyCARR) model for forecasting the dynamic volatility of returns. The main feature of the model is its ability to capture the asymmetric effects of shocks on the conditional expectation of Parkinson (PK) volatility series. The PK volatility series is fitted to the proposed AsyCARR and CARR models, and their parameters are estimated using the maximum likelihood (ML) method. A simulation study is conducted to evaluate the effectiveness of the ML method under different sample sizes and error distributions. Empirical examples are based on five stocks from January 3, 2000 to November 9, 2023. The in-sample model fitting performance of the AsyCARR and CARR models is evaluated using the log-likelihood, Akaike information criterion, and Bayesian information criterion, while the out-of-sample forecasting performance is assessed using the heteroskedasticity mean squared error, heteroskedasticity mean absolute error and quasi-likelihood. Results show that the AsyCARR model with the generalized gamma distribution consistently outperforms the CARR model in terms of the in-sample model fit and out-of-sample forecasts. These findings highlight the importance of models by incorporating asymmetric components and flexible distributions in volatility modeling.
