Financial Econometrics: Volatility Models

Financial Econometrics- Volatility Models
Name
Course
Dr. Dror Y. Kenett
Johns Hopkins University
Date
Which is the best volatility model?
The best volatility model is one that can forecast volatility. Volatility reflects the degree to which asset return (price) moves and fluctuates. However, to determine this, there is a need to have sufficient information on asset returns. Sometimes different models are best at predicting the variability of different types of assets. The models include the ARCH model, GARCH, integrated GARCH, the exponential GARCH (EGARCH), and GARCH –M, threshold GARCH, and the stochastic volatility model. The ARCH/ GARCH models take into account the prediction information about the previous variance. The GARCH models are more practical in financial modeling, and while stochastic volatility methods are also beneficial, the GARCH modeling is the best.
GARCH modeling represents the most time series as it has infinite memory, and the process also captures the autocorrelation of the adjusted square returns. The GARCH model is better than the autoregressive conditional heteroskedasticity (ARCH) model. The latter is most useful when many parameters describe the volatility, even as the ARCH model is suitable for modeling the volatility of a time-series (Kenett). However, the Generalized AutoRegressive Conditional Heteroskedasticity (GARCH) fails to reflect the nature of volatility of some financial assets, as these models do not consider the asymmetric behavior, characteristic of the volatility of the asset return of a series. In the GARCH models, volatility behaves differently depending on positive and negative changes of the asset returns, is closely linked to the sq...