Modeling and Forecasting the CBOE VIX With the TVP-HAR Model

Abstract

This study proposes the use of a heterogeneous autoregressive model with time-varying parameters (TVP-HAR) to model and forecast the Chicago Board Options Exchange (CBOE) volatility index (VIX). To demonstrate the superiority of the TVP-HAR model, we consider six variations of the model with different bandwidths and smoothing variables and include the constant-coefficient HAR model as a benchmark for comparison. We show that the TVP-HAR models could beat the HAR model with constant coefficients in modeling and forecasting VIX. Among the TVP-HAR models, the rule-of-thumb bandwidth would be better than the cross-validation bandwidth. Meanwhile, VIX futures-driven coefficients could also provide more accurate predictions and smaller capital losses than the other two variables. Overall, the VIX futures-driven coefficients TVP-HAR model with the rule-of-thumb bandwidth obtains the optimal result for investors in forecasting the market risks and shaping their hedging strategies.