Two are better than one: volatility forecasting using multiplicative component GARCH-MIDAS models
A preliminary version of this paper circulated under the title “On the Statistical Properties of Multiplicative GARCH Models” (2016)
58 Pages Posted: 21 Mar 2016 Last revised: 3 Apr 2019
Date Written: March 30, 2019
We examine the properties and forecast performance of multiplicative volatility models that can be decomposed into a short- and a long-term component. Our leading example for such a model is the GARCH-MIDAS of Engle et al. (2013). We derive certain properties of multiplicative volatility models such as the kurtosis of returns, the autocorrelation function of squared returns, the R^2 of a Mincer-Zarnowitz regression and evaluate these models in a Monte-Carlo simulation. Most importantly, we compare the forecast performance of GARCH-MIDAS models with a wide range of competitor models such as HAR, Realized GARCH, HEAVY and Markov-Switching GARCH. For S&P 500 data, our results show that the GARCH-MIDAS based on housing starts as explanatory variable significantly outperforms all competitor models at forecast horizons of two- and three-months-ahead.
Keywords: Forecast evaluation, GARCH-MIDAS, Mincer-Zarnowitz regression, volatility persistence, volatility component model, long-term volatility, model confidence set.
JEL Classification: C53, C58, G12
Suggested Citation: Suggested Citation