Abstract
The value of overseas information within a volatility forecasting setting is examined. The article innovates by considering the forecasting performance of an augmented version of a popular realized volatility model in which this information is incorporated. An application based on realized volatility data from 13 international stock markets demonstrates that this volatility model delivers improved out-of-sample forecasts.
Notes
1 Jung and Maderitsch (2014) use a similar model to investigate structural breaks in international stock index volatility. They do not impose the cross-sectional dependence restriction as they consider only three indices.
2 We also present a summary of results pertaining to all realized volatility measures.
3 The log component reflects the fact that we model logged realized volatility.
4 Note that Asian markets do not close before the opening of any other market within the same trading day. Consequently, they cannot experience the intraday effect.