Dynamic Integration of Time- and State-domain Methods for Volatility Estimation

Time- and state-domain methods are two common approaches for nonparametric prediction. The former predominantly uses the data in the recent history while the latter mainly relies on historical information. The question of combining these two pieces of valuable information is an interesting challenge in statistics. We surmount this problem via dynamically integrating information from both the time and the state domains. The estimators from both domains are optimally combined based on a data driven weighting strategy, which provides a more efficient estimator of volatility. Asymptotic normality is seperately established for the time damain, the state domain, and the integrated estimators. By comparing the efficiency of the estimators, it is demonstrated that the proposed integrated estimator uniformly dominates the two other estimators. The proposed dynamic integration approach is also applicable to other estimation problems in time series. Extensive simulations are conducted to demonstrate that the newly proposed procedure outperforms some popular ones such as the RiskMetrics and the historical simulation approaches, among others. Empirical studies endorse convincingly our integration method.