Novas transformations: Flexible inference for volatility forecasting

Abstract

In this chapter we present several new findings on the NoVaS transformationapproach for volatility forecasting introduced by Politis (Model-Free VolatilityPrediction, UCSD Department of Economics Discussion Paper 2003–16

Recentadvances and trends in nonparametric statistics, Elsevier, North Holland

J FinancEcon 5:358–389, 2007). In particular: (a) we present a new method for accuratevolatility forecasting using NoVaS

(b) we introduce a “time-varying” version ofNoVaS and show that the NoVaS methodology is applicable in situations where(global) stationarity for returns fails such as the cases of local stationarity and/orstructural breaks and/or model uncertainty

(c) we conduct an extensive simulationstudy on the forecasting ability of the NoVaS approach under a variety of realisticdata generating processes (DGP)

and (d) we illustrate the forecasting ability ofNoVaS on a number of real data sets and compare it to realized and range-basedvolatility measures. Our empirical results show that the NoVaS -based forecasts leadto a much ‘tighter’ distribution of the forecasting performance measure. Perhaps ourmost remarkable finding is the robustness of the NoVaS forecasts in the context of structural breaks and/or other nonstationarities of the underlying data. Also strigis that forecasts based on NoVaS invariably outperform those based on the benchmarkGARCH(1,1) even when the true DGP is GARCH(1,1) when the sample size. © Springer Science+Business Media New York 2013.