Forecasting Realized Volatility : Does Anything Beat Linear Models?

We evaluate the performance of several linear and nonlinear machine learning (ML) models in forecasting the realized volatility (RV) of ten global stock market indices in the period from January 2000 to December 2021. We train models using a dataset that includes past values of the RV and additional predictors (composed by lagged returns and macroeconomic variables), and compare them to widely used heterogeneous autoregressive (HAR) models. Our main conclusions are that (i) the additional predictors improve the out-of-sample forecasts of 1-day-ahead RV; (ii) we find no evidence that nonlinear ML models can statistically outperform linear models; and (iii) all ML models show a tendency to underestimate the RV in high-volatility periods, and overestimate the RV in low-volatility periods