Forecast Evaluation in Large Cross-Sections of Realized Volatility

TL;DR

This paper evaluates the predictive accuracy of realized volatility forecasts in large cross-sections, comparing standard and augmented HAR models with LASSO, considering model sensitivity and cross-sectional dependence.

Contribution

It introduces a framework for forecast evaluation using equal predictive accuracy tests in large cross-sections, incorporating measurement error correction and jump components.

Findings

Augmented HAR models with LASSO improve forecast accuracy.

Measurement error correction affects forecast sensitivity.

Forecast evaluation methods are effective in large cross-sectional data.

Abstract

In this paper, we consider the forecast evaluation of realized volatility measures under cross-section dependence using equal predictive accuracy testing procedures. We evaluate the predictive accuracy of the model based on the augmented cross-section when forecasting Realized Volatility. Under the null hypothesis of equal predictive accuracy the benchmark model employed is a standard HAR model while under the alternative of non-equal predictive accuracy the forecast model is an augmented HAR model estimated via the LASSO shrinkage. We study the sensitivity of forecasts to the model specification by incorporating a measurement error correction as well as cross-sectional jump component measures. The out-of-sample forecast evaluation of the models is assessed with numerical implementations.