Density Forecasts and the Leverage Effect: Some Evidence from Observation and Parameter-Driven Volatility Models

TL;DR

This paper investigates how incorporating the leverage effect into volatility models improves the accuracy of density forecasts for equity returns, comparing different model specifications using extensive financial data.

Contribution

It provides empirical evidence that models with the leverage effect outperform those without in generating accurate density forecasts, highlighting the importance of modeling choices.

Findings

Models with leverage effect produce more accurate density forecasts.

The way leverage effect and volatility are modeled significantly impacts forecast accuracy.

Inclusion of leverage effect improves forecast performance across various financial time-series.

Abstract

The leverage effect refers to the well-established relationship between returns and volatility. When returns fall, volatility increases. We examine the role of the leverage effect with regards to generating density forecasts of equity returns using well-known observation and parameter-driven volatility models. These models differ in their assumptions regarding: The parametric specification, the evolution of the conditional volatility process and how the leverage effect is accounted for. The ability of a model to generate accurate density forecasts when the leverage effect is incorporated or not as well as a comparison between different model-types is carried out using a large number of financial time-series. We find that, models with the leverage effect generally generate more accurate density forecasts compared to their no-leverage counterparts. Moreover, we also find that our choice with regards to how to model the leverage effect and the conditional log-volatility process is important in generating accurate density forecasts