A Forest Full of Risk Forecasts for Managing Volatility

Research output: Working paperPreprintAcademic

Abstract

We propose a hybrid approach to modeling stock return volatility in a large panel of stocks that combines the machine learning principle random forest with ordinary least squares regression models. Our model's time-varying parameters are assumed to be data-driven functions of idiosyncratic stock information and changing market conditions. The empirical analysis demonstrates our model’s superior risk forecasting performance across multiple forecast horizons and 186 S&P 500 constituents, resulting in significantly higher utility for volatility-managed investments. Furthermore, this superior forecast performance is observed uniformly across firm characteristics.
Original languageEnglish
DOIs
Publication statusPublished - 13 Jul 2022

Bibliographical note

JEL Classification: C32, C53, C55, C58, G17

Erasmus Sectorplan

  • Sectorplan SSH-Breed

Fingerprint

Dive into the research topics of 'A Forest Full of Risk Forecasts for Managing Volatility'. Together they form a unique fingerprint.

Cite this