Forecasting Financial Risk Using Quantile Random Forests

38 Pages Posted: 17 Jan 2023 Last revised: 10 Jul 2023

See all articles by Robert James

Robert James

The University of Sydney

Jessica Wai Yin Leung

Monash University - Department of Econometrics and Business Statistics

Date Written: July 10, 2023

Abstract

This paper designs a financial risk forecasting model that can successfully exploit information from a large set of economic and financial predictor variables. The model is built using Generalized Quantile Random Forests, a non-parametric machine learning method that naturally permits variable interactions and non-linear relationships. We use a model-free variable screening technique and a robust cross-validation approach to minimize the risk of over-fitting. Our risk model produces competitive Value-at-Risk and Expected Shortfall forecasts at both a one-day ahead and a 10-day ahead horizon. A dynamic portfolio insurance strategy that uses the VaR and ES forecasts from our risk model generates attractive Sharpe, Sortino, and Omega ratios, particularly at the 10-day forecast horizon. We further provide a detailed analysis of the dynamic importance of our predictor variables. Our findings demonstrate the utility of effectively combining large datasets with tree-based algorithms for financial risk forecasting.

Keywords: Machine Learning, Risk Management, Value-at-Risk, Expected Shortfall

JEL Classification: C58

Suggested Citation

James, Robert and Leung, Wai Yin, Forecasting Financial Risk Using Quantile Random Forests (July 10, 2023). Available at SSRN: https://ssrn.com/abstract=4324603 or http://dx.doi.org/10.2139/ssrn.4324603

Robert James (Contact Author)

The University of Sydney ( email )

University of Sydney
Sydney, NSW 2006
Australia

Wai Yin Leung

Monash University - Department of Econometrics and Business Statistics ( email )

900 Dandenong Road
Caulfield East, 3145
Australia

Do you have a job opening that you would like to promote on SSRN?

Paper statistics

Downloads
264
Abstract Views
731
Rank
219,329
PlumX Metrics