Tail-GAN: Learning to Simulate Tail Risk Scenarios

40 Pages Posted: 16 Mar 2022 Last revised: 19 Apr 2024

See all articles by Rama Cont

Rama Cont

University of Oxford

Mihai Cucuringu

University of Oxford - Department of Statistics; The Alan Turing Institute

Renyuan Xu

University of Southern California - Epstein Department of Industrial & Systems Engineering

Chao Zhang

University of Oxford; University of Oxford - Oxford-Man Institute of Quantitative Finance

Date Written: March 3, 2022

Abstract

The estimation of loss distributions for dynamic portfolios requires the simulation of scenarios representing realistic joint dynamics of their components. We propose a novel data-driven approach for simulating realistic, high-dimensional multi-asset scenarios, focusing on accurately representing tail risk for a class of static and dynamic trading strategies. We exploit the joint elicitability property of Value-at-Risk (VaR) and Expected Shortfall (ES) to design a Generative Adversarial Network (GAN) that learns to simulate price scenarios preserving these tail risk features. We demonstrate the performance of our algorithm on synthetic and market data sets through detailed numerical experiments. In contrast to previously proposed data-driven scenario generators, our proposed method correctly captures tail risk for a broad class of trading strategies and demonstrates strong generalization capabilities. In addition, combining our method with principal component analysis of the input data enhances its scalability to large-dimensional multi-asset time series, setting our framework apart from the univariate settings commonly considered in the literature.

Keywords: Scenario simulation, Generative models, Generative adversarial networks (GAN), Time series, Universal approximation, Expected shortfall, Value at risk, Risk measures, Elicitability.

Suggested Citation

Cont, Rama and Cucuringu, Mihai and Xu, Renyuan and Zhang, Chao, Tail-GAN: Learning to Simulate Tail Risk Scenarios (March 3, 2022). Available at SSRN: https://ssrn.com/abstract=3812973 or http://dx.doi.org/10.2139/ssrn.3812973

Rama Cont (Contact Author)

University of Oxford ( email )

Mathematical Institute
Oxford, OX2 6GG
United Kingdom

HOME PAGE: http://www.maths.ox.ac.uk/people/rama.cont

Mihai Cucuringu

University of Oxford - Department of Statistics ( email )

24-29 St Giles
Oxford
United Kingdom

HOME PAGE: http://https://www.stats.ox.ac.uk/~cucuring/

The Alan Turing Institute ( email )

British Library, 96 Euston Road
96 Euston Road
London, NW12DB
United Kingdom

Renyuan Xu

University of Southern California - Epstein Department of Industrial & Systems Engineering ( email )

United States

HOME PAGE: http://renyuanxu.github.io

Chao Zhang

University of Oxford ( email )

University of Oxford - Oxford-Man Institute of Quantitative Finance ( email )

Eagle House
Walton Well Road
Oxford, Oxfordshire OX2 6ED
United Kingdom

Do you have negative results from your research you’d like to share?

Paper statistics

Downloads
1,189
Abstract Views
2,802
Rank
33,216
PlumX Metrics