Deep Hedging of Derivatives Using Reinforcement Learning

21 Pages Posted: 27 Jan 2020 Last revised: 21 May 2020

See all articles by Jay Cao

Jay Cao

University of Toronto

Jacky Chen

University of Toronto

John C. Hull

University of Toronto - Rotman School of Management

Zissis Poulos

University of Toronto - Rotman School of Management

Date Written: December 20, 2019

Abstract

This paper investigates how reinforcement learning can be used to derive optimal hedging strategies for derivatives. We assume that the objective is to minimize a function equal to the mean hedging cost plus a constant times the standard deviation of the hedging cost. The paper illustrates the approach by showing the difference between using delta hedging and optimal hedging for a short position in a call option when there are transaction costs. Two situations are considered. In the first, the asset price follows a geometric Brownian motion. In the second, the asset price follows a stochastic volatility process. The paper extends the standard reinforcement learning approach by using two different Q-functions.

Keywords: Deep hedging, derivatives, reinforcement learning

JEL Classification: C45, G10

Suggested Citation

Cao, Jay and Chen, Jacky and Hull, John C. and Poulos, Zissis, Deep Hedging of Derivatives Using Reinforcement Learning (December 20, 2019). Available at SSRN: https://ssrn.com/abstract=3514586 or http://dx.doi.org/10.2139/ssrn.3514586

Jay Cao

University of Toronto ( email )

Toronto, Ontario M5S 3G8
Canada

Jacky Chen

University of Toronto ( email )

Toronto, Ontario M5S 3G8
Canada

John C. Hull (Contact Author)

University of Toronto - Rotman School of Management ( email )

105 St. George Street
Toronto, Ontario M5S 3E6 M5S1S4
Canada
(416) 978-8615 (Phone)
416-971-3048 (Fax)

Zissis Poulos

University of Toronto - Rotman School of Management ( email )

105 St. George Street
Toronto, Ontario M5S 3E6 M5S1S4
Canada

Here is the Coronavirus
related research on SSRN

Paper statistics

Downloads
71
Abstract Views
355
rank
348,379
PlumX Metrics