The profitability of trading on large Lévy jumps

• Published date: 01 June 2021
• Author: Kam Fong Chan,Phil Gray,Zheyao Pan
• Date: 01 June 2021
• DOI: http://doi.org/10.1111/irfi.12279

LETTERS SECTION

The profitability of trading on large Lévy jumps

Kam Fong Chan

1

| Phil Gray

2

| Zheyao Pan

3*

1

UWA Business School, The University of

Western Australia, Perth, Western Australia,

Australia

2

Department of Banking and Finance, Monash

Business School, Monash University, Clayton,

Victoria, Australia

3

Department of Applied Finance, Macquarie

Business School, Macquarie University,

Sydney, New South Wales, Australia

Correspondence

Zheyao Pan, Department of Applied Finance,

Macquarie Business School, Macquarie

University, Room 737, Building 4ER,

Macquarie Park, Sydney, New South Wales,

Australia.

Email: terry.pan@mq.edu.au

Abstract

While past research has studied the profitability of trading

based on jump signals, the notion of differentiating between

jumps according to their magnitude has received relatively little

attention. We utilize the approach of Lee and Hannig (2010) to

identify Lévy jumps and classify them as small and large. The

empirical analysis shows that the arrival of large Lévy jumps

provides a strong trading signal in five major equity markets. In

contrast, the signal from small Lévy jumps is negligible.

KEYWORDS

high-frequency trading, jump trading strategy, Lévy jumps

JEL CLASSIFICATION

C58; E44; G11; G15

1|INTRODUCTION

The importance of accommodating discontinuous price variations in models of asset-price dynamics has been recog-

nized since Merton (1976), who proposed a model which augments continuous Brownian noise with a Poisson jump

process. A sizeable empirical literature has subsequently documented the existence of Poisson jumps in a variety of

asset classes.

1

Accordingly, Poisson jumps have been incorporated into option-pricing models (Bakshi, Cao, & Chen,

1997; Ball & Torous, 1985), recognized in asset allocation and portfolio construction (Das & Uppal, 2004; Zhou,

Wu, & Wang, 2019), and influenced risk-management practices (Das & Sundaram, 1999; Duffie & Pan, 1997).

More recently, the financial econometrics literature demonstrates that jumps can take different forms distin-

guished by their magnitude (Aït-Sahalia, 2004; Aït-Sahalia & Jacod, 2012; Carr & Wu, 2004; Rachev, Kim, Bianchi, &

Fabozzi, 2011). In particular, discontinuous price variations can be partitioned into: (a) infrequent large jumps that

reflect severe perturbations of asset price and (b) relatively frequent small jumps of a magnitude that, while notably

less than a large jump, is too big to be attributable to Brownian noise. Consequently, the use of a general Lévy

*We gratefully acknowledge the constructive comments and suggestions of an anonymous reviewer. The paper has also benefited from thecomments of

Christopher Neely, Stafford Johnson, Zhuo Zhong and participants at the 2017 Auckland Finance Meeting. We thank Jan Hannig, Suzanne Lee and Hanlin

Shang for helpful discussions on the implementation of the Lee-Hannig test.

Received: 11 February 2019 Revised: 25 June 2019 Accepted: 28 June 2019

DOI: 10.1111/irfi.12279

© 2019 International Review of Finance Ltd. 2019

International Review of Finance. 2021;21:627–635. wileyonlinelibrary.com/journal/irfi 627