BitMEX bitcoin derivatives: Price discovery, informational efficiency, and hedging effectiveness

• DOI: http://doi.org/10.1002/fut.22050
• Author: Heungju Park,Jaehyuk Choi,Carol Alexander,Sungbin Sohn
• Date: 01 January 2020
• Published date: 01 January 2020

J Futures Markets. 2020;40:23–43. wileyonlinelibrary.com/journal/fut © 2019 Wiley Periodicals, Inc.

|

23

Received: 31 July 2019

|

Accepted: 4 August 2019

DOI: 10.1002/fut.22050

RESEARCH ARTICLE

BitMEX bitcoin derivatives: Price discovery, informational

efficiency, and hedging effectiveness

Carol Alexander

1,2

|

Jaehyuk Choi

3

|

Heungju Park

4

|

Sungbin Sohn

3

1

University of Sussex Business School,

Falmer, Brighton, United Kingdom

2

Peking University HSBC Business

School, Boars Hill, Oxford, United

Kingdom

3

Peking University HSBC Business

School, Nanshan, Shenzhen, China

4

Sungkyunkwan University SKK Business

School, Jongno‐gu, Seoul, Korea

Correspondence

Sungbin Sohn, Peking University HSBC

Business School, University Town,

Nanshan District, Shenzhen, 518055,

China.

Email: sungbin.sohn@phbs.pku.edu.cn

Abstract

BitMEX is the largest unregulated bitcoin derivatives exchange, listing contracts

suitable for leverage trading and hedging. Using minute‐by‐minute data, we

examine its price discovery and hedging effectiveness. We find that BitMEX

derivatives lead prices on major bitcoin spot exchanges. Bid–ask spreads,

interexchange spreads, and relative trading volumes are important determi-

nants of price discovery. Further analysis shows that BitMEX derivatives have

positive net spillover effects, are informationally more efficient than bitcoin spot

prices, and serve as effective hedges against spot price volatility. Our evidence

suggests that regulators prioritize the investigation of the legitimacy of BitMEX

and its contracts.

KEYWORDS

bitcoin, BitMEX, market efficiency, price discovery, spillover

JEL CLASSIFICATION

G13; G14

1

|

INTRODUCTION

Bitcoin was the first digital currency to operate through a peer‐to‐peer network without a central authority (Nakamoto,

2008). Since inception over 10 years ago it has become actively traded through both decentralized and centralized exchanges.

Decentralized exchanges are, in effect, smart contract‐based limit order books run without central authority, where the

change of ownership is made directly between the counterparties’wallets and is recorded on a publicly available blockchain.

Although they may be an intriguing research subject in the future, decentralized exchanges are still in their infancy. In this

paper, therefore, we only investigate the microstructure of bitcoin on centralized exchanges (CEXs). Trading activity is still

highly anonymous and difficult to trace on a CEX because only cryptoasset transfers to and from the exchange’shotwallet

are recorded on the blockchain, unless the exchange is regulated otherwise. A very large number of CEXs have started

operating during the last few years and several of them trade bitcoin derivatives contracts, such as swaps, futures, and

options. Since they provide an application programming interface (API) in the hope of attracting automated high‐frequency

traders, trade‐level data may be freely collected using this API. Hence there is a plethora of data on CEX markets and

academic research on the microstructure of bitcoin markets has become prolific.

Price discovery is one of the most important research questions. As Baur and Dimpfl (2019) point out, two properties of

bitcoin—the absence of a commonly accepted valuation model and the trading of the same asset in numerous venues—not

only make this question particularly challenging but also make empirical evidence more important. The launch of bitcoin

futures on the Chicago Mercantile Exchange (CME) and the Chicago Board Options Exchange (CBOE) in December 2017

triggered much academic discussion on the role of futures in bitcoinpricediscovery.Hale,Krishnamurthy,Kudlyak,and

Shultz (2018) argue that the bitcoin price collapse after these futures are launched is not a coincidence—similar patterns

have been observed in other asset classes. This argument implies that a leader role for those futures contracts in bitcoin spot

price discovery. However, using the accumulated price data from both futures exchanges, Corbet, Lucey, Peat, and Vigne

(2018) and Baur and Dimpfl (2019) independently report that the futures markets neither exercise a price leadership nor

serveasaneffectivehedgeagainstspotmarket,possiblyduetolowtradingvolumeinfuturescontractscomparedwiththe

spot. Yet, this is contrary to general findings in mature futures markets of other asset classes, where futures markets play a

dominant role in price discovery (Bohl, Salm, & Schuppli, 2011; Cabrera, Wang, & Yang, 2009; Hauptfleisch, Putniņš,&

Lucey, 2016; Rosenberg & Traub, 2009). Moreover, using much higher frequency data and a longer time span than previous

studies, a recent paper by Alexander and Heck (2019) has demonstrated a relatively high role of both CBOE and CME

futures. In fact, after the CBOE contract was withdrawn, the CME futures started to play a particularly strong price discovery

role, particularly on expiry dates or around price jumps.

Market efficiency is another active area of research. Some conclude that, although not efficient at first, the bitcoin

market is now highly efficient (Bariviera, 2017; Sensoy, 2019; Tiwari, Jana, Das, & Roubaud, 2018; Urquhart, 2016;

Vidal‐Tomás & Ibañez, 2018). However, this conclusion is not unanimous as the methods, data source, and observation

periods vary considerably. For example, a recent study that uses different efficiency indices argues that the bitcoin

market was inefficient even until 2018 (Jiang, Nie, & Ruan, 2018). For a comprehensive review of the bitcoin efficiency

literature to date, see Bundi and Wildi (2019). Note that some research on efficiency has been extended to

cryptocurrencies beyond bitcoin (Brauneis & Mestel, 2018; Wei, 2018).

We make a novel contribution to both strands of the bitcoin literature by examining the bitcoin derivatives traded on

BitMEX.

1

This is one of the largest bitcoin exchanges by trading volume, as of April 2019, and counting all bitcoin

products its volume is an order of magnitude above CME, CBOE, and major spot exchanges. Therefore, BitMEX data

should provide a comprehensive view about bitcoin market microstructure and, in particular, a study on the price

leadership between spot and derivative trades. BitMEX launched their derivatives contracts before the CME and CBOE

but they have many different features, which aim to attract small but crypto‐focused traders than mainstream financial

institutions: 1 USD contract size, no regulation, bitcoin‐based contract design, minimal margins, and much lower

trading costs. Following Admati and Pfleiderer (1988) we know that all these features should attract the participation of

informed traders. While the exchange had been known among cryptocurrency insiders since its launch in 2014, the

importance of BitMEX has only recently attracted the attention of regulators, who are concerned about the potential for

market manipulation.

2

Our paper aims to inform both traders and regulators about the microstructure of BitMEX, and

its crucial role in the cryptoasset ecosystem.

To learn more about bitcoin market microstructure we (trivially) confirm the cointegration between prices on the

BitMEX perpetual swap, a synthetic spot similar to futures, and three major spot exchanges (Bitstamp, Coinbase, and

Kraken) as a prerequisite for estimating a four‐dimensional vector error‐correction model (VECM), which provides the

foundation of our price discovery measures. We examine the extent to which each exchange contributes to the common

efficient price, measured by the modified information shares (MIS) of Lien and Shrestha (2009) and the component

shares (CS) of Gonzalo and Granger (1995). We also measure the extent to which innovations in one exchange are

transmitted to the others, using the gross and net spillover effects of Pesaran and Shin (1998) and Diebold and Yilmaz

(2012). Moreover, we investigate the extent to which the current return is associated with the past information as

measured by the returns autocorrelation (AC) and variance ratio (VR) (Comerton‐forde & Putniņš, 2015).

We find that, among the four markets studied, the BitMEX perpetual swap takes almost half of both the MIS and the

CS. This very dominant price discovery role is stable and robust throughout the sample period. This finding indicates

that the speed with which new information is incorporated into the bitcoin price is more rapid in BitMEX than in these

other exchanges; that is, the BitMEX perpetual swap plays the price leadership role. In addition, we document that the

strength of price discovery in BitMEX is positively (negatively) associated with the relative bid–ask spread (trading

volume) of the spot markets, consistent with findings in equity derivatives markets (Chakravarty, Gulen, & Mayhew,

2004; Chen & Chung, 2012). We also find that the magnitude of price spreads between exchanges also affects the role of

BitMEX in price discovery. Moreover, the effect of interexchange spreads depends on whether the market is bull or bear

mode. Further analysis shows that BitMEX has significantly positive net spillover effects, meaning that innovations in

BitMEX have a disproportionately larger influence on the other three markets.

These strong price discovery and spillover effects from BitMEX derivatives suggest higher informational efficiency

and hedge effectiveness. We indeed find much supporting evidence for this. When measured by AC and VR, BitMEX is

1

https://www.bitmex.com.

2

See Twomey and Mann (2019) and https://www.bloomberg.com/news/articles/2019‐07‐19/u‐s‐regulator‐probing‐crypto‐exchange‐bitmex‐over‐client‐trades.

24

|

ALEXANDER ET AL.