Bitcoin intraday time series momentum

• Published date: 01 May 2022
• Author: Dehua Shen,Andrew Urquhart,Pengfei Wang
• Date: 01 May 2022
• DOI: http://doi.org/10.1111/fire.12290

DOI: 10.1111/fire.12290

ORIGINAL ARTICLE

Bitcoin intraday time series momentum

Dehua Shen1Andrew Urquhart2Pengfei Wang1

1College of Management and Economics,

Tianjin University, Tianjin, China

2ICMA Centre, Henley Business School,

University of Reading, Reading, UK

Correspondence

PengfeiWang, College of Management and

Economics,Tianjin University, Tianjin, China.

Email:pengfeiwang@tju.edu.cn

Wethank Michael Goldstein (the Editor),

SriniKrishnamurthy (the Editor) and two

anonymousreferees for their invaluable and

insightfulsuggestions that improved this paper

substantially.We would also like to thank

participantsat the 2019 Financial Inclusion

andFintech Conference, SOAS, University

ofLondon for their constructive comments.

Thiswork is supported by the National Natural

ScienceFoundation of China (71790594 and

72071141).

Abstract

This study examines intraday time series momentum in Bit-

coin. Unlike stock markets, Bitcoin trades 24 h a day and

therefore has not got a clear opening and closing period.

Therefore, we use trading volume as a proxy for the mar-

ket trading time and show that the first half-hour positively

predicts the last half-hour return. We find that the first trad-

ing sessions with the highest volume or volatility are associ-

ated with the greatest predictability for intraday time series

momentum. We also show that intraday momentum-based

trading yields substantial economic gains in terms of market

timing and asset allocation, especially in periods of a mar-

ket downturn in Bitcoin. Consistent with the finding in for-

eign exchangemarkets, our results also show that the Bitcoin

intraday momentum is driven by liquidity provision rather

than late-informed trading.

KEYWORDS

bitcoin, cryptocurrencies, intradaypredictability, liquidity provision,

time series momentum

JEL CLASSIFICATION

G12, G13, G15

1INTRODUCTION

Momentum is a well-known effect in financial markets and is broadly the idea that assets that haveperformed well in

the past, continue to perform well in the future. The seminal work in this area, by Jegadeesh and Titman (1993), shows

that winners (losers) over the past 6 months to a year tend to continue to be winners (losers) over the next6 months

to a year.This is known as cross-sectional momentum which has been confirmed by numerous studies in global stock

markets (Barroso & Santa-Clara, 2015; Daniel & Moskowitz,2016; Griffin et al., 2003; Rouwenhorst, 1998), currency

markets (Burnside et al., 2011; Kroenckeet al., 2014; Menkhoff et al., 2012; Raza et al., 2014) and commodity futures

(Fuertes et al., 2010; Miffre & Rallis, 2007; Narayanet al., 2015; Q. Shen et al., 2007).

Financial Review. 2022;57:319–344. wileyonlinelibrary.com/journal/fire ©2021 The Eastern Finance Association 319

320 SHEN ET AL.

Related to cross-sectional momentum is time series momentum, first proposed by Moskowitz et al. (2012), who

show that the previous 12-month return of an asset positively predicts futures returns. This finding has been strongly

supported in the literature by Asness et al. (2013), Georgopoulou and Wang(2017), and Lim et al. (2018). Conversely,

Huang et al. (2020) question the strength of time series momentum and that its investment performance is weak,

especially for a large cross section of assets. Recently,however, intraday time series momentum has been proposed by

Gao et al. (2018) who show that the first half-hour return on the S&P 500 ETF predicts the last half-hour return. They

show that this effect is also present in other US ETFs, robust to transaction costs and other time frames. Consistent

with this finding, Elaut et al. (2018) find that the first half-hour return can be used to predict the last half-hour return in

the RUB-USDFX market during the financial crisis. Recently, Baltussen et al. (2021) provide strong evidence of market

intraday momentum everywhere by using intraday returns on over60 futures on equities, bonds, commodities, and

currencies coveringmore than 40 years. Z. Li, Sakkas, et al. (2021) also document that intraday time series momentum

is economically sizable and statistically significant in 16 developed markets. Jin et al. (2020) and Zhang et al. (2020)

show significant evidence of time series intraday momentum in Chinese commodity markets. Therefore, there is a

modest but growing literature of intradaytime series momentum in financial markets.

Growing stunningly in recent years, the market capitalization (539.05 billion US dollars) of Bitcoin at the end of

2020 is still 57.32 times as large as it was at the beginning of 2014 (9.40 billion US dollars). Given the high media

attention and dramatic price swings, there has been an explosion of studies examining cryptocurrencies, with spe-

cial attention devoted to Bitcoin.1Some characteristics make Bitcoin different from traditional assets, for example,

unclear intrinsic value, frequent “pump-and-dump” schemes and low barriers to entry. The abovefeatures could lead

to intensive trading in the Bitcoin markets.Advanced cryptocurrency exchange’s API also accelerates high-frequency

trading booming, raisingissues on intraday patterns of Bitcoin markets.

The existing literature has reported the existence of bubbles (Cheah& Fry, 2015; Corbet, Lucey, et al., 2018), the

inefficiency of Bitcoin (Nadarajah & Chu, 2017; Tiwari et al., 2018; Urquhart, 2016), the hedging and diversification

benefits of Bitcoin (Borri, 2019; Corbet, Meegan, et al., 2018; Guesmi et al., 2019; Urquhart & Zhang, 2019), the

volatility dynamics of Bitcoin (Katsiampa, 2017, 2019; Katsiampa et al., 2019; Klein et al., 2018), the attention from

media outlets (Philippas et al., 2019) and the manipulations and illegal activity of Bitcoin (Foley et al., 2019; Gandal

et al., 2018). There is also growing literature on the potential benefits of actively trading Bitcoin. Forinstance, Brière

et al. (2015) show that the inclusion of Bitcoin dramatically improves the risk-adjusted returns of portfolios, while

Kajtazi and Moro (2019) examine the role of Bitcoin in portfolios of US, European and Chinese assets and support

previous findings in showing that Bitcoin improves portfolio performance by increasing returns, and not be reducing

risk. Recently, Platanakis and Urquhart (2020) conduct a comprehensive study on the benefits of Bitcoin in a well-

diversified portfolio and show that even during periods of turmoil, the inclusion of Bitcoin substantially improvesthe

risk-adjusted returns. Regarding other forms of trading, Hudson and Urquhart (2021) show that employing a wide-

range of technical trading rules to a broad range of cryptocurrencies generates significant returns to investorswhile

Corbet,Eraslan,etal.(2019) show that simple moving average rules and variable-length moving averagerules gen-

erate significant returns using high-frequency Bitcoin returns, which is supported by Grobys et al. (2020).2However,

Grobys and Sapkota (2019) show no evidence of significant cross-sectional momentum profits in a set of 143 cryp-

tocurrencies. Y.Li, Urquhart (2021), Liu and Tsyvinski (2021), Liu et al. (2021), and Zhang et al. (2021) document fac-

tors that affect the cross-section returns, including size, momentum, extreme returns and etc. Although the above

papers explore potential profitability of Bitcoin trading,none of them examine the time series momentum using high-

frequency data. Given the Bitcoin’s importance, characteristics and literature gap there is a need to document the

intraday momentum in the Bitcoin markets.

1Bitcoinhas been adopted widely. For example, CFAadds topics related to Bitcoin and blockchain, which can be used as test materials; Some company accept

Bitcoinas a payment method; The top 50 universities offer Bitcoin courses.

2Fora recent review of the literature regarding cryptocurrencies, see Corbet, Lucey, et al. (2019).