SSRN-id3088472.pdf

(1940 KB) Pobierz
Bitcoin Time-of-Day, Day-of-Week and Month-of-Year
Effects in Returns and Trading Volume
Dirk G. Baur
∗1
, Daniel Cahill
1
, Keith Godfrey
1
, , Zhangxin (Frank) Liu
12
1
The
University of Western Australia
December 20, 2017
Abstract
There is a large literature that reports time-specific anomalies in equity markets such as the
Monday effect, the January effect and the Halloween effect. This study is the first to report
intra-day time-of-day, day-of-week, and month-of-year effects for Bitcoin returns and trading
volume. Using more than 15 million price and trading volume observations from seven global
Bitcoin exchanges reveal time-varying effects but no consistent or persistent patterns across
the sample period. The results suggest that Bitcoin markets are efficient.
Keywords:
Bitcoin; blockchain; day-of-week effects; January effect; Halloween effect; arbi-
trage; market efficiency
JEL classification:
G12, G14
Address: UWA Business School, The University of Western Australia, 35 Stirling Highway,
Crawley WA 6009, Australia. Email: dirk.baur@uwa.edu.au
∗
Corresponding author.
Introduction
Bitcoin and the blockchain technology was introduced as a peer-to-peer cash system in Nakamoto
(2008). Less than 10 years later, the market capitalization of Bitcoin has erupted, with the price
of Bitcoin doubling several times in 2017 alone and increasing by more than 50,000% in less
than 5 years. Bitcoin’s popularity has endorsed the blockchain technology, leading to the cre-
ation of a plethora of alternative crypto or digital currencies. Given the price increase, volatility,
and the emergence of more than 1,000 alternative cryptocurrencies, it is not surprising that the
underlying blockchain technology has received increased attention by investors, regulators and
academics. However, Bitcoin has also attracted powerful critics.
1
For others optimism resides in
the blockchain technology, which provides the ability to maintain decentralized records for many
business purposes, rather than a role as a currency (Boehme et. al., 2015).
The literature on Bitcoin is relatively new. Yermack (2013), Kristoufek (2013), Böhme, Christin,
Edelman, and Moore (2015), Dwyer (2015), and Selgin (2015) were among the first to study Bit-
coin. Kristoufek (2013) studies the relationship between search queries and the price of Bitcoin
and argues that the Bitcoin market is driven by short-term investors, speculators, and trend fol-
lowers and is not based on fundamentals. Cheah and Fry (2015) support this idea and find that
bubble characteristics exist in Bitcoin and propose that its fundamental price is zero. Although
the sentiment surrounding Bitcoin has been likened to historical events such as the Dutch tulip
bubble in the 1600’s and the dot com bubble in the late 1990’s, there is also evidence that Bitcoin
prices are efficient (e.g. see Bariviera (2017) and Urquhart (2016)). More specifically, Urquhart
(2016) finds that average Bitcoin returns are inefficient from the 1 August 2010 to 31 July 2013
and efficient from the 1 August 2013 to 31 July 2016.
The aim of this paper is to analyze market efficiency by testing for time-dependent anomalies
in returns and trading volume. For example, do Bitcoin investors trade differently when major
stock exchanges are open compared to when they are closed? Do they trade less on weekends
and during the Northern hemisphere summer months (i.e., June - August)? To answer these types
of questions, we analyze time-of-the-day (ToD), day-of-the-week (DoW) and month-of-the-year
(MoY) patterns in Bitcoin returns and trading volume.
1
JPMorgan’s
CEO, Jamie Dimon branded Bitcoin as a “fraud”, which was followed by a 2.7% decrease in price.
1
Given the conflict between efficient pricing and the seemingly irrational demand for Bitcoin, we
look for any consistent patterns in returns that would contradict the efficient market hypothesis
(Malkiel and Fama, 1970). Since Bitcoin is a relatively new and unregulated asset, it is possible
that the market has been dominated by retail investors. This suggests that we can expect to find
inefficiencies and return anomalies in Bitcoin pricing. The fact that Bitcoin is continuously and
globally traded makes an analysis of time-of-day, day-of-week, and month-of-year effects particu-
larly interesting to study. Moreover, since Bitcoin is traded in different currencies and in different
geographic locations, the market provides an additional layer of complexity similar to foreign
exchange and commodities.
The analysis can also help to answer the question whether Bitcoin is a currency or rather a spec-
ulative asset (e.g. see Yermack (2013) and Baur, Hong, and Lee (2015)), if time-dependent price
and trading volume patterns resemble currency rather than equity markets, or vice-versa. Our
analysis of time-of-day, day-of-week and month-of-year patterns shows evidence of time-specific
anomalies such as a lower weekend volume effect and a higher Monday return effect which are
more consistent with currency markets. We find increased trading activity on Bitcoin exchanges
at times when U.S stock exchanges are open and lower trading activity between midnight and the
early morning on most exchanges. Bitcoin exchanges denominated in USD display stronger pat-
terns compared to exchanges denominated in Japanese yen and Chinese yuan. We use heatmaps to
illustrate patterns in returns and trading volume both across time and across exchanges. The results
support the view that Bitcoin markets are weak-form efficient because we do not see any consis-
tent price pattern that could be exploited based on historical information. We also use statistical
tests to check the robustness of the heatmap analysis and to determine the statistical significance
of the effects.
1
Data
We obtained Bitcoin price data in USD, CNY, JPY, and EUR from kaggle.com, which provides
intra-day prices from several exchanges at a 1-minute frequency. We investigate and analyse the
time-of-day, day-of-week and month-of-year effects. Table 1 provides descriptive statistics of the
exchanges, the sample periods for the price and volume data, and the numbers of observations
2
which in total exceed 15 million. The time stamps are converted into universal coordinated time
(UTC).
Insert Table 1 about here
We calculate the log returns and log volume for 1-min, 1-hour, and 1-day intervals for seven
Bitcoin exchanges from 31 December 2010 to 20 October 2017 and present the summary statistics
in Table 2.
Insert Table 2 about here
We report different frequencies to illustrate the volatility in Bitcoin on a more granular level.
2
Empirical Analysis
We use heatmaps to analyze specific time-of-day (ToD), day-of-week (DoW) and month-of-year
(MoY) anomalies in returns and trading volume. The heatmaps illustrate similarities and dif-
ferences in returns and trading volume across time and across exchanges. The heatmaps help to
identify both consistent and transient patterns over time and between exchanges. We then calculate
the differences in means and investigate the statistical significance for robustness.
2.1
Time-of-Day (ToD) Effects
Time-of-day price returns do not show any obvious and persistent pattern across years in our
sample. In contrast, we observe a persistent ToD effect in volume traded across the sample period.
Insert Figure 1 about here
Insert Figure 2 about here
Some exchanges appear to be more active during specific periods of the day, which often coincides
with trading in stock exchanges and periods when potential investors are more likely to be awake.
3
Such periods are 0800-1600 hours at BTCE exchange in US dollars, 1200-2400 hours at Coinbase
in USD and 0800-2400 hours at Kraken. The pattern for the Japanese and Chinese exchanges is
less clear suggesting that trading levels are relatively constant on Japanese and Chinese exchanges
throughout the day.
2.2
Day-of-Week (DoW) Effects
The heatmaps in Figure 3 show no consistent pattern across years and exchanges. Mondays and to
a lesser extent Tuesdays are the only days that show higher than average returns in 2013 and 2017
across exchanges.
Insert Figure 3 about here
Insert Figure 4 about here
In contrast, Figure 4 shows a noticeably higher trading volume on weekdays, reflected in the darker
shades between Monday and Friday for most exchanges except the Chinese and the Japanese
exchanges. The lower weekend trading volume effect is consistent across the sample period.
2.3
Month-of-Year (MoY) Effects
There is no apparent pattern in MoY returns or trading volume across time and Bitcoin exchanges.
There is some evidence for lower trading volume in Northern hemisphere summer months for Bit-
stamp and BTCE consistent with the findings reported in Hong and Yu (2009) for equity markets.
If market participants are on holiday, there is less trading volume. The heatmaps also indicate
increased trading activity in selected years and exchanges in January but the evidence is too in-
consistent to indicate a “January effect”.
Insert Figure 5 about here
Insert Figure 6 about here
4

Plik z chomika:

Inne pliki z tego folderu:

Inne foldery tego chomika:

Zgłoś jeśli naruszono regulamin