Are Cryptocurrencies A Future Safe Haven For Inves

Economic Research-Ekonomska Istraživanja
ISSN: (Print) (Online) Journal homepage: https://www.tandfonline.com/loi/rero20
Are cryptocurrencies a future safe haven for

investors? The case of Bitcoin
Audil Rashid, Walid Bakry & Somar Al-Mohamad
To cite this article: Audil Rashid, Walid Bakry & Somar Al-Mohamad (2022): Are cryptocurrencies
a future safe haven for investors? The case of Bitcoin, Economic Research-Ekonomska
Istraživanja, DOI: 10.1080/1331677X.2022.2140443
To link to this article: https://doi.org/10.1080/1331677X.2022.2140443
© 2022 The Author(s). Published by Informa

UK Limited, trading as Taylor & Francis
Group.
Published online: 02 Nov 2022.
Submit your article to this journal
View related articles
View Crossmark data
Full Terms & Conditions of access and use can be found at

https://www.tandfonline.com/action/journalInformation?journalCode=rero20
ECONOMIC RESEARCH-EKONOMSKA ISTRAŽIVANJA
https://doi.org/10.1080/1331677X.2022.2140443
Are cryptocurrencies a future safe haven for investors?

The case of Bitcoin
Audil Rashida , Walid Bakryb and Somar Al-Mohamada
a
College of Business Administration, American University of the Middle East, Kuwait, Kuwait; bSchool
of Business, Western Sydney University, Sydney, Australia
ABSTRACT ARTICLE HISTORY

The paper seeks to determine whether Bitcoin behaves differently Received 30 October 2021
from forex markets and Gold, and whether it offers any diversifi- Accepted 17 October 2022
cation, hedging, or safe-haven potential. A Markov regime-switch-
KEYWORDS
ing regression model is employed to determine the relationship
Cryptocurrency; Bitcoin;
between Bitcoin, the real economic activity, foreign exchange blockchain; digital
markets, financial markets, Energy, and Gold. The results indicate currencies; finan-
that, unlike USD/EUR and Gold, besides other variables, Bitcoin cial innovation
exhibits significant deviations in terms of its association with
other financial and economic variables. Bitcoin appears to be JEL CLASSIFICATION
strikingly positively associated with equity markets in both G1; G11; G15; F31
regimes. This may limit its potential to either act as a hedge or a
safe-haven for US Equity markets. Bitcoin also deviates consider-
ably from Gold and USD/EUR as it is not affected by the same set
of variables as Gold or USD/EUR are under either regime.
Moreover, while Gold appears to offer considerably weak safe-
haven properties, particularly against equity, Bitcoin fails to be a
safe-haven for any of the assets under study. The results, how-
ever, indicate that the properties of Bitcoin may range between a
diversifier and a hedge, however, such potential of Bitcoin must
be viewed with caution owing to the large volatility exhibited
by Bitcoin.
1. Introduction
Nowadays, businesses have come to largely rely upon electronic transactions, almost
entirely replacing check and cash-based transactions. The viability of electronic trans-
actions is entirely based on trust, as these transactions are performed over platforms
where financial institutions serve as trusted third parties. This system, however, is
prone to fraud, disputes, mediation costs, hacking, and reversibility of transactions,
thus raising the cost and risk of using such services.
Recent innovations in blockchain technology have changed the entire milieu of
online transactions by introducing a disruptive mechanism that allows payments to
CONTACT Walid Bakry w.bakry@westernsydney.edu.au

ß 2022 The Author(s). Published by Informa UK Limited, trading as Taylor & Francis Group.
This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/
licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is
properly cited.
2 A. RASHID ET AL.
be sent directly from one party to another without relying on a trusted third party.
This mechanism involves using cryptographic proofing instead of third-party trust to
conduct transactions, which is achieved through the use of cryptocurrencies. This sys-
tem is highly secure as long as the network nodes control more CPU power than a
group of hacking nodes together (Nakamoto, 2008). Digital money has been in exist-
ence since early 1980s, but to scale its operability as a fully distributed system took
more than two decades. Cryptocurrencies, on the other hand, became viral in a short
span of time by making the system practically decentralized, economical, and viable.
After the introduction of Bitcoin in November 2008, it gained prominence in the
financial markets, and investors started comparing it to assets like Gold and other
precious metals, thus creating value in Bitcoin as a first-mover through a network
effect large enough to maintain its dominance over other cryptocurrencies (Gandal &
Halaburda, 2016).
Although the popularity of Bitcoin as a positive disruption in digital currencies is
widely acknowledged, Bitcoin has many critics. Skeptics like ElBahrawy et al. (2017)
argue that Bitcoin is losing its popularity and market share to alternative cryptocur-
rencies (altcoins such as Ethereum, Litecoin, Ripple, Bitcoin Cash, etc.), which arose
as ‘hardforks’ of Bitcoin. Other critics assert that Bitcoin should be outlawed
altogether due to the lack of regulation thus giving it a fraudulent feel. As recently as
2017, the Governor of the Reserve Bank of Australia Philp Lowe stated, ‘when
thought of purely as a payment instrument, it seems more likely to be attractive to
those who want to make transactions in the black or illegal economy, rather than
everyday transactions’. But the most prominent criticism that Bitcoin has received is
that it is prone to fragile price bubbles. Former chief economist of the International
Monetary Fund Kenneth Rogoff (2018), for example, likens Bitcoin to a lottery ticket,
and extremely vulnerable to bubble-like collapse. Klein et al. (2018) also report that
Bitcoin’s price innovations display a large chain of small to medium bubbles as well
as explosive behavior, which demonstrates the highly volatile nature of cryptocurren-
cies. Cheah and Fry (2015) argue that Bitcoin lacks any fundamental value and con-
tains a considerable speculative element that makes it prone to bubble-like collapse.
Baur et al. (2018) tested the hedging capabilities of Bitcoin as compared to foreign
exchange and stock markets throughout different periods in a dynamic framework
and found that Bitcoin should be considered a speculative asset rather than a transac-
tion medium. Recent volatility in the Bitcoin price and the restrictions on Bitcoin
trading and mining (or its use in online transactions) in many countries (especially
China) tend to reinforce this view. There are, however, cryptocurrency enthusiasts
who are on the path to legalizing and formalizing selected cryptocurrencies. CME
Group, for instance, listed Bitcoin futures in December 2017, while NASDAQ and
other exchanges have been working for most of 2018 and 2019 to list Bitcoin futures
on the exchange. In 2021, many recognized brands and platforms started accepting
Bitcoin as a mode of payment; Payment giants—Mastercard, Visa, and Paypal have
opened their platforms for transactions in Cryptocurrencies, Bitcoin in particular1; El
Salvador has legalized the use of Bitcoin as a legal tender in June 20212, encouraging
other crypto-leaning countries to follow the same path. Interestingly, amid the recent
Russia-Ukraine conflict and trade sanctions on Russia, Russia is also considering
ECONOMIC RESEARCH-EKONOMSKA ISTRAŽIVANJA 3
accepting cryptocurrencies, bitcoin in particular, for its oil and gas exports by allow-
ing what it calls ‘friendly’ countries to pay in cryptocurrencies, despite the risks.3
Bitcoin has also emerged as one of the popular modes for crowdfunding and dona-
tions, for instance, the war efforts in Ukraine are being financed through anonymous
bitcoin donations. As of March 3, 2022, almost 10 days after the onset of the Russia
Ukraine War, more than $54 million have been received in donations for Ukraine’s
war efforts.4 Finally, investment firms are also trying to obtain approval from the
Securities Exchange Commission (SEC) to list a Bitcoin ETF on US stock exchanges.
Recent developments and technological innovations in payment systems and plat-
forms have led to the popularization of blockchain technology in which cryptocurren-
cies are built, leading to the entire milieu of online transactions being disrupted.
Cryptocurrencies in general, most notably Bitcoin, became a financial sensation over
a very short period and have generated overwhelmingly large interest from both prac-
titioners and academicians as an alternative mechanism for electronic payment sys-
tems. Amid the increasing popularity and public interest in Bitcoin, mainstream
financial media and financial blogs are abuzz with discussions on the longevity,
future, and sustainability of Bitcoin. The research community has largely focused on
the fundamental, technical, legal, and security aspects of Bitcoin to the neglect of its
possible links to the broader economic and financial determinants. While some argue
that ‘Bitcoin works in practice, but not in theory’, others argue that Bitcoin’s stability
depends on unknown variables in unknown combinations, and thus it is difficult if
not impossible to model precisely, thereby raising doubts about the integrity of
the system.
The present study seeks to determine the extent to which Bitcoin is affected, if at
all, by changes in economic activity, forex markets, financial markets, energy, and
Gold. Moreover, this paper seeks to evaluate whether or not cryptocurrencies can
provide safe-haven, hedging, or diversification benefits. This study is expected to con-
tribute to the existing literature by providing more evidence on the relationship
between Bitcoin and other financial and economic variables. Furthermore, by analyz-
ing the behavior of Bitcoin, this study provides a brief discussion on the diversifica-
tion, hedging, or safe-haven potential of Bitcoin to global investors. The remainder of
the paper is organized as follows. Section 2 provides a comprehensive review of the
literature. Section 3 provides the description of the data. Section 4 describes the
research methodology. Section 5 presents the analysis of the results, and section 6
sums up the conclusions derived from this study.
2. Review of the literature

The demand for cryptocurrencies, bitcoin, in particular, is mainly led by investors
and users with a broad range of objectives, ranging from its utility as a medium of
exchange to a safe-haven. The diversity among the investors in bitcoin has led to a
disagreement among academicians and practitioners about the classification of bitcoin
into a single category. The holding motives for bitcoin can be broadly classified as
transactional, investment, and speculative. While the transactional motives for holding
bitcoin are guided by peoples’ demand for money and a payment system, which in
4 A. RASHID ET AL.
turn is a key function of interest rates, inflation, central bank policy, etc. The specula-
tive demand for cryptocurrencies, on the other hand, is led largely due to herding
behavior, heuristics, clustering illusion, prospect, and contrarianism (Jiang et al.,
2022; Lee et al., 2020). Ma and Tanizaki (2022), for instance, examined the intra-day
price evolution and clustering behavior of bitcoin and discovered the existence of
behavioral price clustering in bitcoin prices, oddly during the active trading hours in
conventional markets. They attribute the anomalous pattern and price clustering in
cryptocurrency exchanges to the domination of retail investors, strategic pricing, and
the availability and influence of relevant active information during the daytime. Luis
et al. (2019) dismiss the demand for bitcoin as a safe-haven commodity or a medium
of exchange while arguing that the short-term demand for bitcoin is mostly specula-
tive and the long-term demand for bitcoin is driven by expectations towards the
medium of exchange. Glaser et al. (2014) suggest that the current interest in Bitcoin
is not due to its currency aspect but due to its asset nature, and that the increasing
transaction volumes are largely due to speculative investors rather than for transac-
tion-related to purchases of goods and services. Schuh and Shy (2016) and Ciaian
et al. (2016) also argue that while transaction demand is a large component of crypto-
currency demand, there is also a substantial speculative demand for cryptocurrency.
Yermack (2014) further argues that Bitcoin fails on many accounts to qualify as a
medium of exchange, and that the volatility in Bitcoin resembles more with that of a
speculative asset than a currency. Kristoufek (2013) finds a two-way relationship
between Bitcoin users’ interest (Google Trends and Wikipedia) and the Bitcoin prices;
they argue that while increases in Bitcoin prices may generate users’ interest, this
interest also seems to generate speculative demand for Bitcoin. Huang (2019) argues
that the perceived risks and value of investing in bitcoin are not understood by the
majority of the investors and that an increased understanding of the perceived risk
and return dynamics in bitcoin will subsequently alter their transaction or speculation
motive for investment in bitcoin. He further argues that the government intervention
is less likely to impact the speculative or transaction demand for bitcoin, would the
investors be aware of the risk and return prospect/perception of bitcoin. Sontakke
and Ghaisas (2017) also support a similar assertion that cryptocurrencies such as
Bitcoin and Ethereum are an evolving asset class, identifying their high-risk prospect
as characteristic of their early stage of evolution (similar to technology-based compa-
nies coming into existence in the late 90 s and early 2000s). Lewis (2017) on the other
hand, suggests that even though the current demand for Bitcoin is mainly driven by
speculation, the parallel demand for Bitcoin as a medium of transaction is also grow-
ing simultaneously. He also suggests that as the speculative demand for Bitcoin
declines, its potential as a transaction tool could further improve. Sahoo’s (2017)
study on the volatility of Bitcoin using a GARCH model attributes the growing inter-
est in Bitcoin to its novel mechanism, and the speculative demand. The study also
concludes that given Bitcoin may not yet be suitable as an alternative payment system
due to its high volatility. However, if its volatility decreases considerably as more
liquidity flows into the asset class, then the demand for Bitcoin might shift more
towards transaction utility, and thus it may rise as a future digital currency. Nguyen
et al. (2018) in a similar study suggest that Bitcoin prices are currently primarily
driven by unconventional factors such as media attention. They argue that Bitcoin
behaves more like a speculative asset particularly due to the manipulation by ‘Bitcoin
whales’, however, the volatility in Bitcoin prices has declined over time due to the
introduction of Bitcoin futures and drastic changes in regulation. Similar results have
been reported by Bouoiyour et al. (2014), who assert that Bitcoin is being considered
more like a speculative asset, rather than a medium of exchange.
Iwamura et al. (2019) explore the potential of Bitcoin to replace or compete with
central bank money, and conclude that even though Bitcoin is extremely efficient,
secure and decentralized, and operates on a ‘proof of work’ system, it doesn’t have
the potential to replace the existing monetary system that runs on central bank
money. They further argue that cryptocurrencies are subject to extreme volatility and
are thus subject to inflationary or deflationary tendencies. Gandal and Halaburda
(2016) argue that a decreasing hash rate limits the production rate of Bitcoins and
puts deflationary pressure on the system. This being so, the belief that Bitcoin can
operate as a currency without inflation doesn’t hold any water in its current legal and
operational standing. Raymaekers (2015) argue that Bitcoin and other cryptocurren-
cies are currently marred by critical regulatory and legal challenges, and thus don’t
seem to offer a potential alternative to the existing payment systems. Furthermore, as
cryptocurrencies are not linked to any monetary policy instruments, the US
Commodity Futures Trading Commission (CFTC) clearly distinguishes cryptocurren-
cies from fiat money by classifying all digital virtual currencies as crypto-assets5 that
fall under commodities. Tiwari (2018) supports the classification of cryptocurrencies
as a commodity, and also provides a legal framework to remove various legal hurdles
that stand in the way of Bitcoin’s operation, but he also recognizes that there is no
general consensus or support for cryptocurrencies replacing the current payment sys-
tem. In another study, Raskin and Yermack (2016) explore the potential lessons for
digital currencies and central banks to learn to digitize currencies. They argue that
the blockchain mechanism has the potential to bring about a positive disruption that
improves the payment and settlements system, and may also help central banks to
issue their own centralized digital currency. The proponents of cryptocurrencies also
argue that a more comprehensive and accommodative operational and legal frame-
work is needed to integrate cryptocurrencies into the formal fold of payment systems
(Harwick, 2016; Yermack, 2014). It is further argued that such integration would not
only improve payment systems but would also bring about confidence and consist-
ency in economic policies (Harwick, 2016; Raymaekers, 2015). Bonneau et al. (2015)
argue that although Bitcoin’s price is driven by a set of unknown economic variables,
it nonetheless provides a mechanism for developing universal payment systems which
are more decentralized, secure, and relatively less costly. They also argue that the rela-
tionship between Bitcoin and the variables affecting it must be more actively and
assiduously examined by the research community to re-assuring this mechanism
for payments.
Besides legal and regulatory challenges, as opposed to fiat currency, Bitcoin is
largely used for speculative purposes rather than for transaction purposes.
Cryptocurrencies in general, and Bitcoin in particular, have seen considerable turbu-
lence during and towards the end of 2018 with Bitcoin breaching the USD 10,000
6 A. RASHID ET AL.
mark, touching USD 19,000, and falling back to USD 3,300. A similar pattern was
once again witnessed in late 2020 when the Bitcoin touched USD 60,000 mark and
then falling down to USD 30,000 in mid-2021. In April 2021, a flash crash in the
cryptocurrency market wiped out about USD 300 Billion in less than 24 hours and
about USD 1 trillion in one week, which is about more than 15% and 50% of the
total cryptocurrency market, respectively6. This exponential volatility remains to be a
cause of substantial concern among cryptocurrency enthusiasts as well as encouraging
skeptics to pitch their dismal predictions and criticisms. Bouoiyour et al. (2014) attri-
bute the turbulent and speculative nature of Bitcoins to the Chinese Stock Market
Index (Kristoufek, 2015) and the increasing hash rate. Yermack (2014), by comparing
the volatility of fiat currencies and Bitcoin, concludes that the current uncertainty
and price volatility of Bitcoin undermines its ability to compete with fiat currency.
Dehouche (2022) examines the volatility of bitcoin by comparing it to Gold and the
S&P 500 and concludes that the volatility of bitcoin shows a considerable diversion in
terms of the concepts of volatility. He further argues that the volatility in bitcoin is
an intra-day phenomenon that gets considerably reduced for longer time-frames
(weekly or monthly), while the opposite is true for Gold. In a recent study by Jiang
et al. (2022), examining the dynamics and spillover among bitcoin and other financial
markets, they conclude that bitcoin transmits volatility to other financial markets,
therefore suggesting the movement of bitcoin towards the center of financial markets.
Although a significant portion of the literature about cryptocurrencies deals with
the operational issues of the blockchain mechanism, considerable interest in the pos-
sible financial aspects has also been generated among researchers. A few studies
addressing the volatility characteristics and other economic and financial aspects of
cryptocurrencies are briefly summarized in this section. A few empirical studies sug-
gest that there is a weak relationship between Bitcoin and conventional financial
assets because each is driven by a different set of causal factors (Bouoiyour et al.,
2014; Kristoufek, 2015). While financial assets are affected by economic and financial
variables, Bitcoin depends on a unique variety of factors like its exotic appeal, illegal
activity, security, and anonymity, among other factors which remain unknown
(Kristoufek, 2015; Yelowitz & Wilson, 2015). Ciaian et al. (2016) found that Bitcoin
prices are mainly attributed to supply and demand generated mostly by investors
departing from rationality (Bouri et al., 2019). The volatility in Bitcoin prices might
be caused by Chinese stock indices (Kristoufek, 2015), Bitcoin volumes (growing
interest in Bitcoin as a speculative asset), the Dow Jones Industrial Average Index,
and hash rate (Bouoiyour and Selmi (2014). Hernandez et al. (2022) argue that the
relationship of bitcoin with economic policy uncertainty (EPU) changes over different
time horizons, they found that the EPU affects the bitcoin negatively in the short-run
and positively in the long run. Mokni et al. (2021) argue that while bitcoin does not
act as a hedge against EPU as an aggregate measure, it provides considerable hedging
potential against fiscal policy, trade policy, and other specific political and economic
uncertainties. In a similar study, Yin et al. (2021) argue that the oil price shocks may
produce an effect similar to macroeconomic events, and therefore, oil prices may be a
key factor/determinant in cryptocurrency uncertainty. Palombizio and Morris (2012)
assess the impact of global macroeconomic indicators on Bitcoin and report an
indirect effect of oil prices on Bitcoin price evolution, whereas Ciaian et al. (2016)
report no impact of oil prices on Bitcoin. Samah et al. (2018), found that Bitcoin is
sensitive to macroeconomic indicators, and thus behaved in a similar fashion to Gold
and the US Dollar. They conclude that since the response of Bitcoin to macroeco-
nomic indicators is more frequent and rapid, it has some hedging to diversifica-
tion properties.
Research also suggests that Bitcoin is a good choice for portfolio diversification
due to its anomalous behavior, its resemblance to Gold, as well as its relationship
with unknown economic and behavioral factors (Bouri et al., 2017a; Dyhrberg, 2016a,
2016b). Pal and Mitra (2019) show that Bitcoin can provide a hedge against volatility
in equity markets, Gold, and commodities. They also indicate that the hedging effect-
iveness of Bitcoin is highest with Gold; a long position in Bitcoin provides a hedge
with a short position in Gold. Dyhrberg (2016a) and Palombizio and Morris (2012)
classify Bitcoin as an asset with price behavior resembling Gold and also bears some
resemblance to the US Dollar. Tschorsch and Scheuermann (2016) argue that while
the investment potential of Bitcoin is huge due to its price behavior resembling that
of Gold, its ability to transform the payment system (and to be one itself) must not
be undermined. Conrad et al. (2018) use a GARCH-MIDAS model to study the fac-
tors affecting Bitcoin volatility and conclude that Bitcoin volatility is positively associ-
ated with real economic variables and negatively related to volatility in the financial
markets. Their results indicate that Bitcoin volatility is less during the ‘flight to safety’
periods and periods of high volatility in the US stock markets (Kristoufek, 2015),
while the volatility has a positive and significant association with real economic activ-
ity. They further argue that although most research compares Bitcoin with precious
metals like Gold (Baur et al., 2018), its behavior is unlike precious metals; precious
metals are counter-cyclical whereas Bitcoin is pro-cyclical. Baur et al. (2018) indicate
that Bitcoin returns are not related to conventional assets like stocks or bonds, and
thus can be a potential asset for portfolio diversification.
During periods of market turbulence, investors seek assets that are uncorrelated or
negatively correlated with conventional financial assets and commodities.
Unfortunately, conventional safe havens like Gold are dissipating fast (Gantori et al.,
2017). Thus, some research has been conducted into the question of whether crypto-
currencies (such as Bitcoin) are capable of serving as either diversifiers, hedges, or
safe-havens. Research suggests that due to the unique risk-return characteristics of
Bitcoin and its uncorrelated nature with other assets, Bitcoin might serve as a safe-
haven against global financial stress, commodities, and energy (Bouri et al., 2017b,
2017c, 2018), a hedge against equities, currencies, commodities, and VIX (Baur et al.,
2018; Bouri et al., 2017a, 2017c; Chan et al., 2019; Dyhrberg, 2016a). Park (2022), for
instance, examines the trading behavior and the impact of market stress on the
dynamics of the relationship among bitcoin futures and trading behavior variables
and concludes that the trading volumes and volatility of bitcoin increased during the
COVID-19 pandemic, while the returns decreased substantially. On the properties of
bitcoin as a safe-haven during turbulent markets, owing to strong positive correla-
tions between Bitcoin and Gold during periods of market turmoil, Jare~ no et al.
(2020) argue that Bitcoin may offer some safe-haven potential. However, Shahzad
8 A. RASHID ET AL.
et al. (2020) assert that there are considerable differences between Bitcoin and Gold
in terms of safe-haven, hedging, and diversification properties, they argue that Gold
is a better safe-haven, hedge, and diversifier compared to Bitcoin. Jiang et al. (2022)
in a similar study discovered that the volatility spillovers in bitcoin are significantly
correlated with other financial markets, and therefore, its potential in acting as a safe-
haven is rather limited. They further argue that bitcoin may provide considerable
potential as a hedge against financial markets and commodities.
At this stage, however, the results are mixed. As defined by Baur and Lucey
(2010), unconventional assets positively correlated to conventional assets can act as
diversifiers. An uncorrelated or a negatively correlated unconventional asset can pro-
vide a hedge for the conventional assets, while an unconventional asset that is uncor-
related or negatively correlated with conventional assets serves as a safe-haven only in
‘flight to safety’ periods. The existing literature supports that Bitcoin can provide
diversification opportunities or act as a short-term hedge (Bouri et al., 2017b; Corbet
et al., 2018; Dyhrberg, 2016a). By applying APARCH and FIAPARCH models to ana-
lyze the leverage effect and long memory, Klein et al. (2018) show that although
Bitcoin has a similar asymmetric response as observed for Gold and silver, the per-
sistence in Bitcoin is fairly different. They further conclude that the connectedness of
Bitcoin to markets is considerably different from that of Gold; there are, however,
short-lived temporal similarities that dissipate after corrections. They further argue
that in terms of being a safe-haven during the ‘flight to quality’ periods, although
Gold’s safe-haven nature is established, Bitcoin shows a positive decoupling effect and
fails to provide a hedge or a safe-haven against equity markets/instruments. Shahzad
et al. (2019), on the contrary, show that the safe haven nature of Bitcoin is market
and context-dependent. By applying the rolling window approach, they find that
while Bitcoin offers a hedge or safe-haven during extreme market movements in
developing economies, it fails to offer a hedge against the developed markets where
the only safe-haven is Gold. They also propose a new definition of safe-haven, and
classify Bitcoin and Gold as ‘weak’ safe havens in some cases. On the other hand,
Charfeddine et al. (2020) found that the diversification or hedging potential of cryp-
tocurrencies is dependent upon the economic and political scenarios, and therefore,
they concluded that the cryptocurrencies can be considered for portfolio diversifica-
tion, however, they are poor hedges, in most cases. Similar results were reported by
Das et al. (2020) for Bitcoin, they argue that Bitcoin does not offer superior hedging
capabilities and that the hedging efficacy of Bitcoin is sensitive to market conditions,
and therefore, must be viewed with caution, particularly during uncertain economic
conditions (Hernandez et al., 2022). Similar results are reported by Shahzad et al.
(2022) during COVID for BRICS economies. They report that while bitcoin and gold
may offer conditional diversification and hedging benefits for BRICS economies dur-
ing the COVID-19 pandemic, gold offers the highest potential for diversification and
hedging, while employing bitcoin as an alternative asset for diversification or hedging
should be considered with extreme caution, despite its huge popularity among the
institutional investors. Symitsi and Chalvatzis (2019) found that Bitcoin can offer sub-
stantial diversification benefits in regime-dependent (bearish and bullish) market con-
ditions. Chemkha et al. (2021) in their study on major stock indices and currencies
during the COVID-19 crisis found that while Bitcoin and Gold can hedge risk in the
international portfolios effectively, Gold only acts as a weak safe-haven and Bitcoin
does not offer any safe-haven potential. Chkili et al. (2021) examine the safe-haven
potential of bitcoin for Islamic stock indices during the COVID-19 pandemic and
found that while the diversification potential of bitcoin considerably increases during
the pandemic, the hedging strategy involving bitcoin requires higher costs during
the crisis.
Corbet et al. (2020), on the other hand, in their study on Chinese financial markets
during the COVID-19 pandemic assert that Bitcoin does not act either as a hedge or
safe-haven. Choi and Shin (2021) also found that bitcoin prices are correlated with
the economic policy uncertainty, particularly during the COVID-19 pandemic, and
therefore do not act as a hedge or a safe-haven. Similar results were reported by
Conlon and McGee (2020) and Karamti and Belhassine (2021) for US markets during
the COVID-19 crisis. They found that Bitcoin and the S&P 500 index were strikingly
positively correlated during the development of the COVID-19 crisis, therefore dissi-
pating any potential of Bitcoin to act as a safe-haven for US Equity markets. In a
similar study, Goodell and Goutte (2021) report that the benefits of Bitcoin in port-
folio diversification for an Equity portfolio are very thin during the COVID-19 crisis,
they argue that the co-movements between Bitcoin and Equity markets gradually
spiked as the crisis progressed, thereby reducing the benefits of Bitcoin to portfolio
diversification. On the other hand, Su et al. (2022) studied the dynamic relationship
between Bitcoin and VIX and found that Bitcoin does not act as a safe haven during
the high fear sentiment. They, however, found a striking signal transmission between
BTC and VIX and concluded that the high fear sentiment can potentially provide a
signal for investors to alter their investment strategies with regard to portfolio diversi-
fication with Bitcoin through portfolio rebalancing. Obeng (2021) reported the exist-
ence of asymmetry in the cryptocurrency markets and concluded that the
cryptocurrency markets are characterized by regimes of bad sentiment and good sen-
timent. While cryptocurrencies strongly respond to the negative market sentiment
represented by a huge rise in volatility, the same is not reflected during the positive
news regime.
3. Data description
The current study aims to examine as to what extent, if at all, is Bitcoin related to
global economic activity, forex markets, equity markets, energy markets, fixed-income
assets (corporate bonds), and commodities (Gold). The study further examines
whether Bitcoin expresses any different behavior than currencies (proxied by USD-
Euro) and Gold in terms of its potential to act as either a hedge, diversifier, or a safe-
haven. The following proxies were employed to model the proposed relationship;
Bitcoin (BTC) for Bitcoin (cryptocurrencies), the Baltic Dry Index (BALTICF) for
real economic activity, USD to EURO (USEURSP) for Forex markets, S & P 500 for
equity markets (S&P), the Dow Jones UBS Energy Spot Subindex (DJUBENS)
for Energy Markets, the iShares Long-Term Corporate Bond ETF (U:IGLB) for
Corporate Bonds, and the CMX-Gold 100 ounce (NGCC.01) for Gold. The data for
10 A. RASHID ET AL.
all variables have been collected from the Refinitiv Eikon database for the period
January 1st, 2014 to July 15th, 2021. The data is of daily frequency and each variable
has been sorted and time-stamped for all variables with a total of 1,967 observations
across all variables.
Bitcoin has remained the most prominent cryptocurrency in the decentralized
digital payments ecosystem despite a phenomenal growth of altcoins in recent times.
As of August 2021, the market capitalization of Bitcoin is approximately USD 920 bil-
lion which is roughly 45% of the total cryptocurrency ecosystem7. The Baltic
Exchange Dry Index (BDI) provides insights into the global supply and demand
trends and is considered an indicator of global economic activity. The index was first
introduced in January 1985 by the London-based Baltic Exchange. BDI is a composite
of Capesize, Panamax, Handysize, and Supramax sub-indices. It measures the changes
in the cost of transporting raw materials across more than 20 different sea routes.
USD/EUR exchange rate is considered the most important indicator across worldwide
forex markets. The importance of the USD/EUR exchange rate is due to the invest-
ment and trade of these two large economic areas with each other. The trade and
investment among the two regions is such that the prices in these economic regions
are arbitraged against the exchange rate8. Introduced in 1957, the S&P 500 index is a
widely followed indicator and is regarded as the best gauge of large-cap US equity
markets. The index includes the top 500 leading companies making up around 80%
of the available market capitalization9. The Dow Jones UBS Energy Spot Subindex
measures the price movements of energy included in the Bloomberg subindices -
crude oil, heating oil, natural gas, and unleaded gasoline. The iShares Long-Term
Corporate Bond ETF is an index of US-denominated investment-grade corporate
bonds with remaining maturities greater than 10 years10. Finally, CMX-Gold
100 ounce is the price of 100 ounces of Gold for spot delivery.
4. Research methodology
As part of preliminary time series testing, we implement a unit root test. The study
employs the Kapetanios and Shin (2008) unit root test to confirm whether the data is
stationary, as non-stationarity in time series data could lead to spurious innovations
in the results. There is a growing consensus among researchers that the traditional
unit root tests fail to consider nonlinearity, fractional integration, and other determin-
istic properties of time series (Bierens, 1997; Gil-Alana & Robinson, 1997; Kapetanios
et al., 2003; Kapetanios & Shin, 2008; Leybourne et al., 1998). Since the Markov
Regime Switching Regression (MRSR) model employed in this study requires the data
to follow a non-linear structure, it is more appropriate to use the Kapetanios and
Shin (2008) unit root test as the traditional tests could lead to spurious results.
To confirm the non-linear structure of the series, the residuals of the estimated
ARFIMA model are generated to perform the BDS (Brock, Dechert, and Scheinkman)
test. The BDS test is employed to check for the stochastic linearity of returns. BDS is
a portmanteau test for independence and identical distribution but has been extended
to test for deviations from independence including non-linear dependence. After
checking for non-linear dependence in the series, MRSR models are employed to
analyze the asymmetric relationship between the regime-dependent variables. The

model, thus, examines the impact of the analyzed variables on Bitcoin prices, and
well as the switching properties of Bitcoin from bull to bear regime and vice-versa.
The specification of the models is presented below.
4.1. The Kapetanios and Shin (K & S) unit root test

The Kapetanios and Shin (2008) unit root model is given by the following analogous
ESTAR specification:

Dyt ¼ cyt1 1 expðhyt1
2
Þ þ et (1)
where et iid ð0, r2 Þ, 2 < c < 0, h 0, yt is the demeaned or detrended time
series of interest, and 1 expðhyt1
2
Þ is the exponential transition function to pre-
sent the nonlinear adjustment. Given c, yt contains a unit root when h ¼ 0, while yt
is globally stationary when h 0: Therefore, under the null hypothesis, the series fol-
lows a linear unit root process whereas, under the alternative hypothesis, the series is
non-linear stationery.
To test the null hypothesis, reparameterization by computing a first-order Taylor
series approximation of Eq.(1) is suggested by KSS, to obtain the following auxiliary
function:
X
p
Dyt ¼ qj Dytj þ dyt1
3
þ et (2)
j¼1
Where the serially correlated errors are corrected with q augmentations, and d ¼
ch, with a t-test hypothesis for d ¼ 0 against d < 0, the asymptotic critical value
of t-statistics is computed as:
t ¼ ^d=s:e: ð^d Þ
where ^d is the OLS estimate of d and s:e: ð^d Þ is its standard error. Kapetanios and
Shin (2008) argue that the OLS-based KSS test demonstrates low power, and therefore
suggest a modified KSS test with GLS residuals. They confirm through a simulation
procedure that the GLS detrending-based KSS test is superior in terms of power to
the OLS-based test.
4.2. The ARFIMA model

Since most financial markets exhibit long memory properties, the systems working
with short memory parameters fail to capture this long memory component (Liu
et al., 2017; Sheng et al., 2011). Most financial time series exhibit long-range depend-
ence that can be classified as I(0) and not I(1). To more accurately capture the long
memory component for time series exhibiting fractional properties, a more flexible
class of process known as autoregressive fractionally integrated moving averages
12 A. RASHID ET AL.
(ARFIMA) model has been introduced by Granger and Joyeux (1980), Granger
(1980), and Hosking (1981). The ARFIMA process of a time series of order (p, d, q)
is given by:
d
/ðLÞð1 LÞ ðyt lÞ ¼ hðLÞet , i:i:d:ð0, r2e Þ (3)
where the fractional differencing parameter is denoted by d, and et is the white noise.
The stochastic process yt is both stationary and invertible if all roots of /ðLÞ and
hðLÞ lie outside the unit circle and jdj < 0:5; the process is nonstationary for d
0 Granger and Joyeux (1980). The autocorrelation function, qð:Þ is proportional to
k2d1 as k ! 1, indicating the slow rate of hyperbolic decay (Hosking, 1981).
4.3. The Brock, Dechert and Scheinkman (BDS) test

The BDS test developed along with the chaos theory by Brock et al. (1987) is one of
the most popular tests for detecting general stochastic non-linearity. It is a nonpara-
metric test originally designed to test for independence and identical distribution but
has been extended to test for deviations from independence including non-linear
dependence (or ‘chaos’). The test can be applied to a series of estimated residuals to
check whether the residuals are independent and identically distributed. The correl-
ation integral at the embedding dimension, m, e > 0, is estimated by the following:
X
2
Cm, T ðeÞ ¼ m m
Ie Xt , Xs (4)
t<s
Tm ðTm 1Þ
where IXtm , Xsm denotes an indicator function and is given by:

1 if jXtm Xsm j<e
IXtm , Xsm ¼
0 otherwise
The correlation integral estimates the probability that any two m-dimensional
points are within a distance of e of each other (Bisaglia & Gerolimetto, 2014):
PðjXt Xs j < e, jXt1 Xs1 j < e, . . . :, jXtmþ1 Xsmþ1 j < eÞ
If the Xt are IID, the probability equals the following:
C1, T ðeÞm ¼ PðjXt Xs j<eÞm
The BDS statistic as defined by Brock et al. (1987) is given by:
pffiffiffiffi Cm, T ðeÞ C1, T ðeÞm

Vme ¼ T
sm, T
where sm, T denotes the consistent estimate of the asymptotic standard deviation.
4.4. The Markov Regime Switching Regression (MRSR) model

Hamilton (1989) introduced a model of order p, where the time series can assume
substantially different characteristics by forming regimes. The model is generally
applicable to economic scenarios such as policy changes (e.g. expansionary vs con-
tractionary fiscal policy stances), as well as to financial structures where the data can
assume multiple structures such as bearish vs bullish regimes, as is relevant to this
study. The distinguishing feature of this model is that it allows for switching between
the bear-bull regimes and allows one to control this switching mechanism through
the unobservable state variable which follows the first-order Markov chain (Kuan,
2002). For two regime data, the model can be presented as:
Pp
a1 þ Pi¼1 ;1, i Xt1 þ e1, t if st ¼ 1
Xt ¼ p (5)
a2 þ i¼1 ;2, i Xt1 þ e2, t , if st ¼ 2
where each ei, t is independent of each other and are normally distributed with a
given mean and variance (0, r2i ). The unobservable state or regime variable is gov-
erned by a first-order Markov chain with the transition probabilities:
Pðst ¼ 1 j st1 ¼ 1Þ Pðst ¼ 2 j st1 ¼ 1Þ p11 p12

P¼ ¼
Pðst ¼ 1 j st1 ¼ 2Þ Pðst ¼ 2 j st1 ¼ 2Þ p21 p22
A small pij value indicates that the model stays longer in state i, with an expected
duration 1=pij in state i:
5. Empirical analysis
As previously stated, this study examines the impact of several financial and eco-
nomic indicators to understand the resemblance (if any) between Bitcoin and the
conventional investment asset classes with particular attention given to determining
whether it could contribute to portfolio diversification, acts as a hedge, or serve as a
safe-haven asset. The study employs a Markov-regime regression model to ascertain
whether or not bitcoin considerably resembles or differs from other assets in its rela-
tionship with other variables in a regime-dependent model, and whether it presents
any potential safe-haven properties.
To carry out the analysis, it is important to employ the variables in their stationary
state. Since the non-stationary time series data could lead to spurious innovations in
the results, the unit root test is used to confirm whether the data is stationary.
Conventional tests such as the Augmented Dickey-Fuller (ADF) Test and the Philips-
Perron (PP) unit root test have largely been used in the existing literature to test for
a unit root in the time series. However, these tests may fail to reject the unit root
hypothesis in the presence of a structural break (Perron, 1989). Zivot and Andrews
(1992) have shown that one can check for unit root by allowing a single structural
break. Moreover, Lumsdaine and Papell (1997) and Kapetanios (2005) have shown
that one can extend the test by allowing anywhere between two and m breaks,
respectively. However, when the time series of interest is non-linear, the traditional
14 A. RASHID ET AL.
Table 1. Kapetanios and Shin (2008) unit root test.

Variable Test Statistic p-value Criteria
Bitcoin 5.526 0.000 Fixed
5.544 0.000 AIC
6.185 0.000 SIC
BDI 7.103 0.000 Fixed
7.427 0.000 AIC
10.157 0.000 SIC
USD/EUR 1.686 0.017 Fixed
1.686 0.025 AIC
3.149 0.000 SIC
S&P 500 6.957 0.000 Fixed
6.957 0.000 AIC
7.641 0.000 SIC
Energy 6.719 0.000 Fixed
6.719 0.000 AIC
7.157 0.000 SIC
Bonds 2.107 0.006 Fixed
2.156 0.007 AIC
3.312 0.000 SIC
Gold 1.342 0.043 Fixed
1.342 0.069 AIC
2.428 0.004 SIC
Source: Authors’ estimations/calculations.
unit root tests fail to consider nonlinearity, fractional integration, and other determin-
istic properties of time series (Bierens, 1997; Gil-Alana & Robinson, 1997; Kapetanios
et al., 2003; Leybourne et al., 1998). Therefore, this study has employed the
Kapetanios and Shin (2008) unit root test to capture the non-linear dynamics of time
series. The test results are presented below:
The results presented in Table 1 indicate that all the variables are stationary at
level at the 99% confidence interval based on the SIC criteria, and stationary at level
at 90% confidence level based on all criteria. Since the results confirm that all the var-
iables chosen for the study are stationary at a level, we can proceed with the raw
returns series at I(0) level. The further analysis is conducted by employing the
Markov-Regime Switching (MRSR) to examine the nature of interactions and impact
of the selected variables on Bitcoin prices, as well as the probability of transitioning
from a bull to bear regime and vice-versa. The model requires the returns series to
have a non-linear structure and to test that the BDS test is employed to check for the
stochastic linearity of returns. Thus, the residuals of the estimated linear models
(ARMA, ARIMA, or ARFIMA, for instance) are generated to perform the BDS test as
demonstrated by Brock and Sayers (1988). This analysis is based on the ARFIMA
structure of the returns, which is presented in Table 2.
The ARFIMA (p,d,q) model considers the long-run dependence or persistence in
the time series with a fractionally (non-integer) integrated differencing parameter.
The concept of fractional integration often refers to defining the long-range depend-
ence or long memory, while a pure ARIMA process represents a short memory,
where the value of past disturbances follows a geometric decay and almost quickly
dies out. Against the reference ranges of 0 < d < 0.5 for long memory (Baillie, 1996),
Bitcoin and Energy autocorrelation function follows a long memory or hyperbolic
decay, which indicates a deviation from the weak form of market efficiency, and thus
non-random chaotic behavior. It may thus be concluded that Bitcoin exhibits
Table 2. ARFIMA structure of the returns series.

USD_EUR
Variable Bitcoin (1,1)ᶧ BDI (1,1)ᶧ (1,1)ᶧ S&P (1,1)ᶧ Energy (1,1)ᶧ Bonds (1,1)ᶧ Gold (1,1)ᶧ
Constant 0.0018 0.0003 0.0000 0.0003 0.0000 0.0001 0.0002
(0.0015) (0.0003) (0.0000) (0.0000) (0.0007) (0.0000) 0.0001)
d 0.0635 0.5692 0.0386 0.8875 0.0856 0.1751 0.0926
(0.0302) (0.0313) (0.0362) (0.0277) (0.0324) (0.0157) (0.0529)
AR (1) 0.0128 0.9565 0.2815 0.9833 0.0753 0.4149 0.6420
(0.2015) (0.0109) (0.3819) (0.0049) (0.1499) (0.0272) (0.1166)
MA (1) 0.0983 0.3492 0.2221 0.2502 0.2304 0.1195 0.5692
(0.2212) (0.0251) (0.3627) (0.0259) (0.1703) (0.0209) (0.1106)
Sigma 0.0022 0.0003 0.0000 0.0001 0.0003 0.0000 0.0000
(0.0000) (0.0000) (0.0000) (0.0000) (0.0000) (0.0000) (0.0000)
Log Likelihood 3233.28 4908.3 7708.44 6130.55 5024.68 6978.10 6452.46
AIC 3.28 4.98 7.83 6.23 5.10 7.09 6.56
SIC 3.27 4.97 7.82 6.21 5.09 7.08 6.54
ᶧp is the order of the autoregressive process, q is the order of the moving averages process.
, and denote the significance at the 90%, 95% and 99% confidence intervals. Standard errors are given in
parenthesis.
AIC: Akaike Information Criterion; SIC: Schwarz Information Criterion.
different volatility characteristics in comparison to ordinary fiat currencies (USD/

EUR exchange rate), Gold, and other selected variables as indicated by their d param-
eter in Table 2. The extracted residuals of the ARFIMA estimations are then used in
the BDS independence test. This test checks whether there is any non-linear depend-
ence in the series and if the residuals of the ARFIMA process (linear fitted model)
follow an independent and identical distribution (i.i.d). Simply put, the observations
of a pure i.i.d series have a constant probability of the distance between any two
points being less than or equal to epsilon (e). The model requires parameters e and
m; e being the epsilon value assigned to the difference between the consecutive m-
histories (the embedding dimension). For the residuals of ARFIMA, epsilon values
between 1 and 2, and embedding dimensions 3, 5, and 7 have been employed (Hsieh
& LeBaron, 1988). The BDS test results thus obtained are presented in Table 3.
The BDS test can be used to detect various possible deviations from independence
(such as chaotic behavior) and can be employed to test against arbitrary dependence
by using the correlation function—the asymptotic distribution of which is known
under the assumption of the null hypothesis (Barnett et al., 1996). A BDS test statistic
above the critical value is required for the null hypothesis to be rejected. As observed
in Table 3, the BDS test statistic for all variables (except Gold at embedding dimen-
sion 3) is indeed above the recommended critical values of 1.960 and 2.575 at 95%
and 99% confidence intervals, respectively (Broock et al., 1996). Therefore, the null
hypothesis is rejected for all the variables at all epsilon values and embedding dimen-
sions. The results indicate deviation from the i.i.d property, and a non-linear depend-
ence in all variables, thus making it convenient to use for the MRSR model.
The MRSR model is useful when the parameters of a data generating process vary
over a set of different observed states. The model can estimate the probability of each
possible state, and the transitions thereof, and thus is useful in application to eco-
nomic cycles (Recession/Expansion), financial markets (Low/High volatility), and
interest-rate regimes (Expansionary/Contractionary), etc. An MRSR model thus ena-
bles the analysis of the asymmetric relationship between the regime-dependent
16 A. RASHID ET AL.
Table 3. BDS test results.

e 3 5 7
Bitcoin 0.5 8.897 12.665 16.917
0.7 7.611 10.452 12.580
0.9 7.297 9.373 10.359
BDI 1 40.110 44.079 52.876
1.5 32.285 31.142 31.656
2 28.220 25.715 24.472
USD/EUR 1 8.049 10.394 13.080
1.5 9.008 11.126 13.047
2 9.306 11.068 12.600
S&P 500 1 19.446 15.518 31.770
1.5 19.019 22.002 24.356
2 19.022 20.506 21.067
Energy 1 8.744 12.688 16.367
1.5 9.846 12.780 14.956
2 10.934 13.255 14.622
Bonds 1 7.672 9.138 9.689
1.5 10.570 11.966 12.297
2 15.077 15.897 15.933
Gold 1 1.982 2.933 3.675
1.5 2.125 4.135 4.655
2 3.162 5.197 5.576
variables. Since, financial markets, and more recently cryptocurrency markets, have
witnessed increased volatility (switching between the bearish and bullish phases). The
current study is an attempt to examine the behavior of Bitcoin across different phases
of volatility by using the MRSR model in comparison to other variables, Gold and
Forex, in particular. The MRSR results are presented in Table 4.
As the MRSR model does not itself label the regimes, the regimes are labeled on
the basis of average returns represented by their constants. Therefore, Regime 1 rep-
resents a bullish regime and regime 2 represents a bearish regime for all variables. In
Regime 1, Bitcoin is positively related to the S&P 500 and Gold, while negatively
related to Energy (p-value < 0.1); whereas in Regime 2, Bitcoin is positively related
to USD/EUR and S&P 500, while negatively related to Energy. Gold, on the contrary,
is positively related to USD/EUR, Energy, and Bonds, and negatively related to S&P
500 in Regime 1, whereas, in Regime 2, positively related to Bitcoin (p-value < 0.1),
S&P 500, and Bonds, and negatively related to BDI, USD/EUR (p-value < 0.05), and
Energy. Similarly, USD/EUR is positively related to Bitcoin (p-value < 0.1) and Gold,
and negatively related to Bonds (p-value < 0.1) in Regime 1, whereas, in Regime 2, it
is positively related to Bonds and negatively related to S&P 500 at 90% and 95% con-
fidence intervals, respectively. These results indicate that the behavior of Bitcoin is
considerably different from both the USD/EUR and Gold in terms of its relationship
with different variables under study in regime-dependent models. The results also
suggest that while Gold and Bonds appear to be significantly positively related in
both the regimes, Gold offers better hedging and safe-haven potential compared to
both the bonds as well as the Bitcoin due to its substantial and negative relationship
with variables across regimes. To further substantiate the argument, it is observed
that Bitcoin is positively (though not considerably) related to USD/EUR (p-value <
0.1) S&P 500, and Gold (p-value < 0.1); and negatively related to Energy and Bonds
in Regime 1 (Bullish Regime) of the underlying variables, whereas, unrelated to these
Table 4. Markov regime switching regression.

Variable Bitcoin BDI USD/EUR S&P 500 Energy Bonds Gold
Regime 1 Constant 0.0057 0.0228 0.0004 0.0042 0.0105 0.0303 0.0003
(0.0009) (0.0025) (0.0003) (0.0015) (0.0113) (0.0016) (0.0002)
Bitcoin – 0.0254 0.0099 0.0926 0.2465 0.0540 0.0072
(0.0198) (0.0049) (0.0186) (0.0330) (0.0129) (0.0039)
BDI 0.0107 – 0.0063 0.1353 0.0279 0.5768 0.0061
(0.0321) (0.0064) (0.0410) (0.3343) (0.0660) (0.0068)
USD/EUR 0.1052 0.0211 – 0.9540 3.4568 6.9706 0.6429
(0.2277) (0.2313) (0.1807) (1.8510) (0.2037) (0.0380)
S&P 500 0.2837 0.0007 0.0639 – 2.6587 0.3068 0.0937
(0.1093) (0.1279) (0.0402) (0.3342) (0.0423) (0.0195)
Energy 0.0922 0.0179 0.0410 0.5888 – 0.0177 0.0437
(0.0527) (0.0607) (0.0392) (0.0441) (0.0324) (0.0100)
Bonds 0.1726 0.3845 0.1314 0.7210 1.3428 – 0.2232
(0.1393) (0.1565) (0.0793) (0.0769) (0.3433) (0.0316)
Gold 0.4636 0.1762 0.3662 0.3911 4.3072 0.0123 –
(0.1209) (0.1104) (0.0952) (0.0863) (1.2306) (0.0518)
Regime 2 Constant 0.1137 0.0117 0.0005 0.0006 0.0004 0.0000 0.0104
(0.0095) (0.0015) (0.0002) (0.0002) (0.0003) (0.0000) (0.0035)
Bitcoin – 0.0073 0.0066 0.0054 0.0062 0.0008 0.0909
(0.0143) (0.0069) (0.0046) (0.0078) (0.0026) (0.0776)
BDI 0.0289 – 0.0003 0.0105 0.0341 0.0032 0.2086
(0.2340) (0.0063) (0.0077) (0.0135) (0.0043) (0.0549)
USD/EUR 3.0065 0.1489 – 0.0493 0.2683 0.0869 1.1057
(1.1265) (0.1394) (0.0477) (0.1132) (0.0259) (0.5138)
S&P 500 3.3455 0.0009 0.0545 – 0.4740 0.0449 0.7621
(0.6477) (0.0731) (0.0271) (0.0604) (0.0123) (0.1064)
Energy 1.2678 0.0628 0.0099 0.1115 – 0.0195 1.1118
(0.4126) (0.0356) (0.0213) (0.0120) (0.0066) (0.1557)
Bonds 0.1217 0.1426 0.0791 0.0454 0.1319 – 1.0509
(0.7202) (0.0902) (0.0421) (0.0348) (0.0907) (0.1359)
Gold 0.4679 0.1050 0.0551 0.1045 0.0141 0.1588 –
(0.4807) (0.0787) (0.0573) (0.0264) (0.0695) (0.0141)
Model LL 3373.61 4512.74 7875.19 6460.86 5248.75 7541.61 6708.01
AIC 3.41 4.57 7.99 6.55 5.32 7.65 6.80
HQ Criterion 3.39 4.55 7.97 6.53 5.3 7.63 6.79
Expected R1 33.19 5.89 2.08 3.48 1.29 1.97 73.15
Duration R2 1.14 16.30 2.60 55.76 38.09 227.39 1.16
, and denotes the significance at the 99%, 95% and 90% confidence interval. Standard errors are given in
parenthesis.
any of those variables in Regime 2 (Bearish Regime) of those variables. Gold, on the
contrary, is positively related to Bitcoin, USD/EUR, S&P 500, and Energy in Regime
1 of the underlying variables; whereas, negatively related to S&P 500 and positively
related to Bonds, and not related to any other variables in Regime 2. This indicates
that while Gold may offer considerable hedge to safe-haven potential (particularly for
the equity markets), Bitcoin can at best offer potential as a diversifier or a hedge.
These results are consistent with Charfeddine et al. (2020), Das et al. (2020), Shahzad
et al. (2020), Bakry et al. (2021), Choi and Shin (2021), Jiang et al. (2022), and
Shahzad et al. (2022) who suggest that Bitcoin can, at best, act as a diversifier or
a hedge.
Moreover, since Bitcoin demonstrates no resemblance with the properties of USD/
EUR or Gold, as it is not affected by the same set of variables as USD/EUR or Gold
are under either regime, it can be concluded that Bitcoin considerably deviates from
both. However, owing to the nature of its relationship with other variables across
regimes, it can be concluded that Bitcoin offers a considerable potential in portfolio
18 A. RASHID ET AL.
Figure 1. Dynamic conditional correlation graphs (DCC-GARCH).

diversification which may extend to hedging, but it would be going too far to classify
‘Bitcoin as the new Gold’. According to the definition of a diversifier, hedge, and
safe-haven proposed by Baur and Lucey (2010), Bitcoin may be classified as a poten-
tial diversifier, or at most, a hedge, particularly against energy and bonds. A safe-
haven requires an asset to have a negative association with other assets, particularly
during negative market movements. Based on this fundamental property of a safe-
haven, Bitcoin does not seem to resemble these properties as it is not related to any
variables in the bearish regime (Regime 2). To further explore the safe-haven, hedge,
or diversifier properties of Bitcoin, we examined the dynamic conditional correlation
of Bitcoin with all assets, which is presented in Figure 1.
As depicted by the graphs above, we follow the approach followed by Ngene et al.
(2018) to investigate the time-invariant interactions in volatility among different assets.
Specifically, we explored the dynamic conditional correlation of Bitcoin with all assets
as shown in Figure 1. The results indicate that, on average, Bitcoin exhibits a very low
dynamic conditional correlation with USD/EUR, the S&P 500, Gold, and Bonds.
However, it exhibits a negative dynamic conditional correlation with Energy and no
discernible pattern of conditional correlation with BDI. These results suggest that
Table 5. Transition probabilities.

Regime 1 2
Bitcoin 1 0.969876 0.030124
2 0.880195 0.119805
BDI 1 0.883553 0.116447
2 0.061335 0.938665
USD/EUR 1 0.520727 0.479273
2 0.384452 0.615548
S&P 500 1 0.712939 0.287061
2 0.017934 0.982066
Energy 1 0.228036 0.771964
2 0.026256 0.973744
Bonds 1 0.493911 0.506089
2 0.004398 0.995602
Gold 1 0.986329 0.013671
2 0.860664 0.139336
Bitcoin may provide considerable potential for portfolio diversification, in general, as

well as some hedging potential against Energy and global economic activity (proxied by
BDI). Moreover, upon careful examination of the shift in the dynamic conditional cor-
relations during the periods of market distress (marked by the COVID-19 crisis in early
2020), we find a sharp rise in the conditional correlation of Bitcoin with USD/EUR,
Gold, S&P 500, and Bonds. These findings are consistent with our earlier analysis sug-
gesting that Bitcoin, at best has the potential to act as a diversifier or a hedge, rather
than a safe-haven. These results are consistent with Corbet et al. (2020), Conlon and
McGee (2020), Karamti and Belhassine (2021), and Goodell and Goutte (2021).
Moreover, it is often argued that Bitcoin may not act as a stable hedge or a safe-
haven due to its extreme volatility, unknown set of determinants, absence of institu-
tional investors, the dominance of retail investors, lack of proper regulation, and
market mechanism, legal, and environmental concerns. With respect to the concern
of extreme volatility, we compare the standard errors of Bitcoin and find that it has a
considerably higher set of standard errors compared to other variables. This may sug-
gest that Bitcoin price discovery is driven by some speculative components and thus
is more volatile. Furthermore, since the standard errors for the bearish regime are
considerably higher compared to the bullish regime, particularly for Bitcoin, it could
be concluded that the volatility in Bitcoin is prone to a downward spiral due to the
dominance of retail investors.
The transition probabilities are presented in Table 5 below. It can be observed
from the transition probabilities in Table 5 and Figure 2, that although most of the
variables move considerably between bullish and bearish trends, for the most part,
the Bitcoin price exhibits a bullish trend, and the probability of transition from bear-
ish to bullish trend is quite large. This transition probability bears a striking resem-
blance with Gold. Specifically, the probability of staying in Regime 1 for Bitcoin and
Gold is 0.97 and 0.99, respectively, while the probability of staying in Regime 2 for
Bitcoin and Gold is 0.12 and 0.14, respectively. The corresponding probability of
transition from Regime 2 to Regime 1 is 0.88 and 0.86 for Bitcoin and Gold, respect-
ively. Nevertheless, this resemblance is not sufficient to argue that Bitcoin carries the
same safe-haven properties as Gold, if not substantiated by other measures.
20 A. RASHID ET AL.
Figure 2. Filtered regime probabilities.

Note: The Figure represents the transition probabilities of all the assets under the study in a two-regime Markov
Regime Switching Regression Model. The graphs indicate that Bitcoin and Gold mostly stay in Regime 1 (Bullish
Regime) with a considerably low transition to Regime 2 (Bearish Regime). USD/EUR demonstrates a constant transition
between the regimes and almost equal durations in either regime. BDI, likewise, exhibits a high probability for both
regimes to exist with a considerable probability to transition between the regimes. On the other hand, S&P 500,
Energy, and Bonds tend to mostly stay in regime in Regime 2 (Bearish Regime) with a considerable probability of
transition from Regime 1 to Regime 1.
6. Conclusions
Bitcoin and other cryptocurrencies (altcoins) have attracted increasing attention over
the last few years. Despite numerous controversies and substantial legal uncertainties,
cryptocurrencies continue to gain interest and legitimacy, besides their entry into the
derivatives market. The study aims to understand bitcoin from the perspectives of a
medium of transaction, a speculative asset, a diversifier, a hedge, or a safe-haven, and
therefore, compares it to currency, commodities, other financial assets, and gold. This
study examines the relationship between Bitcoin, global economic activity, foreign
exchange markets, equity markets, Energy markets, Bonds, and Gold, and explored
the potential for Bitcoin to act as a safe-haven, a hedge, or a diversifier, with a par-
ticular focus on being classified as a medium of transaction and a hedge (or a safe-
haven). The results indicate that Bitcoin behaves differently than both the USD/EUR
and Gold vis-a-vis its relationship to different variables. Our study finds that Bitcoin
exhibits a positive relationship with equity markets in both the bullish and bearish
regimes, a positive relationship with the Forex markets and Gold in the bullish
regime, and a negative relationship with Energy in both regimes. Moreover, the study
suggests that Bitcoin differs considerably from USD/EUR and Gold in terms of its
relationship with the examined variables under both regimes. The results also indicate
that Bitcoin is related to all the variables except BDI in the bullish regime while not
related to any of the variables in the bearish regime, contrary to the properties of
Gold. It may, therefore, be concluded that Bitcoin fails to provide any potential safe-
haven properties as characterized by Gold, rather it may act as a diversifier or a
hedge, at best. Furthermore, while most of the variables frequently switch from one
regime to another and have higher transition probabilities, Bitcoin’s transition prob-
ability is very similar to Gold. We also found that Bitcoin and Gold both largely tend
to stay in a bullish regime with a considerably lower probability of transition to a
bearish regime. While these results indicate that Bitcoin may offer some hedging to
diversification potential in global portfolio investments, it must be viewed with cau-
tion due to the speculative component in Bitcoin which can cause excessive volatility
in this evolving asset class.
Nevertheless, it must be noted that our results are based on a short sample period;
and it will be worth examining how Bitcoin is related to yet other financial and real
variables. Future research could also examine whether derivatives trading has affected
the volatility and maturity of Bitcoin for larger sample periods. This could provide
further insights into whether the diversification potential still exists in Bitcoin, or
whether it tends to diminish in the long run. The results of this study could be fur-
ther extended by examining the dynamics of bitcoin with other altcoins during the
turbulent market conditions to examine its safe-haven/hedging potential. It would
also be interesting to see how bitcoin differs from conventional safe-haven assets for
different asset classes and portfolios by conducting event studies during turbulent
market conditions. The recent research in the area of cryptocurrencies has established
that the prominence and popularity of bitcoin are decreasing fast leading way to alt-
coins and Non-fungible tokens (NFTs) like Ethereum, Litecoin, Ripple, etc., it would,
therefore, be interesting to study the evolutionary direction of the diversification and
hedging dynamics of cryptocurrencies in the evolving NFT and metaverse landscape.
22 A. RASHID ET AL.
Notes
1. Reuters 2021
2. Forbes 2021
3. https://www.bbc.com/news/business-60870100
4. https://www.cnbc.com/2022/03/03/ukraine-raises-54-million-as-bitcoin-donations-surge-
amid-russian-war.html
5. https://twitter.com/CFTCquintenz/status/1422912721637580803?s=20
6. Forbes 2021
7. Data accessed from https://coinmarketcap.com/ on August 28th, 2021.
8. Forbes 2018
9. https://www.spglobal.com/spdji/en/indices/equity/sp-500/#overview
10. https://www.ishares.com/us/products/239423/ishares-10-year-credit-bond-etf
CRediT authorship contribution statement

All authors contributed equally in carrying out the study.
Disclosure statement
No potential conflict of interest was reported by the authors.
ORCID
Audil Rashid http://orcid.org/0000-0003-3168-8007
Walid Bakry http://orcid.org/0000-0002-6049-9662
Somar Al-Mohamad http://orcid.org/0000-0001-6191-8892
References
Bouri, E., Gupta, R., & Roubaud, D. (2019). Herding behaviour in cryptocurrencies. Finance
Research Letters, 29, 216–221. https://doi.org/10.1016/j.frl.2018.07.008
Baillie, R. T. (1996). Long memory processes and fractional integration in econometrics.
Journal of Econometrics, 73(1), 5–59. https://doi.org/10.1016/0304-4076(95)01732-1
Bakry, W., Rashid, A., Al-Mohamad, S., & El-Kanj, N. (2021). Bitcoin and portfolio diversifica-
tion: A portfolio optimization approach. Journal of Risk and Financial Management, 14(7),
282. https://doi.org/10.3390/jrfm14070282
Barnett, W. A., Barnett, W. A., Kirman, A. P., & Salmon, M. (Eds.). (1996). Nonlinear dynam-
ics and economics: Proceedings of the Tenth International Symposium in Economic Theory
and Econometrics (Vol. 10). Cambridge University Press.
Baur, D. G., & Lucey, B. M. (2010). Is gold a hedge or a safe haven? An analysis of stocks,
bonds and gold. Financial Review, 45(2), 217–229. https://doi.org/10.1111/j.1540-6288.2010.
00244.x
Baur, D. G., Hong, K., & Lee, A. D. (2018). Bitcoin: Medium of exchange or speculative assets?
Journal of International Financial Markets, Institutions and Money, 54, 177–189. https://doi.
org/10.1016/j.intfin.2017.12.004
Bierens, H. J. (1997). Testing the unit root with drift hypothesis against nonlinear trend statio-
narity, with an application to the US price level and interest rate. Journal of Econometrics,
81(1), 29–64. https://doi.org/10.1016/S0304-4076(97)00033-X
Bisaglia, L., & Gerolimetto, M. (2014). Testing for (non) linearity in economic time series: A
Monte Carlo comparison. Working Paper Series, Department of Statistical Sciences, pp.
1–24.
Bonneau, J., Miller, A., Clark, J., Narayanan, A., Kroll, J. A., & Felten, E. W. (2015). Sok:
Research perspectives and challenges for Bitcoin and cryptocurrencies. In 2015 IEEE
Symposium on Security and Privacy (pp. 104–121). IEEE.
Bouoiyour, J., & Selmi, R. (2014). What does crypto-currency look like? Gaining insight into Bitcoin
phenomenon. MPRA Paper No. 58133. https://mpra.ub.uni-muenchen.de/id/eprint/58133
Bouoiyour, J., Selmi, R., & Tiwari, A. (2014). Is Bitcoin business income or speculative bubble?
Unconditional vs. conditional frequency domain analysis. MPRA Working Paper 59595.
http://mpra.ub.uni-muenchen.de/59595
Bouri, E., Gupta, R., Lau, C. K. M., Roubaud, D., & Wang, S. (2018). Bitcoin and global financial
stress: A copula-based approach to dependence and causality in the quantiles. The Quarterly
Review of Economics and Finance, 69, 297–307. https://doi.org/10.1016/j.qref.2018.04.003
Bouri, E., Gupta, R., Tiwari, A. K., & Roubaud, D. (2017c). Does Bitcoin hedge global uncer-
tainty? Evidence from wavelet-based quantile-in-quantile regressions. Finance Research
Letters, 23, 87–95. https://doi.org/10.1016/j.frl.2017.02.009
Bouri, E., Jalkh, N., Molnar, P., & Roubaud, D. (2017b). Bitcoin for energy commodities before
and after the December 2013 crash: Diversifier, hedge or safe haven? Applied Economics,
49(50), 1–11. https://doi.org/10.1080/00036846.2017.1299102
Bouri, E., Molnar, P., Azzi, G., Roubaud, D., & Hagfors, L. I. (2017a). On the hedge and safe
haven properties of Bitcoin: Is it really more than a diversifier? Finance Research Letters, 20,
192–198. https://doi.org/10.1016/j.frl.2016.09.025
Brock, W. A., & Sayers, C. L. (1988). Is the business cycle characterized by deterministic
chaos? Journal of Monetary Economics, 22(1), 71–90. https://doi.org/10.1016/0304-
3932(88)90170-5
Brock, W., Dechert, W. D., & Scheinkman, J. (1987). A test for independence based on the cor-
relation dimension. Economics Working Paper SSRI-8702. University of Wisconsin.
Broock, W. A., Scheinkman, J. A., Dechert, W. D., & LeBaron, B. (1996). A test for independ-
ence based on the correlation dimension. Econometric Reviews, 15(3), 197–235. https://doi.
org/10.1080/07474939608800353
Catania, L., & Grassi, S. (2017). Modelling crypto-currencies financial time-series. https://ssrn.
com/abstract=3028486 accessed on 21/03/2019.
Chan, W. H., Le, M., & Wu, Y. W. (2019). Holding Bitcoin longer: The dynamic hedging abil-
ities of Bitcoin. The Quarterly Review of Economics and Finance, 71, 107–113. https://doi.
org/10.1016/j.qref.2018.07.004
Charfeddine, L., Benlagha, N., & Maouchi, Y. (2020). Investigating the dynamic relationship
between cryptocurrencies and conventional assets: Implications for financial investors.
Economic Modelling, 85, 198–217. https://doi.org/10.1016/j.econmod.2019.05.016
Cheah, E. T., & Fry, J. (2015). Speculative bubbles in Bitcoin markets? An empirical investiga-
tion into the fundamental value of Bitcoin. Economics Letters, 130, 32–36. https://doi.org/10.
1016/j.econlet.2015.02.029
Chemkha, R., BenSaïda, A., Ghorbel, A., & Tayachi, T. (2021). Hedge and safe haven proper-
ties during COVID-19: Evidence from Bitcoin and gold. The Quarterly Review of Economics
and Finance, 82, 71–85. https://doi.org/10.1016/j.qref.2021.07.006
Chkili, W., Rejeb, A. B., & Arfaoui, M. (2021). Does bitcoin provide hedge to Islamic stock
markets for pre-and during COVID-19 outbreak? A comparative analysis with gold.
Resources Policy, 74, 102407. https://doi.org/10.1016/j.resourpol.2021.102407
Choi, S., & Shin, J. (2021). Bitcoin: An inflation hedge but not a safe haven. Finance Research
Letters, 46, 102379.
Ciaian, P., Rajcaniova, M., & Kancs, D. A. (2016). The economics of Bitcoin price formation.
Applied Economics, 48(19), 1799–1815.
Conlon, T., & McGee, R. (2020). Safe haven or risky hazard? Bitcoin during the COVID-19
bear market. Finance Research Letters, 35, 101607. https://doi.org/10.1016/j.frl.2020.101607
Conrad, C., Custovic, A., & Ghysels, E. (2018). Long-and short-term cryptocurrency volatility
components: A GARCH-MIDAS analysis. Journal of Risk and Financial Management, 11(2),
23–12. https://doi.org/10.3390/jrfm11020023
24 A. RASHID ET AL.
Corbet, S., Larkin, C., & Lucey, B. (2020). The contagion effects of the COVID-19 pandemic:
Evidence from gold and cryptocurrencies. Finance Research Letters, 35, 101554. https://doi.
org/10.1016/j.frl.2020.101554
Corbet, S., Meegan, A., Larkin, C., Lucey, B., & Yarovaya, L. (2018). Exploring the dynamic
relationships between cryptocurrencies and other financial assets. Economics Letters, 165,
28–34. https://doi.org/10.1016/j.econlet.2018.01.004
Das, D., Le Roux, C. L., Jana, R. K., & Dutta, A. (2020). Does Bitcoin hedge crude oil implied
volatility and structural shocks? A comparison with gold, commodity and the US Dollar.
Finance Research Letters, 36, 101335. https://doi.org/10.1016/j.frl.2019.101335
Dehouche, N. (2022). Revisiting the volatility of bitcoin with approximate entropy. Cogent
Economics & Finance, 10(1), 2013588. https://doi.org/10.1080/23322039.2021.2013588
Dyhrberg, A. H. (2016a). Bitcoin, gold and the dollar–A GARCH volatility analysis. Finance
Dyhrberg, A. H. (2016b). Hedging capabilities of Bitcoin. Is it the virtual gold? Finance
ElBahrawy, A., Alessandretti, L., Kandler, A., Pastor-Satorras, R., & Baronchelli, A. (2017).
Evolutionary dynamics of the cryptocurrency market. Royal Society Open Science, 4(11), 170623.
Gandal, N., & Halaburda, H. (2016). Can we predict the winner in a market with network
effects? Competition in cryptocurrency market. Games, 7(3), 16. https://doi.org/10.3390/
g7030016
Gantori, S., Donovan, P.Ganesh, K.DeMichiel, M.Dennean, K.Trussardi, F., & Klien, M.
(2017). Cryptocurrencies Beneath the Bubble. UBS. https://www.ubs.com/content/dam/
WealthManagementAmericas/cio-impact/cryptocurrencies.pdf
Gil-Alana, L. A., & Robinson, P. M. (1997). Testing of unit root and other nonstationary
hypotheses in macroeconomic time series. Journal of Econometrics, 80(2), 241–268. https://
doi.org/10.1016/S0304-4076(97)00038-9
Glaser, F., Zimmermann, K., Haferkorn, M., Weber, M. C., & Siering, M. (2014, April 15).
Bitcoin-asset or currency? Revealing users’ hidden intentions. ECIS.
Goodell, J. W., & Goutte, S. (2021). Diversifying equity with cryptocurrencies during COVID-19.
International Review of Financial Analysis, 76, 101781. https://doi.org/10.1016/j.irfa.2021.101781
Granger, C. W. (1980). Long memory relationships and the aggregation of dynamic models.
Journal of Econometrics, 14(2), 227–238. https://doi.org/10.1016/0304-4076(80)90092-5
Granger, C. W., & Joyeux, R. (1980). An introduction to long-memory time series models and
fractional differencing. Journal of Time Series Analysis, 1(1), 15–29. https://doi.org/10.1111/j.
1467-9892.1980.tb00297.x
Hamilton, J. D. (1989). A new approach to the economic analysis of nonstationary time series
and the business cycle. Econometrica, 57(2), 357–384. https://doi.org/10.2307/1912559
Harwick, C. (2016). Cryptocurrency and the problem of intermediation. The Independent
Review, 20(4), 569–588.
Hernandez, J. A., Hasan, M. Z., & McIver, R. P. (2022). Bitcoin, gold, and the VIX: Short-and
long-term effects of economic policy uncertainty. Applied Economics Letters, 1–5.
Hosking, J. R. M. (1981). Fractional differencing. Biometrika, 68(1), 165–176. https://doi.org/
10.1093/biomet/68.1.165
Hsieh, D., & LeBaron, B. (1988). Small sample properties of the BDS statistic, I. Graduate school
of business. University of Chicago.
Huang, W. (2019). The impact on people’s holding intention of bitcoin by their perceived risk
and valuec. Economic Research-Ekonomska Istrazivanja, 32(1), 3570–3585. https://doi.org/10.
1080/1331677X.2019.1667257
Iwamura, M., Kitamura, Y., Matsumoto, T., & Saito, K. (2019). Can we stabilize the price of a
Cryptocurrency?: Understanding the design of Bitcoin and its potential to compete with
Central Bank money. Hitotsubashi Journal of Economics, 60(1), 41–60.
Jare~no, F., de la, O., Gonzalez, M., Tolentino, M., & Sierra, K. (2020). Bitcoin and Gold price
returns: A quantile regression and NARDL analysis. Resources Policy, 67, 101666. https://doi.
org/10.1016/j.resourpol.2020.101666
Jiang, S., Li, Y., Lu, Q., Wang, S., & Wei, Y. (2022). Volatility communicator or receiver?
Investigating volatility spillover mechanisms among Bitcoin and other financial markets.
Research in International Business and Finance, 59, 101543. https://doi.org/10.1016/j.ribaf.
2021.101543
Kapetanios, G. (2005). Unit-root testing against the alternative hypothesis of up to m structural
breaks. Journal of Time Series Analysis, 26(1), 123–133. https://doi.org/10.1111/j.1467-9892.
2005.00393.x
Kapetanios, G., & Shin, Y. (2008). GLS detrending-based unit root tests in nonlinear STAR
and SETAR models. Economics Letters, 100(3), 377–380. https://doi.org/10.1016/j.econlet.
2008.02.033
Kapetanios, G., Shin, Y., & Snell, A. (2003). Testing for a unit root in the nonlinear STAR
framework. Journal of Econometrics, 112(2), 359–379. https://doi.org/10.1016/S0304-
4076(02)00202-6
Karamti, C., & Belhassine, O. (2021). COVID-19 pandemic waves and global financial markets:
Evidence from wavelet coherence analysis. Finance Research Letters, 45, 1–9, 102136.
Klein, T. (2017). Dynamic correlation of precious metals and flight-to-quality in developed
markets. Finance Research Letters, 23, 283–290. https://doi.org/10.1016/j.frl.2017.05.002
Klein, T., Thu, H. P., & Walther, T. (2018). Bitcoin is not the New Gold–A comparison of
volatility, correlation, and portfolio performance. International Review of Financial Analysis,
59, 105–116. https://doi.org/10.1016/j.irfa.2018.07.010
Kristoufek, L. (2013). Bitcoin meets Google Trends and Wikipedia: Quantifying the relation-
ship between phenomena of the Internet era. Scientific Reports, 3, 3415. https://doi.org/10.
1038/srep03415
Kristoufek, L. (2015). What are the main drivers of the Bitcoin price? Evidence from wavelet
coherence analysis. PloS One, 10(4), e0123923. https://doi.org/10.1371/journal.pone.0123923
Kuan, C. M. (2002). Lecture on the Markov switching model. Institute of Economics Academia
Sinica, 8(15), 1–30.
Lee, A. D., Li, M., & Zheng, H. (2020). Bitcoin: Speculative asset or innovative technology?
Journal of International Financial Markets, Institutions and Money, 67, 101209. https://doi.
org/10.1016/j.intfin.2020.101209
Lewis, N. (2017). What is the fundamental value of bitcoin? Retrieved February 2, 2019, from.
https://www.forbes.com/sites/nathanlewis/2017/12/07/what-is-the-fundamental-value-ofBitcoin/
2/#123d85343c08
Leybourne, S., Newbold, P., & Vougas, D. (1998). Unit roots and smooth transitions. Journal
of Time Series Analysis, 19(1), 83–97. https://doi.org/10.1111/1467-9892.00078
Liu, K., Chen, Y., & Zhang, X. (2017). An evaluation of ARFIMA (autoregressive fractional
integral moving average) programs. Axioms, 6(4), 16. https://doi.org/10.3390/axioms6020016
Lowe, P. (2017, December 13). An eAUD? Address to the 2017 Australian Payment Summit Sydney.
Reserve Bank of Australia. https://www.rba.gov.au/speeches/2017/sp-gov-2017-12-13.html
Luis, P., de la Fuente, G., & Perote, J. (2019). The drivers of Bitcoin demand: A short and
long-run analysis. International Review of Financial Analysis, 62, 21–34.
Lumsdaine, R. L., & Papell, D. H. (1997). Multiple trend breaks and the unit-root hypothesis.
Review of Economics and Statistics, 79(2), 212–218. https://doi.org/10.1162/003465397556791
Ma, D., & Tanizaki, H. (2022). Intraday patterns of price clustering in Bitcoin. Financial
Innovation, 8(1), 1–25. https://doi.org/10.1186/s40854-021-00307-4
Mokni, K., Bouri, E., Ajmi, A. N., & Vo, X. V. (2021). Does bitcoin hedge categorical eco-
nomic uncertainty? A quantile analysis. SAGE Open, 11(2), 215824402110163. https://doi.
org/10.1177/21582440211016377
Nakamoto, S. (2008). Bitcoin: A peer-to-peer electronic cash system. https://Bitcoin.org/Bitcoin.pdf
Ngene, G., Post, J., & Mungai, A. (2018). Volatility and shock interactions and risk manage-
ment implications: Evidence from the US and frontier markets. Emerging Markets Review,
37, 181–198.
26 A. RASHID ET AL.
Nguyen, T., de Bodisco, C., & Thaver, R. (2018). Factors affecting Bitcoin price in the crypto-
currency market: An empirical study. International Journal of Business & Economics
Perspectives, 13(1), 106–126.
Obeng, C. (2021). Measuring value at risk using GARCH model - evidence from the crypto-
currency market. International Journal of Entrepreneurial Knowledge, 9(2), 63–84. https://
doi.org/10.37335/ijek.v9i2.133
Pal, D., & Mitra, S. K. (2019). Hedging Bitcoin with other financial assets. Finance Research
Letters, 30, 30–36. https://doi.org/10.1016/j.frl.2019.03.034
Palombizio, E., & Morris, I. (2012). Forecasting exchange rates using leading economic indica-
tors. Open Access Scientific Reports, 1, 402. https://doi.org/10.4172/scientificreports.402
Park, B. J. (2022). The COVID-19 pandemic, volatility, and trading behavior in the bitcoin
futures market. Research in International Business and Finance, 59, 101519. https://doi.org/
10.1016/j.ribaf.2021.101519
Perron, P. (1989). The great crash, the oil price shock, and the unit root hypothesis.
Econometrica, 57(6), 1361–1401. https://doi.org/10.2307/1913712
Raskin, M., & Yermack, D. (2016). Digital currencies, decentralized ledgers, and the future of
central banking (No. w22238). National Bureau of Economic Research.
Raymaekers, W. (2015). Cryptocurrency Bitcoin: Disruption, challenges and opportunities.
Journal of Payments Strategy & Systems, 9(1), 30–46.
Rogoff, K. (2018, December 10). Cryptocurrencies are like lottery tickets that might pay off in
future. The Guardian. https://www.theguardian.com/business/2018/dec/10/cryptocurrencies-
Bitcoin-kenneth-rogoff
Sahoo, P. K. (2017). Bitcoin as digital money: Its growth and future sustainability. Theoretical
& Applied Economics, 24(4), 53–64.
Samah, H., Wajdi, M., & Regaïeg, R. (2018). Dynamic linkages among Bitcoin, Gold prices
and exchange rates of US Dollar in JPY, GBP and CNY: DCC EGARCH approach. Journal
of Academic Research in Economics, 10(2), 239–247.
Schuh, S., & Shy, O. (2016). US consumers’ adoption and use of Bitcoin and other virtual cur-
rencies. In DeNederlandsche Bank, Conference Entitled “Retail Payments: Mapping Out the
Road Ahead.
Shahzad, S. J. H., Bouri, E., Rehman, M. U., & Roubaud, D. (2022). The hedge asset for BRICS
stock markets: Bitcoin, gold or VIX. The World Economy, 45(1), 292–316. https://doi.org/10.
1111/twec.13138
Shahzad, S. J. H., Bouri, E., Roubaud, D., & Kristoufek, L. (2020). Safe haven, hedge and diver-
sification for G7 stock markets: Gold versus Bitcoin. Economic Modelling, 87, 212–224.
https://doi.org/10.1016/j.econmod.2019.07.023
Shahzad, S. J. H., Bouri, E., Roubaud, D., Kristoufek, L., & Lucey, B. (2019). Is Bitcoin a better
safe-haven investment than Gold and commodities? International Review of Financial
Analysis, 63, 322–330. https://doi.org/10.1016/j.irfa.2019.01.002
Sheng, H., Chen, Y., & Qiu, T. (2011). Fractional processes and fractional-order signal process-
ing: Techniques and applications. Springer Science & Business Media.
Sontakke, K. A., & Ghaisas, A. (2017). Cryptocurrencies: A developing asset class.
International Journal of Business Insights & Transformation, 10(2), 10–17.
Su, C. W., Xi, Y., Tao, R., & Umar, M. (2022). Can Bitcoin be a safe haven in fear sentiment?
Technological and Economic Development of Economy, 28(2), 268–289. https://doi.org/10.
3846/tede.2022.15502
Symitsi, E., & Chalvatzis, K. J. (2019). The economic value of Bitcoin: A portfolio analysis of
currencies, Gold, oil and stocks. Research in International Business and Finance, 48, 97–110.
https://doi.org/10.1016/j.ribaf.2018.12.001
Tiwari, N. (2018). The commodification of cryptocurrency. Michigan Law Review, 117(3), 611.
https://doi.org/10.36644/mlr.117.3.commodification
Tschorsch, F., & Scheuermann, B. (2016). Bitcoin and beyond: A technical survey on decen-
tralized digital currencies. IEEE Communications Surveys & Tutorials, 18(3), 2084–2123.
https://doi.org/10.1109/COMST.2016.2535718
Yelowitz, A., & Wilson, M. (2015). Characteristics of Bitcoin users: An analysis of Google
search data. Applied Economics Letters, 22(13), 1030–1036. https://doi.org/10.1080/13504851.
2014.995359
Yermack, D. (2014). Is Bitcoin a real currency? An economic appraisal. In D. Lee Kuo Chuen
(Ed.), The handbook of digital currency (pp. 31–44). Elsevier.
Yin, L., Nie, J., & Han, L. (2021). Understanding cryptocurrency volatility: The role of oil mar-
ket shocks. International Review of Economics & Finance, 72, 233–253. https://doi.org/10.
1016/j.iref.2020.11.013
Zivot, E., & Andrews, D. W. K. (1992). Further evidence on the great crash, the oil-price
shock, and the unit-root hypothesis. Journal of Business & Economic Statistics, 10(3),
251–270. https://doi.org/10.2307/1391541

Are Cryptocurrencies A Future Safe Haven For Inves

Uploaded by

Copyright:

Available Formats

Are Cryptocurrencies A Future Safe Haven For Inves

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Are Cryptocurrencies A Future Safe Haven For Inves

Uploaded by

Copyright:

Available Formats

Economic Research-Ekonomska Istraživanja

ISSN: (Print) (Online) Journal homepage: https://www.tandfonline.com/loi/rero20

Are cryptocurrencies a future safe haven for

Audil Rashid, Walid Bakry & Somar Al-Mohamad

To link to this article: https://doi.org/10.1080/1331677X.2022.2140443

© 2022 The Author(s). Published by Informa

Published online: 02 Nov 2022.

Submit your article to this journal

View related articles

View Crossmark data

Full Terms & Conditions of access and use can be found at

Are cryptocurrencies a future safe haven for investors?

ABSTRACT ARTICLE HISTORY

CONTACT Walid Bakry w.bakry@westernsydney.edu.au

2. Review of the literature

analyze the asymmetric relationship between the regime-dependent variables. The

4.1. The Kapetanios and Shin (K & S) unit root test

4.2. The ARFIMA model

4.3. The Brock, Dechert and Scheinkman (BDS) test

where IXtm , Xsm denotes an indicator function and is given by:

PðjXt Xs j < e, jXt1 Xs1 j < e, . . . :, jXtmþ1 Xsmþ1 j < eÞ

If the Xt are IID, the probability equals the following:

C1, T ðeÞm ¼ PðjXt Xs j<eÞm

The BDS statistic as defined by Brock et al. (1987) is given by:

pffiffiffiffi Cm, T ðeÞ C1, T ðeÞm

4.4. The Markov Regime Switching Regression (MRSR) model

Pðst ¼ 1 j st1 ¼ 1Þ Pðst ¼ 2 j st1 ¼ 1Þ p11 p12

Table 1. Kapetanios and Shin (2008) unit root test.

Table 2. ARFIMA structure of the returns series.

different volatility characteristics in comparison to ordinary fiat currencies (USD/

Table 3. BDS test results.

Table 4. Markov regime switching regression.

Figure 1. Dynamic conditional correlation graphs (DCC-GARCH).

Table 5. Transition probabilities.

Bitcoin may provide considerable potential for portfolio diversification, in general, as

Figure 2. Filtered regime probabilities.

CRediT authorship contribution statement

You might also like

Pfad - The Proxy pFad of © 2024 Garber Painting. All rights reserved.