Can Bitcoin Become a Viable Alternative to Fiat [629171]

Can Bitcoin Become a Viable Alternative to Fiat
Currencies? An empirical analysis of Bitcoin's volatility
based on a GARCH model
Vavrinec Cermak
Skidmore College, Saratoga Springs, NY
Abstract
This study examines whether Bitcoin, a digital decentralized currency, can
become a viable alternative to at currencies. Bitcoin currently does not
fulll the criteria of being a currency because it does not function as a medium
of exchange, a unit of account, and a store of value. Bitcoin's biggest obstacle
from fullling these functions is the price volatility. A GARCH(1,1) model is
used to analyze Bitcoin's volatility in respect to the macroeconomic variables
of countries where Bitcoin is being traded the most. Bitcoin already behaves
similarly to at currencies in China, the U.S. and the European Union but
not in Japan. There is also evidence that Bitcoin acts as a safe-haven asset
in China. The volatility of Bitcoin has been steadily decreasing throughout
its lifetime. If it follows the trend of its six years of existence, it will reach
the volatility levels of at currencies in 2019-2020 and become a functioning
alternative to at currencies.
Keywords: Bitcoin, volatility, cryptocurrency, GARCH
Preprint submitted to Skidmore College May 13, 2017

Contents
List of Figures 3
List of Tables 3
Abbreviations 4
1 Theoretical part 5
1.1 Origin of Bitcoin . . . . . . . . . . . . . . . . . . . . . . . . . 5
1.2 Brief History . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
1.3 Bitcoin's Supply . . . . . . . . . . . . . . . . . . . . . . . . . . 8
1.4 Schools of Economic Thought . . . . . . . . . . . . . . . . . . 11
1.4.1 Austrian School of Economics . . . . . . . . . . . . . . 11
1.4.2 Keynesian Economics . . . . . . . . . . . . . . . . . . . 14
1.5 Currency Criteria . . . . . . . . . . . . . . . . . . . . . . . . . 16
1.5.1 Medium of Exchange . . . . . . . . . . . . . . . . . . . 16
1.5.2 Unit of Account . . . . . . . . . . . . . . . . . . . . . . 18
1.5.3 Store of Value . . . . . . . . . . . . . . . . . . . . . . . 19
1.5.4 Overview of Bitcoin as a Currency . . . . . . . . . . . 20
1.6 Bitcoin's volatility . . . . . . . . . . . . . . . . . . . . . . . . 21
1.6.1 Government Regulations . . . . . . . . . . . . . . . . . 21
1.6.2 Security Concerns . . . . . . . . . . . . . . . . . . . . . 23
1.6.3 Depreciations of Fiat Currencies . . . . . . . . . . . . . 23
1.7 Bitcoin's Weaknesses and Altcoins . . . . . . . . . . . . . . . . 25
1.8 Literature Review . . . . . . . . . . . . . . . . . . . . . . . . . 28
2 Practical part 30
2.1 Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
2.2 Methodology . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
2.3 Expectations . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
2.4 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
3 Conclusion 47
2

List of Figures
1 Amount of bitcoins in existence . . . . . . . . . . . . . . . . . 10
2 Bitcoin trading volume by currencies . . . . . . . . . . . . . . 22
3 BTC surges when CNY falls . . . . . . . . . . . . . . . . . . . 24
4 Metcalfe's Law of Bitcoin . . . . . . . . . . . . . . . . . . . . . 26
5 Bitcoin Price Index . . . . . . . . . . . . . . . . . . . . . . . . 30
6 Daily log returns of Bitcoin prices . . . . . . . . . . . . . . . . 31
7 Bitcoin trading volume by the most traded currencies . . . . . 32
8 Bitcoin trading volume by currencies without CNY . . . . . . 33
9 A comparison of log returns of BTC, CNY, EUR and JPY . . 35
10 A comparison of log returns of BTC, CNY, EUR and JPY on
the same scale . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
11 A test of clustering volatility in the residuals . . . . . . . . . . 41
12 Conditional variance of daily Bitcoin log-returns . . . . . . . . 45
13 Daily volatility of Bitcoin . . . . . . . . . . . . . . . . . . . . 46
List of Tables
1 Amount of bitcoins in existence in the short term . . . . . . . 9
2 Cryptocurrency Market Capitalizations over $100k . . . . . . 25
3 Explanatory macroeconomic variables . . . . . . . . . . . . . . 34
4 Summary Statistics . . . . . . . . . . . . . . . . . . . . . . . . 37
5 ALM test for ARCH . . . . . . . . . . . . . . . . . . . . . . . 42
6 GARCH(1,1), dependent variable return on bitcoin. . . . . . . 43
3

Abbreviations
ARCH Autoregressive Conditional Heteroskedasticity
BTC bitcoin (unit of currency)
CNY Chinese Yuan
EIA U.S. Energy Information Administration
ETF Exchange-Traded Fund
EUR Euro
FinCEN Financial Crimes Enforcement Network
GARCH Generalized AutoRegressive Conditional Heteroskedasticity
GH/s Gigahash per second – primary measure of a Bitcoin \miner's" per-
formance
IRS Internal Revenue Service
JPY Japanese Yen
MSB Money Service Business
PBoC People's Bank of China
SEC U.S. Securities and Exchange Commission
USD U.S. dollar
Wwatt (unit of power)
4

1. Theoretical part
1.1. Origin of Bitcoin
Bitcoin1was rst mentioned in 2008 in a self-published white paper writ-
ten by an anonymous person (or people) who went by the pseudonym of
Satoshi Nakamoto [1]. In the paper, Nakamoto described Bitcoin as a pay-
ment system, which uses a decentralized peer-to-peer cryptocurrency with a
nite monetary supply. There is a theory that Nakamoto created Bitcoin as
a response to the global nancial crisis in 2008 [2]. Before the introduction of
Bitcoin, online payments relied exclusively on nancial institutions that act
as trusted third parties to facilitate electronic payments [1]. Nakamoto was
the rst developer who solved the problem of replacing the trust model with
cryptographic proof, which ultimately eliminated the need of an intermedi-
ary and replaced it with a peer-to-peer network. Nakamoto also solved the
problem of double spending by recording all transactions on a public time-
stamped ledger called the blockchain [1]. All transactions that happen on
Bitcoin are veried by network nodes, which are essentially computers that
are connected to the Bitcoin network. These network nodes are called the
\miners" because they are compensated in bitcoin for verifying transactions.
Bitcoin compensation is an incentive that keeps the Bitcoin network running.
Therefore, the nodes are theoretically \mining" bitcoin in exchange for their
computing power. One year after publishing the white paper, Nakamoto re-
leased Bitcoin as an open-source software and made it available for anyone
to use2.
1.2. Brief History
Even before Bitcoin, the idea of a digital currency was nothing new.
David Chaum, a pioneer for cryptographic protocols, wrote the rst paper
that outlined an anonymous payment system by using blind signatures [3].
Since then, cryptographers published several academic papers attempting to
improve the security and eciency of the hypothetical digital currencies [4].
Some of these ideas came to life in forms of independent digital currencies –
Digicash, E-Gold, Flooz and Beenz. However, the lack of decentralization,
1For the remainder of this paper, I will use Bitcoin with a capital \B" when talking
about the protocol and payment system. I will use lower-case \b" when talking about the
denomination of the currency
2www.github.com/bitcoin/bitcoin
5

transparency and security ultimately led to the extinction of all of these
attempts.
After the release of the open-source code of Bitcoin, the rst 50 BTC were
\mined" by Nakamoto himself in order to demonstrate the method to online
observers [5]. The rst Bitcoin real world transaction took place in May
of 2010 when Laszlo Hanyecz, a programmer living in Florida, sent 10,000
BTC to a volunteer in the United Kingdom, who then ordered two pizzas
for Hanyecz, which cost him 25 USD [5]. Today, 10,000 BTC have value
of over 10 million USD. The interest in Bitcoin, especially among computer
enthusiasts, led to the creation of Mt. Gox, the rst bitcoin exchange. Prior
to the creation of Mt. Gox, users could only \mine" their own bitcoins.
However, the introduction of a bitcoin exchange meant that the users could
now buy bitcoins in exchange for at currencies [4]. On the rst day of
trading, Mt. Gox sold 20 BTC and oered one BTC for the rate of 4.96 cents
[5].
The rst and only vulnerability was discovered in the Bitcoin verica-
tion mechanism in 2010, which allowed hackers to extract large amounts of
bitcoin. However, the code was quickly patched and the falsely \mined" bit-
coins were deleted from the blockchain. Ever since then, there has not been a
single vulnerability in the Bitcoin verifying mechanism or Bitcoin blockchain.
Moreover, the Bitcoin blockchain has never been hacked according to a for-
mer White House communicator Jamie Smith [6] and it is thought to be
virtually unhackable.
In 2011, Bitcoin started to become more widely utilized. Its popularity
was sparked because Bitcoin is global, anonymous, has low transaction costs,
accounts cannot be frozen, and there are no prerequisites or arbitrary limits
[7]. Controversial organizations such as Wikileaks started accepting bitcoins
as anonymous donations [7]. Financial institutions such as PayPal, Visa,
Mastercard, which were susceptible to government pressure, cut o their
services to Wikileaks because of its controversial political aliations. As
a result, this made Bitcoin a perfect option to use as a donation system.
However, its anonymity also started attracting illegal activities including
black markets that sold illegal goods online. The operation of a darknet
marketplace called \Silk Road" was discovered, which created a nightmare for
the FBI because it was virtually impossible to track the illegal transactions.
Even though \Silk Road" was shut down by the FBI in 2014, there are now
several replacements, which still use Bitcoin as a payment system [8]. On
February 9, 2011, Bitcoin brie
y reached parity with USD on Mt. Gox [9].
6

The interest in Bitcoin skyrocketed after the Time magazine wrote one of the
rst mainstream articles describing the digital currency in April of 2011 [10].
Later in 2011, the rst conferences about Bitcoin took place in New York
City and the rst European conference was held in Prague, Czech Republic
[9].
In 2012, several Bitcoin startups such as Coinbase began to form. They
were aimed at assisting nontechnical users in becoming familiarized with us-
ing Bitcoin [11]. Coinbase created an online, technologically friendly Bitcoin
wallet, which let users purchase bitcoins in exchange for USD and then store
the bitcoins online. Moreover, by the end of 2012, a French company called
Bitcoin-Central became the rst bitcoin exchange to be licensed to operate
as a bank [12]. Bitcoin-Central also functioned within the framework of the
European Union, which meant that customers' funds were held under their
name rather than that of the exchange.
In 2013, FinCEN established regulatory guidelines for \decentralized vir-
tual currencies", which classied Bitcoin \miners" and exchanges in the U.S.
as MSB [9]. All MSBs are required to be registered and are subject to legal
requirements such as disclosing large transactions or suspicious activities. As
a result, Mt. Gox started dealing with serious legal issues. In May of 2013,
the U.S. Department of Homeland Security seized more than 5 million USD
from Mt. Gox's U.S. accounts because Mt. Gox had not registered as a MSB
with FinCEN [9]. Mt. Gox received an MSB license in June of 2013 but
ever since then, its users started experiencing long delays of their USD with-
drawals [9]. Mt. Gox, was handling 70% of all Bitcoin trading [13], which
means that even a slight issue with Mt. Gox resulted in signicant price
uc-
tuations of bitcoin. When diculties with Mt. Gox became apparent, more
and more people tried to withdraw their funds. Nearly one million customers
lost trust in Mt. Gox resulting in a digital bank run. There were specula-
tions that these issues were caused by a system of fractional reserves and
that Mt. Gox simply did not have enough reserves to satisfy all withdrawals
[14]. Towards the end of February, Mt. Gox announced that the company
led for bankruptcy protection. The company claimed that an error allowed
hackers to steal more than 850,000 BTC, worth around 480 million USD
[15]. Mt. Gox had 127,000 creditors, a debt of 6.5 billion CNY ( 63.7 mil-
lion USD), and assets of 3.84 billion CNY [16]. Because Mt. Gox was located
in Tokyo where bitcoin exchanges were unregulated, customers had no legal
protection. Moreover, digital currencies are not backed by the central banks.
This event had crippled faith in Bitcoin for many investors and the price
7

plummeted.
On a more positive note, the rst Bitcoin ATM was launched in Vancouver
in 2013 [17]. The ATM let customers withdraw their bitcoins in Canadian
dollars and it also let them deposit Canadian dollars, which were quickly
converted to their Bitcoin wallets. Currently, there are 956 Bitcoin ATMs
in 55 countries3. Additionally, several retailers began accepting Bitcoin as
a means of payment; notably Overstock.com, Reddit, OKCupid, WordPress,
Microsoft, Expedia and Virgin Atlantic [18]. By the end of 2014, the IRS an-
nounced that it would be treating Bitcoin as an asset rather than a currency
meaning that any capital gains from selling or \mining" bitcoin are treated
the same way as selling shares4.
1.3. Bitcoin's Supply
In a centralized economy, at currency has, theoretically, an unlimited
supply. It is at a central bank's discretion to inject money into or withdraw
money from the banking system in order to match the growth of the economy.
The central banks have distinct instruments of monetary policy, by which
they control the monetary base. The most common methods are the open
market operations, quantitative easing, modifying the reserve requirements
and changing the federal discount rate.
In a decentralized monetary system, no central authority can regulate the
monetary base; therefore, monetary policy would be ineective. Bitcoins are
created by the \miners" around the world as opposed to the central bank.
The Bitcoin's algorithm denes, in advance, at what rate the currency will
be created [1]. A new bitcoin is created when a \miner" discovers a block.
A block is a le where transaction data is permanently recorded. Therefore,
a blockchain is a perpetual store of records. New blocks are discovered when
`miners" solve cryptographic mathematical problems, by which they also
verify the transactions. Every year5, the number of blocks discovered is
constant at 52,500. However, the number of bitcoins generated per one block
halves every 210,000 blocks (roughly four years). This decreasing-supply
algorithm is supposed to replicate the rate at which commodities such as gold
are mined. Because of the reward halving every four years, the maximum
3www.coinatmradar.com
4https://www.irs.gov/pub/irs-drop/n-14-21.pdf
5On average 343 days. The length varies depending on the mining power and network
diculty. In the future, it could take signicantly less time.
8

amount of bitcoins that can ever exist is 21,000,000. At the time of writing,
there were approximately 16,250,000 bitcoins in existence. The nal amount
of Bitcoin can be calculated by Equation 1.
P32
i=0210000
(50
108
2i)
108(1)
Table 1 shows the number of bitcoins in existence in the short term.
However, it is more dicult to estimate the amount of bitcoins in circulation,
which are the bitcoins that are physically used to conduct transactions rather
than stored somewhere. In 2012, Ron and Shamir analyzed the full available
history of Bitcoin transactions and found that 78% of all bitcoins were stored
in accounts that never made a transaction. That implies that only 22% of
existing bitcoins were in active circulation [19]. In 2014, Swanson conducted
a similar blockchain analysis and found that 70% of all bitcoins had not
moved in six or more months implying that about 30% of bitcoins were in
active circulation [20].
Year Date Blocks BTC/Block BTC added BTC in circ.
1 1/2009 0 50 2,625,000 2,625,000
2 4/2010 52,500 50 2,625,000 5,250,000
3 1/2011 105,000 50 2,625,000 7,875,000
4 12/2011 157,500 50 2,625,000 10,500,000
5 11/2012 210,000 25 1,312,500 11,812,500
6 10/2013 262,500 25 1,312,500 13,125,000
7 8/2014 315,000 25 1,312,500 14,437,500
8 7/2015 367,500 25 1,312,500 15,750,000
9 7/2016 420,000 12.5 656,250 16,406,250
Table 1: Amount of bitcoins in existence in the short term
Currently, it is estimated that 99% of all bitcoins will be \mined" by 2040
and the remaining 1% will be \mined" in the next 80-100 years [5]. Figure 1
shows the rate, at which the remaining bitcoins will be added.
9

Figure 1: Amount of bitcoins in existence
However, it is impossible to predict the exact year without knowing future
advances in mining power and network diculty. The diculty of solving the
cryptographic problems increases when more \miners" are competing for a
xed amount of bitcoins. In 2009, an individual with a personal laptop could
\mine" 200 BTC in a few days but in 2014, it would take about 98 years
to \mine" just one bitcoin on a personal computer [21]. Thus, most of the
mining right now happens in specialized mining pools with power ecient
supercomputers situated in a place with cheap electricity. Because of these
reasons, over 70% of Bitcoin mining pools are located in China [22]. The
protability of new \miners" entering the market decreases because of the
reward halving and the lack of economies of scale, which creates a barrier
to entry that could lead to a centralization of Bitcoin [23]. Bitcoin has a
so-called positive feedback loop, which means that when the price of bitcoin
decreases or the halving event occurs, \mining" becomes less protable and
some \miners" will be forced out of business. When there are less miners,
the network automatically adjusts to decrease the diculty of the crypto-
10

graphic problems and therefore makes \mining" protable again. In other
words, Bitcoin is designed to make \mining" barely protable on average.
Valfells and Egilsson found a break-even point, at which it makes economic
sense for independent \miners" to enter the market. For the current reward
(12.5 BTC/block), which will be in eect for at least the next three years,
the breakeven point is 600 USD/BTC [23]. In 2016, The New York Times
reported that three Chinese mining companies processed over half of the Bit-
coin's computing power, which technically gave them an ability to veto over
any changes to the Bitcoin software and technology [22]. If the leading min-
ing pools decided to cooperate6, they could theoretically alter the original
code and make macroeconomic decisions similar to the central banks. How-
ever, such cooperation is unlikely to be formed because Bitcoin's users would
stop using the altered version of Bitcoin in favor of the older version and
\miners" would lose all the prots. Any alteration of Bitcoin's code would
result in a fork, which means that the network would split into two separate
Bitcoin networks. At the time of writing this has yet to occur, but it is a
possibility in the future.
Because bitcoins are added at a decreasing rate and there is a nite
amount that will ever exist, rational people will expect the value to increase
in the future. Therefore, they will tend to hold the bitcoins in the short term
rather than spending it. Bitcoin is theoretically disin
ationary, which means
that the in
ation rate is slowing down every four years until all the remaining
bitcoins are \mined" and then it becomes de
ationary. De
ation in at cur-
rencies causes economic activity to slow down because it encourages saving,
and discourages consumption. Therefore, Bitcoin can act as an alternative
to at currencies but it cannot fully replace at currencies because it would
not be able to match the growth of the economy.
1.4. Schools of Economic Thought
In order to determine whether bitcoin can become a viable alternative to
at currencies, I will look at two schools of economic thought – Austrian and
Keynesian. I will analyze how these frameworks view Bitcoin in theory.
1.4.1. Austrian School of Economics
In 2012, the European Central Bank (ECB) published a study of the
decentralized cryptographic money, in which they say: \ the theoretical roots
6Assuming that they would have more than 51% of Bitcoin's mining power combined
11

of Bitcoin can be found in the Austrian school of economics " [24].
One of the cornerstones of the Austrian School is the Austrian business
cycle theory, which originated from economists Friedrich Hayek and Ludwig
von Mises. The theory suggests that the occurrence of business cycles in the
economy is not a natural phenomenon but rather that it is caused by the
central banks and fractional reserve banking. According to this theory, the
business cycle is the consequence of the government's manipulation of the
money supply and the interest rates. The idea behind it is that when the
central banks keep the interest rates articially below where the free market
rate would have been, it sends a false signal to the productive segments of
the economy. According to Ludwig von Mises, the false signals result in
in badly allocated business investments – malinvestments [25]. According
to Mises and Hayek, the only way the interest rates would decrease in a
free economy is if there was an increase in savings [26]. This would cause
businesses to have more money available to nance capital investments for
long term purposes. Moreover, Mises and Hayek believed that if a central
bank starts stimulating the economy and articially lowers the interest rates,
it creates an incentive for businesses to nance capital investments for long
term purposes even if there was not an increase in savings. This causes
malinvestment and overconsumption in the economy, which results in an
economic boom, followed by a crisis, and the economy would ultimately try
to correct itself.
In 19th and the beginning of the 20th century, the world's most successful
currencies were easily exchangeable for a xed amount of gold or a precious
metal. Mises formulated an argument as \the regression theorem" of money,
which says:
\People will only accept a medium of exchange if they observe that it has
value, and can actually be exchanged for things. The only way to observe that
is by looking at whether it was so used in a preceding time period. Thus, this
chain of observations can be followed back until the rst instance in which a
particular type of money was used as a medium of exchange, and in order for
those rst adopters to accept it, it must have had value independent of its use
as a medium of exchange, or in other words, be a commodity. Paper money,
especially that with no commodity backing, is only adopted when governments
force it upon people. " [27]
However, the gold standard collapsed in most economies between 1920's
12

and 1970's because the economic growth outpaced the production of gold
and the governments could not keep their promise to exchange the currency
for the xed amount of gold. Therefore, the vast majority of currencies in
the world today are at, which means that they are not backed by a physical
commodity. The value of at currencies relies directly on the public belief
that the central bank will not increase the supply too rapidly and therefore
overin
ate the value of the currency to worthlessness [5]. The Austrian school
claims that currency is only possible if governments monopolize the issuance
of a at currency [28].
In 1976, Hayek published an essay, in which he described his vision of
competition in currency markets; not just competition between the countries
but also within the countries [29]. He believed people should be free to choose
any currency they wish to use. Hayek argued that a free-market monetary
system would provide a check against in
ation. His reasoning was that the
government would have to keep the in
ation low, otherwise people would
simply switch to a dierent currency.
In theory, the Austrian School of economic thought should nd Bitcoin to
be an intriguing currency because it introduces a possibility of disrupting the
monopolization of the issuance of at currencies and because it has potential
to weaken the power of central banks. If Bitcoin continues to attract more
interest and if it becomes more widely used, it could introduce a viable alter-
native currency to people, who could switch to Bitcoin if the in
ation of their
domestic currency was too high. In fact, we already see this phenomenon,
but on a much smaller scale. The predetermined issuance of new bitcoins
and their nite monetary base is also attractive to Austrian economists be-
cause it presents certainty of the monetary supply, which they prefer over
the uncertainty of central banks' manipulation of the interest rates.
However, even the bitcoin exchanges (e.g. Mt. Gox) could sell more bit-
coins than they actually have and hope that there will not be a digital
bank run. Fractional reserve banking could theoretically increase the Bit-
coin money supply but not the amount of bitcoins on the blockchain. It is
impossible to increase the bitcoin supply on the blockchain because it is pre-
determined and created by the \miners". It is important to note, however,
that people would rather keep their bitcoins in private wallets rather than
allowing fractional banking because of Bitcoin's disin
ationary nature. The
only reason why Mt. Gox was able to sell more bitcoins than they had is
because their customers were not aware of it. All of the future eorts to
create a transparent service of fractional lending with bitcoins failed because
13

the Bitcoin community is skeptical of fractional reserve banking; especially
without the oversight of an authority.
Despite these factors, the majority of economists supporting the Aus-
trian school of economics are critical of Bitcoin and other cryptocurrencies
[30][31][32]. The criticism stems from the belief that Bitcoin violates Mise's
regression theorem of money because it is not backed by a commodity. How-
ever, there are some Austrian economists who are pro-Bitcoin such as Konrad
Graf [33] and Peter Surda [34]. Surda and Graf have a dierent interpreta-
tion of the regression theorem. They believe that even though Bitcoin is not
backed by a tangible commodity, it is a scarce intangible good, which satis-
es the same requirements as a commodity. Therefore, according to Surda
and Graf, Bitcoin satises the Austrian denition of a currency because it
functions as a medium of exchange. I will examine whether Bitcoin satises
the criteria of being a currency in the next chapter. It is interesting to note
that most of the Bitcoin enthusiasts are supporters of Austrian economics,
while the majority of Austrian economists are critical of Bitcoin.
1.4.2. Keynesian Economics
Central banks around the world use Keynesian models in order to set
appropriate interest rates. Essentially, the Keynesian view of the business
cycles is that the
uctuations are caused by the changes in the aggregate
demand. When aggregate demand decreases, producers are not able to pro-
duce as much as previously and they must adjust in the short run. However,
adjusting by cutting the production costs is dicult because wages are said
to be sticky-down, which means that they can move up easily but it is harder
to move them down. The producers are therefore forced to re their work-
ers rather than to decrease their wages, which translates to higher levels of
unemployment. If less people are working, spending also decreases and ag-
gregate demand is pushed down even further. The Keynesian model says
that when the unemployment rate is higher than the natural unemployment
rate, change in scal and monetary policies can help reduce the economic

uctuations. The natural unemployment rate is the lowest unemployment
rate attainable in the economy. The relationship between unemployment
and in
ation in the short run is described by the Phillips curve, which indi-
cates that there is an inverse relationship between unemployment rates and
in
ation.
In his book The General Theory of Employment, Interest and Money,
John Maynard Keynes rst developed the concept of liquidity preference. He
14

argued that individuals value money for \ the transaction of current business
and its use as a store of wealth " [35]. Keynes also suggested that individuals
give up their interest in order to spend money in the present, but also as a
precautionary measure. Moreover, he saw interest as more of a reward for
giving up liquidity rather than as a reward for saving. The three motives,
which aect the demand for liquidity are: the transaction motive, the pre-
cautionary motive, and the speculative motive. Because of these motives,
the demand for money
uctuates based on the current interest rate.
It is accepted by economists that the majority of users currently treat
their bitcoins as a speculative asset rather than as a means of payment
[36][19][37]. In a study that looked at blockchain data from 2009 to 2012, Ron
and Shamir found that only about a half of bitcoins were spent during the
rst three months after being received. The same nding was accomplished
by Baek and Elbeck who report strong evidence that Bitcoin volatility is
internally (buyer and seller) driven, leading to the conclusion that the Bit-
coin market is presently highly speculative. The interest rates are at historic
lows, which incentivizes riskier investments that have higher potential re-
turns. Therefore, we can assume that when the interest rates are lower,
risk-tolerant investors will invest more in Bitcoin. When the value of Bitcoin
starts to decrease, the speculative investors will try to minimize their losses
and sell their Bitcoin, which causes the value to drop even lower. When the
value is low, other speculative investors will enter the Bitcoin market, which
snowballs the value back up. A large Bitcoin infrastructure is needed where
more businesses will accept Bitcoin and individuals will use it as a payment
system rather than just as an investment to hedge in
ation. As I previously
mentioned, Bitcoin is decentralized with a predetermined nite supply, which
makes it impossible to be aected by a monetary policy and therefore, aect
its in
ation. Thus, we can assume that Bitcoin's demand for liquidity is
largely driven by speculative motives.
Most of the modern Keynesian economists are skeptical of Bitcoin but
there is not a unanimous opinion. Although Bitcoin is a relatively new con-
cept, it has already garnered the attention of economists. Nobel memorial
prize laureate Paul Krugman famously wrote a blog post titled \Bitcoin
is Evil", in which he questioned Bitcoin's ability to be a reliable store of
value [38]. Other economists such as DeLong and Cowen question whether
Bitcoin can survive when the cost of producing a Bitcoin clone is virtually
zero [39][40]. DeLong is also skeptical because Bitcoin is not backed by a
\too-big-to-fail entity" that could buy it back if necessary. Cowen and other
15

Keynesian economists are concerned about the eventual de
ationary nature
of Bitcoin.
1.5. Currency Criteria
A currency is generally considered by economists to be an instrument
that serves as a medium of exchange, a unit of account, and a store of value.
In the next three subchapters, I will discuss whether Bitcoin satises these
criteria.
1.5.1. Medium of Exchange
A medium of exchange is an instrument to facilitate sale, purchase, or
trade of goods between multiple parties to avoid the inconvenience of a barter
system. For a currency to function as a medium of exchange, it must repre-
sent a standard of value accepted by all parties. The most essential function
of a medium of exchange is to measure value. A medium of exchange should
have a constant intrinsic value and a stable purchasing power on average.
Bitcoin is denitely an instrument that achieves facilitating sale, pur-
chase, or trade of goods between multiple parties. I would even argue that
Bitcoin facilitates transactions more elegantly without an intermediary that
often charges unreasonable fees. However, in order to become an eective
medium of exchange, bitcoins would have to be accepted by a sucient
amount of merchants. Merchants have several reasons motivating them to
accept bitcoin; the biggest reason being the avoidance of expensive, 2%-3%,
credit card fees for each transaction. This has been especially useful for mi-
crotransactions that cost a minimum of 30 cents, which could end up being
30% on a 1 USD transaction [41]. However, the transaction fees of Bitcoin
have recently been increasing because of the blocksize limit, which makes
microtransactions completely unfeasible. Bohme et al. [42] found that be-
cause of the blocksize limit, the transactions could be delayed for up to an
hour, which diminishes the liquidity possibilities and increases risk associated
with accepting bitcoins. I will discuss the blocksize limit issue further in the
\Bitcoin's Weaknesses and Altcoins" chapter.
Moreover, a big disadvantage of accepting Bitcoin payments is a minimum
10 minute waiting time until a conrmation is cleared. This means that a
merchant is technically taking on a risk that in 10 minutes, the transaction
will not clear. The customer can therefore double spend the same bitcoin
again and the transaction would not clear before 10 minutes. However, there
are now services such as BitGo instant and Coinbase, which allow merchants
16

to basically insure every transaction for a 0.1% fee and get the payment in-
stantly. Another disadvantage of accepting bitcoins is the extreme volatility
of the bitcoin's price. Just in the 10 minutes that it takes to clear the transac-
tion (without a service such as BitGo or Coinbase), the value of Bitcoin could
potentially fall by more than the fee that is normally paid to the credit card
companies. It is important to note that most of the merchants that accept
bitcoins will only show prices at the register or checkout and exchange them
for a at currency immediately after the transaction clears in order to mini-
mize the volatility risk. Another diculty that merchants have to deal with
are the returns. Bitcoin transactions cannot be reversed and all of the rev-
enue is immediately converted to USD. The merchants solve this by oering
in-store credit denominated in USD. However, even with all of these disad-
vantages, some merchants oer a discount for paying with bitcoins, which
indicates that transactions in bitcoin are on average more protable than
credit card transactions [43].
As Bitcoin's volatility decreases and more businesses realize that Bitcoin
has smaller transaction costs when compared to the credit cards, the number
of merchants accepting Bitcoin is increasing. In 2015, 160,000 merchants ac-
cepted bitcoins [44]. Several large companies such as Overstock.com, DISH,
Microsoft, Dell, Expedia, Newegg, Zynga, TigerDirect and Virgin Galactic
accept Bitcoin in 2017. Nonetheless, all of these companies do not accept
bitcoins directly; instead they partner with a middleman (usually either Bit-
Pay or Coinbase), which takes the customers' bitcoin, immediately converts
it to cash, and deposits the cash into the company's bank account [45].
Bitcoin can also be used for remittances especially because of the low
fees associated with sending bitcoins across dierent countries. A report
by Goldman Sachs from 2014 measured a cost of sending remittances using
Bitcoin, which ended up being 1% on average. The average cost of sending
remittances using traditional methods is notoriously high, 7.7% on average
in 2015 [46].
Overall, Bitcoin works best as a medium of exchange when both par-
ties are willing to negotiate in Bitcoin rather than just using Bitcoin as an
intermediary to convert their at currencies from and then immediately af-
terwards. The reason why there are not more parties willing to do that is be-
cause of Bitcoin's volatility, which makes trading ineective and risky. How-
ever, infrastructure has certainly been improving and this motivates more
merchants to begin accepting bitcoins as a payment. When more merchants
accept Bitcoin as an alternative payment, it has the potential to disrupt a
17

very protable monopoly on payments. Even if that means that payments
in Bitcoin will be immediately converted to a at currency until Bitcoin's
volatility stabilizes. Transactions in Bitcoin are cheap, reliable, and trans-
parent, which makes it attractive for both the merchants and the customers.
When Bitcoin becomes less volatile and the blocksize limit issue is solved, it
could become a very eective medium of exchange.
1.5.2. Unit of Account
A unit of account is a numerical monetary unit of measurement of the
market value of goods, services, and other transactions. In other words, it is
something that can be used to value goods, services, and other transactions
and make calculations. A unit of account must also be divisible, countable,
and fungible.
Bitcoin's use as a unit of account is so far entirely dependent on its
medium of exchange function. Bitcoin's biggest obstacle in becoming a useful
unit of account is its high volatility. As previously discussed, merchants
who accept bitcoin will either use a middleman or convert bitcoin to a at
currency immediately after the transaction clears. Merchants will also post
their prices in at currencies and only show the converted price in bitcoins
at the register or checkout. These are clear indications that Bitcoin is not
sucient as a unit of account yet. However, that does not necessarily mean
that it will never be. If Bitcoin's volatility ever reaches the levels of other
at currencies, it would be entirely possible for countries to use bitcoin as a
unit of account.
Bitcoin is perfectly divisible and countable. In fact, some refer to Bitcoin
as having an \innite" divisibility because the level, at which Bitcoin can
be divided can be adjusted as time goes on. The current level selected in
the code is 8 decimal places. The smallest unit is thus 0.00000001 BTC.
However, one of the issues is that Bitcoin's price is relatively high when
compared to goods and services. If I were to buy a Big Mac, which costs
3.99 USD, it would equal to 0.00379 BTC in today's prices. This creates a
confusion for a consumer and makes price comparisons fairly complicated.
One of the possible solutions in the future is to start using smaller units such
as milli-bitcoins (mBTC) or micro-bitcoins ( µBTC).
Moreover, bitcoins are not fully fungible. Any two bitcoins have the same
exact value. But because all transactions are publicly available, it is common
for bitcoin exchanges to discriminate between bitcoins based on the owner or
their history. For example, some exchanges will attempt to block bitcoins,
18

which have been conrmed as stolen or obtained illegally. This becomes an
issue because when not every exchange accepts the so called \dirty" bitcoins,
the \dirty" bitcoins become less valuable. Vorick, a Bitcoin Core developer,
is worried that if \dirty" bitcoins become more common, the only way to
know if bitcoins are clean is to go to a centralized service and ask for a
background check. Suddenly the value of bitcoins would be decided by a
centralized party, which would go directly against Bitcoin's core values [47].
Another aspect that is important to discuss is an occurrence of price
fragmentation across dierent bitcoin exchanges. At the time of writing, the
bitcoin price of two leading bitcoin exchanges BitStamp and BTC-E diered
by 19 USD. The ask price at BitStamp 1048 USD and the bid price at BTC-
E is 1029 USD. This creates a clear violation of the law of one price. The
price disparities technically create an opportunity for arbitrage. However,
because of slow bitcoin verication process and fees (exchange, deposit and
withdraw), arbitrage is rarely protable. The prices between exchanges vary
greatly because of the reliability of exchanges, dierent fee policies, dierent
cashout methods, and dierent standards of accepting \dirty" bitcoins. Since
Bitcoin does not serve as a unit of account yet for the reasons above, most
of the regular users will need to convert back and forth from a at currency.
These price fragmentations and uncertainties also hinder Bitcoin as a store
of value.
1.5.3. Store of Value
A store of value can be any asset that is not perishable or subject to
depreciation over time. Essentially, a store of value must retain purchasing
power into the future. The most liquid assets are the best store of value
because they can easily be exchanged for other goods and services.
Bitcoin's utility as a store of value is dependent on its utility as a medium
of exchange. That means that in order for Bitcoin to maintain its store of
value, it has to rely on an expectation of others' willingness to accept it in
the future. Bitcoin has no intrinsic value7, which means that its value lies
in people's willingness to keep using it. Similarly to gold, bitcoin is limited
in quantity, easily transportable, easily divisible, impossible to counterfeit
and people accept it as barter. However, if network eect gives bitcoin its
7Economists agree that bitcoin has no intrinsic value but Bitcoin enthusiasts claim that
its intrinsic value is the invention of eliminating trust from the payment system.
19

only real value and the network is still relatively small, it remains a question
whether a new cryptocurrency similar but better than Bitcoin could com-
pletely replace it. Bitcoin has a rst mover advantage along with popularity,
network eect, investment and trust; but so did Yahoo and MySpace. On
the other hand, hundreds of alternative cryptocurrencies exist and none of
them were successful at replacing or even threatening Bitcoin so far. I will
talk about the largest alternative cryptocurrencies in more detail in the latter
chapter. Bitcoin enthusiasts claim that Bitcoin's invention is comparable to
the invention of the Internet and that if there was a cryptocurrency that was
threatening Bitcoin with superior features, Bitcoin could possibly implement
those features in its own protocol as well. However, this assumption has been
very questionable lately as Bitcoin is being threatened by its block size limit
and a solution has yet to be implemented.
Bitcoin serves as the best store of value for individuals who are looking
for an unregulated and relatively anonymous store of wealth. However, if
an individual is not necessarily looking for non-regulation and anonymity,
he must be averse to possibility that a superior Bitcoin clone will emerge
and diminish Bitcoin's biggest value maker, which is the network eect. To
summarize, if an individual has a positive outlook for Bitcoin in the future, it
could serve as a store of value. However, Bitcoin is currently not established
enough to function as an eective store of value.
1.5.4. Overview of Bitcoin as a Currency
As of now, based on the reasons above, Bitcoin is not suitable as a cur-
rency. Currently, it could be categorized as a scarce digital commodity. Bit-
coin acts more as a payment system to avoid high credit card fees rather than
a medium of exchange. Businesses are not willing to hold bitcoins and, as
a result, they immediately convert it to at currencies. This should change
when volatility decreases. Bitcoin does not serve as a unit of account for
many reasons but it is primarily a result of its volatility again. Cheah and
Fry [48] argued that if Bitcoin were to function as a unit of account and
a store of value, it would not display such volatility expressed by bubbles
and crashes. Bitcoin is also not fully fungible and frequently violates the
law of one price. Lastly, Bitcoin does not serve as store of value because of
its volatility and uncertainty whether Bitcoin will be able to hold its value
without having any intrinsic value.
The biggest obstacle that comes up while evaluating whether Bitcoin can
act as a medium of exchange, unit of account, and store of value, is its volatil-
20

ity. If bitcoin's volatility reaches the levels of other at currencies, there will
not be much stopping it from becoming a fully functional alternative to at
currencies.
1.6. Bitcoin's volatility
As described in the previous chapter, price volatility is bitcoin's Achilles
heel. As of the time of writing, there were roughly 16,250,000 BTC in ex-
istence with one valued at about 1,084 USD. That means that Bitcoin's
market cap was currently approximately 17.6 billion USD, which is still rel-
atively low when compared to at currencies. For example, even a small
currency such as the New Zealand Dollar had an M1 money supply of 41.9
billion USD in February 2017. Due to Bitcoin's small market cap, it is easy
to aect the price with large orders. Because Bitcoin is decentralized, its
price is strictly determined by supply and demand. In combination with
Bitcoin's speculative nature, its price volatility is mostly aected by external
events. Investors react to these events either positively or negatively, which
often causes a panic selling or panic buying. The large price
uctuations
are caused mainly by the sudden government regulations, security breaches
of the third party wallets or exchanges and depreciation of at currencies.
There is one internal event that aects volatility, which is the reward halving
that I discussed before.
Bitcoin's volatility has been trending downwards as the volume of bit-
coins in existence and circulation increases. Another factor that helps limit
bitcoin's volatility is an emergence of bitcoin derivatives exchanges8where
customers can hedge or short sell their positions by using futures contracts.
Bitnex also oers peer-to-peer lending and marginal trading in Bitcoin,
which technically allows customers to short sell. As the infrastructure im-
proves and the volume increases, the volatility will decrease.
1.6.1. Government Regulations
In December of 2013, the People's Bank of China (PBoC) banned Chinese
banks from handling Bitcoin transactions citing risks of money laundering
and protecting nancial stability [49]. The PBoC stated that individuals were
free to buy and sell bitcoins at their own risk. This warning had triggered
the bitcoin bubble to burst and introduced volatility levels of more than
8Deribit and BitMEX (BitMEX does not work in the United States at the time of
writing)
21

13%. Nonetheless, these factors did not stop the existence of Chinese bitcoin
exchanges. According to bitcoinity9and as displayed in Figure 2, in 2016,
over 90% of Bitcoin's volume was traded in CNY. On top of that, over 70%
of Bitcoin mining pools are located in China [22].
Figure 2: Bitcoin trading volume by currencies
One of the reasons why China has taken over Bitcoin is because it is
not controlled and regulated by the PBoC. Therefore, we can assume that
external events aecting trading in China will be the most detrimental to
price
uctuations. Indeed, in January and February of 2017, Bitcoin has
experienced the highest levels of volatility since February of 2015. This sud-
den spike has been attributed to the attempts of the Chinese government to
limit the out
ow of the CNY. In early January, the PBoC announced that
all nancial institutions will be required to notify the PBoC of any transac-
tions over 50,000 CNY ( 7,250 USD), down from a ceiling of 200,000 CNY
(29,000 USD) [50]. China's foreign exchange reserves have hit a 5-year low
in early February of 2017 [51]. Later in February, China banned the users of
the two largest bitcoin exchanges from withdrawing any bitcoins amid con-
cerns that the cryptocurrency is being used to facilitate capital
ight [52].
9https://data.bitcoinity.org/markets/volume/
22

In the following month, the PBoC will inspect the exchanges and if seri-
ous violations are discovered, the platforms will be shut down. The users
of these two exchanges are still allowed to sell their bitcoins and withdraw
CNY. These actions by the PBoC resulted in a huge sello of Bitcoin, which
increased the volatility greatly.
1.6.2. Security Concerns
Even though Bitcoin itself is unhackable according to security experts,
third party wallets and exchanges are prone to hacks or even scams. Bitcoin
platforms are attractive to hackers because of bitcoin's high price, anonymity,
and lack of involvement from the government authorities. This has been an
issue especially in early days of Bitcoin as these third party platforms were not
as transparent and as secure as they are now. Security was a major concern
in the early days and the perceived trust of investors re
ected the price
movements quite accurately. Whenever a security breach was announced and
a large amount of bitcoins were stolen, the volatility increased. The most
publicized scam was Bitcoin Savings and Trust, which ran a Ponzi scheme
and achieved to steal 700,000 BTC from its investors. The most signicant
hack was Mt. Gox, which was handling over 70% of all bitcoin transactions
at the time. It went bankrupt in 2014 and claimed that 850,000 BTC worth
around 480 million USD were stolen by hackers. Mt. Gox customers lost their
savings and were left with no recourse. Not only did people lose their savings,
they lost their trust in Bitcoin and this drove the price down signicantly.
Since the insolvency of Mt. Gox in early 2014, there were two more hacks,
resulting in thefts of over 19,000 BTC from Bitstamp in 2015 and 120,000
BTC from Bitnex in 2016. The price has dropped by more than 10% in
both of these instances but it has recovered fairly quickly. Nonetheless, these
breaches resulted in spikes of volatility. Because it is so dicult to avoid
getting hacked, most of the widely used exchanges store their virtual currency
in oine storage and keep only a small percentage of funds online. The funds
that are stored online are often insured, which means that the customers of
these exchanges are not risking anything.
1.6.3. Depreciations of Fiat Currencies
The price of Bitcoin grew by 130% in 2016 (435 USD to 998 USD) and
outperformed all of the at currencies and commodities. Drastic booms and
crashes have become a standard for bitcoin but what is interesting is that
it grew relatively organically for the rst time in its history with volatility
23

levels of under 5%. This steady growth is largely attributed to the economic
uncertainties in the world such as devaluation of CNY, hyperin
ation in
Venezuela, Indian banknote demonetization, depreciation of the GBP caused
by the BREXIT, and other events. Investors rely on Bitcoin when they
want to minimize the eects of in
ation and economic uncertainty, which
indicates that Bitcoin could qualify as a safe-haven asset. However, even
though Bitcoin is often referred to as \digital gold", there is actually very
little correlation between the two when geopolitical events are factored out
[53]. Gold is a much larger and more mature asset – the demand needed
to boost the gold prices is much larger than the demand needed to boost
bitcoin.
As mentioned before, China has the largest eect on driving the bitcoin
prices. In fact, there was an apparent inverse correlation in 2016 between
CNY and USD as seen in Figure 3 [54].
Figure 3: BTC surges when CNY falls
24

1.7. Bitcoin's Weaknesses and Altcoins
It is relevant to talk about Bitcoin's weaknesses because with the lack
of intrinsic value, its sole value is based on trust and people's willingness
to keep using it. There are currently hundreds of alternative cryptocurren-
cies (called Altcoins in the Bitcoin community), which try to capitalize on
Bitcoin's weaknesses as well as innovate by adding new functions. Bitcoin's
shortcomings currently include long verication times, relatively high fees,
power ineciency, non-fungibility and centralization. Nonetheless, the ma-
jority of these Altcoins are rarely used on a regular basis. At the time of
writing, there were two cryptocurrencies that had a market capitalization
larger than 1 billion USD – Bitcoin and Ethereum. Ethereum is an appli-
cation platform with a cryptocurrency built-in called Ether. Ether already
has conrmation times of 12 seconds instead of 10 minutes and it does not
have a block size limit, which means that it could theoretically process an
unlimited amount of transactions per second. Moreover, there are seven more
cryptocurrencies with a market cap of over 100,000 USD. The Table 2 shows
all the cryptocurrencies with a market cap of over 100,000 USD at the time
of writing.
Name Available Supply Price (USD) Market Cap (USD)
Bitcoin 16,248,737 1,084 17,610,868,623
Ethereum 90,306,627 49.348 4,456,144,390
Ripple 37,388,960,792 0.021562 806,173,295
Dash 7,199,514 68.23 491,227,135
Litecoin 50,439,682 6.50 327,616,834
Monero 14,217,500 20.24 287,743,710
Ethereum Classic 90,261,50 2.79 251,598,527
NEM 8,999,999,999 0.017586 158,271,300
Augur 11,000,000 12.82 141,002,400
Table 2: Cryptocurrency Market Capitalizations over $100k
Bitcoin's biggest asset over Altcoins is its network eect. At the time of
writing, the total number of unique addresses used on the Bitcoin blockchain
was 450,00010and the number of conrmed Bitcoin transactions per day was
300,00011. In fact, Bitcoin's market cap appears to be obeying Metcalfe's
10https://blockchain.info/charts/n-unique-addresses
11https://blockchain.info/charts/n-transactions
25

Law, which states that the value of a network is proportional to the square
of the number of connected users of the system. Because the number of
users in the Bitcoin network is not measured, a proxy of unique addresses
used daily can be utilized. The result of the correlation can be seen in Figure
4.
Figure 4: Metcalfe's Law of Bitcoin
Currently, Bitcoin is threatened mainly by the imperfections in its code,
which was written nine years ago and is becoming outdated. Since then,
Bitcoin has evolved into a widely-used network, which now requires dierent
approaches to some problems. Currently, the biggest constraint is the block
size limit of 1MB. The block size limit was designed in 2009 in order to
protect the network from attacks. This means that Bitcoin can only tolerate
7 transactions per second [55], which is a drop in the ocean when compared
to payment processors such as Visa. Visa handles 1,750 transactions per
second on average but is capable of processing up to 24,000 transactions per
26

second12. Bitcoin users have the freedom to decide if they want to pay a fee
and how much, meaning that whoever pays the highest fee gets prioritized in
the verication process. Because there are currently more than 7 transactions
per second at any given time, transactions become delayed and this causes
the average transaction fee to increase. Before the network was as widely
used, most of the transactions were veried for free. However, the current
average fee in order to avoid delays is 0.37 USD13, which hinders one of
the key attributes of Bitcoin. Even the most prominent Bitcoin developers
are convinced that scalability is an urgent problem that is threatening the
existence of the platform because of the inconvenience of waiting times and
larger fees [56]. In order to overcome a scaling problem like this, the code
needs to be modied, which has been the source of controversy in the Bitcoin
community. As I have mentioned before, any alteration in Bitcoin's code
results in a fork and Bitcoin \miners" are hesitant about this potentially
risky step. Opponents of increasing the block size argue that it could damage
decentralization by driving up the amount of resources needed for mining.
They suggest to implement the so-called o-chain transactions, which means
that the smaller transactions would happen outside of the blockchain. No
matter which solution is implemented, if the scalability is not gured out
soon, users will migrate to other Altcoins such as Ethereum with lower fees
and shorter verication times.
Another critical issue is Bitcoin's power ineciency. At the time of writ-
ing, the \miners" were performing approximately 3,286,700,000 GH/s14. On
average, mining companies use 0.352W for every GH/s of computing power15.
Following that assumption, the Bitcoin network runs at 1,156,918,400W or
roughly 1,156MW. According to EIA, that's enough to power 930,000 aver-
age American households' daily electricity usage. With about 345,000 Bitcoin
transactions per day at the time of writing, that works out to 2.7 households'
daily usage of electricity per one Bitcoin transaction. The power ineciency
is caused by the extremely demanding cryptographic calculations that the
\miners" have to compute on top of verifying transactions. The transaction
costs are thus subsidized by the Bitcoin rewards in exchange for Bitcoin's
in
ation. As the bitcoin rewards get smaller, the power consumption will de-
12https://usa.visa.com/run-your-business/small-business-tools/retail.html
13https://bitcoinfees.info/
14https://blockchain.info/charts/hash-rate
15http://digiconomist.net/beci
27

crease. Another, more immediate, solution would be solving the scalability
issue as discussed in the previous paragraph. This would result in increasing
the amount of transactions per second and therefore, result in a more energy
ecient system. On top of long verication times, relatively high fees, and
power ineciency, Bitcoin is also not fully fungible and not as decentralized
as before.
1.8. Literature Review
In addition to the technical issues such as scalability, power ineciency
and non-fungibility, which could hinder Bitcoin's value alone, economists re-
main skeptical of Bitcoin because of its lack of attributes of a useful currency
[5][38], operation in a legally gray area [7], threat of a superior Bitcoin clone
[39][40], lack of government backing and oversight [7][57][39][58], eventual
de
ationary nature [7][40] and vulnerability to speculation [57][37].
The focus of this study is to determine whether Bitcoin can become a
viable alternative to at currencies. I will evaluate the economic feasibility
of Bitcoin without considering the risks of the technical issues as described
in the previous chapter. From my research and analysis, I have found that
Bitcoin's biggest obstacle from fullling the functions of a currency is its
price volatility. In the past decade, Bitcoin has proven that it can eec-
tively operate in the legally gray area, that it can function eciently without
government backing and that the currently disin
ationary nature has not af-
fected its functioning. The conventional models of measuring volatility such
as standard deviation and moving average deviation, among others, do not
take into account that the variance-covariance of returns may be volatile
during crises. These methods only account for constant volatility, whereas
a GARCH model accounts for time-varying volatility, which is much more
common.
Gronwald [59] used an autoregressive jump-intensity GARCH model and
concluded that Bitcoin prices are strongly characterized by extreme price
movements, which is an indication of an immature market. Dyhrberg [60]
explored the nancial asset capabilities of bitcoin using GARCH models. By
using the asymmetric GARCH model, she found evidence that bitcoin may be
useful in risk management and ideal for risk averse investors in anticipation
of negative shocks to the market. Bouri et al. found that Bitcoin can serve
as an eective diversier but it does not function as a hedging instrument
[61]. Urquhart [62] examined Bitcoin's volatility and the forecasting ability of
GARCH and HAR models in the Bitcoin market. He found that the realized
28

volatility is quite high in the rst half of the sample but has decreased in
the recent years. He also concluded that there is no evidence of the leverage
eect. In the following practical part, I will examine the dynamics and
driving factors of Bitcoin's volatility by using a GARCH(1,1) model.
29

2. Practical part
2.1. Data
The daily Bitcoin data is collected from Coindesk16. Coindesk uses Bit-
coin Price Index (BPI), which is the price of one BTC in USD and uses an
average from the world's leading bitcoin exchanges. This index is particularly
useful when we consider the price disparities amongst dierent exchanges.
Moreover, many exchanges went bankrupt in Bitcoin's relatively short life-
time. All available data at the time of writing is used in order to avoid
creating a bias by choosing a smaller sample. The data available ranges from
August 18, 2010 to March 17, 2017, which is approximately six and half years
or exactly 2404 observations. The Bitcoin price chart during this timeframe
can be seen in Figure 5.
Figure 5: Bitcoin Price Index
16http://www.coindesk.com/price/
30

As a response variable, I am using the log returns of the daily Bitcoin
price. The usage of the log returns is well documented by the empirical
nance literature because prices are believed to be nonstationary. The ad-
vantage of using log returns is that the data is normalized and normally
distributed. The log returns are dened as the rst dierence of the natural
logarithm of the prices as seen in Equation 2.
rt= ln (Pt)ln (Pt1) (2)
where:
rt= log returns at time t
Pt= price of BTC in USD at time t
The daily log returns of Bitcoin prices can be seen in Figure 6.
Figure 6: Daily log returns of Bitcoin prices
31

Most of the previous studies had focused on forecasting the exchange rate
levels rather than their volatility. There is no consensus in the literature on
exactly which factors aect exchange rate volatility of at currencies, but
most studies agree that exchange rate volatility can generally be explained
by macroeconomic variables such as in
ation, interest rates, money supply,
exports and GDP [63][64][65][66]. Because Bitcoin is decentralized and could
technically be considered an international currency, its exchange rate volatil-
ity cannot be explained by macroeconomic variables of just one country. In
order to determine which variables to use, we have to consider the countries
in which Bitcoin is traded the most. To determine these countries, we can
use a proxy statistic that tracks the currencies in which Bitcoin is traded the
most. The data was taken from Bitcoinity17and the results of the traded
volume can be seen in Figure 7.
Figure 7: Bitcoin trading volume by the most traded currencies
17https://data.bitcoinity.org/markets/volume/
32

As expected from the research in the theoretical section, CNY has domi-
nated the traded volume in the past two years. The same chart is generated
without CNY (Figure 8) to determine other most traded currencies. From
Figure 8, it is apparent that the second most traded currency in Bitcoin is
USD, which is then followed by EUR and JPY most recently. Other curren-
cies do not have a signicant presence. As a result, it can be assumed that
China is the most signicant external factor that in
uences Bitcoin's volatil-
ity, followed by the United States, and then the European Union and Japan.
Therefore, I will be using macroeconomic indicators from these three coun-
tries and the EU to explain the exchange rate volatility of Bitcoin. On top of
these macroeconomic variables, I will also use the price of gold because of the
shared qualities with bitcoin such as limited quantity, easy transportation,
easily division, the inability to counterfeit, and acceptance as barter.
Figure 8: Bitcoin trading volume by currencies without CNY
The macroeconomic indicators with daily observations recommended by
the literature are the currency exchange rate, the stock market index, 10-year
33

government bond yield, and a three month interbank rate. The exchange
rate represents the maroeconomic strength of a country, the stock market
index represents the stock market, the government bond yield represents
the bond market and the three month interbank rate represents the bank
sector. The explanatory exogenous variables will therefore be: CNY/USD,
EUR/USD, JPY/USD, Gold Price per Ounce, Shanghai Composite Stock
Market Index, S&P 500, Euro Stoxx 50, Japan NIKKEI 225 Stock Market
Index, China Government Bond 10Y, United States Government Bond 10Y,
Germany Government Bond 10Y (as a proxy for a 10Y bond of the European
Union), Japan Government Bond 10Y, China Three Month Interbank Rate,
US Dollar LIBOR Three Month Rate, Euro LIBOR Three Month Rate and
Japanese Yen LIBOR Three Month Rate. All the explanatory macroeco-
nomic variables can be seen in Table 3.
China U.S. EU Japan
Exchange rate
ln CNY t1
lnEUR t1
lnJPY t1
Stock market index
ln ShanghaiStockIndex t1
lnS&P500 t1
lnEuroStoxx t1
lnNIKKEI 225 t1
10Y gov't bond yield China 10YRGovtBondt t1U:S: 10YRGovtBondt t1Germany 10YRGovtBond t1Japan 10YRGovtBondt t1
3M interbank rate China 3M InterbankRatet t1U:S: 3M InterbankRatet t1EU3M InterbankRatet t1Japan 3M InterbankRatet t1
Table 3: Explanatory macroeconomic variables
All the data is taken from the Federal Reserve Economic Data (FRED)
and Trading Economics. All the explanatory variables are missing observa-
tions because nancial markets are closed over the weekend and holidays.
Summary statistics can be observed in Table 4. A comparison of log returns
of BTC/USD, CNY/USD, EUR/USD and JPY/USD can be seen in Figure
9 and Figure 10.
34

Figure 9: A comparison of log returns of BTC, CNY, EUR and JPY
35

Figure 10: A comparison of log returns of BTC, CNY, EUR and JPY on the same scale
2.2. Methodology
In order to understand how the GARCH model functions, it is important
to introduce the ARCH model rst. The ARCH model, which stands for Au-
toregressive Conditional Heteroskedasticity, was rst developed by Robert
Engle [67] as a method of estimation that allows for time varying volatility.
The concept of the ARCH model is that the volatility estimation will be
more precise when the data from the previous periods is taken into account;
this means that this period's volatility is conditional on information previous
period. When the variance of residuals is constant (homoscedastic), we can
use the method of the ordinary least squares (OLS). However, if the variance
of residuals
uctuates (heteroskedastic), which is the case with Bitcoin, we
must use the method of the weighted least squares (WLS) to estimate the
regression. The ARCH model transforms the OLS residuals into an endoge-
nous process. When the data is heteroskedastic, the standard OLS regression
would fail the Gauss Markov assumptions because Var("t) =2would not
36

Variable N Mean SD Min Max
Time 2404 1202.5 694.1 1 2404
Bitcoin Price Index 2404 271.3 291.3 0.059 1290.8
CNY/USD 1648 0.1571 0.00531 0.1437 0.1656
EUR/USD 1648 1.257 0.1189 1.038 1.488
JPY/USD 1648 0.01043 0.001676 0.007963 0.013207
Gold OZ (USD) 1731 1377.7 201.3 1051.7 1898.3
Shanghai Stock Index 1602 2702.2 602.9 1950 5166.4
S&P 500 1657 1716.6 360 1047.2 2396
Euro Stoxx 50 1675 2905.8 382.1 1995 3838.8
NIKKEI 225 1629 13887.9 3923.5 8160 20868
China 10YR Govt Bond (%) 1706 3.61 0.48 2.62 4.85
U.S 10YR Govt Bond (%) 1732 2.26 0.5 1.36 3.72
Germany 10YR Govt Bond (%) 1708 1.33 0.9 -0.19 3.49
Japan 10YR Govt Bond (%) 1614 0.6 0.39 -0.3 1.35
China 3M Interbank Rate (%) 1718 4.43 1.14 2 9.89
U.S. 3M Interbank Rate (%) 1718 0.4 0.22 0.22 1.15
EU 3M Interbank Rate (%) 1718 0.32 0.55 -0.36 1.56
Japan 3M Interbank Rate (%) 1718 0.12 0.08 -0.08 0.24
Table 4: Summary Statistics
hold: Both ARCH or GARCH models consist of two equations; a conditional
mean equation and a conditional variance equation. Both equations must be
estimated simultaneously since the variance is a function of the mean. The
conditional variance equation of an ARCH(1) model presented by Engle is
seen in Equation 3.
2
t=0+1"2
t1 (3)
where:
2
t= conditional variance this period
0= constant
"2
t1= squared residual return in the previous period (ARCH term)
0>0 and10 to ensure positive variance and 1<1 for stationarity
If the residual return in the previous period is large, the forecast for
this period's conditional volatility will also be large. The ARCH process of
order \p" or ARCH(p) signies how many lags (noted as \p") of the squared
37

residual return is present. When \p" is 1 such as in ARCH(1), it means that
it lags on previous period's squared residual return. The conditional variance
equation for ARCH(p) is seen in Equation 4.
2
t=0+1"2
t1+:::+p"2
tp (4)
However, in the ARCH model, this period's conditional volatility only
depends on the square of the residual return in the previous period, which
means that a crisis with a large residual would not have the same persistence
as is usually observed during an actual crisis. In order to correct this and
capture crises with large residuals, an extension of the ARCH model was
developed by Bollerslev [68]. Bollerslev called this extension the Generalized
Autoregressive Conditional Heteroskedasticity or GARCH. The most com-
monly used model in the literature is the symmetric GARCH(1,1), which is
seen in Equation 5.
2
t=0+1"2
t1+12
t1 (5)
where:
2
t= conditional variance this period
0= weighted long run average variance
"2
t1= squared residual return in the previous period (ARCH term)
2
t1= variance in the previous period (GARCH term)
1+1<1 is the stationary condition; 0>0,1>0,1>0 must hold
According to GARCH(1,1), this period's forecast of volatility is a func-
tion of the weighted average long-term variance, previous period's squared
residual return, and previous period's volatility. The GARCH process of
order \p" and \q" or GARCH(p,q) signies how many lags of the squared
residual return (noted as \p") and how many lags of variances (noted as \q")
are present. When both \p" and \q" is 1 such as in GARCH(1,1), it means
that it lags on previous period's squared residual return and previous pe-
riod's variance. The conditional variance equation for GARCH(p,q) is seen
in Equation 6.
2
t=0+1"2
t1+:::+p"2
tp+12
t1+:::+q2
tq (6)
38

The key statistic in a GARCH model is the sum of and. This gives
us a parameter of persistence, which tells us how fast large volatilities decay
after a shock. In the literature, GARCH models are used to exhibit time-
varying volatility clustering, which is when time series data show continuous
periods of high volatility and continuous periods of low volatility. GARCH
models are not only used for modeling the historical process of volatility
but also in forecasting multi-period volatility. GARCH models are mostly
used to analyze and forecast the volatility of stock returns, interest rates
and foreign exchange data, which is helpful in portfolio allocation, dynamic
optimization and option pricing [69]. In my study, I will use a GARCH model
precisely because of its ability to analyze and forecast volatility. Moreover,
GARCH(1,1) model has been continuously proven to be a superior model for
currency volatility estimation [70][71][72].
The rst step in using a GARCH(1,1) model is to set a conditional mean
equation. As stated before, the conditional mean equation is estimated si-
multaneously with the conditional variance equation because variance is a
function of the mean. In stationary time series, the conditional mean equa-
tion must always be an AR, ARMA, or an MA model. The AR model is the
most used in the literature when analyzing the volatility of nancial returns.
I decided to use the rst order autoregression, written as AR(1), which means
that the returns in the previous period will be used to predict the returns of
the current period. The conditional mean equation while using the AR(1)
model is seen in Equation 7.
rt=0+1rt1+"twith"ti:i:d:(0;2) andjj<1 (7)
In order to nd whether the explanatory variables have an eect on Bit-
coin's volatility, both AR(1) and GARCH(1,1) models can be modied by
adding the explanatory variables to both the conditional mean and condi-
tional variance equation as described by Vlastakis and Markellos [73].
The modied mean equation is seen in Equation 8.
39

rt=0+1rt1+2
lnCNY t1+3
lnEUR t1+4
lnJPY t1
+5
lnGold t1+6
lnSSI t1+7
lnSP 500 t1+8
lnStoxx t1
+9
lnNIKKEI t1+10
ChinaB t1+11
USB t1
+12
GerB t1+13
JapB t1+14
ChinaI t1
+15
USI t1+16
GerI t1+17
JapI t1+"t(8)
The modied variance equation is seen in Equation 9.
2
t= exp(0+1
lnCNY t1+2
lnEUR t1+3
lnJPY t1
+4
lnGold t1+5
lnSSI t1+6
lnSP 500 t1+7
lnStoxx t1
+8
lnNIKKEI t1+9ChinaB t1+10USB t1+11GerB t1
+12JapB t1+13ChinaI t1+14USI t1+15GerI t1
+16JapI t1) +1"2
t1+12
t1+"t(9)
It is recommended by the literature to modify the mean equation with
internal explanatory variables and the variance equation with exogenous ex-
planatory variables. However, Bitcoin does not have any internal variables
other than the intervention events such as the bitcoin exchange hacks, de-
preciation of at currencies, governmental regulations, and halving periods;
which cannot be used as a function of returns. This is because Bitcoin's value
is driven by market forces and has no intrinsic value. There are two possible
solutions to this problem – either use the unmodied AR(1) mean equation
or modify the mean equation by exogenous explanatory variables. Dyhrberg
[60] recommends to modify the mean equation by exogenous explanatory
variables and my study follows that recommendation.
Before estimating the GARCH(1,1) model, the statistical properties of
the mean equation are investigated. The two preconditions that must be
met in order to estimate GARCH(1,1) model are clustering volatility and
ARCH eect in the residual. The results of the test of clustering volatility
in the residuals can be seen in Figure 11. Figure 11 shows that the periods
of low volatility of the error term are followed by the periods of low volatil-
ity and vice versa. This indicates that large returns are followed by large
returns and small returns are followed by small returns. This phenomenon
40

is called clustering volatility, which means that the residual is conditionally
heteroskedastic. Therefore, the rst precondition is met and we can move to
the second precondition: whether there is an ARCH eect in the residual.
Figure 11: A test of clustering volatility in the residuals
An ARCH eect determines whether there is serial correlation of the het-
eroskedasticity. In order to determine whether there is an ARCH eect, we
have to conduct an LM ARCH test. The null hypothesis in this test is that
there is no ARCH eect and the alternative hypothesis is the opposite. The
results of the test can be seen in Table 5. The LM test for ARCH provides
strong evidence that there is an ARCH eect in the mean equation because
we can reject the null hypothesis (0% <5%). Therefore, the mean model sat-
ises both of the criteria and we have the validity to run the GARCH(1,1)
model.
It is also important to note that there is multicollinearity between the
individual macroeconomic variables of Japan and the European Union. There
is no multicollinearity between the individual macroeconomic variables of
41

lags(p) chi2 df Prob >chi2
1 105.484 1 0.0000
H0: no ARCH eects vs. H1: ARCH(p) disturbance
Table 5: ALM test for ARCH
the U.S. and China. However, the presence of multicollinearity between
the macroeconomic variable is an expected outcome since macroeconomic
variables often correlate based on economic developments of such country. In
this study, multicollinearity does not present a problem since it is analyzing
the larger macroeconomic picture of a country rather than the individual
macroeconomic variables.
2.3. Expectations
Based on the previous literature, it is expected that both the previous
day's return information and the previous day's volatility will be signicant
in explaining today's volatility [59][60][61][62]. Moreover, because China has
the largest eect on driving the bitcoin prices, it is expected that it will be
the most signicant variable in predicting next day's volatility of Bitcoin.
Furthermore, CNY has an inverse relationship to BTC [54], which indicates
that the relationship between BTC's volatility and CNY's volatility should
be negative.
There has been no previous literature that looked at the relationship be-
tween macroeconomic variables and Bitcoin's volatility to determine whether
Bitcoin is starting to behave similarly to at currencies. If the macroeco-
nomic variables of countries, where Bitcoin is being traded the most, have a
signicant relationship with Bitcoin's volatility, it would indicate that Bitcoin
is starting to react to the same variables as its at currency counterparts.
Moreover, the results will also show which countries are the most signicant
in predicting next day's volatility of Bitcoin.
2.4. Results
The results are shown in Table 6. GARCH(1,1) was also estimated
with an unmodied AR(1) mean equation to verify the accuracy of the
GARCH(1,1) estimation while using a modied mean equation. The same
statistically signicant variables and the same signs for these variables veri-
ed the validity of the modied mean equation.
42

Variable Mean Equation Variance Equation

lnCNY t1 -.6109801 (.856853) -199.9299 (69.57488)***

lnEUR t1 -.7133233 (.2609827)*** 61.09854 (13.86623)***

lnJPY t1 .2073412 (.2561578) 3.604117 (19.09725)

lnGold t1 .1737769 (.129766) -21.24307 (7.515253)***

lnShanghaiStockIndex t1.0820254 (.0738187) 39.90785 (9.025129)***

lnS&P500 t1 .0801723 (.1970611) 36.0822 (8.322125)***

lnEuroStoxx t1 .1361042 (.1368653) -38.94776 (8.221662)***

lnNIKKEI 225 t1 -.1217739 (.1025617) -7.737607 (7.615126)
China 10YRGovtBondt t1 -.00446 (.0056461) 1.159014 (.3773523)***
U:S: 10YRGovtBondt t1 -.0029994 (.0057765) -2.100576 (.4388985)***
Germany 10YRGovtBond t1.0106237 (.0067866) 4.549737 (.5396294)***
Japan 10YRGovtBondt t1 -.009015 (.0221015) -9.660189 (1.227745)***
China 3M InterbankRatet t1-.0015031 (.0014874) -.2687734 (.1006237)***
U:S: 3M InterbankRatet t1 -.0094399 (.0088782) -1.725574 (1.009195)*
EU3M InterbankRatet t1 -.0011845 (.0065199) 2.041789 (.2874844)***
Japan 3M InterbankRatet t1-.0302514 (.0848204) 3.605629 (4.809523)
L.ar .0195859 (.0271997)
L.arch .1924326 (.034369)***
L.arch .3662459 (.0298018)***
Constant .0309313 (.0178738)** -6.789791 (1.334359)***
Observations 1041 1041
***p<0:01,**p<0:05,*p<0:1
Table 6: GARCH(1,1), dependent variable return on bitcoin.
In the mean equation, most of the explanatory variables are statistically
insignicant, which is an expected outcome. All of the explanatory variables
are lagged by one period, which indicates that if the variables were statis-
tically signicant, it could present an opportunity for arbitrage. The only
statistically signicant variable is the log returns of EUR/USD, which in-
dicates that when the log returns of EUR/USD increase, the log returns of
BTC/USD decrease in the following period. However, a separate regression
was run to verify or refute this relationship and it showed that the relation-
ship is no longer signicant. Therefore, we can conclude that none of the
explanatory variables in the previous period are signicant in forecasting the
current period's log returns of Bitcoin. In other words, Bitcoin's returns are
independent from the in
uence of all of analyzed explanatory variables and
there is no arbitrage opportunity.
In the variance equation, both ARCH ( ) and GARCH ( ) terms are
43

statistically signicant, which means that the previous day's return infor-
mation of Bitcoin does aect today's volatility of Bitcoin (ARCH) and also
that the previous day's volatility of Bitcoin does in
uence today's volatility
of Bitcoin (GARCH). It is also important to note that because  >  , we
can conclude that past volatility eects are superior to past shock eects and
that past volatility eects should be used when forecasting Bitcoin's volatil-
ity. In contrast with the mean equation, most of the explanatory variables
are signicant in explaining the volatility of Bitcoin, which is a favorable
outcome.
The most signicant nding is that macroeconomic explanatory variables
of China, U.S. and European Union are all signicant to forecast next day's
volatility of Bitcoin. This indicates that Bitcoin is starting to react to the
same variables as its at currency counterparts in China, U.S. and European
Union. Moreover, it shows that Bitcoin is maturing to become an alternative
currency especially in these countries and the EU. However, it also points
to the centralization of Bitcoin, which is not a favorable outcome because
macroeconomic shocks from China, the U.S. and the EU aect Bitcoin dis-
proportionately more than shocks from other countries. If Bitcoin were to
act as an alternative international currency, it must become independent
from large countries and become more widely used worldwide. Therefore,
the macroeconomic shocks and intervention events would spread out and
not aect Bitcoin as greatly as now. For example, when China suspended
the users from withdrawing bitcoins in early 2017 amid concerns that the
cryptocurrency is being used to facilitate capital
ight, the price of bitcoin
dropped by over 30% in just a week. Since then, the price has recovered and
continued to increase to the highest levels in Bitcoin's lifetime. But right
after it hit its peak, the SEC has denied the application for the Bitcoin Trust
ETF and the price crashed by 24% again. These are just two examples of
how external shocks coming from either China or the United States greatly
aect Bitcoin's volatility.
On the other hand, macroeconomic variables from Japan are not suitable
to forecast next day's volatility. This is an expected outcome because Japan
has just recently seen a growth of Bitcoin adoption and media coverage. If
we only looked at the last year of Bitcoin's price development, it is possible
that Japan's macroeconomic variables would be more signicant in predicting
Bitcoin's volatility.
The coecients on the returns of exchange rates suggest that Bitcoin's
volatility is the most sensitive to the price changes in CNY. This is also
44

an expected outcome because the vast majority of Bitcoin was traded in
CNY in the last two years. The returns of gold are signicant in forecasting
Bitcoin's volatility but the relatively small coecient means that Bitcoin's
volatility is not very sensitive to its price changes. Another signicant nding
is that a positive volatility shock to CNY decreases the volatility of bitcoin.
The possible explanation could be that Bitcoin acts as a safe-haven asset in
China. When there is an increased risk in holding CNY, more people convert
their at currency to bitcoin and therefore increase the traded volume, which
decreases Bitcoin's volatility. Estimated conditional variance of daily Bitcoin
log-returns can be seen in Figure 12.
Figure 12: Conditional variance of daily Bitcoin log-returns
In Figure 13, the daily volatility of Bitcoin is shown with an average
trendline and with a 30-day peak trendline. It is apparent that the volatility
is trending downward but it is still far away from the volatility levels of
other at currencies. However, an important takeaway from Figure 14 is
that the economic properties for Bitcoin to become a viable alternative to at
currencies are developing quickly. The dotted line indicates a continuation of
45

the trendline as a prediction of volatility levels in the future. If the volatility
levels follow the same trend as in the last six and a half years, Bitcoin will
reach the at levels of volatility in 2019-2020. However, it is improbable that
the volatility will continue to reduce at the current rate because of the absence
of central banks and their ability to make macroeconomic adjustments to
stabilize the exchange rate. An important task of the central bank is to
minimize the systematic (undiversiable) risks. Bitcoin does not have any
way of minimizing the systematic risk and therefore, it is possible that Bitcoin
will never reach the volatility levels of stable at currencies. Nonetheless,
volatility will surely approach the at levels when the traded daily volume is
sucient.
Figure 13: Daily volatility of Bitcoin
46

3. Conclusion
Currently, Bitcoin does not satisfy the criteria for being considered as a
currency because it does not act as an eective medium of exchange, unit
of account, or as a store of value. Bitcoin presently acts as a scarce digital
commodity with a nite supply. The nite supply and de
ationary pressures
make Bitcoin highly speculative. When compared to other cryptocurrencies,
which are also decentralized, anonymous and unregulated, Bitcoin's sole value
lies in its network eect and a rst mover advantage. This makes Bitcoin
vulnerable to potentially superior alternatives. Apart from the technical
issues such as the scalability, the biggest obstacle preventing Bitcoin from
becoming an alternative to at currencies is its volatility. Bitcoin's volatility
is far larger than even at currencies with a small supply.
Within the last two years, Bitcoin has mostly been traded in China. His-
torically, China is followed by the United States, the European Union, and
Japan. Therefore, it is fair to assume that China is the most signicant exter-
nal factor that in
uences Bitcoin's volatility, followed by the United States,
the European Union, and Japan. The macroeconomic explanatory variables
of China, the United States, and the EU are all signicant to forecast the
next day's volatility of Bitcoin. This indicates that Bitcoin is starting to
react to the same variables as the at currencies in these countries. Japan's
macroeconomic variables are not signicant, however.
Bitcoin's volatility is the most sensitive to the price changes in CNY,
which further points to the centralization of Bitcoin. Currently, macroe-
conomic shocks from China, the United States, and the EU aect Bitcoin
disproportionately more than shocks from other countries. In order for Bit-
coin to truly become an international alternative to at currencies, it must
become less dependent on larger countries.
However, the volatility levels have historically been trending downward
and if Bitcoin were to follow the same trend as it has been for the previous six
years, it would reach the volatility levels of at currencies approximately in
2019-2020. However, the absence of a central bank that acts as a minimizer of
the systematic risk and a stabilizer of the exchange rate means that Bitcoin
might never reach the volatility levels of stable at currencies. Nonetheless,
volatility will surely approach the at levels when the traded daily volume
is sucient. If Bitcoin approaches the volatility levels of at currencies, it
will satisfy the criteria of being a functioning currency and therefore, will be
ready for mass adoption.
47

References
[1] S. Nakamoto, Bitcoin: A peer-to-peer electronic cash system (2008).
[2] M. Bustillos, The bitcoin boom, 2014.
[3] D. Chaum, Blind signatures for untraceable payments, in: Advances in
cryptology, Springer, pp. 199{203.
[4] S. Barber, X. Boyen, E. Shi, E. Uzun, Bitter to betterhow to make
bitcoin a better currency, in: International Conference on Financial
Cryptography and Data Security, Springer, pp. 399{414.
[5] D. Yermack, Is Bitcoin a real currency? An economic appraisal, Tech-
nical Report, National Bureau of Economic Research, 2013.
[6] D. Golumbia, Trump, clinton, and the electoral politics of bitcoin, 2016.
[7] R. Grinberg, Bitcoin: An innovative alternative digital currency (2011).
[8] J. Redman, Openbazaar is here but darknet markets will remain, 2016.
[9] M. Miller, The ultimate guide to Bitcoin, Pearson Education, 2014.
[10] J. Brito, Online cash bitcoin could challenge governments, banks, 2011.
[11] R. Stross, What's coming out of silicon valley, 2012.
[12] A. Santos, Bitcoin-central becomes rst bitcoin exchange licensed to
operate like a bank, 2012.
[13] P. Vigna, 5 things about mt. goxs crisis, 2014.
[14] D. Howden, Bitcoin bank run, 2014.
[15] O. Williams-Grut, Mt.gox les for bankruptcy protection, 2014.
[16] R. Sidel, E. Warnock, T. Mochizuki, Almost half a billion worth of
bitcoins vanish, 2014.
[17] P. Liljas, Worlds rst bitcoin atm launched in canada, 2013.
[18] S. Acharya, J. Dunn, Overstock.com ventures into digital currencies,
Journal of Business Cases and Applications 12 (2014) 1.
48

[19] D. Ron, A. Shamir, Quantitative analysis of the full bitcoin transaction
graph, in: International Conference on Financial Cryptography and
Data Security, Springer, pp. 6{24.
[20] T. Swanson, Approximately 70% of all bitcoins have not moved in 6 or
more months, 2014.
[21] O. Beigel, Is bitcoin mining protable in 2017?, 2016.
[22] N. Popper, How china took center stage in bitcoin's civil war, 2016.
[23] S. Valfells, J. H. Egilsson, Minting money with megawatts [point of
view], Proceedings of the IEEE 104 (2016) 1674{1678.
[24] Virtual currency schemes, European Central Bank (2016).
[25] L. V. Mises, B. B. Greaves, Human action: A treatise on economics,
Yale University Press New Haven, 1949.
[26] R. W. Garrison, Overconsumption and forced saving in mises-hayek
theory of the business cycle, History of Political Economy 36 (2004)
323{349.
[27] L. v. Mises, The theory of money and credit, Indianapolis, IN: Liberty
Fund, Inc., 1912.
[28] A. G. Clegg, Could bitcoin be a nancial solution for developing
economies, University of Birmingham (2014).
[29] F. Hayek, Choice in currency: a way to stop in
ation, volume 48, Ludwig
von Mises Institute, 1976.
[30] N. Gertchev, The money-ness of bitcoins, 2013.
[31] P. Korda, Bitcoin: Money of the future or old-fashioned bubble?, 2013.
[32] F. Shostak, The bitcoin money myth, 2013.
[33] K. Graf, On the origins of bitcoin: Stages of monetary evolution, Konrad
S. Graf Investigations and Observations (2013).
[34] P. Surda, The origin, classication and utility of bitcoin (2014).
49

[35] J. M. Keynes, General theory of employment, interest and money, At-
lantic Publishers & Dist, 2016.
[36] F. Glaser, K. Zimmermann, M. Haferkorn, M. C. Weber, M. Siering,
Bitcoin-asset or currency? revealing users' hidden intentions (2014).
[37] C. Baek, M. Elbeck, Bitcoins as an investment or speculative vehicle?
a rst look, Applied Economics Letters 22 (2015) 30{34.
[38] P. Krugman, Bitcoin is evil, The New York Times 28 (2013).
[39] B. DeLong, Watching bitcoin, dogecoin, etc, 2013.
[40] T. Cowen, How and why bitcoin will plummet in price, 2013.
[41] G. Varriale, Bitcoin: how to regulate a virtual currency, International
Financial Law Review 32 (2013) 43.
[42] R. Bohme, N. Christin, B. Edelman, T. Moore, Bitcoin: Economics,
technology, and governance, The Journal of Economic Perspectives 29
(2015) 213{238.
[43] S. Lo, C. Wang, et al., Bitcoin as money?, 2015.
[44] C. Burniske, Bitcoin: A disruptive currency (2015).
[45] J. Davidson, Bitcoin not really being accepted by major companies,
2015.
[46] C. Burniske, All about bitcoin, Global Macro Research (2015).
[47] D. Vorick, Ensuring bitcoin fungibility in 2017 (and beyond), 2016.
[48] E.-T. Cheah, J. Fry, Speculative bubbles in bitcoin markets? an em-
pirical investigation into the fundamental value of bitcoin, Economics
Letters 130 (2015) 32{36.
[49] G. Mullany, China restricts banks use of bitcoin, 2013.
[50] G. Smith, Bitcoin is melting down as china cracks down on capital out-

ows, 2017.
50

[51] G. Wildau, China forex reserves dip under $3tn to touch 5-year low,
2017.
[52] G. Wildau, Major chinese bitcoin exchanges halt withdrawals after
crackdown, 2017.
[53] C. Bovaird, Are bitcoin and gold prices correlated?, 2016.
[54] G. Wildau, China probes bitcoin exchanges amid capital
ight fears,
2017.
[55] M. Vukoli c, The quest for scalable blockchain fabric: Proof-of-work
vs. bft replication, in: International Workshop on Open Problems in
Network Security, Springer, pp. 112{125.
[56] T. Simonite, The looming problem that could kill bitcoin, 2015.
[57] F. Brezo, P. G. Bringas, Issues and risks associated with cryptocurren-
cies such as bitcoin (2012).
[58] M. Van Alstyne, Why bitcoin has value, Communications of the ACM
57 (2014) 30{32.
[59] M. Gronwald, The economics of bitcoins{market characteristics and
price jumps (2014).
[60] A. H. Dyhrberg, Bitcoin, gold and the dollar{a garch volatility analysis,
Finance Research Letters 16 (2016) 85{92.
[61] E. Bouri, P. Molnar, G. Azzi, D. Roubaud, L. I. Hagfors, On the hedge
and safe haven properties of bitcoin: Is it really more than a diversier?,
Finance Research Letters 20 (2017) 192{198.
[62] A. Urquhart, The volatility of bitcoin (2017).
[63] N. Antonakakis, J. Darby, Forecasting volatility in developing countries
nominal exchange returns, Applied Financial Economics 23 (2013) 1675{
1691.
[64] P. De Grauwe, Exchange rate variability and the slowdown in growth
of international trade, Sta Papers 35 (1988) 63{84.
51

[65] T. G. Andersen, T. Bollerslev, Answering the skeptics: Yes, standard
volatility models do provide accurate forecasts, International economic
review (1998) 885{905.
[66] C. Engel, K. D. West, Exchange rates and fundamentals, Journal of
political Economy 113 (2005) 485{517.
[67] R. F. Engle, Autoregressive conditional heteroscedasticity with esti-
mates of the variance of united kingdom in
ation, Econometrica: Jour-
nal of the Econometric Society (1982) 987{1007.
[68] T. Bollerslev, Generalized autoregressive conditional heteroskedasticity,
Journal of econometrics 31 (1986) 307{327.
[69] T. Bollerslev, J. M. Wooldridge, Quasi-maximum likelihood estimation
and inference in dynamic models with time-varying covariances, Econo-
metric reviews 11 (1992) 143{172.
[70] S.-H. Poon, C. Granger, Practical issues in forecasting volatility, Finan-
cial Analysts Journal 61 (2005) 45{56.
[71] P. R. Hansen, A. Lunde, A forecast comparison of volatility models:
does anything beat a garch (1, 1)?, Journal of applied econometrics 20
(2005) 873{889.
[72] C. T. Brownlees, R. F. Engle, B. T. Kelly, A practical guide to volatility
forecasting through calm and storm (2011).
[73] N. Vlastakis, R. N. Markellos, Information demand and stock market
volatility, Journal of Banking & Finance 36 (2012) 1808{1821.
52