How Bitcoin Might Slip Into an Economic Downward Spiral

Bitcoin is a digital cryptocurrency based on a peer-to-peer network, which is designed to allow to quickly transfer arbitrary amounts of currency units (i.e., money) to anybody in any region of the world without having to pay excessive amounts of fees to payment processors. During March and April 2013, the exchange rate of bitcoins (short: BTC) dramatically increased from about 25 EUR/BTC to over 180 EUR/BTC. The market capitalization of bitcoin recently hit the $2 billion mark on April 10, just before the bubble burst.

In this article, I will discuss people’s motivation to consider bitcoins as a worthwhile investment and why many proponents consider cryptocurrencies superior to traditional fiat currencies. Finally, I will point out a technical design flaw in bitcoin, which I consider to be crucial enough to deem bitcoin currently an unsafe long-term investment.

Motivation to engage

People engaging in bitcoin have various reasons to do so, most notably:

  • Mining: Mining is the process of solving cryptographic puzzles which are used to secure transactions taking place within the bitcoin peer-to-peer network. By solving these kinds of puzzles, miners get a chance to be rewarded with newly generated bitcoins. Hence, while miners invest money in hardware and energy in order to earn bitcoins, they keep the network up and running and – most importantly – they secure it by their collaborative, immense computing power. The more miners participate, the more difficult the puzzles become to solve, and the more becomes bitcoin restitant against frauds.
  • User: The users of bitcoin want bitcoin become a widely accepted payment method. As consumers, they prefer BTC over traditional payment processors. There are more and more merchants who accept BTC as an alternative currency in addition to their official currency. The users desire a rather steady growth of BTC value such that trading is not too much disrupted by fluctuating exchange rates.
  • Trust: There are people who trust bitcoin more than their official currency, e.g. due to imminent expropriation by their governments. This might probably have been a motivation for some Cypriots to buy bitcoins during March 2013, resulting in safe haven flow out of EUR and into BTC. BTC can actually be transferred more easily somewhere else than cash money or bullions. As no bank is involved during the processing, bitcoin may (perhaps mistakenly) be perceived as the ideal means of money laundering.
  • Value storage: Some people have a great expectation towards the long-term market trend of bitcoin and want to grab their share as long as it’s cheap. They rather hoard coins than use them for payments.
  • Speculation: Last but not least, there are financial speculators, which might see bitcoin purely as a means to make quick money, but they don’t take bitcoin too seriously as an effective currency.

Note that a single person may act in several of the above mentioned roles at a time. E.g., a merchant might be a supporter by accepting bitcoins, but he might also decide to save the value rather than exchanging his revenue back to the official currency.

A deflating currency?

Libertarian proponents of bitcoin and often point out that bitcoin is a chance to evade the tentacles of governments in the sense that bitcoin cannot be inflated by central banks like fiat currencies. As opposed to monetary systems with a central issuing authority, bitcoin runs as a decentralized peer-to-peer network on computers in many different countries around the world, which is not under control of any single person or institution. So, unlike fiat currencies, many expect bitcoin to deflate rather than to inflate due to the following reasons:

  • Limitation: The number of total bitcoins that will come into circulation is strictly limited to 21 million BTC. Miners generate so-called blocks in a blockchain which can be considered as units of transactions in the network. In bitcoin, a block is generated roughly every ten minutes and with each block, a specified amount of bitcoins is generated and given to the miner who found the block (which may also be a pool of collaborating miners). When bitcoin came into existence as of January 2009, 50 BTC were generated with each block. This amount is halved every four years until the limit of 21 million BTC will be reached in 2140.
  • Growing acceptance: Supporters of bitcoin promote the acceptance as an alternative currency for goods and services. This process pushes the real value of bitcoins as you can afford more goods and services by a bitcoin, essentially enhancing its purchasing power.
  • Economic growth: The economic growth of the businesses represented by bitcoin is expected to be reflected in a growing value of the currency, as the currency cannot be inflated.

However, a deflating currency is detrimental to an economy as people would rather save their money than spend it for goods and services. This in turn has a negative impact on the economies represented by that currency, effectively reducing the currency’s value, which is in contradiction to the premise of a deflating currency. Matthew Yglesias describes this course of hoarding cycles here and finally concludes that expectations about the price level will be “unanchored”.

Comparison to fiat currencies and gold

When it comes to comparison to fiat currencies and gold, bitcoins have some remarkable advantages over them, which makes them so fascinating. Unlike gold, bitcoins are easy to transfer, easy to secure, easy to verify and easy to granulate. Furthermore, unlike electronic fiat currency systems, bitcoins are predictable and limited in supply, not controlled by a central authority, not dept-based, freeze-proof as well as faster and cheaper to transfer [source]. They may also be considered to be anonymous, unless you publicly link your identity with your pseudonyms. Nevertheless, there are also disadvantages in bitcoin, which physical stores of value (like noble metals) do not have.

Firstly, noble metals are unique, bitcoin is not. Imagine what would happen if a gold2 and probably also a gold3, gold4, gold5 and a gold6 was going to come into existence in addition to traditional gold and silver. Each of those metals would be distinguishable from the others, but they would have security features quite similar to traditional gold. You could start using them to store value besides gold, potentially causing the effective value of gold to drop. In analogy, even though nobody can practically create counterfeit money within the bitcoin ecosystem, people can quite easily create a fork of bitcoin and run an equivalent network in parallel to the existing one. In fact, some of these clones are already in existence. Some of their names are Litecoin, Namecoin, IXCoin, PPCoin, Terracoin and Devcoin. Most of them are considered worthless right now, but that may possibly change in the future. Note that with the boom of bitcoin during March 2013, Litecoin has already drawn attention in the cryptocurrency community, making it a potential alternative to the alternative currency. There is also a rumour that Mt. Gox, one of the largest exchange markets for bitcoins, is going to adopt Litecoin.

Secondly, bitcoins do not preserve their value by themselves. Their value and their security depends on the collaborative computing power of miners. The latter spend their money for energy and competitive hardware in the expectancy of being rewarded with valuable bitcoins. Every single individual is mainly driven by its own greed, not by the effect that it preserves value for other people. As opposed to bitcoins, gold is self-preserving. It is one of the chemically least reactive elements and is practically indestructible. You don’t have to spend money or rely on others to save it. At most, you spend money to secure it.

Bitcoin’s exchange rate correlates to its resistance against fraud

As mentioned before, in order to continue operation, the network must attract miners. Mining a crypto-currency is perceived profitable by miners if they expect their monetary investment in form of hardware and energy to be rewarded, probably also at a later point in time. The aggregate computing power of all honest nodes in the network in terms of calculated hashes per second directly translates to the reward expected by the participating miners in form of the amount of bitcoins that they earn. Hence, if bitcoins are perceived as more valuable (i.e., if they have or are expected to have a higher exchange rate), the hashing rate will increase as mining becomes more lucrative. You can very well observe this correlation between the exchange rate at Mt. Gox or and the networks aggregate hashing rate. Also, the aggregate computing power of the network is in proportion to the security of the network, in that it dictates the effort that an fraudulent attacker has to take to outrun the network (e.g., in order to withdraw a transaction). Putting these two relationships together, the perceived value per bitcoin on the exchange market is related to the network’s resistance against fraudulent attackers.

Attacking the network by supporting it

The bitcoin network adjusts its difficulty level every 2016 blocks in order to keep the average block generation rate constant. This happens about every two weeks. The adjustment of the new difficulty is based on the time it took to generate the previous 2016 blocks.
Hypothetically, if many miners suddenly lose interest in the bitcoin network and discontinue operation just right after the difficulty level has been adjusted, the network might become slower in that transactions may take somewhat longer than usual. This is due to the design of discrete steps of difficulty adjustment rather than continuous control.

Consider an attacker with immense computing power, which joins the network to honestly participate in the mining process for a period of 2016 blocks. After that period, the network will realize that it runs too fast. It will adjust its difficulty level accordingly by extrapolating its current performance such that the next 2016 blocks will be found within approximately two weeks. Then the attacker will leave the network right after the adjustment. The difficulty level will stay at a high level but the network’s computing power will drop, resulting in an average block generation time greater than 10 minutes. If it gets significantly greater, transactions will take significantly longer to be secured by the network, thus compromising liquidity of the network.

Let’s play this through. The difficulty level as of 5th April 2013 was 6695826. Assume the network running at 50 TH/s (i.e., 50 \cdot 10^{12} hashes per second). Thus, according to the formulas stated here, the average time to generate a block is  \frac{6695826 \cdot 2^{32}}{50 \cdot 10^{12} \cdot 60} minutes \approx 9.58 minutes. Now let’s assume that an attacker with 30 TH/s enters the network right after the difficulty level adjustment to 6695826. The network will generate a block every 6 minutes on average. After the next 2016th block, the difficulty will be adjusted to 11175870 such that the network will find a block every 10 minutes on average at a total hashing rate of 80 TH/s. When the attacker leaves the network, the network’s hashing rate will drop to 50 TH/s, such that each block will be found in 16 minutes on average. Thus, the attack introduced an additional delay of 60% to transactions for the next 22.4 days, given that the network’s computing power stays at a constant level of 50 TH/s. This might introduce some inconvenience to the bitcoin markets as liquidity is compromised.

Economic downward spiral

At a first glance, the attack described above does not seem threatening to the bitcoin network as it seems to slow down the network just for a little while. Also, it appears to be quite irrational for an attacker to spend such a huge amount of resources for a nuisance. However, also note that the attack results in a disincentive for miners as the generation of new bitcoins now runs slower, i.e., miners now have to spend 60% more time (i.e., energy) into mining to be rewarded with the same amount of bitcoins due to the higher difficulty level. So, some miners might discontinue operation during this period of 2016 blocks, resulting in even lower network performance and thus an even further decrease of liquidity. Also, the markets might realize that bitcoin liquidity is being compromised, potentially resulting in a drop of exchange rates, which in turn is yet another disincentive for miners to continue mining. The network cannot escape this situation until 2016 blocks have been generated after the attack. Just after 2016 blocks being found, the network adjusts its difficulty level, making mining more attractive again. Nevertheless, in the worst case, the bitcoin network might already be dead before the last miners find the 2016th block. Hence, this sort of support attack still has some destructive potential.

Also note that this run of events does not necessarily have to be introduced by an attacker. It can actually result out of any event which discourages miners from mining. For instance, a drop in hashing power may result from a legal prohibition in a country with many righteous miners who stop mining all at once. Also, a drop of exchange rates might discourage miners from mining, just as an increase of exchange rates attracts them. If a sudden drop of exchange rates occurs just right after the adjustment of the difficulty level, the network might potentially enter into this economic downward spiral, as a lot of miners might consider mining bitcoins becoming unattractive. When miners leave, the average time for each block being found goes up, hence prolonging the unattractive mining phase with a too high difficulty level and compromising liquidity of bitcoins, which might result in a lower perceived value of the currency, i.e., falling rates, and so on. Hence, bitcoins can be considered a more risky investment during times of falling exchange rates.

In the future, other crypto-currencies (bitcoin forks) may also play a more important role than they do today. Once people have started speculating on them, crypto-currencies might effectively be competing for miners. If mining Y-coin is considered more profitable than mining X-coin, miners will switch from mining X-coin to mining Y-coin. By doing so, they potentially contribute to a catastrophe in the X-coin ecosystem.

Visual clarification

The graph on shows bitcoin’s difficulty level in red and its 14-day average window of the network’s aggregate hashing rate in purple. Note that both lines have different scales but they are matched against each other such that their correlation gets clear. Whenever the difficulty is adjusted, the red line quite accurately hits the purple line.

In times of rising hashing rates, miners do outperform the estimated hashing rate during a period of 2016 blocks. During these periods, the purple curve is above the red line. This implies that on average less than ten minutes are needed to find a block. You can see this contraction also directly in the graph, e.g., during 03’13 there were almost three periods of 2016 blocks, while in 12’12 there were about two periods. This also implies that miners are rewarded in less time with new bitcoins than they should be. Hence, when they are given an incentive to increase hashing performance by rising exchange rates, they are overly much rewarded with new bitcoins.

Vice versa, in times of falling hashing rates, the purple curve will be below the red line. This is due to the fact that during the previous period of 2016 blocks, the total hashing rate was still greater than it is at the time at which the difficulty is adjusted. Hence, during the next period of 2016 blocks, each block will take on average longer than 10 minutes and miners will earn less bitcoins per time unit (each miner earns just as much as he would have earned if the estimated hashing rate had been reached). Also, such a period of 2016 blocks will take longer than the estimated two weeks.


Cryptocurrencies based on peer-to-peer networks introduce some new interesting features that might greatly simplify monetary market businesses in the future. Most importantly, they allow for secure transactions without any intermediate payment processors involved, thus allowing for transactions between people in any countries in the world at low or zero processing fees. Bitcoin is the most prominent digital currency nowadays and has served as a basis for some other peer-to-peer-based cryptocurrency systems. While bitcoin’s proponents emphasize its openness and decentralized nature, these same properties make the currency also susceptible to market fluctuations, such that bitcoin establishes as a speculative object and a gambling tool rather than a real currency for trading goods and services.

While bitcoin has a lot of advantages over fiat currencies and noble metals, it still has noteworthy disadvantages. Firstly, bitcoin forks can easily be created, and in fact, there are already various other bitcoin-based cryptocurrencies in existence. These may potentially compete for miners in the future and level down exchange rates of each other. Secondly, liquidity of the bitcoin currency is ensured by miners, which, however, are driven by their own greed in earning bitcoins. As bitcoin’s difficulty level always lags behind the network’s actual hashing power, a sudden drop of hashing power may potentially cause bitcoin to slip into an economic downward spiral. In order to overcome this risk, the update interval of the difficulty level would have to be reduced. In addition, the derivative of the 14-day average of the estimated network’s hashing power could be used to extrapolate the new difficulty level more accurately.

However, while these measures might effectively reduce the risk of entering into an economic downward spiral, bitcoin can still be considered an unproven economic concept and bears many other risks. I recommend reading this article to learn more about some other important aspects.

Leave a Reply

Your email address will not be published. Required fields are marked *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>