19 March 2018
Do cryptocurrencies live up to the hype? Are they sensible economic phenomena or just fraud?
This follows up from my piece last month.
Cryptocurrencies either aim to be a direct replacement for fiat money (like Bitcoin/Litecoin/Monero), or to provide some economic function where then the cryptocurrency underpins the integrity of this function, like (Ethereum/Ripple).
For a good description of the technological fundamentals of cryptocurrencies on YouTube see this, while the most entertaining take on cryptocurrencies is John Oliver.
While the terminology is in some flux with the term “cryptocurrency” usually synonymous with “virtual currency”, while “cryptoasset” is used for a subset of these, I will use the term cryptocurrency to refer to all of them.
The original cryptocurrency, Bitcoin, was envisioned as a replacement for fiat cash. It retains the first mover advantage, with the highest market cap of about $150 billion at the time of writing, see coinmarketcap.com/.
Bitcoin suffers from perceived design limitations, with efficiency (high transaction cost and slow speed) seen as problematic. Recent competitors, such as Bitcoin cash and Litecoin improve efficiency.
Transactions with Bitcoin are not fully private because the underlying blockchain is searchable, and other cryptocurrencies use different technologies to ensure privacy, like Monero’s ring signatures.
The visible blockchain with slow updates is what guarantees Bitcoin’s trust. Improving efficiency by speeding transactions or removing disability reduces trust.
With cryptocurrencies, there is a direct trust vs. efficiency vs. privacy tradeoff.
The main competitor to Bitcoin is Ethereum, aiming to be a globally distributed computer program — world computer — designed to execute smart contracts.
Ripple is number three. Its aim is to be a new type of a fast payment system. In order to achieve that, it gives up on blockchain and its native technological trust, replacing some of it it with trusted institutions — typically banks.
The following figure shows the market cap of the top 10 cryptocurrencies on March 19, 2019 from coinmarketcap.com.
The total volume of cryptocurrencies is typically controlled by an algorithm set up so that some computationally intensive operation is required to create new units of the currency — mining — where the computational problem becomes progressively harder until some other hard limited is reached, what is known as controlled supply. In the case of Bitcoin, the theoretical upper limit is almost 21 million, expected to be reached around 2140, and we now have almost 17 million Bitcoins or 81% of the total. The following figure shows the supply until now.
Most other cryptocurrencies follow a similar set up, but often with a faster mining schedule.
It is the controlled supply that provides the main argument for cryptocurrencies actually being money. This argument deliberately mimics gold as money discussions: A cryptocurrency based monetary system is to be a modern version of the gold standard.
Cryptocurrencies are increasingly controversial. While initially passing without much notice, recently they have made mainstream news and the various financial authorities have expressed increasing concern.
The main arguments in favour of cryptocurrencies are typically some combination of the following list:
The technologies underpinning the various cryptocurrencies are quite elegant, and many of those who argue against them are often called Luddites, or said to otherwise not understand technology.
However, an elegant technology does not imply real world usefulness. Knowing all their mechanical details does not translate to understanding their economic or social function.
Take as an example human beings. I can know all the physics and chemistry and physiology, understand how molecules and organs operate, yet still not know the first thing about an individual.
Therefore, what the advocates of cryptocurrencies should show is how they solve real-world social problems, and not just resort to handwaving or mysticism to justify them. Yes, there are several use cases for cryptocurrencies, but nothing sufficiently substantial to justify their evaluations or supposed impact on the economy.
Someone might retort that cryptocurrencies might be inefficient and insecure today and without large applications, but that will improve in the future.
A marvellous technological solution where the real benefits are promised and not real is just a pie in the sky promise. Otherwise the flying cars and all the fantastic last century technology promises would be with us today.
What do we need money for? The Wikipedia page on money, lists:
My list is slightly different, encompassing the Wikipedia list and adding lending of last resort and the economic considerations:
We have been debating what money is for long time. The lists above are what most mainstream economists would subscribe to. But there are strong opinions to the contrary.
There are two related arguments made.
Over time, we have used a number of assets as money, silver, copper, seashells and cigarettes, just to name a few. However, the asset best associated with money is gold, and the longest period of monetary stability the world has ever seen, the gold standard of 1873 to 1914, was based on money being gold.
Of course, it isn’t quite that simple as money and monetary stability in Europe, 1300-1914 show. There is plenty of room for manipulation.
So, does it make sense to use a real asset, like gold, or a cryptocurrency as money instead of fiat money?
If we want to link money to some real assets, gold is probably the best choice, and many cryptocurrencies advocates propose Bitcoin or something similar in its place.
But there are however several reasons why fiat money is better.
To begin with, transactions with fiat money money are much cheaper and faster than any of the cryptocurrencies. Transactions with cash costless and instantaneous, and so are many electronic transactions. The latter of course depend on the country, it is much more efficient in places like China and Scandinavia than for example United States or Germany. However, that is not a failure of fiat money, its a failure of the financial technology used in these countries.
Blockchain based cryptocurrencies like Blockchain are inherently slow if we want to trust the transaction, at least 10 minutes or even an hour. And it is not costless. Over the past six months, the transaction costs for bitcoin have ranged from $55 to $1.3. If we want to speed this up, or make it cheaper by going to a different cryptocurrencies, trust has to give.
Meanwhile, I can transfer any amount out of my bank account to someone else instantaneously at no cost, via my mobile phone, at least here in the UK.
Fiat money is also much more efficient as a store of value, at least in those countries that follow some principles of monetary policy.
And that leaves the strongest argument against cryptocurrencies as a good form of money, which is the same argument one makes against basing money on real assets like gold.
The concept of money is not simple, it has multiple forms see, M0 is cash, M1 is instantly available money like cash and checking accounts and M3 is a broader definition take into forms of money that is similar but broader and less liquid than M1.
If we compare the economies with the largest money supply to the four largest cryptocurrencies, (see the following picture)
the volume of M1 fiat money is over 100 times that of cryptocurrencies. If cryptocurrencies are to replace fiat money, they need to increase significantly in value. That will of course be highly valuable to existing holders of cryptocurrencies, but at the expense of the rest of society.
Therefore, even if one buys the argument that cryptocurrencies will replace fiat money, we need something more fair than existing cryptocurrencies.
If we want money to make the economy work more efficiently, the supply of money needs to adjusted to best facilitate economic growth. Things like gold or bitcoin just don’t do that. Well managed fiat money does.
The central banks are able to control M0, but the broader definition of money, the less control they have. In a financial crisis, we start seeing people converting higher forms of money into lower forms. For example, we might not trust banks so we cash in our savings and stuff the money under the mattress.
That leads to a reduction in the supply of money, because of how the higher forms of money are created. Suppose I put $1000 of cash into my checking account, the bank lends $900 out of that to someone else keeps it in their checking account, the total amount of M1 is $1900 while M0 remains $1000. If that loan is used to back real economic activity, like a loan to a small or medium-size enterprise, this deleveraging will reduce economic activity.
The following figure shows the supply of money in the United States in the Great Depression 1929 to 1933:
We see a rapid reduction of supply of M2, while M1 moderately increases, signalling that people are deleveraging on a rapid scale, which in turn will slow down the economy and result in a recession.
We have as we have seen the same on a smaller scale since 2008. The following figure shows M1 and M3 in the euro zone. At the height of the European crisis, M1 was growing rapidly while M3 was contracting.
Most importantly, fiat money issued by a credible modern central bank is vastly superior to money based on real assets like gold or cryptocurrencies, not least because the supply of fiat money can be adjusted to best serve the economy
Suppose then we don’t think fiat money does this well. It didn’t in the 1970s, giving us stagflation.
In a free market, the best money would win out, as eloquently expressed by Hayek’s 1977, “Free-Market Monetary System”. A good modern analysis of that is by Jesús Fernández-Villaverde and Daniel Sanches “Can currency competition work?”.
However, monetary policy has improved since then, and we now know that a credible central bank, with sufficient political cover, using inflation targeting, is the best way to achieve price stability and stable growth without too many deep recessions.
Cryptocurrencies do promise freedom, as this article Welcome to Liberland: the nation that Bitcoin built. This relates to the ideas on self sovereign identity.
Fiat money is controlled by the government, and governments are anything but shy in using their powers over money to control their citizens and other countries.
Therefore, for those who resent such government control, forms of money that are outside of the control of government and whose quantity and integrity are guaranteed by a technical solution, are attractive.
This is a long-running debate that predates cryptocurrencies. For example, in the United States in the 19th century (witness the debate over the establishment of the Fed in 1913), and the free market monetary discussions in the 1970s.
And for those for which this matters, cryptocurrencies can make sense. This however is a tiny fraction of society.
And even then, I question the freedom one gets, because even with cryptocurrencies, the governments can and will exercise control.
Many governments are going after profits made by trading cryptocurrencies, which takes away much of the perceived freedom benefits. How can the government monitor that? Easy, it sees every transaction in and out of your bank account, and if there is a cryptocurrencies exchange on one side, the government will know.
If then one says that “in the future I will be able to make all my transactions without needing to have a bank account”, that is only possible if the government allows you to. The government has power to control any economic enterprise, it can prevent them from accepting cryptocurrencies, or require reporting of their use.
The governments have the power to ensure money controlled by them remain Legal_tender and they will certainly do so.
The idea that cryptocurrencies will provide freedom is just a dream. On planet Earth, it will not happen.
We have to trust that all these entities have our best interests at heart and are keeping our money safe. We also have to trust the government not to confiscate or devalue our money.
Cryptocurrencies promise to replace that with algorithms, see e.g. Bitcoin security model: trust by computation. We trust the network because the interaction between market participants is protected by algorithms that are practically impossible to manipulate. If we want to hold an an asset that does not require us to trust any trust third party, an asset that that can not be censored or confiscated, cryptocurrencies might seem attractive.
So long as we trust the algorithm, we are safe.
Except, this is only the theoretic middle part of the transaction. There are a lot of practical implementation details that erode the trust.
Some tiny segment of the population is sufficiently technically adept so that they can implement the entire thing themselves, and trust their own work. The rest of us have to rely on someone else for implementation.
And then we are left with trusting unknown entities. Here is a small list of what can go wrong.
Electronic fiat money and traditional assets are of course subject to all of these to some extent, but on a much smaller scale than cryptocurrencies, not the least because we are protected by securities law, financial regulations and the legal system.
If we take elementary precautions, internet banking and electronic fiat money transactions are quite safe, and we are protected by multiple layers of security. The chance of hacking is very low and in the event, we have recourse.
I am perfectly happy to do online banking without constantly looking over my shoulders or resorting to air gapped laptops.
The concept of trust does not only apply to the algorithm, it applies to the entire transaction. And I trust my bank and my payment system much better than any cryptocurrency intermediary.
So, from the point of view of trust, fiat money is much superior to cryptocurrencies.
Some cryptocurrencies, like Monero, promise privacy. That we can enter into a transaction without anybody knowing about it except we and the person we are dealing with.
The most popular, Bitcoin does not, unless one is really careful in hiding one’s tracks, using skills that are only available to a small group of users. The reason is that transaction records on the blockchain cannot be changed or deleted and are therefore searchable.
However, trust is provided by the blockchain being visible. If we want purely anonymous transactions, trust has to give. The question is to what extent.
There is no such thing as 100% privacy. Fiat cash is fully anonymous, but someone might be monitoring the transaction. If we move to electronic fiat money, we are subject to tracking, both by private companies and government authorities.
Same with even the most privacy focused cryptocurrency, the starting point is the internet, giving scope to monitoring.
Yes, it is conceivable that two entities are able to conduct business by only using a privacy based cryptocurrencies, with correctly implemented end-to-end encryption and no monitoring of the exchange of goods. Even then, the transaction would have to be based on some goods that are outside of the standard economy — like drugs.
And meanwhile, while bitcoin is the most liquid cryptocurrency, it is not exactly what one would call liquid in the sense that fiat money is. Moving to some untested and highly illiquid cryptocurrency that promises privacy and trust, needs a considerable leap in faith and belief.
However, this one area where a central bank issued cryptocurrency might have an advantage, even if, I don’t think they would want to.
Some cryptocurrencies, most prominently the second largest Ethereum, are not really designed to be replacements for money, but provide other services.
The most visible is smart contacts, something that sounds really clever until one extra tries to put in practice, when it becomes distinctively pedestrian.
The starting point is some distributed ledger technology, like blockchains, for holding assets. Then, smart contracts replace the need of having to employ lawyers and even third-party escrow.
This can certainly be beneficial, and is likely to be increasingly implemented, but it can certainly be done without a cryptocurrencies attached. A crypto advocate might say that trust is provided by the cryptocurrency, but there is nothing conceptually preventing us from using trusted institutions.
If we don’t like that, we are back to freedom, as discussed above.
Many advocates of cryptocurrencies argue that while for those living in developed countries with relatively credible central banks and governments creating fiat money there may be little reason to move to cryptocurrencies, some countries have unstable governments and large segments of the population unbanked. Cryptocurrencies then solve that problem. Often cited examples include places like Venezuela and Zimbabwe.
In countries with high inflation, people usually seek out other currencies, typically the US dollar. Transactions might be made entirely in dollars or prices may be quoted in dollars while transactions take place in local currency at the spot rate.
This is called Dollarization or currency substitution.
I can’t see how citizens of countries with unstable monetary policy are better served by cryptocurrencies than the most widely available fiat currency, the USD, or possibly the Euro.
The problem of the unbanked can be solved by fintech, banking via mobile phones, and the like. However, such a solution is currency agnostic. It is a technological solution and one can plug in any currency. The unbanked would be much better served by a stable fiat currency that is accepted everywhere coupled with innovative Fintech solutions.
Any asset can get into a bubble. People will buy it because they expect others to pay a higher price in the future, creating a positive feedback.
A bitcoin was worth $0.06 in 2010, and $9800 on last count. So a 16 million percent return.
Someone who invests early and gets out in time, will make money, just like an early investor in a Ponzi scheme will make money provided she gets out early.
So this leaves two questions:
Investments like stocks and bonds have value because we have expectation of future income.
Other assets have value because we expect people to pay for them in the future.
Collectibles are of the latter category. The Wall Street Journal ran an interesting story recently Sorry, Collectors, Nobody Wants Your Beanie Babies Anymore “Over two decades after the great Beanie Baby craze, speculators are back, hoping someone will finally buy their floppy collectibles.” It is the same with collecting art and stamps. Collectable stamps have scarcity value, some cost more than $200k, just make sure to buy at the right time.
Fiat money also falls into this category. It only holds value if the central bank and the government manage it properly, and in cases where they do not, the use of fiat money can be very costly, and in extremis result in hyperinflation.
Cryptocurrencies fall into the same category as stamps, fiat money and Beanie Babies, not stocks or bonds. Their price is derived from what people will want to pay in the future, not from a revenue stream like stocks and bonds.
That does not mean there is not money to be made. However, one is well advised to keep the lessons from global games in mind. You can see the model in my slides on this in a currency framework here, from page 57.
It is especially important not to be affected by hindsight bias. Just because the price of something increased in the past, does not mean that it will increase in the future.
Those who have made money out of cryptocurrencies have done so out of luck, not because of anything fundamental.
A counterargument might be that the supply of cryptocurrencies is limited by costly mining and may have a fixed asymptote — just like gold. As to the former, sunk cost should not affect the value of assets, but the limit to supply is more relevant.
However, that is only an advantage if the alternative is unstable fiat currency. In countries with credible and well-managed central banks, the ability to adjust the supply of money to meet the needs of the economy is highly valuable, as discussed above.
Most central banks are actively studying cryptocurrencies and have even considered issuing their own. See for example the Bank of England and the more academic BIS paper “Central bank cryptocurrencies” by Morten Bech and Rodney Garratt.
So why should a central bank do that? Besides just keeping up with popular technology, many central banks would like to get rid of cash, like the Bank of England (or at least its chief economist). While they might think they have a good reason to do so, it is a terrible idea.
But why should a central bank issue cryptocurrencies? Bech and Garratt argue that retail clients might benefit from anonymity and the ability to hold accounts directly with the central bank, while wholesale clients might benefit from increased efficiencies. Furthermore, a central bank issued cryptocurrency might have the advantage of a fixed supply and one-to-one convertibility to other forms of money. Such money might solve the problem of the zero lower bound. Furthermore, a central bank cryptocurrency would likely have the price stability that other cryptocurrencies lack.
All of these make sense, but seem to be rather minor advantages, except the zero lower bound and anonymity (and I suspect the central banks could not countenance that very same anonymity,) and the broader discussion of control of money supply. Under the current system, the central banks do not fully control the money supply, except M0. This is why we had all the experimental monetary policies, low interest rates and QE over the past 10 years.
So, if the central bank issues cryptocurrencies, the supply of money, in all its forms, all the way from M0 to M3, can be controlled — in theory.
That might be a good idea because, conceptually, it gives more fine-grained control of inflation, can insulate failing intermediaries in a crisis, and perhaps most importantly, provide control of higher forms of money in a crisis when everybody is rushing to convert M3 to M0.
Except, it is more complicated, as trust gets in the way of the efficiency.
This is because of what happens in a crisis, as discussed above. Suppose the economy is deleveraging rapidly. Theoretically, with a central bank created cryptocurrencies, that process can be prevented. However, if the central bank tried to do so, its trust would evaporate.
That means, trying to control deleveraging in times of crisis could well end up amplifying the same crisis.
It is important to recognise the distinction between fintech generally and cryptocurrencies. Many of the good use cases of cryptocurrencies really are just fintech where one can plug in any type of money. This relates for example to the question of the unbanked and the use of blockchains without a currency attached.
For all its faults, I cannot see how fiat money issued by a credible central bank in the 21st century is worse than any of the cryptocurrencies. There simply is no evidence to the contrary. Saying that fiat money is bad and therefore alternative such as cryptocurrencies must be better does not make any sense unless one can show how. And that should be done in the context of the real world and how people actually use money, instead of some abstract theories of how we should think about money.
I think many of the cryptocurrencies advocates who reference freedom or trust or theories of money might be well advised to keep the following in mind:
“Practical men who believe themselves to be quite exempt from any intellectual influence, are usually the slaves of some defunct economist. Madmen in authority, who hear voices in the air, are distilling their frenzy from some academic scribbler of a few years back”
—John Maynard Keynes
What most cryptocurrency advocates generally miss is that we are not discussing cryptocurrencies in a vacuum. There is entranced incumbent technology that works really well. Show how you plan to beat fiat money, preferably without resorting to mysticism.
It is not surprising that so much of the cryptocurrencies discussion verges into mysticism. Conventional fiat money also has mystical elements, as elegantly shown by John Moore’s Claredon Lecture “Evil is the Root of All Money”:
“money and religion have much in common. They both concern beliefs about eternity. The British put their faith in an infinite sequence: this pound note is a promise to pay the bearer on demand another pound note. Americans are more religious: on this dollar bill it says “In God We Trust”. In case God defaults, it is countersigned by Larry Summers.”
I still think that cryptocurrencies are more like a religion or a cult, not a rational economic phenomena.
I still await my enlightenment.