Bitcoin: promises and problems
Posted in Technology on 2013-06-23 17:18
Future Bright Amusement Park, Macau
Following on from my explanation of how Bitcoin works I wanted to go deeper into to what degree it works as promised, and what its future is likely to be. We'll go through the points one by one.
"Nobody uses it"
Most economists I've read dismiss Bitcoin because, they claim, nobody really uses it. Well, let's look at the numbers. Right now there's 11.3 million Bitcoin in circulation, which at the current exchange rate gives Bitcoin a monetary base of 1.2 billion USD. In the last 24 hours there were 42,663 transactions, at a total value of 546,252 bitcoin, or 59 million USD. That translates to a GDP for the economy of Bitcoinia of 21.5 billion USD. That's roughly the size of the economy of Estonia, which is not bad at all.
Probably some percentage of these transactions will be transfers between public keys controlled by the same person or organizations, so this overstates Bitcoin's GDP a little, but probably not very much. And having an economy the size of Estonia's is not bad at all.
Another data point comes from the London pub chain that accepts Bitcoin. They took in 750 GBP in three weeks after they started accepting Bitcoin. Again, that's not a huge amount, but neither is it trivial, and it shows very clearly that Bitcoin is being used for real.
That story also shows that it took one hour of programming to add Bitcoin support to the pub chain's point of sales systems, which is interesting. It suggests that for a vendor, starting to use Bitcoin is actually very easy. The Bitcoin documentation bears this out. By comparison, accepting Visa payments is both difficult and expensive.
A key aspect of the functioning of a national state is the collection of taxes. Collecting many of these taxes require knowing the salaries of individuals, and (for sales taxes) being aware of every individual transaction in the economy. It follows that governments are not going to accept having substantial parts of their national economies disappearing into an uncontrollable network like Bitcoin. Essentially, if the government demands that you get paid in euros, and that you conduct most of your transactions in euros, and that you pay your taxes in euros, that's what you are going to have to do.
So it seems clear that Bitcoin is not going to replace official currencies, simply because states will not accept it. That said, however, there is no inherent reason why Bitcoin couldn't become a useful "side currency", used for certain types of transactions or for parts of the economy. Provided, of course, that this use of the currency does not interfere too much with tax collection and other goals the government considers essential.
The money we use today is known technically as "fiat money", meaning that unlike previous money (which was backed by gold), it is backed by nothing more than legislation. That is, in most states the law requires you to accept the local currency as payment for debt. In quite a few, you are also required to accept the currency as payment for goods. You are allowed to accept any other form of payment you wish (including foreign currency or Bitcoin), but if someone wants to pay with the local currency, you cannot refuse.
So "fiat money" has no intrinsic value, and people use it basically because everyone else does, and because the government forces them to. Hence the name; "fiat" being Latin for "it shall be". Bitcoin has been described as "the ultimate fiat currency", since it is not backed by anything (not even legislation), and people use it only because other people do.
I would argue that this makes Bitcoin weaker than fiat money, because it's susceptible to a kind of bank run, or panic attack. Basically, if something happens to shake people's faith in Bitcoin, then people would start leaving the currency. This would cause the value of Bitcoin to sink (in fact, it would effectively be inflation), which would encourage more people to leave. This could turn into a self-reinforcing spiral, which need not have any bottom. So, in fact, because Bitcoin is based on nothing, the whole thing could collapse in a very short time.
Contrary to what is commonly claimed, Bitcoin is not anonymous. Instead, as the official documentation says, it's pseudonymous. That is, anyone can see what sums are transferred between what public keys when. So the moment that someone discovers which public keys belong to whom, the whole system is blown open, and every single transaction in the entire Bitcoin economy becomes visible.
The question is whether anyone can work out who owns what keys. Of course, the moment you do a transaction with someone, you inevitably learn who owns the key you transfer the money to. Many people also publish their keys in various ways, to ask for donations, tips, and the like. So part of this information is already public, and quite a bit of it will inevitably become known through the use of Bitcoin.
There are ways to combat this, such as using many public keys to receive and send Bitcoin, so that your identity is not tied to a single key. Many Bitcoin clients automate the creation and use of more keys to make this easier. Vendors are also recommended to create a new public key for every payment they receive. However, there are limits to this approach, since all transactions between any pair of keys are visible.
The research that has been done on anonymity in Bitcoin strongly suggests that it's possible deanonymize Bitcoin keys to a substantial degree. This is consistent with other research that shows how anonymized research data (such as movie ratings) can be effectively deanonymized by careful analysis. Modifications to Bitcoins to fix the problem, such as Zerocoin, have been proposed, but not adopted yet.
For now, it seems that Bitcoin transactions are partly public, particularly to entities with high motivation and considerable resources in data mining. This is troubling, and does not bode well for the longer-term adoption of Bitcoin. People may not care that others can see the photos of their breakfast uploaded to Facebook; everyone knowing every detail of their financial transactions is likely to feel considerably more sensitive.
Fast and slow transactions
As described in the previous blog post, to prevent against double spending, users have to wait for their transaction to be accepted into the block chain. Since the block chain is constantly forking it's not enough to have your transaction accepted into the last block in one of the forks. To be certain, users have to wait until some blocks have been added to the end of the chain. Currently, experts recommend waiting until 5 more blocks have been added.
The trouble with this is that the protocol adjusts the difficulty of creating new blocks to keep the average block creation time to 10 minutes. It follows that on average you have to wait about an hour to be certain that your transaction is accepted.
Researchers have demonstrated that double-spending attacks against transactions that wait only a few tens of seconds are possible. So the problem is definitely real. There have been proposals to fix this problem, but whether they are 100% secure is not clear, and in any case they have not yet been adopted.
So, basically, fast Bitcoin transactions are not safe at the moment, making applications like Bitcoin ATMs, Bitcoin payments in physical stores, etc impractical. Or so one would have thought. The pub I mentioned above sells beer via Bitcoin using fast transactions, basically taking the chance that someone will double-spend. The odds of that are probably pretty low, and as the owner says, they're not taking the beer with them. Five minutes later, the customer will probably still be in the pub, with the beer in front of them.
So perhaps this is not as much of a problem as people have thought. I would think that most transactions that need to be fast are also low-value transactions where sellers may want to accept the risk of being scammed, while high-value transactions are probably transactions where it's OK to wait an hour for them to complete.
No central bank
By design, Bitcoin has no central authority of any kind, which means it doesn't have any central bank, either. The monetary base increases via mining, but the number of new bitcoin in each block decreases over time, and around 2021 all bitcoin which will ever be created will have been created. This means that the Bitcoin monetary base does not respond in any way to demand, and that's what's driven the increase in the value of Bitcoin, which has gone up quite dramatically over time. That sounds like it's great for people who own Bitcoin, and like it makes Bitcoin a good investment. At the moment it seems to be, but there's a problem.
Bitcoin has so far experienced constant deflation. That is, the value of cash has increased all the time. As long as the user base and Bitcoin activity continues to increase, this will go on. The problem is that deflation discourages people from spending, and encourages saving. A little saving is good, of course, but if everyone tries to save money at the same time, it reduces economic activity, causing a recession. In fact, that's what's behind the current economic crisis, where people and companies took up too much debt, and then sharply reduced their spending to pay down the debt.
This problem is well known, and this is why central banks around the world generally set an inflation target around 2%. When inflation goes below this level (as it has in all of the crisis-hit economies), central banks respond by essentially printing money. So, basically, the Bitcoin economy is designed for permanent recession.
That's would be terrible for a national currency, but how badly it will affect a currency that's basically a "side currency" for online transactions is not clear. It's possible that the only effect will be to reduce the number of Bitcoin transaction.
However, the absence of a central bank also makes the currency susceptible to wild swings in exchange rates (which have been seen in practice). This, combined with the long transaction times, is also problematic for the use of the currency. So far that doesn't seem to have stopped people from using it, though.
Immaturity of infrastructure
Another problem is that while the Bitcoin network itself is secure, the infrastructure around it is not. Since the only thing that prevents others from making transactions from your public keys is that they don't know the private keys, it follows that someone who copies the private keys can steal all your money. This happens fairly frequently, as users have their machines compromised in various ways, and malicious intruders steal their bitcoin.
You can avoid this by keeping your bitcoins in an exchange, but the exchange then effectively becomes your bank, and if there are problems with the exchange you may lose your money, or lose access to it. Sadly, these problems also occur with disturbing frequency. For example, at the moment the biggest exchange, Mt. Gox, has suspended all accounts for reasons that are not wholly clear.
It's clear that there is a fair amount of risk associated with using Bitcoin at the moment.
It should be noted that Bitcoin is far from the only digital currency out there. There were a few around before Bitcoin, and after Bitcoin started a whole range of similar currencies have started up. Most of them use a similar design, but some, like Ripple, are fairly different. One of the biggest, PPcoin, is designed for a constant inflation rate of 1%, and it will be interesting to see to what degree that will help the currency compete.
As the use of Bitcoin expands mining has become more difficult, and at the moment it takes about a year to create a block on a normal laptop, and miners have started using custom hardware for mining. In fact, miners spend about 150,000 USD in electricity every day, and earn about 500,000 USD.
However, it might well be that as the difficulty goes up it becomes more profitable to mine other currencies, in which case Bitcoin could find itself outcompeted by another currency. Or, Bitcoin might be damaged by security issues, 51% attacks, problems with exchanges, or legal challenges, causing people to switch to another currency. If any of this happens, the "panic attack" scenario could play out very quickly.
At the moment, the situation is rather unstable, and that's a worry for anyone using Bitcoin seriously.
Distrust of high finance
(Note, this section was added after publication, because I noticed I'd forgotten to address part of Svein's comment.)
Viewing Bitcoin from a political and social perspective it must be understood in light of the financial crisis and the deep mistrust this has created to the established financial system.
This is a view I fully agree with. Over and over on the Bitcoin forums I see people expressing anger not just with the banks, but also with government, which they tend to see as "in the pocket" of the banks. And, of course, this view is not entirely illogical, given what we've seen over the last few years. The question, of course, is to what degree Bitcoin can help.
It helps to understand the cause of the problem first. In the US it was to a large degree deregulation of the banks. That is, the laws that had required banks to not take on too much risk had been watered down, with the inevitable consequence that banks did take on too much risk. One could blame greedy bankers for taking on that risk, of course, but as Daron Acemoglu pointed out, people have always been and will always be greedy. That's why we have laws and institutions: to protect us against the consequences of people's greed.
So the problem here was the insufficient regulation of banks. Clearly, moving substantial parts of the economy into an uncontrollable and untraceable peer-to-peer network is not going to help us keep the financial sector under control. The problems of risk and regulation are effectively the same, regardless of what currency is used. So this is a political problem, and it needs a political solution.
In short, I think Svein is right that distrust of the financial system has fuelled interest in Bitcoin, but I consider this interest misplaced. We're now in the process of recreating classical financial institutions inside Bitcoin, but nothing about the currency or the protocol will protect these institutions from the exact same problems that have plagued their offline counterparts.
Future Bright Amusement Park, Macau
It seems clear that the best Bitcoin can hope for is a role as a "side currency", used for online transactions. In some ways it seems fairly promising for this role, given that it's very cheap to use, easy to integrate, and highly transparent. Some of the issues with using it at the moment, such as a lack of anonymity and security, may be addressed with time. The deflation issue, however, is likely to be a drag on the use of the currency, and the issue of competition from other currencies is also a worry.
However, the biggest concern for me is that it seems that the same role could be filled by an online payment service that takes a similar approach. Imagine an alternative to Paypal which has a simple API, is easy to integrate, and where transactions are non-refundable and cheap. Such a service could operate directly in known currencies, and would be free of many of the problems with Bitcoin.
Granted, a system like this would be privately owned and operated, but so is your bank. That doesn't stop you from using them, nor does it stop people from using Paypal. So it seems clear to me that a service of this kind would stand a good chance of outcompeting Bitcoin. And ultimately this is how the issue will be resolved: people will vote with their wallets, and one of the solutions (quite possibly one that doesn't exist yet) will end up as the winner. It's not at all clear that the winner will be Bitcoin.
Joe economist - 2013-06-23 13:56:36
Your comments about deflation are the often repeated common view, they are contradicted by actual experience. What has actually happened is that spending increases as bitcoin appreciates. When the value of your cash holdings doubles, you are more likely to spend some of it, not less.
Also it does not follow that individuals will ignore their immediate personal interest because of the fear of macroeconomic effects. Individuals will prefer to hold a currency that will increase in value over one that looses value.
Lars Marius - 2013-06-23 14:10:00
@Joe Economist: Bitcoin usage has increased, but that tells us nothing about the effect of deflation, because obviously usage will increase as more users join the currency. Similarly, if the population of Japan had quadrupled, obviously deflation couldn't prevent GDP from increasing.
It's true that having more money makes you more likely to spend, but the thing about deflation is that if you wait longer you'll have even more money, giving you a strong incentive to wait as long as possible. Similarly, investments will have to give a real rate of return higher than deflation to outcompete the return you'd get from simply sitting on the money.
You're absolutely right that individuals will prefer to hold a currency that increases in value, but that's precisely the problem. A currency exists to be used, not to be hoarded. If everyone saves at the same time you get what's called a Minsky moment, and GDP plunges.
Joe economist - 2013-06-23 16:36:30
You are correct about GDP. Creating more money necessarily increases spending as measured in the units of money that were created. Rothbard famously parodied this relationship by saying "the volume of water that runs off the ground is the same as the amount of rain that falls from the sky". While the statement is true, it's not particularly useful for predicting the weather.
Your statement about currency existing to be used also ignores the motivations of individual actors. Individuals hold currency (or "hoard" or "save") in order to shift purchasing power into the future, even when it's only for a short period like from Friday's paycheck until lunch on Saturday. They will tend to choose whatever medium best allows them to do this. Bitcoin will succeed or fail based on how well it allows people to maximize their future purchasing power relative to the alternatives.
I recommend the following post which provides a more in depth exploration of this phenomena, based on Carl Menger's theory of the origins of money:
Lars Marius - 2013-06-24 06:11:17
@Joe: The spending issue is not just about GDP. When GDP falls, unemployment goes up, etc. Recessions have very real effects, and GDP is just a useful health indicator, no more.
I agree with your second paragraph, but note that the longer people hold their Bitcoins, the less Bitcoin will be used. But, as you say, ultimately the outcome will be decided by competition.
I note that this Mencius Moldbug you link to agrees with my comparison of fiat currency versus Bitcoin, but his comments about deflation are totally at odds with accepted economic theory (and, I might add, actual experience).
Svein ělnes - 2013-06-27 09:02:28
I very much agree with your main conclusions, perhaps apart from your concern that Bitcoins' role could be copied or adapted by an online payment service. I would actually welcome that.
Initially I was very intrigued by the way Bitcoin was designed and how it worked technically. But when you view it from the monetary side and compare it to fiat currencies, it becomes clear, at least for me, that it has some serious shortcomings and unwanted effects (deflation probably being the most serious).
I have therefore come to view Bitcoin as an interesting alternative to current monetary systems but most of all I think it has given valuable contribution to fundamental challenges such as trust in a decentralized system and micro payments. Still, I agree with you that Bitcoin will for the most part be a "side currency".
Lars Marius - 2013-06-27 10:39:29
@Svein: Yeah, I guess "concern" was the wrong word. The idea of Bitcoin being supplanted by a normal bank-like service is perfectly OK to me, too. Perhaps even preferrable, as it doesn't leave a chunk of the world economy in an unregulated P2P world.
I agree that the technical design, and the way Bitcoin builds trust, are really interesting, and it's entirely possible that other applications in other areas may build on this design to do something completely different.
Stig - 2014-01-09 03:41:53
Mindgame of the day: StigCoin - a BMW currency Using a cryptocurrency token for value transactions, but with a fictive value not tied to facevalue of the underlying currency.
Some problems with BTC for my usage:
* Volatile facevalue which in time will inflate. Neither a merchant nor a customer wants to take the risk of betting against a currency conversions
* Trusted transactions can take a long time
StigCoin to the rescue; Picture a casino:
When you enter, you pay 1000$ at the cashier, and get a fist full of plastic chips with. You can walk from pokertable to slot machines, win some, loose some and throw a couple at the good looking chap who fills up your with assorted entoxicating beverages and places his bum on your lap.
When the evening is over, you drag yourself back to the cashier, and you expect to get back 122$ from whats left in your pockets. You have a look at some of the pieces of plastic. Most look like they've circulated in this and other casinoes in this chain for ages. You even recognize the first you bet and lost from your teethmarks. Must have lived its own life from when you saw it last. a 1StigCoin chip is all thats left from the sweet 100StigCoin you split up at the second table.
The cashier gives you 102$ and rejects the 20StigCoin which is a fake someone scammed you into accepting.
What happened under the hood?
The cashier has a treasureChest of BTC bought previously at Mtgox, each marked and registered belonging to the casino. When you as a customer "buy" 1000$ worth of StigCoin with your visacard, the cashier does a banktransaction of 1000$ + fees and stores the moneys in their account. The value of StigCoin is controlled by the Cashier. It can be bound 1-1 to USD or NorwegianOilKroners and devalued at will. A tranfer of BTC is done from the cashiers wallet to your authorized wallet. The StigCoin have a facevalue of 1000$, but are actually BTCs worth 50cents. The only way of anyone converting the StigCoins into $$$ is through the Cashier, which looks at the transactionchain of each coin, and verifies that it has only been transacted by authorized BTC wallets and signatures. It also waits long enough to ensure that the chips haven't been spent more than once.
* Trust, trust, trust. Especially that the casino can trust the chips they accept and convert into fiat currency
* Decentralized - can even be used offline and for transactions outside of the intended system
* Transaction history, anti money laundering benefits, anti fraud
* Low in-game transaction fees
* Piggyback on existing infrastructure
* Someone has done the thinking for you
What can our customers use StigCoin for
* Ingame purchases
* Online and offline gambling
* Ecommerce and IRL commerce (peddlers on the street corner)
* Travel- or cantine- purse-on-a-credit card (Ruter Kort)
* Bank transactions, even abroad