My Biggest Doubt About Bitcoin

DISCLAIMER: At the time of writing, I own bitcoin. I like the project and want to see it succeed. I don’t approve of or condone attacks on the Bitcoin network. I am not a financial advisor and this is not financial advice. Do your own research and consult with a qualified professional before investing.

TLDR

I’m concerned about the future of Bitcoin. It worries me that this might be true:

  1. A large percentage of Bitcoin is uninsured and held by people who have significant amounts of money in it.
  2. Bitcoin will be successfully attacked. A significant minority of people will irretrievably lose important amounts of money. Too little will be lost for the network to agree to a hard-fork to reverse the damage.
  3. When news of (2) spreads, there will be a kind of Bitcoin “bank run”. People will realize that personally significant mounts of money can be lost without recourse, and they will rush en-masse to exchange their Bitcoin for fiat.
  4. The value of Bitcoin will plummet. The bank-run, coupled with a liquidity crunch will make Bitcoin untradeable for fiat on exchanges, rendering it effectively worthless.

This argument is gappy, I fully admit that. But I think each of its premises is plausible enough, and its connections tight enough, that it succeeds in justifying some degree of worry. I’ll argue for each of its premises below.

1. Most Bitcoin is Uninsured and Held in Significant Quantities

It probably goes without saying that most Bitcoin is uninsured. Bitcoin wallets are not FDIC protected, and private coverage is either non-existent or hard to come by.1

It’s probably equally obvious that the vast majority of coins is controlled by entities (people/companies/governments) that own a lot of them. At the time of writing, 87% of coins are held by just 0.65% of addresses, and 62% of coins are held by addresses that contain 100 or more coins. Obviously, many of these addresses are controlled by the same entities. Which is to say that these entities own enough coins that it would probably cause them great distress to think they were at risk of losing them.2

If either of these claims seems controversial, or doubtful, check out my footnotes for more details. I think most people would agree with them, though, so I won’t say more here.

2. Bitcoin Will Be Successfully Attacked

I’ve argued at length in a previous post why I think it is almost certain that Bitcoin will be successfully attacked at some point in the future. It is only a matter of time. Briefly, the argument was:

  • the Bitcoin core software has to continue changing indefinitely
  • each change adds a non-zero chance of introducing an exploitable bug
  • after enough changes, the probability of an exploitable bug approaches 100%; it is now ~99.5% even under conservative assumptions

So, I’m just going to assume that at some point Bitcoin will be successfully attacked.

Who cares?

Why should anyone care about a successful Bitcoin exploit? After all, we all saw what happened when the DAO was hacked in mid 2016: the Ethereum community just hard-forked to restore the stolen funds. Wouldn’t the Bitcoin community just do the same thing? Some people, Arjun Balaji for example, find this line of reasoning pretty compelling, and so are unconcerned about bugs. As he says:

If there is a bug in Bitcoin which causes undue inflation I think that we would patch the bug [and] fork to a different version of Bitcoin because that would be the new consensus. I think if there was a massively inflated Bitcoin almost no one who holds Bitcoin would want to participate in it. So I worry less about bugs like that. [Quote begins at 2:30].

It’s a good question whether a fork could save the network in the event of an exploited bug. Maybe the Bitcoin community would rise to the challenge and accept a hard-fork to reverse a bug exploit. Then again, the level of disagreement and partisanship within the Bitcoin community suggests that this is a non-trivial assumption. Mike Hearn’s medium post from early 2016 gives an insider’s view of just how divided the community was (and still is today). As Hearn put it:

[D]espite knowing that Bitcoin could fail all along, the now inescapable conclusion that it has failed still saddens me greatly. Why has Bitcoin failed? It has failed because the community has failed. [Ref].

And let’s not forget that when Ethereum forked it was still in its infancy, still had not reached mainstream attention, and had a network value under $1 billion dollars. Ethereum had much less at stake than Bitcoin does now.

But let’s assume Arjun is right: major exploits to Bitcoin would be reversed. Everyone would behave selflessly and recognize that they shouldn’t allow attackers to get away with other people’s money.

But what about minor exploits?

What if the attack that succeeds is fairly small, effecting just a small fraction of Bitcoin’s total value? Maybe just a hundred people, or a thousand. A couple discreet exploits of the same bug, over the course of a few hundred blocks. A few million dollars are stolen in total. A drop in the bucket. A fraction of a fraction. It takes a while for people to even notice.

The Bitcoin community would have far less incentive to hard-fork to reverse a small attack like the one I’ve described. Very few people would be helped by it. Those who would be helped would have lost relatively little. If the attacker was smart, he/she could target people along party lines, so that only one faction within the community has a stake in forking. Or they could attack people who have no voice in the community at all, who would be hard-pressed to drum up feelings for their plight. Or they could space out attacks over several blocks so that many legitimate transactions would have to be reversed as well, ensuring that forking would be painful for those who have not been stolen from. Perhaps the attack could be timed to coincide with a number of big transactions, so that reversal would be maximally painful, and the collateral damage very large. And remember: hard-forks are inherently very risky. Risking a permanent split to the network – a potential halving of the network’s value and hash power – to save just a few million dollars for a hand-full of people is a tough sell indeed.

Not only would the community be far less likely to agree on a hard-fork for a small attack versus a big one, but a small attack is also much more likely than a big one. Bitcoin has been battle-tested for over 10 years now. It’s arguable that the big issues have already been found and fixed. There are tests for them. Developers have them in mind and look out for them when writing and reviewing code. A small bug, though – that’s the kind of thing you could create with just a little sloppy programming. Suppose it’s at the periphery of the system. The PR is given only cursory review by one or two people before merging, the whole process overshadowed by more ambitious PRs being reviewed at the same time.

Nothing extraordinary is needed for someone to fat-finger a small bug into Core and get it into a release.

3. There Will Be a Bitcoin Bank Run

After an attack is recognized and the community refuses to reverse it via hard-fork, one could imagine stories beginning to come out about the people effected. (Imagine how eagerly the Wall Street Journal would run with a story like this.) You’d hear about how person A lost his life savings, how B won’t be able to send her kids to college, how C’s down-payment is gone and his kids are sleeping 3 to a room, how D lost her retirement and will have to continue working indefinitely. Etc. The stories would be poignant. They’d trend, maybe even go viral. The point is: they’d get mainstream attention (because Bitcoin is in the public eye), and they’d resonate with most people in the cryptocurrency community, especially those who have significant amounts of money in Bitcoin.

We know how people behaved in the past when they believed that personally important amounts of money were at risk: they ran, en masse to withdraw them. These so-called bank runs triggered the Great Depression, and eventually led to the creation of the FDIC.

Bank runs are surprisingly common. And when they happen, they tend to be devastating:

In the 16th century onwards, English goldsmiths issuing promissory notes suffered severe failures due to bad harvests, plummeting parts of the country into famine and unrest. Other examples are the Dutch Tulip manias (1634–1637), the British South Sea Bubble (1717–1719), the French Mississippi Company (1717–1720), the post-Napoleonic depression (1815–1830) and the Great Depression (1929–1939). [Source]

The kind of scenario we’re considering has all the markers of a bank run: (i) a large number of people (ii) realize at roughly the same time that (iii) there is a real possibility that (iv) a significant amount of their money will be lost forever unless (v) they act fast. Given the history, I think the burden of proof is on someone who thinks that people wouldn’t try to move their money out of Bitcoin in circumstances like this. The assumption should be that history will repeat itself unless there is overriding evidence otherwise.

Hence, I say that a mass Bitcoin sell-off is a live possibility in the event of a successful attack like the one I have described.3

4. The Value of Bitcoin Will Plummet

Price is a function of (among other things) supply and demand. It is directly correlated with demand, and inversely correlated with supply. If demand increases, price tends to rise (all else being equal). If supply increases, price tends to fall (all else being equal). This is Economics 101.

In a bank-run situation, there would be a massive surge in circulating Bitcoin (increased supply) coupled with a concurrent drop in demand. This would constitute a double blow to price.

What’s less obvious is that there is a liquidity problem in cryptocurrency exchanges – where people would overwhelmingly go to sell their Bitcoin. Here’s a quote from Preston Byrne’s excellent article (which also considers the possibility of a crypto bank run, albeit for different reasons):

Holders expect that Bitcoin intermediaries/exchangers (exchanges, wallets, traders) will be in a position to provide them with instant dollar liquidity when they wish to margin trade or take their profits (dollar liquidity which, I am reliably informed, is provided in part by mainstream commercial banks on a very short-term basis at very high rates). [. . .] [I]s it realistic to expect that exchanges, trading counterparties, and other market participants are prudently hiving away their profits to stave off a liquidity crunch?

The question is rhetorical, of course. The odds that the collective liquid (fiat) assets and credit lines of the world’s cryptocurrency exchanges currently total the market capitalization of Bitcoin, some $60 billion dollars, seem exceedingly low. Almost certainly they are a small fraction of that number.

A bank run, then, would likely exhaust the available liquidity in these systems very quickly, meaning that – no matter what the quoted spot price of Bitcoin had fallen to in the immediate wake of the run – there would be effectively no money to exchange for it, which is to say effectively no demand for it. With little or no demand for it, it would be valueless.

Conclusion

A coin that miners cannot redeem for fiat – to pay their electricity bills, to pay their staff – is a coin that miners won’t mine, or won’t mine for long. And a network without miners is neither secure nor functional. In this way, then, a bank run coupled with a liquidity crunch could not only drive down the price but actually kill the Bitcoin network permanently.

Whether this scenario will play out, or play out in exactly this way, is anyone’s best guess. I am simply worried that others are not sufficiently concerned about it.


  1. Bitcoin wallets are not FDIC insured. This means that, for most people’s coins to be covered by insurance, they’d need to have private insurance. But private Bitcoin insurance seems hard to come by at this early stage. A quick Google search turns up no obvious, legitimate-seeming providers, which means that people would not only need to seek out insurance, but also would need to do some non-trivial research to get it. As a niche financial product, it likely would also be expensive. Moreover, the libertarian-, privacy-, be-your-own-bank minded folks that have thus far been attracted to Bitcoin are not the sort to seek out trusted 3rd parties to secure their coins, nor to disclose important details about their financial situation (how they store their coins, how many they have, what addresses they use, etc). And there are good reasons to believe that large numbers of people who own Bitcoin do not live in the US, and do not even have access to all of the financial resources we do. All of these factors mitigate against people having insurance for their Bitcoin. Companies like Coinbase and Gemini have insurance for their hotwallets. But, obviously, most coins are not stored in these wallets.

  2. It seems plausible that the vast majority of coins are personally significant to those that own them. By a personally significant amount of money for a person, P, I mean an amount of value that would play an important role in P’s life. Examples would be a down payment, a college savings fund, a retirement account, a wedding fund, an inheritance, an emergency fund, etc. If your coins are greater than or equal in value to any of these, then you have a personally significant amount of value in Bitcoin. Losing this amount of money would probably really matter to you. I say that most coins are probably personally significant for two reasons. First, because of the statistics I cited earlier: 62% of coins are held by addresses that contain 100+ Bitcoin. Second, because people routinely, openly, proudly broadcast that they actually are putting personally significant amounts of money into Bitcoin. People are putting their life-savings, their retirements, their homes into it. A quick Twitter search shows just how common this kind of thinking is. There is accordingly a growing market for financial services that convert retirement accounts into Bitcoin.

  3. It bears emphasizing that the bank run would probably include other decentralized cryptocurrencies as well. This is not only because most of these currencies are still purchased with Bitcoin – i.e. because they must be bought/sold for Bitcoin before they can be bought/sold for fiat currencies – and the loss of a functioning Bitcoin network would effectively lock up their value until a new intermediate currency came along. That’s one issue. But I also think the other decentralized cryptocurrencies would be sucked into the bank run because the exact same concerns I have about Bitcoin apply to them. They are just as prone (if not more so) to attack, and they are uninsured. These are not safe-harbors for one’s personally significant sums of money. They thus would not likely absorb the value exiting Bitcoin. Nor would they escape the psychological flight to safety.