It’s probably a sign of the times that every self respecting value investing blog has to have a post about bitcoin. And when you tell a random person that you invest the first question these days is if you own (or trade) bitcoins or not. So I guess “The Alpha Vulture” can’t stay behind, and share with the world what it thinks about bitcoin, whether you want to hear it or not. As the title of this post implies I’m skeptical about bitcoin, and most other software projects that are based on blockchain technology[1].
To explain why I’m skeptical I’ll first try to explain in the simplest possible way what the whole fuss is about because I suspect that part of the reason why people are so enthusiastic about the blockchain is the fact that they don’t really understand the technology. If you read an explanation about bitcoin you hear cool terms like distributed trust (which is actually cool) and stuff you probably don’t understand without a computer science background (like hash functions). These things are actually not that hard to explain, but not really necessary to understand the key idea behind the blockchain. What you basically do is, instead of trusting security to a single party that (hopefully) has a smartly designed system you substitute that by a large amount of computational power provided by multiple parties. There is a lot more to it than that, but that is just noise and implementation details.
The fact that you don’t rely on a single party is great! I’m not going to dispute that.
The fact that the security is provided by a large amount of computational power is bad.
Just a large amount of computational power would actually not be bad. Chips always get faster and more efficient. But it’s not just a large amount, it’s a relative large amount. The security only works because the amount of computational power required to “take control of the network” is simply too expensive for a malicious third party to acquire (even temporary). So if chips get 10 times faster the computational power required to keep the network secure simply increases with the same factor. If chips get 100 times more energy efficient, the amount of energy required to keep the network secure doesn’t drop at all[2].
Additionally, the amount of computational power required scales with the value of the transactions being done on the network. You need enough computational power to make it unattractive for an attacker to acquire even more computational power and take control of the network. How much computational power this exactly is, is quite an interesting theoretical question (I don’t know the answer). But for sure the amount is high. An attacker only needs a short moment to inflict huge and long-lasting damage on the network. They could for example double spend coins, hurting not only those who would receive them, but also causing a massive loss of trust in the system that would be longer lasting.
So why is this all so problematic? It’s simple: high computational power requirements translate into (relative) high transaction fees. And that’s a problem for a lot of the applications that have been proposed for the blockchain. Using bitcoin as a currency is the biggest obvious problem. Most banks for example process millions of transactions daily, and most of these transactions are almost free because running a nice secure sever that handles a million transaction a day isn’t a lot more expensive than one that handles just a few transactions[3].
If you use bitcoin to speculate on the value of bitcoin transaction costs of a few dollar[4] aren’t very problematic. If you use bitcoin like some sort of virtual gold slash store of value it’s not a real problem. If you use bitcoin for money laundering or other black market transactions it also not a problem[5]. But for the first two to work out I think you need bitcoin to become a broadly used medium for transactions, and just betting on there being a large black market is a questionable proposition as well in my mind. The current hype is mostly driven by speculation, and for most real-world applications high transaction fees are a very serious obstacle. There are probably some niches where it makes sense, but I think broad adaptation of blockchain technology would in most cases be a step backwards. And I’m only looking at the financial angle here, not even mentioning the environmental impact it would have…
Disclosure
Author has zero blockchain related exposure[6]
[1] I’m using the blockchain and bitcoin sort of interchangeably in this post. I know it’s not the same.
[2] I doubt this is true. There is probably some complex relationship between both the cost of new chips, the energy requirement and perhaps even some other tangential factors.
[3] At least not a factor million more expensive. Scaling performance is still hard. It is also going to be hard for bitcoin.
[4] Average transaction fees are now already ~$4, and this is despite the fact that miners are currently also compensated by newly mined bitcoins. Currently miners get 12.5BTC for every block mined. With roughly 2,000 transactions per block and a current BTC price of ~$5500 one transaction actually costs almost $40.
[5] The fact that all transactions are publicly stored in the blockchain might be though…
[6] I used to own a couple of dozen bitcoins in 2011 or something like that back then when mining them on your own graphics card was still economical. Sold them between $20 and $30. Still better than buying two pizza’s for 10,000 BTC though.
First of all great to see someone else experience the pain of conversations about investing invariably ending up at Bitcoin. Even though I’m bullish on it, it’s sort of aggravating to have heard so much complaining about investing, stock markets, stocks and whatever you talk about zeroing in on every little risk in the idea. Then comes Bitcoin and the attitude is just let’s get all-in or alternatively: Bitcoin ran up too much, what coin is only a few cents a piece; all-in. I’m bullish on Bitcoin (select altcoins as well) but your post is very thoughtful and addresses real problems that do exist. Here are my 2 cents:
“Just a large amount of computational power would actually not be bad. Chips always get faster and more efficient. But it’s not just a large amount, it’s a relative large amount. The security only works because the amount of computational power required to “take control of the network” is simply too expensive for a malicious third party to acquire (even temporary). So if chips get 10 times faster the computational power required to keep the network secure simply increases with the same factor. If chips get 100 times more energy efficient, the amount of energy required to keep the network secure doesn’t drop at all[2].”
I don’t think this needs to be true forever. Especially with Bitcoin that runs on some kind of ASIC powered algorithm the time will come, relatively soon, where it is extremely hard to find the hardware outside of the Bitcoin ecosystem. Perhaps the difficulty increases could be throttled back.
“Additionally, the amount of computational power required scales with the value of the transactions being done on the network. You need enough computational power to make it unattractive for an attacker to acquire even more computational power and take control of the network. How much computational power this exactly is, is quite an interesting theoretical question (I don’t know the answer).
But for sure the amount is high. An attacker only needs a short moment to inflict huge and long-lasting damage on the network. They could for example double spend coins, hurting not only those who would receive them, but also causing a massive loss of trust in the system that would be longer lasting.”
I believe this is another argument in favor of Bitcoin especially against Ethereum and blockchains powered by GPU based algos. GPUs are useful for AI which means the cost to the attacker is by definition much lower as compared to investing in ASICs.
Fwiw, I agree way the usefulness of the Bitcoin blockchain as a trading currency is limited. Its primary selling point is security. However, that doesn’t mean there can’t be secondary layers built on top of the blockchain that would be much better to enable commerce and traditional currency functions at lower transaction rates. The store of value / asset / ledger function is really the core value for now and either technology gets built on top or another currency becomes most useful for commerce.
I guess it’s sort of helpful for bitcoin security that it runs best on specialized hardware, but at the same time it doesn’t change the fundamental problem. If it works it is because people have spend a shitload of money on specialized chips. So it’s still security by just making the whole network very expensive to run. GPU-based algo’s don’t have this cost, so it’s both cheaper to create the network and cheaper to attack it. As a result to provide the same level of security it will simply have to have more computational power and use more electric power. I don’t think it changes things fundamentally.
And with regards to bitcoin/blockchain becoming useful by adding secondary layers. I don’t really see it. In most cases that removes the whole thing that makes the blockchain unique and useful, and you just end up reinventing something that already exists.
“First of all great to see someone else experience the pain of conversations about investing invariably ending up at Bitcoin. ”
Says the guy who writes articles titled “15 Reasons To Buy Bitcoin Near Its $3,525 High” for SeekingAlpha ..
@av you could have something settle on a subchain before running a batch of transactions through the Bitcoin blockchain. There are a number of solutions like lightning networks.
@Pietje I enjoy talking about Bitcoin (that article is sent out by email to subscribers to the COIN ticker as well as my followers. ) I just think it’s painful if people have virtually no investments or want to start and start there… I try to talk about the fiscal advantages of something like brand new day and there is just no interest
@Pietje As a sidenote I started that series of x reasons to buy Bitcoin around $300 or so: https://seekingalpha.com/article/3986325-5-reasons-buy-bitcoin but also wrote an accompanying article why NOT to invest in Bitcoin https://seekingalpha.com/article/3986721-6-reasons-invest-bitcoin
But what if Bitcoin would switch to proof of stake instead of proof of work algorithm, or something else that doesn’t have the PoW flaws? That would remedy this concern, right?
I think that eventually a bigger problem is that crypto currencies are extremely intolerant against human error, which is a big thing in the retail banking business.
I think that banks that flag uncommon transactions and generally protect the customer in many other ways as well offer a compelling service; and that they can be effectively controlled by the law is valuable as well.
To compensate for this, businesses will be built on top of cryptocoins that are more forgiving, but I don’t see why they would be fundamentally better or cheaper than non-distributed ledgers.
If someone manages to figure out a proof of stake algorithm that really works, yes (not too up to date on that to be honest, but thought that these have some unsolved flaws as well – but not sure if these are fundamental unfixable flaws, or fixable ones).
And yes, I agree with you that there are lots of practical problems to use bitcoin securely and successful as well. Keeping some cash in your bank account safe is so much easier than keeping your own bitcoins safe, especially if you aren’t using a third party. And as soon as you start using a third party it’s sort of pointless to use bitcoin, since then you could use just as well use a third party and use dollars/euro’s (aka a bank).
Nicely written article, thank you. What really leads me to think this is a bubble (in addition to the reasons you note), is that virtually all of the appreciation seems to be driven by speculation as opposed to anyone actually “using” it. Most of the reasons bulls claim it’s “useful” seem to be bogus.
Yes, I think so as well.
Yeah agreed. I’ve also never talked to anyone who actually uses bitcoins to do legal, boring stuff like buying movie tickets, i.e. like you use cash or your bank account. All the owners says that bitcoins are very useful and that they will change the world but at the same time they don’t buy them for that reason: they only buy them because they hope/expect they go up. Buying a pizza with my virtual currency? Are you crazy? Next week it will be up another 50 percent!
I think for large international transactions or remittances it could be useful. Although that assumes that financial institutions don’t adapt. Because there transaction costs can be up to 10% and headaches of sending money abroad.
But a major overlooked thing for bitcoin speculators is holding period of bitcoins. If bitcoin takes the entire $600 billion remittance market, but holding period is only a day (between buying of bitcoin and selling of bitcoin of the receiver of remittance), one bitcoin would be turned over 365 times a year, so contribution to total market value would be a lot smaller (essentially improving liquidity by a factor 365). Probably several 100-200 times smaller than $600 billion (assuming some hold bitcoins to iron out volatility). So transaction value per year could be in the trillions, and the bitcoin market cap that would justify could only be $10-20 billion.
And an additional concern for everyday transactions is the confirmation period. This can be several hours. So this would either be an additional cost or you have to wait for several hours before your transaction is confirmed.
Seems like your biggest critique is the transaction costs — will tackle that.
The most important and fundamental feature for bitcoin is its censorship resistance and decentralization. Every design choice of the bitcoin network is to ensure bitcoin can continue to be censorship resistant money. The high transaction costs of executing on chain transactions (vs off chain / layer 2) are simply a bi-product of that. To your point, because the system requires high amounts of decentralized computational power, it will always cost a non trivial amount of money to update the base blockchain. However, the media and newcomers to bitcoin seem to make a much bigger deal about this factor than it is and will be.
First off, the world already has great, cheap, centralized transacting systems, especially in the US. Venmo, paypal, apple pay and even certain bank transfers are perfect for low value payments or those where censorship resistance isn’t necessary / beneficial. If I’m sending $100 to a friend, I couldn’t care less if the government spies on it, so am more than fine doing it through venmo for free. The base bitcoin blockchain will never and should never be used for low $$ value transactions within the US. Bitcoin is not here to be a better medium of exchange for low $$ value transactions — at least not yet.
What BTC excels at is being a revolutionary new store of value that is portable, divisible, scarce and highly secure — a digital gold if you will, that is provably scarce and not exposed to the whims of central banks. Through clever public key cryptography and a blockchain that would take many $bns to attack, BTC is the worlds first unseizable asset (if stored properly that is , without exposing your key which is getting easier every day). Let that sink in. Add up all the swiss bank accounts, trust funds, and elaborate financial structures to store wealth securely (that costs significant amounts of money) and that use case alone is in the $tns. Then layer in potential market share BTC can take from gold because it’s better in almost every aspect of moneyness (fungibility, divisibility, portability, even security) except for the fact it hasn’t been around for as long, and the TAM is quite large. Finally, because the bitcoin network is permisionless, it will help provide access to the financial world to the 2-3bn underbanked individuals in countries where banks do not serve today. Just like gold, transacting will likely be a relatively small % of its usage and most of the value will come from the store of value nature.
Lastly, bitcoin developers are working on what’s called layer 2 solutions that will allow people to open trust channels with merchants and individuals they transact with frequently to reduce costs. Think of this like a bar tab where I don’t close my tab until I’ve ordered all my drinks for the night, instead of closing after each one. Instead of committing every transaction to the blockchain, you’ll ultimately only need to do so after xx # of transactions, dramatically increasing speed and reducing costs.
I guess we sort of agree on what Bitcoin is and what you potentially can do with it 🙂
Perhaps it can become some store of value successfully. I don’t think that is driving it’s current success though. People buy Bitcoin because it has gone up a crazy amount in the past, and are hoping that it will continue going up in the future. A good store of value would have a less volatile value. Perhaps Bitcoin can graduate into that in the future.
For me the problem is that if Bitcoin is solely a store of value it has nothing to base that value on. If Bitcoin could become an alternative currency it would be inherently valuable, but if it’s just a store of value I’m not so sure it will work in the long run. Sure, I know there is a limited supply of bitcoins, so the scarcity can make them valuable. But at the same time there is an unlimited supply of alternative crypto currencies. People can just continue to do new ICO’s… and what will happen if speculators withdraw or move in large numbers to something new and exciting?
But perhaps I’m just too skeptical…
I made a Bitcoin transaction yesterday. Transaction cost was 0.0001 BTC, which is about 0.5 USD at current exchange rate.
I suppose that the cost is the same regardless of the amount transferred.
In case it was not clear from the above comments, off chain settlement solutions such as Lightning network (mentioned by Bram) or Thunder network could significantly reduce transaction fees. So could other layer two solutions. Fees could also be reduced by other network connectivities such as those discussed by the Dash cryptocurrency on the below article:
https://medium.com/@eduffield222/how-to-enabling-on-chain-scaling-2ffab5997f8b