This article is a rebuttal of economist and Bitcoin researcher Paul Sztorc’s essay “Security Budget in the Long Run” where he explains why he thinks Bitcoin’s fee model will not work. Following are Paul’s quotes in his essay with my observations bellow each one.
Edit October 1st, 2020: This article on The Block describes how Ethereum 1.x is confirming my argument below. Its success proves that proof of work based Nakamoto consensus, the consensus mechanism pioneered and used by Bitcoin, is viable through fees. This is the link to the article: First time in Ethereum’s history, miners made more from fees than from block rewards.
“Bitcoin’s “security budget” is the total amount of money we pay to miners (or, if you prefer, the total amount spent on mining — they are the same thing).”
I agree with this statement and I think it is very important to have the “security budget” present when analyzing monetary policy (or trying to stop all monetary policy changes in my case). For example many researchers have proposed compensating node operators or developers for their work out of the same budget, but it is obvious that the energy ball, or the total hashing power dedicated to mining and protecting a chain, would be reduced.
“The dual nature of Bitcoin (as both a money-unit, and a payment-rail) has confused people since Bitcoin was first invented.
This is the other part of the essay that I find very useful as it helps clarify that Bitcoin is a system with several components, each one with different mechanics and economics, and that correctly deconstructing them will make their analysis much better and perhaps more rational. As to the duality of “money-unit” and “payment rail” I think that is true, but not exactly as Paul describes it. In particular, that the Bitcoin platform is a transfer “rail” is true, but, as Nick Szabo pointed out, there can be highly secure base layer rails that are used for high value expensive settlement transfers and there can be less expensive and less secure rails that can be used for low value high frequency payments. In the traditional banking system, Fedwire and Visa can be examples of the former and the latter.
While the two are mixed into the same “security budget”, the block subsidy and txn-fees are utterly and completely different. They are as different from each other, as “VISA’s total profits in 2017” are from the “total increase in M2 in 2017”.
Totally agree. The sources of funding of the security budget are very different. The market dynamics of funding the Bitcoin network primarily through fees is different, may have competitors and substitutes in the “pure value transfer” segment of the market, and requires its own analysis and proper unit economics projections. This doesn’t mean though, as I write below, that the economics of the fee market will be insufficient or that any potential solution is fine.
Transaction fees are explicitly priced in BTC. But, unlike the block reward, they do react to changes in the exchange rate.
True. The block rewards are fixed, determine total BTC supply in the long run, and are product-centered, meaning that the market (miners, hashing power, developers, investors, BTC price, etc.) adjusts to them given supply variable, not the other way around as in Ethereum, for example, where the lead developers regularly change the supply to adjust to market conditions or their desired economic or technical agenda. On the other hand, transaction fees are market-centered, meaning that they go up and down adjusting to supply and demand.
As the exchange rate rises, a given satoshi/byte fee rate becomes more onerous, and people shy away from paying it.
I am not sure this concept is true, at least as seen in the last few years: as Bitcoin went up in value, transaction fees (and volume) also went up in value, and vice versa. Of course people will think twice before paying $50 for a single transfer that was $0.50 for the same transaction a few weeks before, but compensating factors are that fees are usually proportional to the value being transferred and that now there are alternatives such as Lighting Network.
But it is the reverse when we consider transaction fees and “btc-block-bytes”: Altcoin-blockspace is a pretty good substitute for Bitcoin-blockspace.
I disagree that block space is a generic commodity, therefore a transaction in Bitcoin is totally replaceable by a transaction in any altcoin or L2 system for that matter.
Up to a certain extent, and in different proportions, users will tend to tradeoff low transaction fees for standards and safety. This is because there are several types of costs that will induce the world to select very few chains in the future: Security costs, exchange costs, and coordination costs.
Security Costs
It is not the same to move value using the Bitcoin network or using any altcoin or L2 solution. This will create a market for high value, highly secure transfers in Bitcoin, e.g. central banks, governments, interbank, corporate and other large value payment users, who will gladly pay very high fees.
For those who don’t need such security or for low value transfers there will be of course cheaper alternatives in other chains or in L2 solutions as mentioned above.
Exchange Costs
When senders and receivers want to store value in Bitcoin, but need to transfer them, it will not be frictionless to move from BTC to an altcoin, send it for a lower fee, and then pass it back to Bitcoin on the other side.
Between exchange commissions and spreads there will be an indifference point beyond which it will be better to pay for Bitcoin fees. Not only that, but the security dynamics mentioned above also change substantially. If Bitcoin is a very good store of value, transfers will occur within Bitcoin precisely for the same reason it is a good store of value.
Coordination Costs
Just because you can create N altcoins does not mean they are all going to survive. Because they are all not going to survive, there will be a limited competing block space market in the world. This is because our minds are limited and we will not think about 250 cryptocurrency names, transfer fees, the subsequent 250 prices, and go selecting the cheapest one each time we move value.
Our brains will only support understanding the value of 3 to 5 coins at most, and we will be comfortable using them interchangeably up to a certain point. If this is correct, then, together with the security and exchange costs considerations above, there will be a limited number of coins in the future, with a limited amount of block space and transaction capacity. This means that the cost of sending value will be distributed between the surviving chains in proportion to value and safety requirements.
We see that BTC’s crossing of the “1 USD per transaction line”, in May of 2017, coincides with the rise of Altcoins. We also see that the “pressure” of late 2017 quickly canceled itself out, and then some. Finally, we see that this release-of-pressure coincided with a sudden (and unprecedented) decline in BTC-transactions.
I disagree that the rise in fees in 2017 was a function of a block space market demand and competition. I think that bubble was actually a race to buy the “store of value” side, which created the BTC rise and subsequent fall when the bubble pierced, which is normal in free markets. Then the fees came down to historical rates again just because the demand for the store of value diminished after the hype.
Similarly, I think that the sudden multiplication of altcoins and ICOs during the period was a race to mimic the wealth creation that happened in Bitcoin, not necessarily millions of people suddenly using blockchains to transfer money around the world and seeking to minimize those fees.
In other words, the coincidence of altcoin and ICO proliferation with the 2017 crypto bubble was a generalized gold rush (and in many cases a fraudulent money grab by many altcoin and ICO creators), but not a rational pricing dynamic and arbitrage of block space through transaction fees.
Going forward, the pattern of seemingly unlimited alternative tokens, and new block space as alternative “rails” in unlimited chains, will be extremely reduced for the reasons explained in the previous points.
To me, this data refutes the theory that users will pay high BTC fees willingly.
As the data is not well interpreted, that is, that the bubble was a “store of value run” and not a block size and cost competition, the above argument is not true, in my opinion.
Users, depending on the value they need to transfer and safety needs, will gladly pay for different gradations of block security and costs. This includes layer 2 systems.
If the consumer is cost-conscious, and will only pay the lowest tx-fees, then how can we get those numbers up?
As explained above, the consumer is cost conscious, but has a coordination need, mental limitations, and, therefore, will count with a limited selection of alternatives. The numbers on the selected few chains that will survive in the future will be distributed according to their economic benefit and function: store of value vs Turing complete, high vs low value, high vs low security, etc.
The real question to raise volumes on Bitcoin and the selected few chains that will prevail will be whether they are competitive with current centralized systems, which they are, but not necessarily whether they will be viable or not due to an imaginary unlimited block space market.
…[merged mining side chains]send all txn-fees they collect to Bitcoin miners. Under Blind Merged Mining, they do this without requiring any users or miners to run the sidechain node software.
Although public blockchains are permissionless, I object to promoting merged mining because it creates “super-miners”. Super-miners are pools or direct miners who provide their block creation and security services for multiple chains with the same equipment at the same time as envisioned by Paul and other entities promoting them.
This will create a situation where the miners will become the Facebooks, Googles, Apples and Amazons of the future. If they mine for multiple chains, and they have the prerogative of selecting what chains can be merged-mined or not, they will have an equal permissioning power as the Apple App Store, or the Google Play Store, for example.
Also, they will be so large as compared to the corresponding node operators in each chain that they will likely be able to dictate rule changes in the future with little resistance, as counter-strategies as UASF will be insignificant compared to their power.
[Conclusion 1] To deter 51% attacks, Bitcoin needs a high “security budget”.
Agreed, but I think there will be proper security budget financing through fees, and 51% attacks are not that catastrophic anyway. Especially to create a potential monster as super-miners. I don’t think that we have to reach astronomic hash rates to compete with the pentagon, for example.
If the Pentagon wanted to do a significant 51% against an enemy, first they would have to buy a substantial amount of BTC to do so, which would be good for Bitcoin, second they would have to use that value to send it to the enemy and then double spend it, third they would have to respond publicly for such an attack as the rest of the world at that point, including American reserves, tax payers and savers, will be invested in BTC.
Lastly, if the enemy is not that stupid, they will wait for several days, weeks or months of confirmations to consider a transfer by their supposed foe finalized.
Remember that the “security budget” or the “energy ball” that protects Bitcoin does so for just the tip of the chain. That is, out of the current ~567,000 blocks that the current chain has, the miners, with their hashing power, secure the chain in 6 or so blocks, or a little more depending on each user’s security preferences, after which point it becomes exponentially difficult, even with all the hashing power in the network, to actually reverse the chain.
The above shows that 51% attacks are not the monster that everybody thinks. In the Ethereum Classic community we know, we actually lived through that! [Please read this to learn from our experience!]
[Conclusion 2] Today’s tx-fee revenues are not high enough; we must ensure that they are “boosted” in the future.
Yes, they are not high enough presently, but I think they will be higher just for the sheer value and volume that will go through Bitcoin, and the few chains that will succeed in the future.
The solution, other than continuing to build the public blockchain stacks as in the current roadmaps, I think will be that the world will move to use blockchains for many functions and the fees will be enough and well distributed between the prevailing networks.
Code Is Law
