Questioning assumptions: do cryptocurrencies need fees?
Ah, cryptoeconomics. A field of study which did not exist a little more than 9 years ago is now full of more self-described experts, CEOs, and crypto OGs (which is basically anyone who was interested in Bitcoin before 2017) than you can count. Everyone has their opinions on what is or is not ‘right’, and each — especially — have their own interpretation of the economics of cryptocurrencies and how everything works.
And so here I am, sharing my own opinions on cryptoeconomics and positing about how I think things should be. In this space, of course, there’s little use trying to guess on what direction things are going, because nobody knows, and anyone who claims to know is a charlatan or time traveler. Cryptocurrencies did not exist 10 years ago, so there’s little history from which we can learn.
What I can say with confidence, however, is that very few people actually understand how Bitcoin actually works (or any other cryptocurrency for that matter), or why they work the way they do. When I hear people talk about “blockchain” rather than Bitcoin (or any other particular cryptocurrency), I usually roll my eyes. The chain of blocks which makes Bitcoin a cryptocurrency is just an implementation detail. It’s also not some magic tech you can apply to solve all the world’s problems. Bitcoin’s novelty comes from the fact that it’s censorship resistant, scarce, and as close to impossible to hack as anything we know of. And the idea of a private blockchain is just plain absurd; you’d be better off using MySQL or PostgreSQL if you want to create your own private virtual currency.
That brings me on to the subject of this post: I want to call into question one of the long-held beliefs in the crypto world, which is this: the idea that cryptocurrencies must have fees. The systems that exist today (including Ethereum, Litecoin, and so on) are almost all entirely based upon Bitcoin, with only minor iterative innovations beyond what was originally created by Satoshi. Ethereum, for example, uses a proof of work system similar to Bitcoin (albeit with different algorithms), but it features a more powerful programming language called Solidity that allows people to build smart contracts. Ethereum offers more bells and whistles than Bitcoin, but that complexity comes at a price: Ethereum has suffered from a number of major attacks (the DAO, Parity multisig wallet).
Why do cryptocurrencies have fees?
The conventional wisdom in crypto is that there should be fees to reward miners (or in the case of proof of stake, stakeholders) for finding blocks. This is nuanced: note that I said finding blocks, not validating blocks. A widely held misbelief of Bitcoin is that miners are responsible for validating blocks. This is not entirely true: all Bitcoin nodes are responsible for validating blocks. It’s subtly written in Satoshi’s original Bitcoin whitepaper:
Miners find blocks by making random guesses until they find a block that fits in place. A good analogy is that it’s like trying to fit puzzle pieces together, but you can’t actually see any of the pieces or how they fit. You have to test every piece until it fits. You might occasionally find one that fits, and if you do, you earn a reward which is the sum of the block reward (currently 12.5 BTC) and the transaction fees.
The original whitepaper made a few assumptions which later proved to be false. One of those assumptions was that mining was something everyone could participate in. Satoshi assumed that Bitcoin would continue to be CPU minable, and the reward would keep nodes (or actors) honest:
Satoshi was wrong about continuing to be able to use a CPU for mining, and Satoshi also overlooked the problem of the upfront cost of mining hardware. The high cost of specialized ASICs, and the amount of energy required to run them, makes it impractical for people to run their own mining rig and fully validating node. You can’t mine Bitcoin on low cost hardware, like a smartphone. It’s still possible for regular users to run a fully validating node, but few do.
If you do want to run your own Bitcoin node, you can head to bitcoin.org and download the core wallet. It might take a few days to download all the blocks and become operational on the network.
Do incentives work?
The big question few seem to be asking is whether or not these incentives are working. If you read through my previous post about Bcash, you might draw a similar conclusion as me, which is that not all miners are acting in the best interest of Bitcoin users or HODLers. In fact, I think the incentives are working against the best interests of Bitcoin users recently due to the scaling debating, and also by creating market confusion and uncertainty. The mining incentives are (in my opinion) responsible for transaction spam, and works as a disincentive for businesses to adopt better technology (notably Segwit and transaction batching, which both reduce TX fees).
Spamming the mempool drives up transaction costs, which goes directly into the pockets of miners. Bitcoin currently doesn’t have a good spam prevention system, aside from the transaction fees. A spammer can create transactions that are cheap enough to never be confirmed but still occupy space in the mempool for a period of time.
Other projects (notably Ethereum) have proposed transitioning to a proof of stake system, which aims to correct the incentive issues, further democratize the system, and address the concern around energy consumption of the Bitcoin network. Proof of stake does this by rewarding people who run fully validating nodes and are willing to lock up a portion of their holdings for a specific period of time. In a full proof of stake system, there are no miners. At this point, my opinion is that there isn’t enough evidence to support claims that proof of stake is better (or worse) than proof of work. If anything, it seems to me to be equivalent: you still need to run a powerful and well-connected computer 24 hours a day 7 days a week to receive income.
An alternate theory
Let’s for a moment go back to first principles and forget about how everything currently works. What if we had a system that didn’t have any transaction fees? Suppose that the only incentive for an actor in the system to remain honest is the appreciation of their own holdings. In other words, if I own 500 units of BremCoin (a fictional cryptocurrency I just invented), what is my incentive for following the rules of the protocol?
If we make the assumption that cryptocurrencies will always be scarce (by design of the protocol), and the GDP of the world continues to increase, then my BremCoin will always appreciate in value. In order to capitalize on the growth of the 500 BremCoin I’m HODLing, I need to follow the rules of the system. I don’t want anyone stealing my coins, or stealing anyone else’s coins. If the currency gets hacked, people lose confidence in the system, its value diminishes, and thus so do my holdings.
Has anyone tried this?
There are a few new cryptocurrencies which are trying to do something like this. RaiBlocks is one worth mentioning. It was created by one person, has a finite supply, and all of the coins have been distributed through a captcha-based faucet. There was no ICO or crowdfunding. RaiBlocks doesn’t use a blockchain (eyeroll) either, it’s a new design which uses a DAG-based ledger, and a consensus system called ‘delegated proof of stake’. When I learned about RaiBlocks, I was quite skeptical. Particularly due to the lack of incentives. How can this work? Does it work? — Honestly, I don’t know, but I read the whitepaper and I find it interesting.
I know I’ll be called out for shilling RaiBlocks, but please allow me to make it perfectly clear that I’m not suggesting you should invest in RaiBlocks, I’m just discussing the incentive system. It’s a new technology and has not been battle-tested or hardened in the way Bitcoin, Ethereum, or Litecoin has.
There’s another (potentially) feeless cryptocurrency that we’re all a little more familiar with: Bitcoin. How can Bitcoin be feeless you ask? Through the Lightning Network, of course. The interesting thing about the Lightning Network (LN) is that LN nodes have the option of processing transactions for free, if they choose.
Minor detail: the fees to send funds to a LN node are not free in the current system, as it constitutes a normal Bitcoin transaction, and a TX without fees will probably never confirm.
With LN, it’s plausible that vendors will operate their own LN nodes to accept payments, and they might completely forgo the network fees. Their main business is probably not the business of operating a Bitcoin network, so much like a cash transaction, fraud risk is low enough that they would probably accept payment without adding extra fees. Vendors sometimes charge additional fees for credit card transactions because the credit card networks charge a fee for the transaction as a percentage of transaction cost, and vendors are also liable for fraudulent transactions. Credit cards are convenient for consumers, but they carry a lot of risks for vendors (guess who pays for it: you do!).
It’s also interesting to think that Bitcoin and LN might provide the best of both worlds: for smaller, frequent transactions, the Lightning Network provides quick and feeless payments, but everything eventually settles on the Bitcoin blockchain providing the security and confidence that Bitcoin offers.
I’m calling into question the incentives that have been accepted as necessary in most cryptocurrencies. I’m not convinced they’re necessary, and in fact, I think it’s possible that the incentives for miners (or stakers) create an environment where miners (or stakers) act against the best interest of users out of a desire for profit. For example, miners or stakers may spam the pool of pending transactions simply to drive up TX prices for those who want their transactions confirmed sooner.
Follow me on Twitter: https://twitter.com/brndnmtthws
Thanks for reading!
Special thanks to the super talented Drea Roberts for help with the artwork.