Originally posted to Substack.
Special thanks to Michael Jordan and Joe Coll for their review.
Despite the title, you'll be hard pressed to find someone more excited about Ethereum than myself. However, I believe the narrative and policy objectives around ETH the asset are in need of some refinement. It appears to be a pretty contentious issue based on my very scientific polling:
This is the first of several upcoming pieces on my framework for ETH. I'll cover the basics here:
Future pieces will go much deeper:
Additionally, I recently shared a simple Ethereum model linked below. Playing with the assumptions helps to provide an intuitive sense for how Ethereum's economic engine operates and how it will change.
This is a detailed functional Ethereum model including long-term roadmap items
Flexible to easily run any scenarios/timing (eg, when 4844, danksharding, rollup data compression rates, staking rates, fees, MEV burn, issuance, etc)https://t.co/iMBzdbJi8M
— Jon Charbonneau (@jon_charb) January 29, 2023
The primary components are transaction fees, other MEV, and issuance.
EIP-1559
EIP-1559 dictates Ethereum's current pricing mechanism. It includes elastic block sizes targeting 50% usage on average. The current gas limit is 30mm gas per block with a target of 15mm gas. Gas is the network's measurement of all resources.
There's currently only one gas limit and fee market - it's a single-dimensional EIP-1559. Whether you're an L1 user (e.g., swapping on Uniswap) or a rollup (e.g., posting calldata to Ethereum), you're directly competing over shared resources.
Research is underway on multi-dimensional EIP-1559 to implement different fee markets for different resources. EIP-4844 will implement the simplest version of this - a two-dimensional EIP-1559:
Each layer will have independent fee markets that float depending on the demand for their respective resources. They won't be competing anymore, so price spikes for either layer will no longer bleed into the other.
Base Fees
The simplest fee market would just be a first-price auction (FPA) - users all bid what they're willing to pay, and the highest bids get included. EIP-1559 instead dictates that every transaction must pay the prevailing reserve price (the "base fee") to be included in a given block. This provides a nice UX for users:
The base fee is deterministically calculated based on previous block sizes, changing by a max of ±12.5% per block:
EIP-1559 then burns the entire base fee. This is not for the deflationary memes. Burning prevents validator + user collusion. As an example:
This ability to collude would effectively remove the reserve price, reverting to a FPA again. However, the game theory does not dictate that you must burn the fee. It only dictates that the base fee must not go to the producer of that block. It would be equally incentive compatible to do something else with the base fee (e.g., pay to future block producers, fund public goods, put into a treasury, etc.).
Priority Fees
Users can also add an additional "priority fee" to get included faster.
This fee goes directly to the proposer, so they're incentivized to include transactions with higher priority fees all else equal. However, base fees and priority fees are both "revenue" to the network. They just differ in who they benefit:
You'll often see gas prices quoted in "gwei." Gwei is a unit of measurement = billionth of 1 ETH. Different transactions use different fixed amounts of "gas." The gas price then says how many gwei you're willing to spend for each unit of gas you use. As a simple example:
Validators also stand to profit from other MEV. When there's an opportunity to pick up $100 somewhere (e.g., an arbitrage), searchers/builders should be willing to bid up to $99.99 to the validator to make sure they're the one who gets to pick it up (in a competitive market).
Note that this MEV payment to validators can be partially (but not entirely) captured under the priority fee revenue line item mentioned earlier. There are several ways to express your bid to a validator for the MEV opportunity you want to capture:
MEV is a bit of a wild card long-term. There are three broad areas which could change drastically.
MEV Market Size
MEV opportunities exposed represent some % of total activity. Let's assume for simplicity that we have perfect definitions and measures for total crypto activity and MEV respectively (we certainly don't, but bear with me). Then we can define on average exactly how much MEV is exposed by activity (e.g., for every $1,000 transacted on-chain there's a $1 MEV opportunity, so MEV = 0.1% of activity).
Then we could have this example if total crypto activity is $1 tn:
If crypto activity 10x's:
In reality - MEV is extremely nonlinear. I'm vastly oversimplifying to highlight the variable.
This spikiness is particularly true when activity increases over short time horizons. Variations also occur with new forms of MEV and different chains' architectures.
It's also unclear how "MEV-rich" activity will be on a longer time scale. The composition of crypto activity is likely to change over time. Let's say high value DEX transactions are more MEV-rich per unit of economic activity than gaming apps. Maybe this DeFi stuff goes away and crypto is just for Web 3 League of Legends (I'm not betting on it, but you get the point).
MEV Distribution
Assume you can predict the exact size of the future "MEV market." However, the way in which that value is distributed is a huge variable. Let's say validators capture 50% today, supply-chain middlemen (e.g., searchers, builders, etc.) capture 45%, and users get 5%. What actually matters to validators is the below:
Market size alone is clearly insufficient to predict future validator revenue. Maybe their share of the pie goes from 50% to 5% in that same time period. Now validators are flat even if the market 10x'd:
The crypto gigabrains are hard at work here - mostly trying to make sure that validators capture less of the MEV. Applications expose large amounts of MEV today, and the users don't get much back. Protocols and apps could try to help users out:
Similarly, apps and protocols want their fair cut. Those useless governance token holders want to be holding a vaguely useful governance token. Why give your MEV away for free if you can internalize it?
For example, Osmosis' ProtoRev module is being developed by Skip Protocol to internalize arbitrage opportunities. Money that used to go to searchers could then be divvied up amongst the Osmosis LPs who are getting picked off, OSMO holders, Skip for designing and building, or however else they want. Similarly, wallets could start doing PFOF. The list goes on - it's a big design space.
Ideally we build towards fair value attribution where:
MEV Burn
Sorry validators - forget about MEV market size and split, we're gonna light it all on fire anyway. % of MEV to validators is now a big fat 0. MEV now benefits all ETH holders equally. That's the proposal at least, and it's even on Vitalik's latest roadmap. The basic idea is:
I have several hesitations regarding the specific proposal and whether it's even preferable to do so. I'll cover this in later pieces.
The Ethereum protocol mints new ETH every block to pay validators. It provides an additional incentive to secure the chain instead of sticking your ETH under your mattress, spending it, or using it somewhere else.
Ethereum issuance is currently stable, predictable, and low. There's a predetermined curve which dictates how much ETH can be printed based on the active validator count. More ETH staked = more issuance, but it grows at a decreasing rate (i.e., increasing the validator count reduces the yield provided to each validator).
Now that we understand Ethereum's value flows, we can reason about ETH as an asset. First, let's define what an asset even is:
They're not perfectly distinct buckets - some assets display some crossover features. Still, this provides a helpful benchmark for categorizing assets.
Many application tokens are pretty straightforward. Lido (LDO) is one of the clearest examples of a sustainable decentralized business model:
LDO controls the protocol and its associated cash flows. It displays clear properties of a "capital asset," and traditional valuation tools (e.g., DCF) are largely appropriate.
ETH is complicated - it displays properties of all three traditional asset superclasses (a Triple Point Asset, some might say). A simple DCF is clearly not sufficient in valuing it. It's only one piece of the puzzle. I'll give a brief explainer of each here in regards to ETH.
ETH holders and stakers have a right to protocol cash flows (fees and other MEV). Stakers also receive issuance (though at the expense of unstaked holders). This displays some similarities with equities, (infinite duration) bonds, and income-producing real estate.
ETH is the gas that makes the Ethereum engine run. Transaction fees must be paid to the protocol in ETH. High demand to use Ethereum → higher demand to hold and consume ETH for transactions.
SoVs retain purchasing power over time and have stable demand for the underlying asset. We're seeing very early signs of ETH as a SoV among its other use cases:
So those general frameworks make sense, but we keep hearing that ETH is "ultrasound money." Is it possible for ETH to really grow into being "money," and is it even a good idea? Maybe ETH is just more like a standard productive asset? This discussion will be the heart of future pieces.
But for starters:
Background in hand, we can attempt to clarify what ETH's objective function actually is. You'll likely get suboptimal economic policy if you don't know what you're even trying to optimize for.
The most coherent vision I've seen has been offered by Justin Drake, who gave an overview at Devcon. Crafting ETH to have "monetary premium" increases economic bandwidth and security, starting a flywheel effect:
The argument posits that blockchains are like companies, and their profit can be determined as follows:
Ethereum was considered "unsustainable" shortly after the Merge because it had slightly positive net issuance:
If Bitcoin's capped supply makes it "sound money," then Ethereum's expected deflation makes it "ultrasound money." (ETH net issuance is indeed negative since the Merge now).
There are two closely related variations of the above argument:
Most people seem to agree with this line of reasoning:
You even explicitly see stuff like this:
I strongly disagree with all of this. I've argued this since I started in crypto, so I'll go on a bit of a rant (again).
For starters, PoS issuance differs from PoW in a critical way:
So I disagree on both major points:
A net inflationary network may be unsustainable depending on many related factors. However, this black and white framing and false equivalence to corporate profitability reveals a fundamental misunderstanding of the system architecture.
(Note that I am only discussing PoS issuance provided to validators. However, some other alt-L1s also issue tokens for purposes such as incentive programs. This type of issuance is more analogous to PoW issuance, or similarly a DeFi protocol that is issuing token incentives to LPs to attract liquidity. This is very different mechanically from issuing the token back to stakers of the underlying tokens themselves. Regardless, ETH is only issued to validators. Ethereum has no other issuance mechanisms.)
For those of you compelled to stuff a blockchain into a corporate analogy, the below is much better (albeit they're never perfect).
Issuance is akin to a company paying their employees in stock-based compensation (SBC). Per GAAP, this is a non-cash expense deducted from earnings (though added back to FCF which is a whole other contentious issue). This:
Even worse, imagine the SBC is in the form of actual shares which vest immediately. This company doesn't make much money, so they pay almost entirely in SBC. It's also going to slowly become illegal to pay out SBC by 2140 (if we analogize to Bitcoin specifically). Every four years, you're allowed to give out half as many shares as the year prior.
It's been working so far:
Stock price goes up → able to pay more $ for better employees even with less shares → stock goes up more, because now they have even better employees → ...
However, they're kinda screwed unless they start making a lot of money by 2140. Hopefully they come up with a business model some time soon (Ordinals?).
This company issues new shares too, but only to existing shareholders. Every shareholder has an equal right to receive them. All you have to do is show up to vote at shareholder meetings. There's one condition - you must bring your physical unencumbered shares to vote (yes I'm ignoring re-staking, leave me alone, it doesn't change the logic).
You can't be lending out your shares or borrowing against them. This issuance:
This might impact your decision to buy that stock depending on factors like:
You might be ok to get diluted if:
Even better, imagine there's a third-party service that will give you synthetic shares (LSTs). You give your shares to them, they vote for you, and they give you ~90% of the share issuance. Now you get most of that issuance, and you can go lend out your synthetic shares or do whatever you want with them. They're a little less liquid and trickier to lend against, but they're pretty good. (Re-staking also offers different but analogous benefits).
Taking the analogy to completion:
You'll notice I also draw no distinction such as Token Terminal has between base fees, priority fees, and other MEV (they only consider base fees to be network revenue):
They're all Ethereum protocol revenue inflows to be allocated however it sees fit. Whether it's burned or sent to stakers refers to the protocol's use of the cash, not the source.
Let's walk through an extreme scenario to understand the logic. A PoS chain decides to offer a crazy issuance rate of 100%, so the supply doubles annually. Now the dilution is way too high even if you love their future prospects, so the unstaked holders dump the token or stake it.
100% of the token is now staked. Realize that now you can inflate it by 0%, 10%, or 100%. It doesn't matter to their "profitability". Everyone ends up exactly as well off in terms of network ownership and rights to cash flows.
Now that high inflation isn't helping to further incentivize stakers, everyone decides to lower inflation to 1%. So now as an example:
As described earlier, many would consider this network to be generating a "profit" of $1 bn dollars. But now, let's say they change inflation back to 100%. Did the network suddenly turn into a gigantic money-burning machine that's operating at a loss again? No, of course not. Everyone is staked, so they're all being diluted and made whole at the same rate. Your share of the network and rights to cash flows do not change at all.
Over the course of the year, the $ price/token should roughly halve. You end up with the same MCAP, but twice as many tokens. So it was basically just a "stock split."
PoS inflation isn't an explicit network revenue or cost in any scenario. You just decide who to benefit:
However, if a $100 bn PoW chain similarly needed to increase its issuance from 1% to 100% because miners demanded a higher income, the dynamics are very different. Now all token holders are diluted because they have no rights to receive the issuance. After a year, current token-holders own half the network, and miners own the other half.
Similarly, the PoW token holders aren't even receiving the fees (unless they're burned). So whether the network captured $1 bn of fees or $100 bn of fees makes no direct difference to them. The miners get that money - holders have no rights to any cash flows. The only benefit is that higher fees likely allow you to keep issuance lower, because miners won't demand such a high subsidy. So PoW token holders here "lose less money" based on fees and issuance, but they're not gaining network share or receiving cash flows.
If you want to think of it differently, imagine "staked ETH" and "unstaked ETH" are totally different assets that can trade at different prices (obviously not the case). They're "companies" as so many of you seem to like, so the value of each is based on a DCF. Now we'll include issuance in their DCF because they're isolated assets.
If you increase issuance, then the "staked ETH" goes up in price based on the discounted rate of this additional future issuance. Conversely, "unstaked ETH" would go down based on the discounted rate of this additional future issuance. You'll realize, these changes in their respective DCFs and prices then perfectly offset.
Well, guess what - there's only one token. So, the network's overall DCF is unchanged. Increasing the value of "staked ETH" doesn't make a separate asset price go up. It just shifts demand on the margin from being a holder to being a staker at whatever the new market clearing rate is.
Staked ETH and unstaked ETH are the same asset. They have exact equal rights to receive future cash flows. It's your prerogative what you want to do with it. High issuance doesn't mean you can sell your "staked ETH" for more than my "unstaked ETH."
Cash-flowing assets are valued based on their rights to receive cash flows. That's why stETH and ETH trade at roughly par (in normal market circumstances). If you choose not to exercise that right, it doesn't make your asset less valuable than another equivalent asset choosing to exercise that right. This would be like saying my stock is worth less than someone else's because last time the company issued a dividend, I said no thanks I don't want it.
If you want to see exactly where unstaked holders break even on purchasing power, we can measure it relative to USD for simplicity. I.e., holders want ETH/USD MCAP % increase ≥ ETH net issuance %, because then ETH/USD price stays flat. Hopefully this should be incredibly obvious.
As a simple example:
If the network is growing at 1% per annum in USD terms (several factors internally and externally go into this), then holders break even in USD terms with 1% net issuance per annum. Logically:
But again, this breakeven is measuring dilution specific to unstaked holders' purchasing power (in USD), not "network profitability." In scenario 2, holders lost some purchasing power which stakers equivalently gained. It's wealth redistribution, not wealth destruction.
Something that the popular "profitability" argument also misses - physical operational costs must be accounted for. This encapsulates validator expenditure on hardware, compute cost, etc.:
This analysis is wrong. Inflation or block reward issuance is purely a fee on unstaked native tokens, or on users not participating in consensus. It’s not a “cost”. “Profit” = (network fees) - (physical cost to run the boxes) https://t.co/N1qtYOUkdZ
— toly 🇺🇸 (@aeyakovenko) March 31, 2022
Let's do a simple example again:
If you're deciding from an outside perspective whether or not you want to be a validator, you have a very simple equation. Assume for simplicity that operational costs scale linearly with your stake weight (they don't, there are some centralizing returns to scale in distributing fixed costs). For any amount of stake I put up, my income will = 10% and my operational costs will = 2%. I net 8% on my staked capital.
At a network level, the analysis here is again simple. Everyone is staked and we have no inflation, so the actual profitability to all token holders is $8 bn. The stakers collectively made $10 bn and spent $2 bn.
Looking ahead: MCAP = NPV (future fees - future physical operational costs)
Again, you could add in inflation if you like and have separate DCFs for stakers vs. holders. However, everything will net out to be the same on aggregate. The extra validator income will = the loss to holders. Similarly, if fees are burned instead of all sent to validators, you just shift the allocation from validators to other holders, but the same amount is captured on aggregate.
Thankfully, operational costs are generally relatively low for staking relative to most traditional businesses. It's a high margin business where the predominant cost is your opportunity cost of capital staked.
YES! A LOT!
But I think most people are wrong about why it's important, and that clouds our ability to address it properly. Issuance is being jammed into a misguided and simplistic notion of "profitability." A blockchain doesn't immediately become "unprofitable" if issuance > fees or "unsustainable" with slightly positive net issuance.
There's significant merit to arguing that ETH should be very deflationary which I'll go into in following pieces, but it's not for the popular reasons I presented earlier.
So, inflation is incredibly important depending on the situation/mechanics and what behavior you want to incentivize. For example, aggressive early PoW issuance was effective in broadly distributing ownership for BTC and ETH. Conversely, excessive dilution eventually becomes destructive. Relative incentivization of holding vs. staking matters a lot. Especially because you can't just value ETH based on a DCF anyway. It's fundamentally different from a traditional stock, and utility can potentially be derived from it in a "money-like" way.
As I mentioned earlier, would you immediately sell or stake all of your ETH if we changed the net issuance to always be 0.00000000001% forever? Of course not! You may derive utility from it as a SoV, for gas fees, because it'll share in price appreciation, or whatever else. And inflation does impact these components of demand (which drive the real market valuation beyond a simple DCF).
As a result, changes in inflation absolutely could impact its market cap in reality. However, not in the way that most seem to think or for the popular reasons.
Issuance affects the balance of utility between unstaked and staked ETH, so we must understand it properly if we're going to maximize the collective utility. Here's a great article by Polynya on the pernicious effects of high issuance. I'll also elaborate further on its potentially destructive effects in upcoming pieces.
Broadly, the "ultrasound money" argument posits that ETH's value is important for two reasons:
ETH is cited as pristine collateral primarily due to its lack of associated risks:
However, ETH has one problem - it's volatile. Hence the argument that we need decentralized stablecoins to be minted against it - we want to create a less volatile asset (USD stablecoins like DAI).
This economic bandwidth argument misses some key points in my view:
It's also stated that making ETH illiquid will directly make it more valuable. The argument is that the "liquid portion" should be valued based on a full DCF of Ethereum, then multiplied by the "illiquidity multiplier." For example, if a full Ethereum DCF implies a value of $1 tn and 80% is locked as collateral or staked, then the market cap should be $5 tn.
I disagree with this line of thinking for several reasons.
I agree that economic security is valuable as described above. I'll also note of course that economic security isn't sufficient in isolation. Even more importantly, a network should be robust and decentralized enough to recover from such an attack, but that is a far longer argument out of scope for the discussion here.
However, as I've written previously:
"This security argument alone also ignores the moral argument and overall need for a robust base layer asset in this new digital world, especially with BTC's potential shortcomings. This is an argument for another day, but an important role that I believe in and many in the crypto community clearly want as well."
Well today is that day, which brings me to my next point. My general view is simple for all Ethereum considerations, economic or otherwise - how can the Ethereum protocol offer the greatest benefit to the world?
That goal is of course highly subjective, so now I'll try to be a bit more concrete in how I think about this. There's broadly two considerations as to what Ethereum can provide:
These two goals are self-reinforcing in my view. Maximizing Ethereum ecosystem activity has several inputs which can form a virtuous cycle:
ETH is really cool, and I think it's valuable (NFA). My goal is to better refine why it's so cool and why it's so valuable (still NFA). I think existing popular arguments fall short. It's tempting to compress a nuanced topic into a couple one-liner memes, but they become detrimental in understanding how to optimize the protocol if they're inaccurate. It's unfortunately much more complicated.
Number go up = Demand for (economic collateral + speculation + staking [driven by MEV, pf] + store-of-value/reserve asset for the economy + unit-of-account for NFTs/ERC20s) - security budget + fee revenue - churn of speculators +/- other d/s vectors like restaking, bridging etc.
— polynya (@apolynya) February 10, 2023
And if you want to be selfish, a better argument will help you shill it better. The bankers aren't gonna buy your bags if we say it's ultrasound money because it's deflationary and get profitability wrong.
Don't worry, I'm down too.
Hopefully you now have a solid grasp of how value flows through the Ethereum protocol and why issuance is often misunderstood. Upcoming pieces will elaborate on how it can be optimized.
That conversation will be framed around what ETH as an asset is and possibly should be. This includes discussion of its place in the broader global economic and monetary backdrop, as well as the role of other crypto-assets such as LSTs and stablecoins.
Disclaimer: The views expressed in this post are solely those of the author in their individual capacity and are not the views of DBA Crypto, LLC or its affiliates (together with its affiliates, "DBA"). The author of this report has material personal positions in ETH; stETH; Layr Labs, Inc.; and Skip Protocol Inc.
This content is provided for informational purposes only, and should not be relied upon as the basis for an investment decision, and is not, and should not be assumed to be, complete. The contents herein are not to be construed as legal, business, or tax advice. References to any securities or digital assets are for illustrative purposes only, and do not constitute an investment recommendation or offer to provide investment advisory services. This post does not constitute investment advice or an offer to sell or a solicitation of an offer to purchase any limited partner interests in any investment vehicle managed by DBA.
Certain information contained within has been obtained from third-party sources. While taken from sources believed to be reliable, DBA makes no representations about the accuracy of the information.
