One really strong argument against self-limiting, at least in my view, is that it opens the door for someone else, like Coinbase, to take majority market share.
A common stated assumption is that CEXs would not self-limit. Is there any way we can test that assumption?
There might also be an opportunity for a conditional self-limiting. For example, the DAO could vote to self-limit at X% share as long as no one else exceeds X% share either, and if they do then it removes the self-limit.
As the market leader, Lido is in a position to set a precedent which others might be likely to adhere to.
Social consensus on a divisive topic (and I think this thread demonstrates itâs not a one-sided one) is not a strong foundation; thereâs more than a few ways a solution can power through it. E.g. shift the perception window on whatâs considered acceptable (like Flashbots changed the perception of MEV); make some sort of sybil attack (e.g. the same big holder assures a silent controlling stake in multiple DAOs over time); make a meta-entity controlling most of the stake because itâs totally a different thing and doesnât have the same risks. On a long enough timescale, commitments from the big players of today mean nothing and incentives mean everything.
Also, an interesting fact is that at this point in time thereâs, to my understanding, 0 staking entities that committed to self-limit. Rocketpool and Stakewise teams did commit, but both are ostensibly DAOs and there havenât been any votes or even discussions on governance forums of these entities. That is not to say they wonât, but to point out the question hasnât been really settled (and that it wonât be resettled when there is a real choice for a dao between shooting for market domination and keeping their commitment). If Lido governance is not to be trusted, why should we trust other DAOs?
That said, the way I see the argument for limiting Lido and pushing hard for an oligopoly looks like this:
Oligopoly market is better for Ethereum than power law market.
Lidoâs domination is circumstantial and not indicative of the market shape in less constrained environment; oligopoly is a natural equilibrium.
A push to oligopoly now will make the market remain in a stable well-divided state forever
If all three are true, it doesnât matter that social push on liquid staking providers equity is temporary or that not everyone agrees to it. Iâm not sure I agree with an absolute form of 1 - I think both options can be positive or negative for Ethereum depending on how they end up on the side of governance risk and validator set management. I do not agree with 2 or 3 at all. Forces that made Lido that big become less pronounced when withdrawals are there, but they will remain in effect.
You cannot exit for 1ETH per stETH until after withdrawals are implemented, which is at best 6-8 months away, which is what cobie means by âuntil every user has the ability to protest the fee by no longer being a userâ.
Instituting this cap in the mean time will drive the ETH/stETH exchange rate down since it will only exacerbate selling pressure (and reduces demand, since youâre basically reducing the EV of the purchase). Your suggested mechanism doesnât work because there is even less reason to buy discounted stETH than right now since it will have even less APR attached to it â especially because selling stETH doesnât actually lower Lidoâs share, as itâs necessary that a) sizeable staking deposits are flow into the Beacon Chain and b) are going to somewhere not-Lido. So, in conditions where deposits are generally low and people are already exiting liquid staking positions, this will create even more stress on the exchange rate.
So not only is it morally dubious, itâs also economically both unlikely to work and impractical that this mechanism can even work pre-withdrawals.
Yes, that is the assumption. It has more or less been true so far, as staking deposits continue to flow in.
Regarding impact of a higher fee vs how many people would exit and how much it would reduce new adds, that really just comes down to the price elasticity, which we donât know and are just speculating.
But the mechanism I proposed works regardless of what that elasticity is. If you only raise the fee by a very slight amount, that will have some impact on exits and new adds. If it doesnât have enough impact, you raise it more. If it has enough impact, you stop. If it has too much impact, you undo it, or partially undo it. No matter what, thereâs some value X for a fee increase which gets you to an equilibrium that makes it so new adds offset the exits and you mint no new stETH but also donât depeg (since the demand for new stETH picks up the slack on the secondary markets).
For #1, a big part of it depends on who the other oligopoly players are. If three CEXs each end up with 22% share, that oligopoly is less desirable than Lido having 66%. But if the other oligopolies are liquid staking pools, each with their own operator set, itâs a better outcome than having it all be in one. Obviously also depends on which staking pools, as theyâre not equal in governance and validator sets.
For #2, I do believe that the large liquidity pool requirements pre-withdrawal are whatâs causing the market to be winner-take-most at the moment. Without that, I donât see strong network effects. Anyone can start a liquid staking service and the peg will be maintained by arbitrage. The barriers to entry arenât very high, nor are the network effects.
For #3, I agree with you. This is a very nascent market undergoing rapid change; I would not expect market shares be stable over the next couple of years regardless of their starting position today. Besides withdrawals, DVT will be especially transformative. As we move more in the direction of permissionless operators, and operating itself becomes easier (better UI, documentation, etc), it will become more commoditized. Whoever can implement DVT the best and fastest will have a lead on the resulting lower costs, creating an opportunity for market share gains.
That being said, thereâs a timing element to all of this. The problem is that Lidoâs market share is already high and rising, the merge is coming in August, and Lido still has a ways to go in terms of governance and expanding its validator set. Saying âLet us keep gaining more share now and we promise weâll improve these things laterâ doesnât exactly sit well.
For those on the fence, I would encourage you to consider how early we are. Self-limiting Lido at this early stage could leave the door open to another (more centralized, or even malicious) player racing past Lido.
One very uncomfortable topic that I did not see discussed above in the monopoly vs. oligopoly debate is what would actually happen in the event of a smart contract exploit of an oligopolist? If 70-80% of ETH is eventually staked and in the âhappy oligopolyâ scenario there are 5 major players, we could still be talking about 10ish% of the network. At that point, are we really more anti-fragile because only 10% was lost vs. simply accepting that an exploit to such core infrastructure requires a hard fork? One must ask if a mature Ethereum could withstand a 10% transfer to a nefarious actor at the social consensus layer. This is the more appropriate ânuclearâ analogy.
Perhaps I am not technical enough to understand but is an infinite mint vulnuerability impossible for stETH or a competitor? If not, doesnât the potential DeFi carnage exceed the value of underlying as was the case with wormhole and cashio exploits on Solana?
To me it appears that the proposed attempts of addressing the âissueâ at hand are pre-matured and not well thought through. Social commitment is weak; there is a business opportunity and less Ethereum-aligned entities will take advantage of it. CEXes are well capitalized, they can subsidize fees, aggressively gaining market share by offering a cost-competitive product.
Overall, limiting the growth of the most successful player to artificially advantage less competitive ones is market inefficient.
We need an on-going debate; thank you team for hosting this discussion on the Lido forum.
I wrote some thoughts on why Liquid Staking protocols exceeding consensus thresholds are inherently dangerous to both the Ethereum protocol and holders of the liquid staking asset.
These risks are fundamental and not directly due to how Lido is designed today. These risks would exist even in the best Lido design down the road.
I think it important to Lidoâs long-term viability as a product to self-limit.
And I think it important for capital allocators to be aware of the distinct risks they are taking on by holding capital in such a system pooled over consensus thresholds.
If the fee is only slightly raised so that only a small % of users exit, then you could also conclude that only a small % of new users would be dissuaded from using Lido since the fee has not changed much and existing users are still happy with it.
So, what is the point? Small raises fee raises that donât dissuade existing users likely also doesnât impact new growth/market share, because the change is too small.
In order to dissuade sufficient growth, you would have to raise the fee such that a reasonable number of new users were unhappy with the fee (>50%?). Existing users would be similarly dissuaded and liquidity cannot support mass-exodus.
Anyway, IMO, itâs not âmoral absolutismâ it is simply respecting of terms of engagement between Lido and users. I think Lido could have legal issues if it held user funds hostage under juiced fees and users couldnât exit without slippage losses.
I agree with your analyis in principal, but risks of cartelization and regulatory censorship arenât exclusive to LSD protocols. They exist with other large entities such as centralized exchanges as well.
In fact, I believe that the big three CEX are the bigger issue with the staking landscape as of today. Before withdrawals and some form of withdrawal address controlled way to exit, the control that lido governance has over validators is limited. Regulators going after three companies seems more feasible than going after LDO holders.
Therefore I believe that Lido increasing their share of validators and thus reducing the share of the big three exchanges is better for Ethereum for now than imposing a limit on Lido and growing CEX share.
Long term, these issues can also arise on the node operator level. If a handful of operators are running a large portion of validators, across a number of different smaller LSD protocols, the cartel and regulatory capture concerns are just shifted over to them instead. So Iâm sceptical that imposing a cap on LSD prtocols is the right tool to address these concerns.
I personally believe that many (not all, but many) for self-limiting are non LDO holders that do not understand the protocol, prescribe to an extreme view of decentralization to an irrational extent and/or have vested interests in other Liquid Staking protocols.
I am ideologically opposed to the idea of self-handicapping because of simply being the better choice.
I view it as more risky to the ecosystem to allow inferior options to unnaturally increase in size.
I believe that the risk of a terrible centralized KYC standard is very real and poses one of the largest threats to stakersâ ETH and the network, and this would be much more likely to materialize and succeed if Lido self-limits.
In a macro sense, I see the value in having a terminal state where stake is more evenly distributed across multiple entities and pools. But the fact that we are hoping to get there by asking the market leader to handicap itself gives me a little pause.
Lido has become the market leader because it was (i) first to market, (ii) itâs design isnât bound by supply side constraints, and (iii) offered a product that demonstrated crazy PMFâas it both collapsed the opportunity cost of capital for stakers significantly, and allowed for smaller and/or less sophisticated would be stakers to participate. So from a product-building perspective (strictly), Lido has done an outstanding job thus far.
Given that and given how we got on to this forum discussing what we are, I fear that should Lido decide to self-limit, the precedent set for successful Ethereum product/protocol builders (and those to be), is a dangerous one. Especially for those not intimate with the timeline, such a precedent makes an interpretation along the lines of âEthereum will let you succeed, but not too muchâ all the more likelyâa precedent that would turn future value from being created on Ethereum, elsewhere.
I am not convinced that Lido (or any other party) sitting at or above critical consensus thresholds is the best outcome for Ethereum. At the same time I also donât think that self limiting, at least in the way Iâve understood the timeline thus far, is the answer.
Note:Were this a broader and more open conversation, with equal participation from the next 1/3 of the stake-mass (less transparent CEX pools), itâs possible that my view would be different.
I really think this is whatâs missing from this discussion. Best outcome, in my view, would be a consortium of all the staking pools and CEXs to mutually agree on limits.
Given some elasticity of demand (regardless of what it is) there exists a value of X such that raising the fee by X% will reach an equilibrium state where existing users exiting and new users coming in (in terms of ETH) are equal to each other. At this point, you stop raising the fee.
Because these two equal each other, existing users who want to exit are able to exit because there is an offsetting demand for their stETH from the new users. And because the two equal each other, the total ETH staked remains flat. At that point, you just let the rest of the market keep growing, thus reducing the market share over time until it reaches the desired target (Vitalikâs suggestion is 15%). Once it reaches this target, you start lowering the fee again so the total ETH staked resumes growing again.
If Lido implements a self-limit 25% market share target, may be it reasonable for the extra fees generated by the mechanism to accrue 100% to LDO and 0% to node operators?
I believe it is reasonable for Lido to get 100% of extra fees for excess market share because the target mechanism acts to convert Lidoâs traction and network effect into extra fee revenue, and I donât see why any of that should flow to node operators.
If Lido gets 100% of the extra fees, itâs possible that LDO could generate a lot of extra cash from this mechanism.
Yes, a market share target comes at the cost of Lido running a majority/supermajority of stake, but is that the road we really want to go down anyway?
I know that Vasiliy has elaborated on his view that thereâs no problem with Lido owning an arbitrarily high amount of stake. But, how do we reconcile that view against the idea that such a majority share is toxic to the network?
Hasu suggested that a hyper-dominant derivative is destiny. He says thereâs no other equilibrium. I strongly disagree. I think that Lidoâs monopoly is circumstantial in the early lack of good competition, and I am concerned that the monopoly may be cemented for years by this groupâs inaction.
If Lido hyper-dominance being cemented sounds good to you, if thatâs the goal being advanced here, I think itâs going to end quite badly for LDO holders and ethereum.
If Lido becomes hyper-dominant, I see there being only two likely outcomes: either the community will eventually socially intervene to curtail Lidoâs dominance, which hurts LDO a lot, or ethereum itself may fail to realize its potential, which hurts LDO a lot since the price of ETH will be way down in USD terms.
Thank you Vasily and Sacha, and everyone else who has provided inputs thus far. This is an important topic not just for Lido, but for Ethereum community as a whole. And, the decision here will be a reference for many other organisations now and into the future. So it is important to arrive at the right decision.
Iâd like to argue that Lido should not limit new Eth staked through the protocol. I base this argument on the following points:
As of 1st June â22 Lido has over 32% of total staked Eth. While Lido has a significant share of the total staked Eth, there are many new and re-emerging competitors in Liquid Staking market such as Stakewise, Ankr, PStake etc. which is a good sign.
New competitors and existing competitors must work hard to gain trust and goodwill of stakers. Competing protocols receiving more Eth just because Lido is limiting its intake, is not a good outcome for the network as it doesnât encourage high performance.
The growth of Lido should encourage more providers to offer attractive, decentralised liquid staking solutions of their own. Success should encourage positive competition in this instance. With only 12% total Eth staked, there is a significant unmet or underserved Eth holders to attract. Lido even offers a good blue-print for other Liquid staking protocols to innovate upon.
When withdrawals are enabled, the staked Eth could be withdrawn and staked with any provider of choice based on who offers the best value to the staker (individual and institutional). This will reduce the appeal of liquid staking protocols and create more even playing field. This also means that the period of risk (on this factor alone) to the network is limited to between now and when withdrawals are enabled. If this really matters, then it should act as a motivation to accelerate work on withdrawals post merge.
Success of Lido and any other staking provider rests on continued improvement and health of the network. Any possible operator collusion, or governance capture will erode trust dramatically, damaging future returns of that staking service â making it non-attractive. This must be addressed with the governance improvements identified by Lido.
I rather encourage more liquid staking protocols to innovate and compete fairly for staked Eth, without Lido having to limit its intake.
If the eventual decision happens to be to âlimitâ, I strongly recommend that this is done dynamically i.e. Lido must be able to remove any such limits instantly if other entities happen to ignore these boundary conditions.