This post takes a focused view on the overlooked yet important factors facing the decentralization of Ethereum’s staking layer.
In this article, I will explore the following topics:
- Ether ETFs will follow Bitcoin ETF approvals
- Ether ETFs with rewards are a natural extension of unstaked Ether ETFs
- Institutional staking will move to liquid alternatives, presenting new challenges to Ethereum decentralization
- Magnitude of institutional flows could be quite large
- Wat mean?
- Lido as an effective counter-balance
I. Ether ETFs will follow Bitcoin ETF approvals
The confirmation of a spot Bitcoin ETF seems to be all but certain.
Naturally, people are turning their attention to a potential spot Ether ETF. The confidence in a spot Bitcoin ETF approval is rooted in the SEC’s apparent inconsistency in approving a futures-based product but withholding approval for a spot-based one. This attention has only been accelerated by Blackrock’s filing for a spot Ether ETF on November 9th.
Considering the existence of CME Ether futures markets and multiple futures-based Ether ETFs, the logic for approval seems quite transferable. Even the regulatory approach and treatment of Ether in the US has been on a non-security basis. The ability for Gensler or future regulators to go back on this previous treatment is unlikely for a number of reasons (see thread with image).
Indeed, the SECs recent legislation against Coinbase for listing securities excluded Ether.
II. Ether ETF with rewards are a natural extension of unstaked Ether ETFs
Pending a spot Ether ETF approval, issuers will race to figure out an implementation that allows them to earn Ethereum staking rewards. Rewards-bearing ETH is strictly better than non-rewards bearing ETH and could entice a new subset of investors who would have remained on the sidelines until now.
Issuers will compete on who can be the first to market to offer staking rewards. Initially, it does not seem rational for issuers to run their own validators given the knowledge barrier, business model challenges of node operation and increased regulatory risks.
To be first to market, issuers must propose a solution that fits within existing regulatory frameworks and can be approved as quickly as possible. Thus, the path of least resistance is for ETF issuers to enter into contractual agreements, including lending agreements, with 3rd party centralized staking providers, who take a small fee.
This is already the solution in place with 21Share’s staked Ether ETF, AETH. 21shares custodies their ETH with Coinbase Custody, and likely loan the underlying ETH to Coinbase Cloud, Blockdameon and Figment.
AETH has attracted $240.77 million in AUM, equaling 121,400 ETH, all of which has been staked with centralized providers. These inflows are completely yield-insensitive as some fixed % is programmatically staked regardless of the yield.
It seems reasonable to assume that US based ETF issuers will likely enter lending agreements similar to that of 21Shares AETH. This programmatic flow has the potential to drive greater market share to centralized providers relative to decentralized providers, as many custodians offer vertically integrated staking products (such as Coinbase Prime → Coinbase Cloud) or have existing SLAs with centralized staking providers such as Figment. As a concrete example, given the ARK filing was done with 21shares, it is reasonable to assume that they will use the same providers as the AETH product.
III. Institutional staking will move to liquid alternatives, presenting new challenges to Ethereum decentralization
Savvy institutions may look to stake with providers outside of the ETF wrapper that have more advantageous cost structures and greater utility. Decentralized protocols such as Lido are already accessible as assets to existing institutional customers in a variety of custodial, QC and regulated environments. As decentralized protocols, they offer a uniform experience and institutional-grade security but are open for all market participants with any number of ETH to stake.
On the other hand, a new trend of projects are positioning themselves specifically to target institutional demand. In particular, companies like the Liquid Collective are building a “liquid staking solution that is designed with institutional compliance needs in mind”. Institutions can mint lsETH, which is staked with one of 3 centralized providers (Coinbase, Figment, and Staked), who also govern the project. The idea here being two fold:
- Liquid staking is a superior product than vanilla staking; you retain the moneyness properties of your ETH and the majority of rewards.
- Institutions must stake with KYC/AML compliant providers or they may be subject to civil and criminal liability.
The first point is quite obvious.
Liquid staking tokens (LSTs) can be used as collateral throughout DeFi, base assets in liquidity pools, and avoid withdrawal queue times.
Furthermore, institutional products such as ETFs benefit from liquidity in order to manage fund redemptions in under a day. For illiquid funds, this is typically managed by holding some portion of the ETH unstaked in custody. This is at the risk of a larger rush of withdrawals precipitating essentially a bank run while the rest of the ETH un-stakes, as well as dragging on rewards rates during normal operation.
Having access to liquid staking tokens would make it possible to manage redemptions more fluidly and also increase the proportion of the fund that could be staked at any given time. For this to be a realistic possibility, the liquid staking token must, obviously, be liquid. Just offering a token is insufficient if this token does not have enough liquidity to be usable. At the moment, the only liquid staking token with any meaningful amount of on-chain and off-chain liquidity for institutional use is Lido’s stETH.
The second point is much less clear. Regulated institutions typically face a number of obligations in order to minimize the risk or possibility of money laundering or facilitating criminal activity. To this end, KYC/AML obligations exist in order to maintain an auditable trail of fund flows between an institution and its customers. Furthermore, higher requirements may exist for so-called “qualified custodians”. That said, qualified custodians and institutions in general should be able to meet their KYC/AML obligations without prejudice to the choice of asset, LST token or staking provider.
Even if a staking provider explicitly enters into a contractual relationship with the owner or custodian of funds, I do not believe regulators will create brand new KYC/AML compliance obligations for Ethereum staking. This is because I believe that regulators will understand, with time, that Ethereum staking is a sui generis computational activity whose characteristics do not match ‘flow of funds’ in a traditional or financial sense. Holders and custodians of LSTs should be able to perform KYC/AML on whatever assets are in their purview and meet their compliance obligations to reduce the risk of money laundering or financing crimes.
Concentration of stake within centralized entities poses a number of challenges of varying urgency to the development of the Ethereum blockchain. To summarize, a range of disruptions to the blockchain are possible and become increasingly likely at various levels of staking market share:
- Block reorganization attacks
- Finality delays
- Fork choice
- Complete coercion
The first two types of attacks disrupt the ordinary functioning of the blockchain and Ethereum has built-in incentives to dissuade attackers from attempting them such as gradually diluting rewards and staking balances for the ‘attacker’. However, at 33% market share or higher in a single node operator, this actor could begin to delay finality even if it became costly to disrupt network operations. At higher levels of market share, such as 50%, attackers can effectively fork the blockchain and pick the fork they ‘condone’. At 67% or higher, the blockchain is effectively a delegated database fully controlled by a single party.
These attacks are not just ideological or theoretical, they are at the heart of what makes Ethereum valuable as a settlement layer at all.
The proof-of-stake mechanism on Ethereum can make it possible for centralized players to accumulate large and potentially controlling market shares of total Ether staking through market forces alone. To wit, centralized entities are already advantageously positioned to capture the institutional market.
For instance, Coinbase, which has multiple lines of business, including custody relationships, can quickly convert these to staking relationships, and cement its place in the staking landscape. It is already the single largest validator on the network with 16% market share, and operates multiple acquisition channels, such as cbETH and Coinbase Earn for retail customers and Coinbase Cloud and Coinbase Prime for institutional customers.
In the wings of this new inflow of capital lies a potentially dangerous challenge to Ethereum’s neutrality, or what the Ethereum Foundation describes as a Layer 0 attack on the ‘social layer’ through overregulation. Adding an artificial construction that only supposedly “KYC/AML compliant” staking providers are “allowed”, even if this compliance is illusory and not grounded in legal or technical fact, would only accelerate adoption of staking within centralized players hoping to grow new market share through regulatory entrenchment and capture.
Centralized entities growing their market share pose a risk to Ethereum’s staking layer. They fundamentally have a different set of obligations compared to decentralized protocols. Decentralized protocols exist as smart contract incentive layers to coordinate activities between many participants.
For instance, Rocket Pool’s rETH has almost 20k holders, 9k RPL holders and over 2.2k staking deposit addresses representing unique node operators. Or Lido, a smart contract layer that coordinates over 39 globally distributed node operators, almost 300k stETH holders and about 41k LDO holders.
However, corporations owe a fiduciary responsibility to their shareholders first and foremost, and have an obligation to answer to their local law authorities and regulators. While the decentralization of Ethereum may be somewhat relevant for the business of a corporation, such as an exchange, it does not come above these two obligations.
There is other surface area available for overregulation, a type of soft Layer 0 attack that would compromise on the neutrality of Ethereum, even if loaded with good intentions. If Ethereum is to become the settlement layer of the world, it must have nuclear grade censorship resistance and credible neutrality. If large enough, these centralized entities will inhibit Ethereum’s core goals.
IV. Magnitude of institutional flows could be quite large
Some might discount the arrival or impact of spot Ether ETFs, given a lack of institutional interest. However, there are useful precedents in commodities that can inform the degree of interest that exchange-traded products (ETPs) bring to new asset classes. As a financial vehicle, ETFs and ETPs are remarkably effective at standardizing and democratizing access for institutions and retail investors alike. When such vehicles enter distribution channels for institutional allocation, pension funds or social security contributions, inflows significantly accelerate into the underlying asset classes.
The same benefits of a spot BTC ETF would apply to a spot ETH ETF as well. For example, the ability for the ~80% of US wealth controlled by financial advisors and institutions to participate — along with a general stamp of legitimacy from regulators and government. We can expect this increased legitimacy and recognition to drive further demand for Ether outside of the ETF wrapper. We are already seeing inklings of increased institutional interest. Intermediaries such as Bitwise are hearing potential allocations being raised from 1% to 5%. Brian Armstrong said Coinbase doubled the amount of institutions onboarded to over 100 in their Q3 earnings report.
Guessing the exact amount of inflows from institutions is difficult. However, Gold’s GLD ETF had $3.1 billion of net flows in year one alone. Buying any amount of Gold pre-ETF was much more difficult than buying BTC/ETH. You needed to physically transport and store it, verify and authenticate its purity, and incur high transaction costs from dealers. The Gold ETF represented an improvement on the asset’s underlying properties and democratized its value proposition to millions of individual investors through giant fund allocations.
Bitcoin and Ether are digital-first assets in which billions of dollars can be transported in minutes. Physical Gold’s barriers to entry do not necessarily exist here. Although ETFs do not necessarily improve Bitcoin or Ether’s fundamental value proposition — they may threaten to hinder them — ETFs offer similar degrees of democratization and access to both asset classes.
The reality is that the overwhelming majority of individuals on the planet may well never own any physical crypto themselves, but content themselves to some degree of financial exposure through, for instance, pension allocations or private savings or investments. Regulated financial channels in the developed world already have a high degree of penetration and capillarity. ETFs can help demystify an asset class to an individual who might otherwise be on the sidelines and makes it easy to introduce new exposure within a framework individual investors are already familiar with (i.e. a bank or brokerage).
V. Wat mean?
So what does this all mean for Ethereum’s staking layer?
Large, yield-insensitive flows from institutions will likely drive the total amount of ETH staked higher than economic rationale would suggest, or probabilistic targets based on crypto endogenous variables. The lion’s share of these flows could disproportionately accrue to centralized providers inside and outside the ETF wrapper. A greater concentration of stake within centralized entities would degrade Ethereum’s censorship resistance and credible neutrality without a credible and successful counter-balance.
VI. Lido as an effective counter-balance
Much of the public conversation today revolves around Lido — whether or not they control too much stake, and the possible attack vectors this could introduce in worst case scenarios. This is an important and worthwhile conversation. At risk of rehashing tired debates, here are some quick resources that outline i) how much control Lido governance actually has over node operators, if any, and ii) how the DAO is approaching governance risk and iii) how the DAO is considering expanding the node operator (NO) set and decentralizing the overall validator set.
- Governance risk: Hasu’s GOOSE submission + Dual Governance Proposal
- Expanding the NO set: The staking router + DVT Module
- Decentralizing the validator set: Jon Charbs PoG + Grandjean paper
Please also see this great article on the practical risks of Lido dominance from Mike Neuder of the Ethereum Foundation.
Much of the Lido criticism is founded in a static view of the staking market. It does not take into account how the staking market is set to grow and its practical realities. To adequately assess the staking market, future growth and market forces must be accounted for:
- Future growth: Institutional adoption could drive concentration to centralized entities
- Market forces: (Liquid) staking has strong winner take all-or-most dynamics
The majority of this article outlined future growth, as it is under-discussed and an important aspect. Winner take most dynamics have been argued at length in our digital public arena, but often lack the context of future growth. Rational market incentives, including an incentive to maintain a decentralized Ethereum network, might not prevent institutions from going down the path of least resistance towards onboarding new capital onto Ethereum.
The only effective counterbalance is to grow the market share of decentralized liquid staking protocols at the expense of centralized market share. While there may well be a possibility for multiple decentralized protocols to gain substantial enough share to make an effective buttress, Lido is the only current viable option to keep Ethereum’s staking layer robust and decentralized:
- It is objectively successful at attracting new Ether from holders, as Lido smart contracts have routed over 30% of all staked ETH and stETH has almost 300k holders
- It is objectively successful at rate-limiting the growth of individual node operators, as they each use Lido as a successful acquisition channel for new Ether stake but are unable to grow their individual market shares within Lido more than the other node operators
- Its governance is objectively minimal at the moment and is further minimizing with initiatives such as dual governance
Although the Ethereum base layer is designed to function with ‘no governance’ or very constrained fork-choice governance for example, having an intermediated layer of one or multiple decentralized protocols could fill a necessary gap that Ethereum is unable to. Jon Charbonneau describes it as:
“In particular, LST governance can manage the additional subjective incentives required to decentralize operators (e.g., different modules may receive different fee rates). Free market economics do not lead to a long-tail of solo stakers or uniform stake distribution in the long-run. The Ethereum core protocol is largely built on the notion that it should be objective and un-opinionated whenever possible. However, subjective management and incentives will be required to achieve a decentralized operator set.
While governance minimization is often desirable, some minimal form of governance will always likely be necessary for LSTs. Some process is needed to match the demand for staking with the demand to run validators. LST governance will always be needed to manage the objective functions of the node operator set (e.g., targets on stake distribution, different module weighting, geography targets, etc.). This fine-tuning can be infrequent, but this high-level goal setting is critical for monitoring and maintaining the decentralization of the operator set.”