Staking digital assets often requires tokens to be locked up as collateral as a way to secure the network in exchange for rewards. As a result, tokens remain locked and inaccessible, and subject to “warm up” and “unbonding” periods. The alternative is “liquid staking” which has become more popular as it opens up opportunities to efficiently utilize staked assets as collateral to trade, lend, and provision more quickly.
Liquid staking generally requires minting staked tokens which are initially valued 1-to-1 with the native token. Lido in particular mints stETH representing the ETH staked on the Beacon Chain, where the yield earned and paid in ETH is similarly accrued in stETH (minus fees). While most staking derivatives typically trade below par, stETH had been the exception up until early May 2022, when we saw fund liquidations associated with the TerraUSD (UST) depegging incident.
In our view, the price divergence between stETH and ETH reflects important liquidity, yield, credit and even collateral risks. For example, liquid staked ETH on Ethereum’s Beacon Chain represents only a small fraction of the total outstanding ETH supply (3.8%), about 90% of which is staked on Lido. Conversely, expanding illiquidity windows on the base asset (activation and exit queues) also have an impact on how liquid assets like stETH are valued.
Technological risk is another consideration unique to the crypto space, as a finite set of available liquid staking protocols may limit decentralization and pose security risks for the underlying networks (e.g. Ethereum). That could potentially affect the value of the underlying or lead protocols like Lido to self limit supply. On the upside, there are solutions underway that may help to remediate some of these concerns.
Very recently, we have seen a decoupling between Lido’s liquid staked stETH and underlying ETH prices driven by forced liquidations from margin calls on leveraged positions. Part of that is a reflection of the poor trading in the crypto markets following the TerraUSD (UST) incident. Lido had a stETH version on Terra called bonded ether (bETH), which could be pledged as collateral on Anchor. As a result, the outflows from Anchor put sell-pressure on stETH. A more recent catalyst for the spread widening between stETH and ETH may have also been the reorganization on Ethereum’s Beacon Chain (May 25) as well as the Ropsten testnet failure on the same day.
1. Widening discount between stETH and ETH
For some context, Lido is one of the most popular liquid staking pools for ETH ahead of the merge, accounting for 90% of the liquid staking market on Ethereum. It has a 32.1% market share of all staked ETH in the Beacon chain. (See chart 2.) Lido staked ETH is also a relatively unique case of a liquid staked derivative having a history of trading at or near parity to its underlying (at least prior to May 2022), perhaps because of its dominant market share of Beacon Chain validators.
Lido’s service allows users to stake one ETH in return for one stETH, which represents a claim on one ETH staked on the Beacon Chain. Staked ETH on Lido currently offers a ~4% APR and can be used in DeFi protocols. In order to compound this yield, users can implement recursive lending (or “folding”) in order to borrow incremental ETH on a lending/borrowing platform like AAVE with stETH as collateral, then stake that ETH, and repeat this process in order to generate up to ~10% APR.
After trading at or close to parity to ETH between January to early May, stETH is currently trading at an average ~2.2% discount to ETH, as market players continue to test the solvency of consensus trades. We think there are several factors contributing to the price divergence between stETH and ETH, which were made apparent in this most recent period of volatility.
2. Lido staked ETH is 32% of all ETH staked
First and foremost is liquidity risk, which refers to a market players’ ability to sell their liquid staked tokens at a more-or-less efficient price regardless of market conditions. In our view, the liquidity risk for liquid staked tokens is very high, and perhaps the biggest component of the discount to the underlying. The supply of all staked ETH tokens as a percentage of total outstanding ETH supply is 3.8% (stETH specifically is 3.4%), which is a limiting factor for the liquidity of these assets. Our analysis suggests that stETH volume is only 2% of ETH’s onchain volume over the last 90 days. (Also, stETH liquidity on centralized exchanges is negligible compared to ETH liquidity.)
Also, the fact that stETH cannot be redeemed on Lido makes stETH a more risky asset, as buyers of stETH need to be compensated for the possibility of the merge being delayed (for our analysis of Ethereum’s merge timeline, please take look at our report here.)
If such an event risk occurred, this could potentially leave stETH holders with a position that they may find challenging to offload, likely at deeply inefficient price levels. Whether ~6% (that is, the 4% yield offered by Lido plus the 2% discount for stETH vs ETH) is sufficient compensation for bearing this liquidity risk depends on (1) your assigned probability of market stress or volatility spikes that can impact the value of the liquid staked assets and (2) the likelihood of gaining additional yield via reinvesting stETH in DeFi applications. Chart 3 suggests only about half of stETH supply is taking advantage of the latter.
3. How much stETH is locked in DeFi?
Financial decision making is a relative game. Buying stETH at 4% APR is not a decision that can be considered in isolation, but rather needs to be measured against its alternatives. With the U.S. Federal Reserve hiking rates and shrinking its balance sheet, nominal returns on traditional assets will start to rise. Meanwhile, stETH’s yield is benchmarked to the rewards earned for staking ETH on the Beacon Chain. Validators earn inflationary rewards that are inversely proportional to the amount of ETH staked in the chain. As more ETH are staked, the rewards decline.
That said, when the merge happens, yields on staked ETH should rise as validators receive the transaction fees previously paid to miners. As such, post-merge validators and subsequently liquid staking asset holders earn transaction fees (gas) and MEV (maximal extractable value) which are not inversely proportional to the amount of ETH staked in the chain.
Keep in mind too that there could also be credit or counterparty risk that a protocol may fail to return the underlying assets backing its supply of liquid staked tokens. Smart contracts can help mitigate that risk somewhat, as in the case of Lido deposits made after July 2021, although protocols may still be able to block withdrawals.
There could also be collateral risk, as stETH is accepted within DeFi structures, but institutional investors cannot use the token to post as collateral on centralized exchanges.
Finally, the technological risk of liquid staked tokens is unique to the asset class and refers to the security concerns created by increased centralization. A recent blog post on ethereum.org discusses “the cartelization of block space” in which “staked capital becomes discouraged from staking elsewhere due to outsized cartel rewards.” This poses an existential threat to the network as theoretically operators could be coerced into censorship activities, removing node operators, capitalizing on multi-block MEV or even perform reorganizations of the blockchain. In other words, Lido’s continued growth could potentially pose a threat to the value of ETH itself.
On the other hand, if a protocol like Lido does self limit its stETH supply to contain its governance power, that reduction in stETH supply could potentially increase the liquidity risk. Lido discusses other arguments against limiting its size in a blog post here. In our view, however, a more optimized outcome would be to have healthy competition among multiple liquid staking protocols.
Benefits of liquid staking
That said, we believe that the price discrepancy between stETH and ETH also needs to be considered against the benefits offered by liquid staked assets.
First, liquid staked derivatives pay out rewards based on the yield of the underlying asset minus the costs associated with maintaining the expensive hardware that network validators would otherwise require. This is particularly relevant for ETH, where getting access to rewards via staking currently takes 16 days (the activation queue) which is expected to increase as more staked ETH comes online.
Second, it can eliminate minimum staking requirements like the need for validators on Ethereum’s Beacon Chain to have 32 ETH as a precondition for staking. These protocols also provide and maintain the machinery required to continuously participate in network consensus duties, whereas individual validators may incur opportunity losses if they are unable to keep 100% uptime.
Third, liquid staking allows users to exit their positions, which is particularly relevant for ETH staked on the Beacon Chain, which is currently not available for withdrawal until after the merge. The proposal to enable withdrawals is part of the Shanghai hard fork which may not be implemented until end 1Q23 at its earliest, based on our own estimates. That said, stETH cannot currently be redeemed for ETH on Lido, only swapped on exchanges.
Finally, it offers the possibility of additional yield in DeFi applications. For example, stETH has a total outstanding supply of 4,211,049 tokens as of end-May (for a market cap worth ~US$7.35B). About 50.3% (2.1M stETH) of that supply is used in DeFi, mostly in lending (1.4M stETH or $2.7B.) Lending in AAVE specifically is where a third of the total supply of stETH is currently parked – something that only picked up starting in March 2022.
To be clear, we believe liquid staking tokens like stETH have their shortcomings, which in our view makes trading at a small discount to ETH appropriate. Indeed, buyers are effectively suppliers of stETH liquidity, so in an efficient market, we could expect to see a discount on the bid price versus parity on the offer price for stETH. Many of these shortcomings are also related to the state of the Ethereum network itself. Post merge, when withdrawals are enabled for staked ETH, we would expect liquid staking tokens to have fewer tradeoffs.
This is also still a fairly nascent market, and the space is developing quickly to remediate or address some of these issues mentioned above. For example, Coinbase recently announced their support for Alluvial, an industry standard of enterprise-grade liquid staking across multiple protocols, which will require contributors to enable embedded KYC/AML processes. Liquid staking tokens via Alluvial can be thought of in terms of wBTC (wrapped BTC); gated minting/redeeming, but permissionlessly useful in DeFi, making them more accessible for institutional adoption.
Lastly, new liquid staking solutions coming to market are built for the post-merge world, meaning they don’t carry technological debt from supporting early versions of Ethereum staking. Ultimate token custody, rewards attribution, and other processes can be made more secure and decentralized from the start, rather than as a patchwork process.
Overall, liquid staking can provide a more capital efficient solution for market players whose staked tokens would otherwise remain locked and inaccessible. For example, stETH can earn a higher yield than comparably risky ETH investments, which in preceding months may have helped close the gap to its par value. Of course, liquid staked tokens do come with their own set of risks, and there can be price divergences akin to the one observed between stETH and ETH.
Decomposing those risks is not easy, particularly given the lack of available history for liquid staked assets, making methods like principle component analysis unviable. Moreover, it’s important to note that stETH is sensitive to shifting duration depending on the likelihood of the merge and ETH withdrawals from the Beacon Chain. Overall, we expect that liquid staking solutions will continue to mature in the future while changing protocol dynamics plus new entrants may change the cost-benefit analysis for token holders.