What is the Shanghai Upgrade?
In March or April (exact date TBD), the Shanghai upgrade will permit validators to initiate withdrawals of their staked ETH on the Beacon Chain. Since the launch of ETH staking and the Beacon Chain in December 2020, staked ETH has been locked and cannot yet be converted back into circulating ETH. This inherent illiquidity helped create the rise of liquid staking derivatives (LSDs), such as Lido stETH, Coinbase cbETH, and Rocketpool rETH. By providing a tradable representation of staked ETH, derivatives allowed it to be widely used in DeFi and gave people who do not run validators a way to access the staking yield. In this post, Gauntlet examines the Shanghai upgrade and gives our initial views on:
1. Withdrawal mechanics
2. Possible impacts on staking derivatives
3. Possible impacts on broader DeFi markets
We note that there are still details pending in many protocols preparations for the upgrade, so this post will stick to the basics and may be outdated once complete plans are known.
What will ETH withdrawal look like?
Withdrawal of ETH is a two-tiered process. Validators withdrawing their full balance must first go through the exit queue before entering the withdrawal queue.
The exit queue processes validators at a pace defined by the churn limit. Currently, there are roughly 500k distinct validators, and the churn limit is 7 validators per epoch. This means 7 validators can go through the exit queue every 32 blocks (approximately every 6.5 minutes). At the current growth rate of about 4k validators per week, the churn limit is expected to reach 8 by the time of the Shanghai upgrade.
The withdrawal queue then processes a maximum of 16 withdrawals per block, which results in a limit of around ~115k withdrawals a day. This suggests the upper bound on the withdrawal queue time will be about 5 days if it includes all 500k validators. Validators looking to withdraw partially (for example, only their staking rewards) can skip the exit queue and take a spot in the withdrawal queue immediately.
We can also estimate an upper bound for how much staked ETH can be withdrawn daily. Looking at the distribution of validator ETH positions, we find that roughly 250 (0.05%) have between 64 and 68 ETH, and the rest have 37 ETH or less. This is because validators typically stake 32 ETH to join the network, but a few have chosen to start with twice that amount. Assuming each validator attempts to withdraw their full balance, we estimate 60K ETH can be withdrawn per day (220 epochs x 8 validators per epoch x 36 ETH each), with an additional 10K ETH if all the larger validators withdraw on the same day.
Effects on Staking Derivatives
Assuming a smooth upgrade, we expect liquidity for staking derivatives to improve with the enablement of withdrawals. Though it may seem counterintuitive that redemptions of an asset could attract liquidity, the reduced friction of converting between staking derivatives and ETH will enhance the utility of such derivatives. We expect participants who are now hesitant to invest in derivatives due to non-convertibility to reevaluate the benefits against a reduced risk. As of this writing, derivative pools hold about 33% of all staked ETH, of which roughly 80% is in the two largest pools: Lido stETH and Coinbase cbETH. After the Shanghai upgrade, existing stakers and newcomers will be able to move their ETH more freely between various staking options, and we expect some will gravitate toward derivatives for convenience. Overall, we see ETH staking continuing to grow and likely accelerating with convertibility, especially in derivatives.
One risk of all derivatives, especially those with frictions in convertibility, is price deviating from the underlying asset it is intended to represent. For example, if the market perceives a staking derivative cannot deliver the underlying ETH, this derivative could trade at a large discount. Previously, depegging events of over 20% have occurred in staking derivatives, but for the last 6 months, we have seen a period of stability. The discount for Lido stETH versus ETH, which reached over 6% during the market-wide credit crunch last summer, is less than 1% and generally trending lower as of this writing. Though there is a slight risk that technical issues during the upgrade could result in a widening of discounts, we do not see reasons for concern at this point. Our current view is that it would take a historically large depegging to cause significant disruption in staking-related DeFi markets, and such a move is highly unlikely based on what is known about the upgrade.
Effects on DeFi More Broadly
As mentioned earlier, we expect demand for stETH and other staking derivatives to grow once there is a clear path to redemption. With the recent greater use of staking derivatives in Defi lending, we highlight that utilization may increase leading up to and after the upgrade. This is especially true if protocols continue to enable profitable recursive borrowing (i.e., supply the derivative, borrow ETH, convert it to the derivative, supply more, etc.). Assuming an ETH staking yield of 5% and typical lending parameters, this strategy can yield over 20% if looped repeatedly. While heavy use of recursive strategies may increase utilization and risk, we do not see an immediate problem here. In our work with risk management clients, we will monitor utilization as redemptions begin and propose new lending parameters if needed.
Finally, on many Defi protocols, price feeds are dependent on oracle networks like Chainlink or Pyth. While there are always risks of oracle exploits or manipulation, we do not see any reason for the Shanghai upgrade to affect this, given the current info. Price volatility can also impact the performance of these oracles, but we already covered volatility in the discussion of depegging risks. Overall, our analysis suggests the upgrade-related risk of manipulation or malfunction in staking derivative price oracles is quite low.
