[RFC] Adjusting the ETH+ Collateral Basket To Adjust Liquidity Constraints To Protect ETH+ Holders From Slippages

0xSleepy · August 12, 2024, 7:53pm

Summary

To change the collateral basket of ETH Plus (ETH+) into a more liquid structure allowing for easier redemptions, issuance, and adjustments on its collateral basket as it continues to scale.

Current ETH+ Collateral Basket

33.4% wstETH

33.4% sfrxETH

33.4% rETH

Proposed ETH+ Collateral Basket

61% wstETH

24% rETH

15% sfrxETH

Abstract

This proposal aims to adjust the collateral basket for ETH Plus (ETH+) to enhance liquidity and reduce slippage risks as the protocol scales. The current basket comprises equal allocations to wstETH, sfrxETH, and rETH. However, liquidity assessments indicate that this structure could lead to significant slippage during large redemptions or collateral adjustments, especially beyond 15,000 ETH, posing risks to ETH+ holders and RSR stakers.

The proposed new basket increases the allocation to wstETH to 61%, reduces rETH to 24%, and sfrxETH to 15%. This shift aims to improve the basket’s liquidity, allowing ETH+ to handle adjustments up to 30,000 ETH with lower slippage risks. The adjustment addresses immediate liquidity concerns while leaving room for future diversification as the protocol grows.

Problem Statement

The IP to adjust the ETH+ Collateral basket did not reach quorum. As the community and ABC Labs were performing further testing and assessment the new recommendation was that the proposed basket change introduced potential risks to ETH+ holders and RSR stakers thus a no vote was recommended based on the current IP.

As seen in this ABC announcement, cbETH may not be the best choice to include in the ETH+ basket for the time being. A new proposal must be created to address ETH+ liquidity concerns as it continues to scale. Once liquidity concerns are fully addressed, new collaterals will be added to diversify ETH+ further.

Rationale

This proposal is looking to adjust the following ETH+ basket into a diversified index that scales well with the onchain liquidity available for each collateral.

Current Collateral Basket

33.4% wstETH

33.4% sfrxETH

33.4% rETH

Proposed Collateral Basket

61% wstETH

24% rETH

15% sfrxETH

Each LST plugin is designed to consolidate redemptions and issuances via onchain liquidity and then through redemption and minting. This is to reduce slippage when adjusting the ETH+ RToken market cap or changes within the collateral basket.

ETH+ Slippage Data

The ETH+ backingBuffer is currently at 0.5% to cover loss from events such as slippage before ETH+ could receive a loss which would in turn be covered by RSR stakers. Based on the available onchain liquidity, ETH+ would incur slippage greater than the backingBuffer from sfrxETH at 5,000 ETH collateral adjustment, rETH would incur slippage greater than the backingBuffer at greater than 7,500 ETH adjustment. This means with large redemptions, issuance, and collateral basket changes greater than those amounts from the underlying liquidity, ETH+ would incur a loss affecting RSR stakers.

Projected Slippage Based on Current ETH+ Basket

Based on the current collateral basket, ETH+ will incur slippage greater than the supporting backingBuffer at a redemption size greater than 15000 ETH. Slippage of this size could be incurred via a large redemption or a collateral basket adjustment.

Projected Slippage Based on Proposed ETH+ Basket

The new proposed collateral basket aims to reduce slippage risk by adjusting the basket into a more liquid structure allowing ETH+ to operate under a liquid basket that will be able to scale more easily with its increased market cap while allowing new collaterals to more easily integrate into the ETH+ basket as they are introduced.

Adjusting to a new proposed basket of 61% wstETH, 15% sfrxETH, 24% rETH slippage affects ETH+ price at 30,000 ETH adjustment instead of 15,000 ETH adjustment reducing loss from slippage significantly as ETH+ scales.

This assesses the ETH+ liquidity concerns as it continues to scale while leaving a considerable amount of room for it to diversify as potential new collaterals are added to its basket.

Risks

The proposed basket almost doubles exposure to wstETH, so an exploit event occurring related to wstETH would have a significantly greater effect on ETH+. If liquidity is reduced drastically this could affect the projections on price slippage for ETH+. Though ETH+ becomes more liquid to avoid slippage it significantly becomes less diversified.

I am in favor of the proposed change in the ETH+ collateral basket
I am not in favor of the proposed change in the ETH+ collateral basket

0 voters

rspa_StableLab · August 12, 2024, 8:07pm

Thank you for this well researched proposal!

tbrent · August 13, 2024, 9:14pm

This is an excellent piece of research! I can imagine this kind of analysis becoming table stakes over time for large basket changes. Really cool to see.

I think 61% is too much for any single token. It makes it hard to present ETH+ as a diversification option. This is made worse by the small overcollateralization layer.

Maybe 49/50% is the upper limit that any single token should have, generally? I don’t know, I’m just spitballing. I admit I’m using aesthetics here to inform this number, but I think other people will too so it’s not entirely ungrounded.

Here’s the slippery slope I worry about:

Premise 1: The LST market has winner-take-all / power-law dynamics
Premise 2: Liquidity (for the most part) follows market share
Therefore: if the ETH+ basket optimizes for liquidity it will also have a power-law distribution

My guess is that the “liquidity optimized” basket is the right starting place but it should be discounted by some amount. Maybe that’s already how the 61% number was produced?

Westwood · August 14, 2024, 2:41am

What would slippage look like with a 60/20/20 weighting? From the chart it looks like rETH has the most slippage and sfrxETH in the middle. What if you met them in the middle instead?

0xJMG · August 14, 2024, 4:13pm

Agree this is a terrific analysis and RFC. Definitely sets the bar high on respectable analysis to rationalize changes.

Suggestion @0xSleepy if possible please only use a single color once on the graphs. You’ve got a few blues on same graph and its difficult to discern. Less vested readers won’t take the time to solve the puzzle.

Agree with @tbrent comments above.

Suggestion: What would enhance the RFC is to include the ABC Labs calculus on why not cbETH? re: sharing a thoughtful public calculus can (a) circulate to the cbETH teams AND community to increase accountability (especially in light of them now launching cbBTC - will that have same limitations? and (b) to teach other protocols what are optimal liqiuidity calculus for being includied in an RToken or in other composable applications across DeFi. cc: @griffpeer

griffpeer · August 14, 2024, 7:00pm

Good points Taylor - important to consider diversification as safety mechanism as well.

RE: cbETH
cbETH’s liquidity can be seen in the chart here - it is basically a vertical line, showing almost no onchain liquidity. There is significant slippage at even a 1500 cbETH trade

When ABC Labs reached out to the cbETH team to understand their strategy, they mentioned that growing onchain DEX liquidity is not a priority for them. I imagine their team is trying to push users to the coinbase app as the entry/exit point – cbETH actually has much deeper liquidity on base, although that does not help in this discussion.

0xSleepy · August 14, 2024, 7:50pm

60% wstETH
20% rETH
20% sfrxETH

This looks like a viable alternative with limited affect on liquidity

0xSleepy · August 15, 2024, 7:07pm

50% wstETH
25% rETH
25% sfrxETH

0xJMG · August 16, 2024, 1:44am

I find @0xSleepy’s calculus and charting to be really useful! @nagaking @starl3xx I wonder if some of that could be included in this blog post to further advance and educate other deployers and governors on the “building resilient RTokens” mission? Building Resilient RTokens: The Art of Asset Basket Design | Reserve

StableScarab · August 16, 2024, 11:17am

Hey everyone. StableScarab from the Frax community. Big fan of what Reserve has built and it’s awesome to see your TVL grow, ETH+ being a key part of it.

I think ETH+ is the ideal mix:

sfrxETH has the highest yield
stETH is the most lindy
rETH is the most decentralized

I think @tbrent raises an important point, the ETH+ mandate is:

Maintain an Ethereum-aligned Liquid Staking Token basket.
Positively impact the Ethereum staking distribution.
Provide value to ETH+ holders through diversification.

A product which was created to improve staking decentralization, risks losing holders by balancing a majority weight to the leading staker. However, liquidity risk is also important and @0xSleepy has done an excellent job in surfacing that through analysis! I believe the goal here should be to keep ETH+ as diversified as possible without adding unnecessary risk.

Without being able to interview the holders of ETH+, the best thing Reserve can do for them is be consistent and transparent. To that end, I think it’s important to try to proceduralize liquidity-based basket changes that negatively impact the core tenet of diversification. That logic could look something like:

For large redemptions: On a quarterly basis, ETH+ will adjust basket weighting if needed so that an instant full redemption from the largest holder will not incur slippage greater than the backing buffer.
For changing the collateral basket: On a quarterly basis, ETH+ will adjust basket weighting if needed so that the addition/removal of a collateral would not incur slippage greater than the backing buffer.

@0xSleepy can you expand on your choice of the 15k ETH threshold? Can basket changes be introduced gradually so slippage is of lesser concern? Doubling this limit to 30k ETH would be 3/4 of current TVL. Understanding the rationale behind these thresholds would help guide future liquidity-based changes when ETH+ continues to grow

0xSleepy · August 17, 2024, 8:28pm

Understanding the 15k ETH threshold

15k ETH is the current redemption threshold before ETH+ becomes at risk of depegging. When redeeming or minting, ETH+ opts for purchasing the underlying basket via through online liquidity. When minting through online liquidity, minting and redemptions don’t have directly interact with the underlying mechanisms of each LST product, but through Uniswap and Curve, reducing overall design friction.

When purchasing and selling, there comes a point where users can receive a loss through purchasing via slippage. ETH+ has a backing buffer Reserve Protocol Docs | RTokens that protects ETH+ holders from taking a loss from this slippage. If the total slippage is greater then the backing buffer, then ETH+ holders will incur a loss that is covered through RSR stakers.

Reserve protocol recently did research on one of the larger holders.

This multisig currently hold 5,000 ETH+ as can be seen in their DeBank Profile

At 30K ETH threshold, this would maximize the runway as ETH+ grows without encountering liquidity constraints anytime soon. As the the LST index scales, the size held by large holders will increase and they will need to assurances that ETH+ is designed to handle the amount of ETH+ they hold in case they need to redeem.

This means that measures must be put in place, in advance for large holders to feel comfortable to mint more. If large holders already holds 33% of the total redemption value that will affect the backing buffer, they may be hesitant to mint more.

The model above takes into a consideration a full backing buffer, if the backing buffer is not full due to mass redemption already occurring or changes underlying liquidity of the LSTs backing ETH+, this could directly affect large holders and whales and keep them from redeeming at the full underlying value. If in an event like this occurs and they still choose to redeem, all ETH+ holders could see significant loss.The idea of this decision to address liquidity constraints is at the current basket, there have already been liquidity concerns with products like rETH as ETH+ continues to grow. This concern will only continue to remain a problem until addressed.

If the community looked to adjust the basket dynamically (every quarter) here are a few things to consider
Double trading: the community must not uniformly lower all collaterals at once, at least one of the collaterals need to stay constant. This means as new collaterals are added (in the future) and the collateral basket is adjusted ideally one collateral should not be adjusted. wstETH, would likely be the candidate to take the large share as it is the most liquid.

More Governance: More governance will be required to monitor and deliberate on this. If the market conditions change rapidlly and ETH+ enters a liquidity crunch due to redemptions. The governance process may not be able to adequately address these concerns in time. Which could lead to loss greater then the backing_buffer if not adequately prepared.

Concerns with designing ETH+ fully around liquidity
As noted by @tbrent and @StableScarab

There are questions on whether this does not follow the ETH+ mandates and concentrates too much risk on wstETH. If the community opted for 50% wstETH 25% sfrxETH and 25% rETH this could double the potential ETH+ that could be redeemed before surpassing the backing_buffer.

One way to affectively address this concern is look to begin adding new LSTs to further diversify the product. This will introduce greater counter parity risk to that LSTs liquidity and their protocols infrastructure. However, it would further diversify ETH+.

tbrent · August 17, 2024, 9:05pm

Some clarifications about behavior during mass redemptions:

at the protocol level, there are no bank run dynamics; large redemptions do not impact the remaining ETH+ holders (redemptions are capped to be prorata)
the backing buffer is strengthened by redemptions, not depleted; the marginal extra overcollateralization becomes unnecessary and becomes recognized as revenue (see USDC+ price history it is an interesting case of mass redemptions causing RToken appreciation)
mass redemptions do cause a loss for the marginal zap redeemer. these people are choosing to not just redeem their RToken at the protocol level but also swap the components via pools for a single asset, which can come with lots of slippage if markets are temporarily de-pegged

StableScarab · August 19, 2024, 9:06pm

Appreciate the detailed response @0xSleepy!

I was hoping the procedural approach I was advocating for could actually result in less governance changes and more clarity for ETH+. Basket changes would only need to be made if previously agreed upon liquidity risk criteria weren’t met. If concerned about rapid changes, quick reviews could be held more frequently (often without any changes). Also, to reduce change frequency, the constraints could have a healthy range i.e. ETH+ will adjust weighting so that the largest holder doesn’t hold more than 20% +/- 10% of the total redemption value.

With the above guideline in place, the decision for the liquidity-based basket changes becomes more clear for the holders in cases like this:

“33% > 20%+10% so we will be reducing ETH+ diversification to address liquidity risk to 50% wstETH 25% sfrxETH and 25% rETH to bring largest holder total redemption value back to 20%.”

I’m not familiar enough with the backing_buffer to know the probability of it not being full. Based on @tbrent’s comments it’s unclear to me how probable the 15k redemption being hit is, and balancing that risk with diversification is the tricky part. Looking forward to what you end up deciding! Growth is a good problem to have

knight_mayr · August 22, 2024, 9:19am

Hey folks! Kranthi here from the Staderlabs community. We have been following the protocol for a while now and are impressed by its trajectory

With the need to solve for over-exposure to current LST assets in the basket, we would like to propose the addition of ETHx, Stader’s Liquid Staking token on Ethereum to diversify the risk profile of the index

We have been live for over a year now and are one of the fastest growing ETH LSTs with 125K ETH in TVL. ETHx is integrated across all the DeFi hotspots (Curve, Balancer, Aave, Pendle, Eigenlayer…) and is also live on Ledger, OKX and several other wallets

And from a slippage standpoint, we target a ~4K ETHx swap under 1% slippage (based on current lending market requirements) and are already focused on scaling up the liquidity further

Coming to the weights, we provide the following break-up (expanding along the lines of the 50% threshold suggested by tbrent)

stETH - 50%
rETH - 25%
sfrxETH - 15%
ETHx - 10%

Happy to hear the community feedback and recommendations on the idea. Will incorporate them and draft a detailed proposal for the same

0xSleepy · August 22, 2024, 8:23pm

Stader labs has posted an RFC to add their collateral to address liquidity concerns

MEVCapital · August 31, 2024, 6:42am

Hello, thank you for your proposal @0xSleepy.

The composition of ETH+ is crucial for ensuring the asset’s security and successful growth.

Considering the various discussions on this topic, it would be appropriate to reduce rETH's share in ETH+ in favor of wstETH.

Here is our initial proposal (with more compositions at the end):

50% wstETH
33% sfrxETH
17% rETH

Why is it relevant to reduce rETH's share to 17% in the current basket? Here is our reasoning:

Growth & Expansion issues

rETH frequently trades at a premium. This is due to the protocol's intrinsic workings, which operate with its mini-pool system. This can pose problems, especially when the premium is significant.

Currently, the rETH premium is around 0.15%.

The liquidity for buying rETH could be more optimal: a swap of 4,000 ETH would result in a slippage of 0.46%.

Added to the rETH premium (currently 0.15%).

Due to the slippage and premium of rETH, this asset significantly increases the price impact of minting ETH+. A user who mints 12,000 ETH+ would take over 20 days to recover his price impact loss.

Which is also a clear bottleneck to the growth of ETH+.

Actual ETH+ basket composition price impact

Under the previous composition, it took 21 days to absorb the price impact for a mint of 12,000 ETH (based on the current ETH+ APR)

New ETH+ basket composition price impact

Under the new one, it only takes 10 days (based on the current ETH+ APR)

Redemptions Liquidity

The impact of redemption price is also essential, as it could affect money market risk parameters for any arbitrageur or liquidation bot that would add the redemption as a route.

RocketPool has a liquidity pool of 31,000 ETH, allowing for pegged redemptions. This pool is called for large redemptions via a 1inch route. Interestingly, the redemption price impact for rETH increases and decreases again.

Since 1inch aggregator uses this route, redemptions on this asset have greater liquidity than for wstETH (on substantial swaps). With lower liquidity, the price impact is slightly worse.

Therefore, it is important not to increase the proportion of sfrxETH too much, as this would weaken global liquidity.

Here is how ETH+ would perform during redemptions: