⭕ [RFC] Should we consider adding ETH mAsset (mETH)?

[Dimsome]

Due to popular demand and a very fruitful discussion in our Discord, we would like to bring this idea to the forum for a more formal discussion.

Topic:

Initially, mStable launched with one mAsset (mUSD). Earlier this year, a second mAsset (mBTC) was added with a guarded launch.

Recently the idea of adding a third mAsset was discussed in the discord. That third asset being mETH.

ETH is an interesting asset for various reasons:

• It powers the Ethereum network so everyone using the network has to have some ETH, usually more than is immediately needed.
• It is popular as a store of value and/or is an investment into the future of the Ethereum network.
• More than 20% of the ETH is being utilized in smart contracts, showing that it is popular to use it in DeFi but still has a big untapped userbase and liquidity.
• Holding ETH alone does not generate a return in ETH. Other products that earn a yield exist but either introduce complexity or risk to the user.

Open Questions

• Would the addition of mETH be beneficial for mStable?
• Would mETH add value to potential holders?
• What assets should we consider for the mETH basket?
• What underlying risks do these assets have?

The initial response in the discord was positive. However, we need to allow the whole community to voice their opinion.

This post is just to start a discussion in our forum.
Please let us know what do you think.

[shubidoobi]

Thanks a lot @Dimsome for kicking this off, you’re the hero we need for this

Just wanted to use this to formally tackle the question of ETH derivatives that were considered in the chat and leave a brief blurb on each to kick the discussion off, and narrow down the list to suitable basket & fPool candidates:

alETH = cutting edge (soon the be released) derivative from Alchemix, a definitive consideration for the basket
stETH = Lido’s take on their stakedETH derivative, probably the best for ETH.2.0 right now (very centralized though)
ankrETH = Ankr’s take on staked ETH2.0 and definitely also quite high up the potentials (not sure how centralized)
CRETH2 = C.R.E.A.M. Finance’s version of a staked ETH 2.0 derivative (might be cool to explore more in-depth opportunities with them after the Iron Bank collaboration, what do you think?)
rETH (Ribbon Finance, once they deploy a safer strategy other than their call/put vaults)
bETH (Binance), but for love of decentralization this very much seems more suitable to a feeder pool implementation down the line, as we have already done with the USD version
rETH (Rocket Pool), which is not yet live, but a definitive consideration for the basket
vETH2 (SharedStake), not really familar with this one

I think if we could narrow down the list of candidates that could natively make their way into the basket (not causing any hiccups due to rebasing or other shenanigans), we can move this part forward nicely.

Overall, I think having a mETH derivative makes a lot of sense, due to points mentioned by @Javalasers and @Cold_Summer previously in the Discord. To summarize, these can drive swap volumes, align on easily entering and exiting one system for another, and overall giving a similar experience to our mUSD product.

If you want to own a piece of Ethereum future, mETH is probably going to be your go to product, as you can just hold it, stake it in a pool, or put it to work in our SAVE. As an advanced user, you can use it to swap between different staked ETH 2.0 derivatives, arbitrage easily, as well as have the security you’d expect from our meta asset classes.

I honestly don’t think there is anything else out there on the market today, and we would capture a large audience with this, especially as we’re ramping up to the ETH 2.0 merge later this year.

Once we narrow it down to 3 or 4 mandatory bAssets, I’m more than happy to dive into each one in-depth, and then compare the risks with the rest of the community to ultimately decide on a vote that makes sense on the basket composition.

Really extremely excited for this moving forward, and let’s make history by being the first crafting a metaETH

[Javalasers]

I would recommend starting with alETH, sETH [from Synthetix], and stETH [though this may need to be feeder-pool-only depending on how mStable could handle the constant stETH rebasings].

And what about good ol’ WETH? It’s in a class by itself of course, and I don’t immediately understand the pros/cons of including it in an mETH basket. But certainly has widespread adoption, well-understood properties, and very low peg risk.

As mentioned on discord, I think that interest-bearing assets like rocketpool rETH will not be good fits for bAssets, as by their nature they’re not pegged 1:1 with ETH. Could imagine expanding the mStable model to accommodate this, but seems like it would be a major change. [Nothing against rocketpool, to be clear; but it just doesn’t seem to be “shaped” like a bAsset].

[lonetree]

I fully support the building and release of an mEth meta asset, and I doubt anyone here is going to disagree with the proposal to do so (though I admit I would love to hear a contrarian opinion on that if there is one)

Where I think the discussion will get interesting is how the community and DAO will decide on what the composition of underlying assets should be that prop up mEth. Unlike tokenized BTC, there’s a long list of synthetic or eth derivative tokens that are available to choose from, and they all have different purposes, philosophical ideals, technical considerations, and more. I think we have before us a hefty sifting process, which in my opinion, should start with first collectively deciding upon what are absolute must-haves/have-nots before a token is to even be considered.

Here are some examples that come to mind:

• Maintains an exact 1:1 peg with Eth, or has a reputable track record for not deviating by x% from its peg
• The protocol issuing the synth/deriv has $X amount of TVL (perhaps a valid metric for considering reputation and/or likelihood of not rugging)
• The token has at least X number of other integrations with other projects
• The protocol operates in a (de | centralized) manner
• The token has currently at least $X in volume in aggregate across the ecosystem
• The token does/does not use a rebasing mechanism
• The token is/is not an Eth2.0 staking derivative
• The token is always redeemable for ETH (if not apparent this is different than maintaining peg, as it could maintain peg, but the protocol may not allow 2-way swaps)

And I’m sure there are others that folks can think of.

Point being though, I think it makes sense to decide what the ethos of the basket should be: Is it strictly an eth2 staking basket? Is it a mixture of the top volume synthetic eth-pegged assets? Is it a mixture of what we’ve collectively decided as the “safest” options to create a basket of?

Considering the eth2.0 staking derivatives option alone, there are nearly 6 or 7 protocols with tokens that could be possible contenders, with a few not even having released yet.

[source: defillama]

My submission for the above attributes would be prioritize safety, decentralization, and (perhaps less so due to arb opportunities?) the asset being redeemable for eth at all times.

Either way though, love this idea, and I hope this helps drive the conversation in a constructive way!

[Dimsome]

Really enjoying the conversation and also the more fundamental questions that we raised in the discord!

To rehash the points:

• Staked ETH2 validator tokens are different than other representations of native eth (WETH, sETH, etc.), therefore have not the same risk profile and not suitable to put into the same basket
• Staked ETH2 basket will need a separate pool to swap in and out of this asset.
• Possibility to create 2 meta assets for staked ETH2 and representations of native ETH.
• Basket of staked eth 2 would add a good value proposition of distributed risk to multiple staking providers, but current staking providers are very centralized (e.g. Stakehound disabling transfers of their ERC-20, such a measure would be detrimental to the basket).
• Usability and UX remain a priority to comply with current product logic.

Here is a table of current ETH-like assets:

Scary thing is that stakehound is the second-largest staking provider according to the marketcap and yet disabled transfers of their ETH2 ERC-20 token.

Possible to create a basket of staked ETH2 assets and buy insurance on the basket at the same time?

[shubidoobi]

Thanks a lot for the fantastic recap and summary, this is extremely helpful and brings this discussion up to par.

One other thing we mentioned was that it’s probably a lot easier to integrate the native ETH version before the staked version, simply because the deployment would look very similar to our current portfolio, and not much additonal work might be needed, whereas the staked version would most likely require more fine-tuning (as well as research), so not sure how feasible it is in the short-term right now, especially with the recent Stakehound incident and work being put on v2.

One other super interesting point that was brought up by the Metanauts was that two baskets allow for very flexible and interesting swap opportunities that are currently not possible in the ecosystem in this way, as well as comfortable portability of a set of assets to L2 or sidechains, potentially saving a lot of gas and interactions in the process for users.

Really curious to bring this along, especially as alETH is almost ready from the looks of it, and we could be one of the first to leverage this opportunity.

In regards to the staked derivatives, it feels the next logical step would be to assess and compare them in regards to centralization in order to decide which ones are potential candidates for the basket, and which might be better suited to the feeder pools, or not at all, what do you think?

[james.simpson]

Hey team

Some awesome discussion points here. On a high level I think ETH is becoming the perfect asset for us to create an mAsset for.

The main open question seems to be how to deal with staked varieties and pegged varieties. I’m not sure if this makes sense, but could we just do feeder pools for each staked variety (i.e. stETH / mETH etc.) and have the core mAsset as the 2-4 key pegged versions (WETH, sETH, etc.)? The LP token in the feeder pools may make it easier from a technical perspective. I know Curve have stETH pools so presumably it works. @james.lefrere does that make sense?

Second question is which bAssets to choose. This seems like a derivative question to the above, so would suggest we leave that for consideration until after we have determined how mETH would be built.

Excited to get this launched

[Cold_Summer]

Have mentioned some of this in discord, so sorry for any redundancy, but I wanted to make sure it was in line with this conversation/@james.simpson’s points above.

Staked eth will likely be the primitive/base money of eth defi once eth 2.0 rolls around because it doesn’t make sense not to hold something that’s going to give you 5%+ in risk-free yield. I want make sure we’re positioned ahead of that.

Providing a meta staked eth has a two-fold value proposition: (1) Given the yield will be variable for each staked derivative, people might want exposure to a broad-based basket of staked eth to earn a “composite” staked eth yield that they won’t need to manage (yield will be variable based on validator performance/ability to extract MEV). (2) Because the yields offered by each staking service will be variable, many will likely swap between staking solutions depending on which is offering better yield (or new pools open up with higher APY for a different staked eth than what they’re currently holding, etc.), there is value to be captured in swaps as well.

mstkETH (and a subsequent “i” version) sets us up to provide yield 3x over: (1) native staked eth yield (2) swaps (3) lending idle bAssets on curve and/or convex.

Given the above, I’m not sure there’s ever been a better product-market fit for an mAsset. Let’s focus our efforts on building this early and establish a market position to ride this secular trend.

While regular mETH is great, we need to use our resources as efficiently as possible and also with consideration for how the market will look months down the line when this is actually shipped. Building mstkETH positions us ahead of the trend and its delivery (assuming green light from a technical perspective) will time well with eth 2.0 hype cycle. Furthermore, this asset positions well cross chain, whereas regular eth will be seen as the antiquated asset of the past (given yield opportunity cost mentioned above).

An additional consideration on regular mETH is that the reason sETH/ETH does considerable volume on Curve is due to their cross-asset swap service with Synthetix. We wouldn’t benefit from those tailwinds.

mstkETH is where we can be first and where we can win. mStable was built for this opportunity.

For bAssets for mstkETH, we would need accrual versions, not rebasing. Rebasing wouldn’t work from a technical perspective (to my understanding). Shoutout to @Dimsome for his amazing work on that comparison table. Based on that table and what Curve has onboarded, wstETH, aETH, and rETH seem like leading choices. Once we agree on a pathway, however, we can discuss bAssets more in depth.

[james.simpson]

Great summary there.

Why exactly will the staked eth variations have variable yields? Are there any resources you can point me towards?

Given it is yielding, we need to think of how we make this mAsset. We could of course just have an LP token, but that does away with a core piece of our capital efficiency. We need advice here from @alsco77

[Cold_Summer]

The yields on staked eth variations will be variable based on their ability to extract MEV (i.e. re-order transactions). This will vary for a variety of reasons: (1) ability to collude with other validators to re-order transactions in the most profit-maximizing way (eth 2.0 will allow the current validator to know the next 2 epoch’s validators) (2) MEV can only be extracted by validators who are proposing blocks.

The 45 minute mark to the 55 minute mark on this is a good overview of #2. This also mentions the potential for “meta staked eth” pools to smooth out variability of validator returns (wink wink) MEV Roast | Scaling Ethereum Edition - YouTube.

If you have an hour and a half in the car some day, this will give you a more holistic view from Cobie, who’s helped to found Lido, the leading staked eth service (wstETH). You can hear directly from the horse’s mouth the incentives + game theory at play here, which relate to collusion and MEV extraction optimization I refer to in #1
#24: Ethereum Bull Case with Cobie and Su Zhu by Uncommon Core • A podcast on Anchor.

Our capital efficiency would not be hindered. We could still LP with idle bAssets on Curve (or Convex), which has multiple staked eth pools. As I mentioned in my wall of text (apologies), we could actually earn yield from 3 places: native staked eth yield, swaps on mStable, and LPing on Curve/Convex with idle bAssets.

[james.simpson]

Thanks for this awesome summary and resources Cold Summer. Super helpful.

Separately, I was (blue sky) thinking that we could only accept WETH as our bAsset for mETH which we could then deposit in staking protocols. The other pegged ETHs could be feeder pools, along with the other staked ETH varieties. Could this make sense?

[james.simpson]

Notes about this discussion from the protocolDAO signers call yesterday

• Topic to discuss: mETH
  • staked ETH assets and ETH pegged assets are different in their risk profile
  • Swapping between pegged ETH derivatives using mETH
  • mETH as a bridge asset → mETH with pegged ETH derivatives and feeder pools for staked ETH assets
  • mETH as the value accruing internet bond asset → after the merge of Ethereum 1 and 2?
  • mETH as the staked ETH2 asset itself → partner up with staking providers
  • Next Step:
    • More visibility and research on what pairs are traded now (pairs, volume, TVL) → Commission report via GrantsDAO
    • Forum post for the main talking point to discuss with the community

[penguin]

Hey, I am following up here with some ideas and comments on mETH that I made in Discord. This discussion around mETH (for staked liquid ETH) is largely prompted and inspired by a conversation I had earlier with @Cold_Summer . Looks like he’s also shared many of these ideas in the comments above.

TLDR - we think that an mETH basket, specifically for staked liquid ETH could be huge because it allows MEV (a major source of validator revenue after the merge) to be distributed more evenly.

Intro:
As the Eth2 merge approaches, there will be more and more fragmented staked liquid ETH assets. On top of this, all major staking pools are having discussions and exploring how to properly distribute MEV to liquid stakers.

Problem:
Once the merge happens, APY for staked ETH will increase dramatically. Staked ETH will start capturing value from both from tx fees + MEV (miner-extractactable value). MEV is estimated to earn validators 70% more than network rewards alone and expected to be a major source of validator revenue (and will continue to grow). This is partly because of the way attestations and block proposals will work in Eth2. In short, “assignments are known 2 epochs in advanced for attesters and 1 for proposers”. This means that validators will be able to collude and effectively reorder txs and extract MEV for multiple epochs (32-96 blocks).

What does this mean for mETH?
MEV can only be extracted (and therefore profited) by validators who are proposing new blocks. This means that if you are staking with Lido and they are proposing a lot of new blocks, you will earn significantly higher APY (a lot a lot higher). If, however, most block proposals go to other staking pools, then you will earn almost nothing from MEV.

By creating a basket of staked ETH assets, we can effectively democratize and more equally distribute MEV to liquid stakers.

Again, MEV will be a huge source of revenue after the merge, and mETH could help liquid stakers extract more of it by not having to rely on a single staking pool.

[james.simpson]

Awesome summary @penguin! You’ve explained why mETH is a great asset really clearly. As an update to everyone in this thread, mETH is in the current sprint and we will have an MIP up at the end of the second week of this sprint for your and other MTA governor review

[Dimsome]

Another option to increase APY quite a bit is to use Lido’s new referral program: https://twitter.com/LidoFinance/status/1411963121271455747

The vote finishes soon: Snapshot (most likely it will pass).

With this program, every 1 ETH that is staked would earn 15 LDO:

1 ETH on 5 July 2021: $2214
1 LDO on 5 July 2021: $1.63
15 LDO: $24.45
Which is 1.1% APR instantly.

We could use this very much to our advantage. Depending on the stETH price exchanging WETH for stETH could be more price efficient, or with this referral program, staking directly in case the price does not allow for a price advantage. Either way, win win for Users.

[mZeroNine]

Changed this to a RFC, and with the release of Rocket Pool, it might be a very opportune time to bring this topic up again, and discuss how an asset like this could also serve as the base layer basket for many other mAssets.

[Jeshli]

I do not think that ETH and staked ETH are of the same kind as is the case for USDC/DAI/USDT and wBTC/renBTC/tBTC/bBTC so I am against a basket which includes ETH, rETH, and other forms of staked ETH. I would rather see an mBasket of just staked ETH types and then a feeder pool of mstETH:ETH using the RAI non-fixed peg as an example case for this. I am not sure of the implications for cost effectiveness of such a feeder pool and perhaps the best strategy would be to outsource the pool to Balancer and start with 80% mstETH and 20% ETH until liquidity increased.