[Discussion] Add additional collateral types

Proposals
1

Hello Rafters,

As we strive to continue our expansion and fortify our position within the DeFi space, it is necessary to explore new avenues of growth, flexibility, and security for Raft. Given the substantial growth and positive reception of Raft in its initial days, it has become evident that our community is ready to support this journey. With this in mind, we propose a significant advancement for our protocol: the expansion of the collateral asset pool used for generating R.

Introduction

To build a more diverse and resilient asset base for minting R, we propose the introduction of four new assets into our collateral pool: WETH, WBTC, sDAI, and rETH. These assets have been carefully selected to provide a balanced mix of decentralization, capital efficiency, and price stability, while also reflecting our commitment to engaging with the broader DeFi ecosystem.

Proposed Collateral Assets

1. WETH and WBTC: These represent two of the most decentralized and censorship-resistant assets, aligning perfectly with our core values. Their integration will not only deepen our engagement with the Ethereum and Bitcoin ecosystems but also fortify the stability of R.

2. sDAI: As a yield-bearing version of DAI that capitalizes on the returns generated by MakerDAO through its RWA investments, sDAI can enhance the capital efficiency of our stablecoin, all while adding another layer of yield-earning potential for Raft users though the DAI Savings Rate (DSR).

3. rETH: RocketPool’s liquid staking token, rETH, carries properties similar to stETH and can further diversify our asset base while accommodating users who prefer to hold and stake their Ethereum via RocketPool.

Benefits

Broadening the scope of collateral assets carries several advantages:

  • Liquidity and Resilience: A diverse asset base can make R more resistant to market fluctuations and provide additional liquidity.

  • User Base Expansion: It increases the appeal of R to users who hold these assets, expanding our user base.

  • Risk Reduction: It reduces the concentration risk associated with single-asset-backed stablecoins.

  • Future Potential: The integration of these assets into our system represents a step towards our broader vision for R as more than a simple LSD-backed stablecoin.

Every collateral onboarded on Raft will have a specific set of parameters that governs its functioning. In particular, the minimum collateralization requirement, the borrowing and redemption fees, and the liquidation rewards, will be decided based on the specific characteristics of the collateral and its interaction with the existing collateral base. The rationale is to ensure that future collateral assets provide a synergistic effect to the Raft protocol without exposing users to market manipulation or arbitrage opportunities.

Next steps

As this proposal marks a significant shift in our strategy, we would highly appreciate the community’s insights and feedback on the matter. We want to ensure we strike the right balance between growth, stability, and risk management while maintaining our commitment to decentralization and community governance.

We encourage you to share your opinions on these assets, any potential risks you foresee, and other assets you believe should be considered. Our objective is to make decisions that are best suited for the long-term success of Raft and the satisfaction of our community.

Thank you for your continued support. We eagerly anticipate your feedback and involvement in shaping the future of Raft.

12 Likes
2

That’s great guys, good job. Personally I see the LSDfi space as overcrowded. I’d love to see more diversification towards new trends. sDAI goes in the right direction but also exploring LP tokens as collaterals would be great (maybe leveraging Uni v4?)

2 Likes
3

Hey - Degentrading here.

I think this is a great proposal in general - adds a ton of utility into R via credit creation into the system and also stabilizes the asset mix via diversification.

On WBTC - i think this is good via diversification and opens up to wbtc holders
On WETH - given the similarity to wsteth - should this be introduced as an option but weth gets converted into wsteth on deposit so yield still gets generated? - no strong opinions on this
On sDAI - Generally good
On rETH - Same theory as WBTC

Of this - I think wbtc probably opens up the field for R the most.

2 Likes
4

I think WETH is a no-brainer and would be a great addition, doesn’t seem to be any downside to this at all.

WBTC does have a little centralization risk, but nothing major - probably similar to stETH. IMO the benefits far outweigh this minimal risk.

I like the sentiment, but I do have some concerns here:

AFAIK, sDAI is not a receipt token; would users deposit DAI, and then Raft auto-deposits into the DSR? How would this impact gas intensivity of redemptions/liquidations?

More importantly, I think this could be too attractive during the early stages of Raft while R has not grown to sufficient levels of sticky liquidity.

Leverage shorting R to go lever-long sDAI is a purely extractive position - there is no directional bias happening here, so people will only do so if they are able to extract value from the system. Unlike stETH, where people are directionally long ETH but just able to be more capital efficient by collecting yield in the meantime.

The only argument that could possibly be made for [long sDAI, short R] being a directional trade would be where users believe R will lose its peg to the downside - and, in that case, the system would want to minimize the size of these positions in order to minimize its potential losses.

Further, I believe that offering the possibility for users to take this position creates indirect issues for the health of Raft’s system: if it’s possible to lever up on the DSR in this way, I posit that Raft’s cost of liquidity increases drastically. [3.5% * maximum leverage subject to collateral ratio] would be the expected APR for running this strategy, and given that R and sDAI are tightly correlated in price, I would expect a low collateral ratio to be chosen, and thus the APR of the strategy to be extremely high. I would therefore expect capital that is both aware of Raft and trusts its smart contracts to choose this strategy ahead of providing R-DAI liquidity. Not only would this put extreme downward pressure on the price of R due to swaps from R to DAI, but it would actively reduce the liquidity available for swaps (both to accomodate this strategy and also ensure that healthy and orderly liquidations of volatile collateral can occur) in a double blow.

Finally, as with all protocols which have a minting fee, there is some debate over whether the revenue really ‘counts’ until the loan is repaid and the position closed - the future demand for R that the fee creates is deferred until the loan is repaid. The majority of loans opened using sDAI collateral would likely never be repaid, certainly would never be liquidated, and therefore I believe they would be a huge value drain of the system.

While it’s a great idea for stimulating expansion, it would also put a huge stress on the system at this stage for all the reasons listed above, and therefore I would be against this addition.

I am curious though, what would the implementation look like here w.r.t. redemptions? Are redemption spreads set individually for collateral or is that a system parameter? i.e. would there be possibility for a 0.25% redemption spread just for sDAI (because of its correlation to R & the Chainlink deviation threshold being much lower), such that in exchange for extracting huge value from the system, sDAI collateral owners also act as a tight peg defence mechanism, and collectively have to self-regulate their behaviour so as to not dump R too aggressively else they’ll just unwind their own positions?

I really like this idea - I see rETH as well-established, important to Ethereum’s decentralization efforts, and underserved in the LSDfi market.

Some general thoughts:

While it is true that a diversification of collateral assets reduces concentration risk, and I definitely want to see Raft supporting a range of undeniably safe collateral types, I believe there is some risk of further pressure on the R peg and its stability, and would like to see R trade closer to $1.00 for an extended period before being too aggressive towards supply expansion. Having said that, I can’t fault the actions to date of the Liquidity Committee at all in the face of current downward pressure to R - every reaction so far has been on point, so this is more precautionary thoughts than anything prescriptive.

Great first proposal, and really looking forward to participating in governance further!

wen token

4 Likes
5

When I first learnt about Raft, my first question was about the risks involved in depending on a single asset, Lido’s staked ETH. Diversifying the collateral with the addition of assets mentioned could strengthen R and make it more resilient to market fluctuations by reducing concentration risk, increasing liquidity, and less dependent on a single asset.

2 Likes
6

+100 to additional collateral types (& rETH mainly)

:fire:_ :fire: approves :DDD

3 Likes
7

Great proposal.

I recommend the team focus on LSDs (particularly rETH) and yield-bearing stables (sDAI) to differentiate from competing protocols in the market.

For sDAI, suggest a conservative mint cap to gauge behavior. @rfv’s concerns above are valid and worth exploring how this dynamic plays out live with small stakes first.

I suggest we postpone wBTC implementation for future proposals. It is not clear how big the demand will be here, and as others mentioned above, wBTC has non-negligible centralization/counterparty risks. I suspect the benefit will still outweigh the risk in the long term, but because R is still in infancy, worth taking the time to assess here.

1 Like
8

How would you use Lp tokens from v4 as collateral? you talking about stable pairs or volatile assets?

1 Like
9

Stable pairs would probably be more suited. By stable pairs I don’t mean only stablecoins, having a stETH / ETH LP token as collateral would be great too in my opinion

10

got your point, what about generating high yields with volatile base assets lps and pay higher pay rates to lenders of the protocol through earnings? if v4 with limit orders solves somehow IL problems wouldn’t it be more profitable for the protocol (with higher collateralization rate)? stable pairs fees and trading volumes cant be comparable in this regard.

11

Gm everyone! Just crunched some numbers about a potential sDAI onboarding as collateral. Find below my thoughts:

  • Minimum CR at 105%: this means that DAI has to depeg lower than $ 0.952 to cause liquidations for Positions with the minimum CR

  • Max R supply mintable with sDAI at $ 15 mln: it would be good to cap it to slightly less than half of the R/DAI TVL in the Balancer Pool

  • Max Position size at 10% of total R supply minted with sDAI

  • Borrowing fee at 1%

This would result in the following:

  • Theoretical maximum leverage at 21x which is simply 1/(CR-1) + 1

  • Theoretical maximum leveraged APR (net of borrowing fee) at 52.5%

  • Theoretical slippage to liquidate a maximum leveraged Position size at 0.44% based on the current R / DAI TVL

Happy to hear what the community thinks about that!

1 Like
12

Thanks for these numbers so quickly, adds some more colour to the sDAI conversation.

I’d like to see this a little higher and then lowered over time. Not because I think DAI is at risk of a significant depeg, but more to tame the leverage at the beginning - especially significant to the issue of competing against our own R/DAI liquidity which has wider implications for system health. 110% would halve the maximum leverage possible, and along with it the APR. We don’t want the strategy APR to dwarf the R/DAI pool’s APR too much such that the system pays significantly more to be short R than to be long R when seeking a delta-neutral farming position - the incentive alignment would be off IMO. It can always be modified later, but offering a stablecoin strategy with scalable 52.5% APR in this environment would all but guarantee aggressive leveragers willing to accept much less than $1 per R sold.

Looks good

The other significant parameter I’d be interested in would be the redemption spread for sDAI - would this be different from other collaterals (if so, what would be the proposed value? I imagine it would need to be extremely small to avoid permanently open debt positions), or is this a global parameter for all collateral types?

Thanks again for such prompt clarifications!

13

Thank you for your comments!

As far as the CR, I think it wouldn’t make too much sense to impose the same CR as a volatile asset like stETH. Also, I think the problem of the leverage is the absolute size, not the relative one. The combination between the 10% maximum Position size and the total cap of R minted through sDAI is what we could use to reduce the absolute size of leveraged positions. If small fish want to 21x ape in with positions that are easily liquidatable, so be it.

As far as redemption spread, it should closely follow what happens on the borrowing spread size, otherwise you’re asking people to pay more to take on debt without being proportionally compensated if their positions are redeemed. Hence, I’d imagine that sDAI would have both a higher borrowing and redemption spread compared to stETH. Don’t know the value but imo it could be around 2%.

14

stETH is 120%, so 110% would be lower.

This is kind of my problem here - the positions wouldn’t be easily liquidatable at all, if I am not mistaken. DAI is extremely unlikely to go down to $0.95. Add to this that the proposed redemption spread is higher than usual, and these positions are likely to stay open forever once opened.

As I mentioned previously, minting fees are irrelevant until they are repaid, so positions that are never closed actually have a net negative value to the system. The intended future demand stimulation for R is never realized.

This is why I would propose a 0.25-0.50% redemption spread for sDAI, if we want to go this route. You can’t give sDAI users the free lunch of scalable 25-52%+ APR, unlikely to ever be liquidated (DAI < 0.95), without some kind of trade-off, else it will be abused to the detriment of R holders. The trade-off here should be that they are providing a 0.995-0.9975 price floor to R, and if they partake in extreme leverage, then R holders will just unwind their positions for them. I think this is crucial for ensuring a much higher rate of these positions eventually being closed, and some churn - otherwise, the protocol’s minting fee is never realized. I understand that you are concerned about sDAI depositors’ compensation for being redeemed against here, but I would emphasize that they are being handed a 25-52% APR strategy during periods of non-redemption (plus at least half of the 0.25% redemption fee!) - I think that is their compensation. Otherwise, the deal is far too good for them, and far too bad for those long R.

I’d be far more in favour of a 0.1% minting fee for sDAI, with a 0.25% redemption spread - this:

  1. ensures that positions DO get closed eventually, which is crucial for R’s peg to be maintained properly (if positions are never closed, increasing the size of the minting fee is pointless, it’s never rebought)
  2. routes more volume through R/DAI as redemptions happen and farmers relever sDAI
  3. sets a limit order to support the R peg at 0.9975, increasing trust in the stablecoin’s ability to act as solid collateral in DAO treasuries etc. (again, sDAI users’ compensation for this is the high APY they are getting during times of non-redemption).

and most importantly, takes away any free lunch for sDAI depositors. It would make the decision for stable farmers much more interesting as it stops no-brainer capital flight from R/DAI liquidity happening. Users will either LP R/DAI for more passive APR, or a slightly more advanced user could go the [long sDAI, short R] route for amplified yield at the cost of possibly churning more minting fees and gas to relever their yield when redemptions do occur. This wouldn’t be as much of an obvious choice, so we wouldn’t see as much reduction in R/DAI liquidity. Further, this strategy would be uniquely appealing for yield aggregators to build out strategies for utilizing keepers and other on-chain automation, which increases synergies and partnerships within the DeFi ecosystem.

Putting some numbers to this, let’s assume the 105% CR originally put forward:

call it 20x leverage for easy numbers, using 0.1% borrow spread & 0.25% redemption spread.

I deposit 100k sDAI, and my position in Raft becomes around 1.995MM sDAI (slippage, trading fees), 1.902MM R debt (borrow spread 0.1%)

I’m earning 0.14% per day on my 100k - which means that I have a payback period of 2 days on fees before I’m in the clear and earning 52% APR. Each time my collateral is redeemed against to act as a 0.9975 price floor, I will just relever again only if I expect R to remain above 0.9975 for more than 2 days.

This approach aligns incentives much more effectively. With it, it only makes sense to use this Raft feature if you believe R is resilient and stays above 0.9975 (non-toxic users), whereas the other approach of high minting fees & high redemption spreads allows users to earn 52% APR by being permanently short R (toxic users), while also contributing significantly less in ecosystem value (less trading volume, and less realized minting fees when positions are closed).

4 Likes
15

You may want to split the sDAI proposal / voting from the rest.

Nearly any cap you put on sDAI will be instantly filled. It will put immense sell pressure on R. If you do this, please ensure you select appropriate fees. Please learn from Abracadabra’s Stargate cauldrons - a popular product but it soaked up any MIM they could mint for nearly a year, the product was so popular that the fee values they chose basically made it a loser for the protocol because they couldn’t release new cauldrons as the sell pressure on MIM was so substantial.

Yes, you’ll grow Raft’s TVL, but you might grow to regret this. Be careful, think long term. Would be happy to discuss on Discord further.

2 Likes
16

One more thought: if you allow ETH collateral - it may be interesting for the protocol to stake it with Lido or Rocket Pool and use the resulting yield to bribe R gauges or otherwise strengthen the protocol. As long as you don’t increase risk for the WETH depositor substantially (and IMO simply staking it with Lido or Rocket Pool does not) then it’s a win/win. You’d still allow users to deposit wstEth directly and keep the yield so that option remains.

17

I agree with this approach in general, but I do think we need to do a more careful analysis of taking yield-bearing stablecoins as collateral before we do decide on any integration.

One alternative approach would be to keep all (or part) of the sDAI yield as the protocol, and treat sDAI as DAI. This would also ensure a tighter peg and more price elasticity without the strong downward pressure. It would also incentivize people to close their loans earlier. I think a more careful analysis is needed on this side.

That being said, I think a more natural extension to the current stETH collateral is the addition of rETH. A proposal has been made to support this: [Proposal] Add rETH collateral

One rETH is added (which lays well within the existing proven model with stETH), we can have more thinking and concrete proposals on other types of collateral, including sDAI.

4 Likes
18

Thinking about it, there aren’t that many compelling ways to borrow against BTC:
https://defillama.com/borrow?borrow=USD_Stables&lend=BTC

There’s Aave, which I think Raft could beat when looking at LTV and ongoing costs (interest rate). Other than that it’s just a bunch of small money markets. There are many places to borrow against WETH, but I think that WBTC might give Raft a unique value proposition.

2 Likes
19

Thanks for putting this together @david, Peter from Origin Protocol here. Building off your rETH proposal, I’ve submitted a proposal to add OETH as R collateral: [Proposal] Add OETH collateral

1 Like