[Signal Request] Offboarding MATIC, COMP, AAVE, and BAL

I’d like to poll the sentiment from the community to begin the offboarding process for MATIC, COMP, AAVE, and BAL.

The MATIC token is fairly clear cut. There was a deal in place with Polygon to on-board the MATIC token in exchange for them monetizing a significant part of their treasury. Discussions ranged from 30M Dai all the way to 100M Dai. Polygon did not hold up their end of the bargain, and after being confronted both digitally and in-person recently in ETH Lisbon gave no indication they have any intention of doing so. The community should off-board MATIC on principle for reneging on the deal irrespective of the fact that the vault has been performing extremely poorly (3,627 annual stability fees).

The COMP, AAVE, and BAL vaults have very attractive rates (1% SF) compared to competitor platforms and have had ample time to build up a user-base. However, no usage has materialized. The stability fees generated are quite poor (see below) and don’t come close to covering the Oracle gas-costs. As such, I’m recommending the community offboard the COMP, AAVE, and BAL vaults.

Token Dai Minted Annual Stability Fees
COMP 13,722 137
BAL 171,470 1,715
AAVE 1,065,355 10,654
Offboard MATIC
  • Yes
  • No
  • Abstain

0 voters

Offboard COMP
  • Yes
  • No
  • Abstain

0 voters

Offboard AAVE
  • Yes
  • No
  • Abstain

0 voters

Offboard BAL
  • Yes
  • No
  • Abstain

0 voters

EDIT 10/30/21
The polls will close on 11/6/21

10 Likes

Would be nice to have a hard and fast offboarding rule. Sth like, if a token has been around for X amount of time and generated < Y in revenue, automatically start the offboarding process.

3 Likes

@NikKunkel can you recall here the total operational cost required for running each one of those Vaults (and therefore to break-even)? That would probably enlighten the discussion.

4 Likes

It is quite variable due to gas price fluctuations as well as the type of Oracle used (Medianizer/OSM vs UNI LP Oracle vs DSValue). All of these collateral types utilize the Medianizer/OSM configuration and have historically cost around $150,000 per year per Oracle.

6 Likes

@NikKunkel Maybe like an annualised mean cost by oracle type (with some assumptions) would do the job. I think that can help all teams, including RWA, think how to evaluate onboardings and min utilisation rates required on the opex side.

3 Likes

The problem of trust that MATIC will generate from now on when listing new collateral is very serious.

I think it is important to note that it would be interesting to see the history and everything that is taken into account when incorporating new collateral.

If those behind the collateral wish to make an agreement as or signal intentions, block the capital for it, we are not school children to be believing in false promises.

1 Like

Dumb question: But how are we paying for gas costs for oracles now? Is the Foundation still picking those up? I’m fairly certain it doesn’t come out of the current Oracles CU budget.

2 Likes

For my edification, did MakerDAO and Polygon formalize this agreement anywhere, besides what was written on the forums in this topic?

Off-boarding collateral with trust issue such as MATIC is fine.

Compound low usage, ok

But Aave has 1M DAI generated, perhaps it just needs time to get more users to on-board. DeFi is still very new to majority of people. Off boarding this collateral will push people away from Maker and go to other platforms, and once people go, it’s hard to come back. I incline to have more users as a priority. Same goes with BAL.

3 Likes

Thanks for the post @NikKunkel I am pleased that we are trimming some of the bloat from the system.

Currently 4 ilks - ETH-A, ETH-B, ETH-C and WBTC-A - generate 94.6% of our stability fees per makerburn. Perhaps we should look at this and think about whether WBTC-B and WBTC-C could generate some decent profit too without adding too much expense before adding new ilks that would require new oracles and other costs.

I support offboarding of these assets. COMP an AAVE users are normally using their native platforms to borrow, despite Maker offering lower rates. BAL users also prefer to farm on Balancer.

Truth be told, there is not a lot of demand for other vanilla ERC20 tokens used as collateral, plus we have few limiting factors:

  • either they are very illiquid on-chain and we can not offer high debt ceilings
  • those who are liquid and interesting are some recently hyped and heavily farmed tokens that are very risky on its own and haven’t really went through a stress test (OHM for instance)
  • Maker doesn’t offer cross collateralization, which hurts users if they need to maintain every single vault. It costs >$400 to maintain your vault right now (withdraw collateral and repay).

It is probably only LINK, UNI and YFI where we truly saw some serious demand in the past.

I think Maker should mainly focus on three categories when it comes to onboarding new crypto assets:

  • ETH and WBTC derivatives (different LR, fixed rates, stETH type of products, etc.)
  • LPs (again mostly ETH and WBTC based)
  • D3M type of modules

D3M type of modules are particularly interesting in my opinion. It is the easiest way to scale, at some cases we get other protocol’s backstop and offboarding is easiest. At Rari Capital Maker could even create a pool with its own set of rules. Of course there are also downsides to it, one is that we need to monitor them closely all the time and we start increasing Chainlink oracle dependancy.

16 Likes

Maker Community,

We stand by what we had committed. We initiated the process for it a long time ago. The challenges since then were mainly around chasing the multi-sig signers given that there are a series of transactions that require to be initiated to open a vault. During this time, we have always been in touch with the Maker team; now finally we got the first step done by doing a test transfer of depositing 100,000 MATIC, 16 hours ago.

In the next few hours, we will execute the full amount transaction given the test txn is successful.

Polygon is long-term committed to supporting Maker, and DAI. There are parallel conversations from the Maker team happening with different members of the Polygon team for the growth and support of DAI, and Maker across the ecosystem. We would request the community to consider this vote because DAI liquidity is dependent on MATIC.

Still apologies for the delay, and thank you to the Maker team for supporting us at every step.

14 Likes

It looks like they are up against the temporary DC. Let’s see how it goes.

It would also be great to get a breakdown of how we pay for oracle gas right now + some specifics on what the current oracle gas costs look like,

Thanks, everyone!

1 Like

That’s why my vote was NO for MATIC’s Offboard, even though this act will set a precedent when it comes to making “deals”, it doesn’t cross my mind that Polygon would want to damage their reputation with a bad action with MakerDAO.

Thank you, @sourcex, for your willingness and clarification.

3 Likes

This should be enough not to offboard Matic.

Is sad to have to offboard collaterals. Are there plans to lower the costs of running oracles?

5 Likes

Constantly. The previous version of the Oracles (V1) would have cost about $1.5M per Oracle per year…
Rather, I think the the solution is more getting sharded vaults on L2 online b/c Oracle costs will only be a fraction of what they are on L1.

5 Likes

Yeah that sounds great :+1:

I second this. Given how much oracle gas costs seem to be informing the decisions regarding onboarding/offboarding collateral types, I think whatever clarity can be provided here (even if it’s a range) would be helpful.

I assumed the costs were covered by the Oracles CU budget…is that correct?

1 Like

I’ve mentioned this in the Governance & Risk call multiple times. Oracle Gas costs are currently not stipulated in the Oracle Core Unit budget. Although there is plenty of surplus OCU funds to do so, I don’t think it’s right to spend funds on items governance hasn’t specifically allocated budget for. I’ve been writing a MIP I hope to publish soon which will request funding from governance specifically for Oracle Gas Costs.

Right now there is a gentleman’s agreement between the OCU and the Foundation. The OCU has taken over responsibility for the costs as of August 1st, however the remnants of the Foundation are fronting these costs until the OCU can get funding from governance to retroactively pay them back. This is a relationship we hope to dissolve before the end of the year as it’s one of the last remaining items blocking the Foundation shutdown.

2 Likes

Can you ballpark how much that has been from Aug 1 to now (or end of September or some other milestone)? Is there a reason why we need the Foundation to continue to bear these costs?

1 Like