eIP 25 Decrease UNI and LINK Collateral Factors

[eIP] Euler Improvement Proposals Voted Proposals
1

Simple Summary

This proposal offers to decrease UNI and LINK collateral factors from 0.3 to 0.

Abstract

This proposal offers to decrease UNI and LINK collateral factors to 0. While these assets still pass the thresholds on smart contract risks, oracle security, decentralization and volatility, the current lack of activity for these assets creates a situation where there is limited upside to maintain their collateral status on Euler.

To echo what @seraphim mentionned in eIP 12 there appears to be limited activity derived from maintaining capital-efficient collateral factors on assets beyond USDC, USDT, DAI, WETH, WSTETH and WBTC.

Motivation

The objective of this proposal is to decrease systemic risks on Euler. We believe that UNI and LINK have not met the expectations in terms of activity generated. We believe that the risks stemming from keeping these assets as collateral outweigh the benefits.

Recent events have once again highlighted the risks of accepting less liquid tokens as collateral. Although the cost of manipulating UNI or LINK would be substantially higher than manipulating a token like Mango leaving these assets as collateral increases the risk profile of Euler with no clear upside for the protocol as these assets generate a minimal amount of fees.

Implementation

Contract Method Token Token Address Updates
governance setAssetConfig LINK 0x514910771af9ca656af840dff83e8264ecf986ca collateralFactor:0
governance setAssetConfig UNI 0x1f9840a85d5af5bf1d1762f925bdaddc4201f984 collateralFactor:0

Liquidation risk assessment

The proposed decrease in collateral factors will lead some accounts to become undercollateralized. We ran an analysis to assess how these changes would impact accounts holding UNI and LINK as collateral. Our analysis aims at determining how many accounts would become at risk of liquidation if we update these parameters.

LINK analysis

Address LINK collateral Account at risk of liquidation - Post update
0xbdfa4f4492dd7b7cf211209c4791af8d52bf5c50 47,316.4 No
0x3f60008dfd0efc03f476d9b489d6c5b13b3ebf2c 1,801.5 No
0x9c0d1f4a029c46265831d120dee9cdc72f0ab3c3 349.3 No
0xd95689c3ffca42f91fd347164c4eb6558ba96d18 12.4 No
0x9f60699ce23f1ab86ec3e095b477ff79d4f409ad 0.9 No
0x0bf93ea5a1fa4ce3dd22c7ffd314462d3869777f 0.0 No

Six accounts currently borrow against their LINK collateral on Euler. The proposed collateral factor update would not lead to any of these account becoming eligible for liquidation.

UNI analysis

Address UNI collateral Account at risk of liquidation - Post update
0x139776871ee95f55d20b10d9ba5a0385451066cd 16,751.68 Yes
0x3f60008dfd0efc03f476d9b489d6c5b13b3ebf2c 1,390.26 No
0xa932e2c7d88497fdee9d87e5a450bae3874ff1a1 1,029.23 Yes
0x68db91d7be097c96a7e4034105bb4861434f7710 178.93 No
0x0003657abb17ede8c28bb40c81d20a6df35c9cb2 12.87 Yes
0x53695fae79a11f1e91aacb73ea37e36a8c09baba 0.02 Yes
0x0bf93ea5a1fa4ce3dd22c7ffd314462d3869777f 0.00 No

Seven accounts currently borrow against their UNI as collateral on Euler. Of these 7 accounts, 3 accounts would become at risk of liquidation due to the collateral factor update. These 3 accounts hold 17,793 UNI ($107K).

We propose to execute the collateral factor update at least 2 weeks after the snapshot vote closes. This delay will give time to affected users to update their positions accordingly.

Conclusion

Given the limited upside from maintaining collateral status and possibility of systemic risk stemming from bad debt, we think this change is appropriate given current market conditions.

If passed, the governance proposal will lead to lower health factors amongst some users, but according to our analysis no significant funds would be pushed into liquidation territory.

Snapshot

Link to voting here:

Euler proposal: eIP 25: Decrease UNI and LINK Collateral Factors (snapshot.org)

4 Likes
2

I fully support this as it hasn’t generated any meaningful activity apart from minting UNI and LINK to receive EUL rewards. The interest paid to the DAO is also a tiny portion of the overall reserves given the liquidity risks.

While I am not against taking risks, it doesn’t seem like the risk taken is remunerated with enough upside.

4 Likes
3

Collateral factor of UNI is 0.3 and its even lower for LINK, doesn’t this factor into ur analysis? Why would an attacker want to choose these assets with such low collateral factor.

Also there is a typo in the UNI table i think u meant UNI not LINK.

4

their liquidity is…ah hope it’s not a vicious circle that more bear less collateral factors

5

Collateral factors are 0.3 for both UNI and LINK and are considered in our analysis.

The current collateral factors make it inefficient for borrowers to use these assets as collateral. Given the low level of activity generated by these assets, we think it is better to lower their collateral factors to 0.

You’re right that it is unlikely an attacker would be able to attack the protocol using UNI or LINK. But there is just no clear upside to keeping them as collateral assets at the moment. As shown in our analysis, almost no one is using them as collateral assets.

6

Thx but i disagree. You cant just say “oh usage is small so lets give up.” Have we even tried to make these assets more competitive? Their collateral factors are so low compared to other apps, just take a look at compound… 0.75 and 0.79 respectively.

7

I’m predisposed against proposals that would prematurely liquidate an honest account without their clear consent. I recognize that using the protocol is tantamount to consenting to any changes that governance implements (up to and including seizure of all funds) but I think governance should aspire to induce zero surprise liquidations due to parameter changes.

I would support this proposal if no accounts were at risk of liquidation post-update. I plan to vote no, but I am open to changing my mind if there is a way (and if folks think it’s worth amending the proposal over) for any accounts that get liquidated through this governance action to be made whole through a transfer of in-kind reserves equivalent to the incurred liquidation penalty. Otherwise, at the very least, I think the notification window should be longer than 2 weeks, perhaps more like 2 months.

9

I think this is a fair comment. I’d be in favour of amending the notification window for a longer one. I think 1 month is probably appropriate. I agree that the protocol should aspire to induce zero surprise liquidations due to parameter changes.

I’m open to the idea of compensating users that will get liquidated as a result of this change if this is what the community prefers. Given the collateral at risk of liquidation and historical liquidation penalties on Euler the cost of compensating affected users, in this case, would be very acceptable. If the DAO wants to go down that road I would advocate repaying only a portion of the liquidation penalty to make sure that users are incentivized to rebalance their positions.

If the DAO wants to proceed with the repayment of liquidation penalties we need to acknowledge that it will set a precedent when updating parameters in the future. If we amend the proposal I think we should make it clear that the decision of compensating users will have to be taken every time we update collateral factors based on the expected cost of doing so.

10

Hey! Just one cosmetic suggestion, would not it be better to change This proposal offers to decrease UNI and LINK collateral factors from 0.3 to 0 to something like to exclude UNI and LINK from the collateral tier? I mean it would look strange to have collaterals with 0 collateral factors. But that’s just a stylistic proposal anyway…

As for the possible consequences for the current users I tend to agree with althecolors that we need a consent of the users affected. Your proposal about compensating users looks like a good way out. It would be interesting to have a look at numbers, how much will the DAO have to reimburse ?

11

I personally disagree. As a DAO we should totally be commercially minded and if there isn’t much activity derived from UNI and LINK while we are taking bad debt risks, we shouldn’t do it. The same thing was done for MATIC when liquidity dried up and we saw literally 0 activity from MATIC as collateral. Funnily enough, the Polygon team were really understanding and understood Euler’s perspective well.

I also DO NOT agree with reimbursing anyone as long as we give them enough time to close their positions. It is their responsibility to monitor governance, same way in TradFi you need to monitor futures rolling and various changes in margin requirements when trading.

1 Like
12

Tbh, the situation with TradFi is slightly different, from the experience I have people working in brokers’ and banks’ call centers are quite insistent trying to reach u via phone, email and even traditional post. So it’s quite difficult to miss something in tradfi

13

The first time UNI and LINK collateral factors were decreased, the affected users were given a pop-up on the UI whenever they’d check their accounts. Sadly it’s all we can do but good enough.

1 Like
14

Fully agree with this point. It lowers UNI and LINK from collateral cross-tier assets.

Are we considering a supply cap for them? Thus it will limit potential negative for the protocol or make any attack unprofitable. But Euler will retain the opportunity to gain a share of long-tail assets

15

I agree with this proposal. If there isn’t much to gain then why take the extra risk? This can always be adjusted as conditions change down the line.

Regarding the accounts at risk, I agree with @allthecolors on the timescale, we should definitely give longer than two weeks. A month is a good length imo.

And I am very much against compensating any forced liquidations from this change. There will be ample time to monitor and adjust positions.

1 Like
16

Totally agree. We will amend the notification window to 1 month when posting the eIP. We are also against compensating users as there will be sufficient time for them to adjust.

17

0 CF effectively means cross-tier, just to be clear. Implementing caps is actually quite tricky, would take time and the dev time is better used on the isolated environment aka “Baby Euler” where users will be able to use anything as collateral w/ limited risk.

It’s important to note that the dynamic isn’t as much dictated by collateral factors so much as by EUL rewards. If UNI and LINK borrowers get EUL, they can still mint at 19x leverage using self-collateralisation. I haven’t seen any other usage for these assets.

18

@seraphim yeah i agree that we should not reimburse users with the liquidation penalty. That sets an absurd precedent for the protocol and users should know that the protocol configs are able to be changed from decentralized governance

And i think its good that we are giving people time but doesnt this put the dao at even more risk? By waiting longer? Are there things we can do in the meantime to reduce risk as we “wait”?

Another thing we can do i think is reduce the liquidation penalty which reduces loss to the liquidated users? And maybe in an extreme case maybe we can set the liquidation penalty to zero and start a DAO liquidation bot to make sure those positions actually get liquidated, although the bot will incur some costs in doing that. Happy to help out here lmk

19

Thanks to everyone for the feedback. We’ll update the proposal to include a 1 month notification window for affected users. We’ll post the snapshot vote for this proposal in 24h.

2 Likes
20

Snapshot vote: Snapshot