GIP-62: Allocate GFI token from treasury for principal loan losses

Proposals
21

We very much appreciate the discussion here. There are a number of good points being raised.

First some clarifications and general notes:

  1. There seems to be a misconception that Warbler “bailed out its friends” with the Stratos situation. We want to be very clear it was actually the opposite. Warbler and Stratos were the only lenders that were not bailed out, and this is verifiable onchain. We lost $6 million on the Stratos deal, and Stratos also lost a significant amount.
  2. There may be a misunderstanding regarding the “moral hazard” argument in the initial post. The “moral hazard” argument was not about lenders. It was about borrowers, as noted in the title of that section, “a major moral hazard with borrowers”. If Borrowers see that the Treasury will just repay all lenders, then they would be more likely to simply not pay.
  3. We acknowledge the oversight in the original GFI post about using GFI for loan defaults. However, that one line about the hypothetical use of GFI to cover loan defaults is the full extent to which that path has been discussed. It’s not in the white paper, nor any community proposals or discussions. That’s even after two default scenarios which already played out without the use of GFI. So I think it’s fair to say that the general expectation was that it would not be used.
  4. Also, Warbler does not need a financial incentive to assist the protocol. We helped on the Tugende situation, reaching a settlement that will recover 50% of the principal value to the protocol, and we received nothing. Warbler also helped with the Stratos situation, at great cost to ourselves, and helped shepard Stratos to pay back $13M in principal to the protocol nearly 18 months early. We have also already been asked to act as agent by some Backers in this LendEast situation, have retained counsel, and are actively exploring recourse options.

Now, more to this proposal at hand. We generally stand behind the following points, specifically,

  • GFI holders were not expecting GFI to be used as default protection.
  • Further, lenders have already received a significant amount of GFI. Lenders have received over $40M worth in today’s prices, which is ~40% of the principal value of all money put into Goldfinch.
  • Any system that tries to cover all losses for an inherently risk-bearing asset class is doomed. That can never work. There should be an acknowledgement that this asset class has inherent risk, and no one can guarantee zero loss.
  • We believe the moral hazard (for borrowers) is very real. The borrowers are all watching this. If we pre-guarantee any kind of loss recovery, we think there’s a strong chance borrowers will just not fully repay.

However, 1.) we think the community should do something to help Backers and LPs, and we are working on such a proposal now; and 2.) we aren’t dismissing the idea of ever using any GFI, but we don’t think this proposal is it, and we don’t think this is the right time. This proposal is specifically proposing way too much, tries to entirely remove risk, and is doing so way too soon before we know enough about this particular case.

Lastly, it’s worth reiterating that ultimately this is, as it always has been, the community’s decision to make. Warbler Labs has never voted on a proposal, and we will not vote on this proposal either. Although we are vehemently opposed to this proposal, if the community votes “yes” on it, we will support to the best of our ability.

22

Hi @blakewest ,

So we have ~10 community members here participating in the discussion, all of them saying Goldfinch x Warbler Labs should take responsibility, including some people sliding their opinion from “Nay” to “Yes”, and yet your second comment here is just a reassertion of the previous one. This is totally incomprehensible.

Your general notes are eluding major points raised by the community here.
=> 1) “There seems to be a misconception that Warbler “bailed out its friends” with the Stratos situation”
This is not the point. The question is why have every default cases been covered by Goldfinch so far and suddenly Lend East doesn’t get that treatment, even though there are documented shortcomings in the underwriting and follow up of the project. Reducing the problem to “bail out its friends” is eluding the main rationals.
=> 2) “There may be a misunderstanding regarding the “moral hazard” argument in the initial post”
There is no misunderstanding. The argument is why would you post emphazises the danger of borrower moral hazard, and all the while dodges the lender moral hazard problem which constitutes an even more existential threat to the protocol (and especially Senior Pool and Backers).
=> 3) “We acknowledge the oversight in the original GFI post about using GFI for loan defaults”
You said it all, it’s in your doc. So it is not “fair to say the general expectation was that it would not be used”
=> 4) “Warbler does not need a financial incentive to assist the protocol”
Also from the doc: “4.4% is allocated to Warbler Labs, a separate organization spun out from the early Goldfinch team that will contribute to the Goldfinch community and broader DeFi ecosystem”. There is the financial incentive. As a matter of fact, Goldfinch and Warbler are pretty much the same team, so it’s a bit farfetched to portray Warbler as a selfless org.

Hence I can’t agree with the conclusions you draw from these remarks. It is also outright malicious to imply that Backers shall be expecting 100% loss on the kind of financial product Goldfinch offers, that is large collateralized loans to established, US based financial institution.

Finally, the main topic here is not an out of pocket “bail out” of Backers and Senior Pool, but a purchase of the IOUs to

  1. ensure Goldfinch has an incentive to recover the lost funds and;
  2. provide working capital to Backers while the case is taken to court.

Again, your new post eludes the main arguments and conclusions drawn after two weeks of back and forth between members.

2 Likes
23

Much has been said and I understand the frustration of people who may lose their money. I still support the idea of goldfinch and clearly recognize that the approach outlined in the proposal will set a negative precedent that could very badly affect the future of the protocol.
As blakewest said, the team is now working on an alternative proposal and I believe there is another way and the team will find it, as throughout the history of Goldfinch the team has tried to consider the interests of the community without sacrificing the protocol. I will wait with interest and hope for a proposal from the team.

24

Hi Blake,

We understand the need for the team to think about the different parameters carefully, but I strongly disagree with your analysis.

  1. “GFI holders were not expecting GFI to be used as default protection” is a bit of a joke when it’s written in black and white in the Goldfinch tokenomics, and using treasury to cover hack/frauds/loss has been a very common endeavor for crypto protocols. It’s pretty much the default system in the crypto industry.

  2. “Any system that tries to cover all losses for an inherently risk-bearing asset class is doomed […] There should be an acknowledgment that this asset class has inherent risk” => The private credit asset class in South East Asia is relatively safe and it’s very, very, unusual to lose a significant share of principal, and 60% loss of principal in unheard for a diversified platform such as LendEast. The main competitor of LendEast in South East Asia, Helicap, which has or is still lending to the same companies as LendEast (i.e SAAVi, Akulaku, Oriente etc.) had ZERO default in 5 years and ZERO loss of principal for LPs. So you may wonder, if both platforms lent to the same companies and share the exact same credit risk, why LendEast finished with 60% loss in principal for Goldfinch vs Helicap not losing a single cent of investor money? It’s basically because (i) LendEast tried to get out too late (ii) didn’t enforce its rights when covenants were breached (iii) committed fraud by diverting Goldfinch money to other LPs. In other words, the LendEast loss is not a matter of economic risk but fraud and negligence of LendEast, and frankly also strong negligence from Warbler that should have been monitored much more closely. In other words, I was fine taking the risk related to the asset class, but I am not willing to bear the consequences of the incompetence (and fraud for LendEast) of two paid layers of allegedly “credit professionals”.

  3. “We believe the moral hazard (for borrowers) is very real. The borrowers are all watching this. If we pre-guarantee any kind of loss recovery, we think there’s a strong chance borrowers will just not fully repay.” I agree that all borrowers are watching this and for now what is coming out as a key message is “Warbler is spinning the narrative that Goldfinch is an entirely decentralized protocol so no one is responsible in case of negligence and fraud, so frankly we can milk the system, ghost and rekt Goldfinch retail lenders, and/or prioritize other LPs in case of issues”. What Warbler needs to do is step in as a responsible adult and buy some IOU from lenders, recognize its shortcomings with LendEast, and show to borrowers that Warbler will be ruthless in court because they have aligned incentive and a lot of money at stake

  4. Lastly, I find your comment about having already distributed too many GFI tokens to lenders quite concerning. You perfectly know that in fact, it’s not USD40m that has been distributed to lenders because today’s price is not representative of the price of past 2 years. On the other hand, the USD60-80m in GFI in the treasury is here at today’s price. It’s also fun that you forget to mention the circa USD130m attributed to the team and Warbler “at today’s price”. I don’t know if it’s ivory tower syndrome or greed from Warbler, but trying to spin a narrative that Goldfinch already gave too much money to lenders that are experiencing massive losses for investing in an asset class that is generally very safe, while Protocol treasury, Team, Warbler, are all sitting on millions of dollars of token appreciation is completely out of touch.

1 Like
25

I understand those who may suffer losses due to this whole situation, but I do not think that using GFI is appropriate in this situation, as this in turn may negatively affect GFI holders.

I believe that backers, providing first-loss capital, should have taken into account all possible risks. Even those risks that lie outside the economic risks. All this should have been considered by the backers initially, and that’s why they got an increased interest rate in comparison with the senior pool.

Therefore, I cannot support this proposal, but I think that other possible options should be considered.

26

Hi everyone,

The general vibe is that the community seems to understand the long-term value of providing support to lenders but the team is not very receptive.

Rejecting support to avoid potential sales of GFI tokens seems to be quite short-sighted frankly. Not only the amounts at stake are small compared to treasury size and if Warbler-appointed legal counsel does a proper job at collecting it will cost nothing to the treasury (treasury will just provide a working cap to lenders while the court process is moving).

I get the moral hazard concerns but frankly, right now Goldfinch seems like a very inefficient intermediary that is bad at monitoring and collecting so the long-term question is really “why would people keep paying a 10% fee on top of the 2% annual GP fee on principal already going to LendEast/Stratos/Alma etc.” if Warbler and/or Goldfinch protocol doesn’t take any responsibility for major fails.

Sorry for asking abruptly the question and you’ll deny it @blakewest @mikesall but after reading the whole discussion I am kinda wondering whether Warbler and team concluded that the current Goldfinch lenders are in any case captive and locked in, and the sole remaining objective is to maintain a proper GFI token price so that team and Warbler can dump their c.$100m of GFI tokens once fully vested.

1 Like
27

I believe your point that “Lenders have received over $40M worth in today’s prices [of GFI]” and implying this helps cushion the principal loss is wrong and extremely disingenuous because:

  1. It is not fair to take the total number of tokens distributed and multiply by today’s price after a 400% appreciation. The correct calculation would be to multiply number of tokens distributed on any given day * close price on that day, and likely would lead to a number 4x smaller, ~$10mm
  2. Much more importantly, these token distributions were NECESSARY to attract stablecoin capital, as the competitive yields on crypto stables were in the 15-25% range. It is common practice for ALL protocols to bootstrap liquidity by distributing tokens to “boost” or “enhance” yields. The senior pool would NOT have been able to attract sufficient capital to close deals at the native 5-8% interest rate, for the clear amount of risk being taken.

Therefore, these existing distributions should in no way be thought of as protecting against principal loss, but just a natural part of the interest rate / yield to attract capital to bootstrap liquidity.

The goal with this proposal was to begin raising USDC capital in the protocol treasury using available GFI tokens to take advantage of the recent price appreciation. The track record of underwriting investments in the Goldfinch protocol will follow Warbler Labs and carry over to the Heron Finance protocol. How can you expect to raise funds for a blind private lending fund (Heron), when your public on-chain underwriting is currently experiencing much higher than industry average loss of principal? The future growth of Heron Finance, and by proxy, the future price of GFI tokens are dependent on successfully protecting principal, even if that means dipping into treasury assets to protect against BLATENT FRAUD by borrowers, since the team was unsuccessful in enforcing any covenants during earlier signs of trouble.

Hiding behind the idea that “Goldfinch is decentralized” and “lenders are interacting directly with borrowers” is ridiculous, because Warbler Labs did all the sourcing and underwriting of the deals that came to the platform.

1 Like
28

To add on your last point of “Hiding behind the idea that “Goldfinch is decentralized” and “lenders are interacting directly with borrowers” is ridiculous, because Warbler Labs did all the sourcing and underwriting of the deals that came to the platform.”, even LendEast never really considered having a direct relationship with Goldfinch lenders. LendEast ignored and ghosted any direct request/question from lenders and only communicated to Warbler, which they identified as the actual counterparty.

So as a lender, I can’t help feeling that there’s a LOT of bad faith from Warbler when I hear that the “GFI has never been meant for coverage” (while tokenomics say it does) , “Warbler and Goldfinch protocol are not responsible, lenders are interacting directly with borrowers” (while even LendEast considered Warbler as its counterparty), or “helping LendEast lenders will doom the protocol and a terrible precedent” (while already 2 pools have been assisted, the cost to the protocol is negligible, and the key long term risk is the credibility of the protocol more than anything else)

29

I am voting “NO” in this proposal. Why should GFI token holders rescue the backers when it dilutes the value significantly? Not taking any emotional argument here.

As far as I remember it was clearly mentioned the onus is on backer to properly conduct due diligence. Also Goldfinch was evolving protocol at a time and risks were shared. “Flight Academy”, a kind of training program with huge incentives, was held to educate the backers and investors. Preventative measures were followed.

As far as decentralization is concerned, I tend to agree with Variant - There is no perfect way to achieve sufficient decentralization. Regulations are not sufficient yet. We know this space is novel in its kind, and I as a GFI holder and backer vote “NO” with the given risk known to me.

As Warbler is the major team contributing to Goldfinch, I would expect decent proposal from them.

30

This proposal seeks to use GFI to cover for a default. There is no such thing as a financial mechanism with rewards but zero risk. Also, this cannot be framed as a one-off event because there was nothing happening this time that is guaranteed not to happen in the future. Whether Warbler did “all the sourcing and underwriting” matters not because it also made no guarantees. What was, in fact, guaranteed was that backers would be the main bearers of loss in case of default and not all GFI holders.

Approving this proposal will place the entire protocol in jeopardy because it sets the terrible precedent that GFI holders pay for decisions they did not make.

I have done extensive research on Goldfinch in an attempt to figure out what are its major risks in disrupting financial markets at large. Warbler has done an amazing job at framing a robust protocol. The only risk factor I could identify is the DAO model: whether the community has the capacity to impart effective governance. This proposal is the first manifestation of governance risk. No doubt a well-intentioned proposal, but ultimately a misguided one. Instead, we should be focusing on long-term security: how do we translate the lessons learned from this mishap into protocol code that will minimize known and controllable risks in the future? What is done is done, lets move on.

I urge others to vote NO or at least delay until Warbler has time to come up with a rational proposal. They have shown a consistent stream of good ideas in the past and I am sure they will again. To the Warbler folks: take your time to get this right.

31

Certainly, the choice is indeed difficult, and I understand the concerns of the Backers community members in the current situation, as well as extend my moral support to them. Nonetheless, I believe that the task at hand in this GIP must be approached pragmatically, focusing solely on the Goldfinch mechanism and not on the interests of any one side.

Therefore, I cannot support this proposal because it seems to shift the financial burden from one group of affected community members to another — specifically, to GFI holders and governance participants. This doesnʼt seem like a balanced or fair solution under the current circumstances.

Since junior tranche participants inherently assume higher risks, which they knowingly accept when making their investment decisions, the idea of transferring the consequences of these risks onto the Treasury also disrupts the logic of the current version of Goldfinch as a credit protocol.

I also donʼt believe that Goldfinch, as a platform neutral to both sides, carries reputational risks in the event of losses by any party. However, as both the Warbler team and our entire community are still taking initial steps in creating a fundamentally different space for lending, seeking compromises for further optimization of the protocol is beneficial. Therefore, Iʼm completely open to considering other GIPs on the subject, but for now my vote is ‘No’.

// No.

32

The rationale is that it seems there is no real economic issue with the LendEast portfolio, beyond a small loss with Oriente. As per Goldfinch feedback, LendEast chose to use at least USD750k of Goldfinch money to pay other LPs.

The default is not due to the backers not doing their DD, but the infrastructure layer (Goldfinch protocol) failing to create a safe structure, doing proper monitoring, and let LendEast perform fraud.

In other words, if you had invested directly with LendEast, you would have most of your money back, but if you invested through Goldfinch and without protocol stepping in, you have lost 100% of your money. The loss is due to poor infrastructure performance, not lack of DD about the assessment of the economic risk. So the infrastructure should step in and take responsibility for its own failure, otherwise no future lender in Goldfinch or Heron will be willing to pay a 10% cut to Goldfinch if the outcome is that they are losing more money than going direct.

33

It looks like there has been a really substantive discussion here, as well as an alternative proposal shared in GIP-63. Since it has been over a week for both proposals, we should start bringing these to some resolution.

For this proposal, @nanojohn , there are a few options:

  1. Keep the proposal as is
  2. Make edits, and depending on how substantive, submit them for vote or further discussion.
  3. Close the proposal (such as in favor of GIP-63)

What would you like to do? Also, since it’s similar to GIP-63, if you keep the proposal open, what do you think of combining it with GIP-63 into one vote with 3 options? (a. no action, b. GIP-62’s approach, c. GIP-63’s approach)

34

Hi @RP2743 , I think that looking at a vote with 3 options as you proposed (a. nothing, b. GIP-62, c. GIP-63) would be best at this point given the similarities between 62 & 63, and if none of the options gets a 50+% share vote, then do a run-off between the top 2 options. Let me know what you think

35

The issue raised in the discussion has already been covered many times from different sides, different points of view have been expressed, various proposals have been put forward, the pros and cons of each. Similar precedents (covering losses from certain sources) have already been considered earlier. And even then, of course, there were fears that possible situations could repeat themselves in the future, which is what is happening now. Some of the proposals already put forward are more short-term oriented, while others seek to consider trade-offs between the interests of investors and the further development of the protocol.

In my opinion, the current situation is in some ways uniquely indicative, and how it is resolved will determine further confidence in the protocol. I believe that Lend East (if there is indeed fraud on their part) should be fully prosecuted. Only in this way will it be possible to show other borrowers that Goldfinch knows how to count money and is very sensitive to each specific case.

As follows from the messages above, Lend East did not provide information about its status for a year (!). How could it happen that we find out about this after the fact? And I would like to know who is responsible for communicating with borrowers and checking their current status (creditworthiness).
It is necessary, if possible, to prevent similar situations from occurring in the future. We need people (or a department) that will deal exclusively with this work for an appropriate salary.
It may be worth allocating some (small) portion of funds from the Treasury for this purpose.

The size of the reserve to cover possible losses and the maximum share of losses covered by this reserve were discussed above. I think that, first of all, a reserve should be provided to pay for the legal costs of prosecuting unscrupulous substitutes who deliberately fail to fulfill their obligations. How this will be done is another question: perhaps it will be GFI in the Goldfinch protocol, perhaps legal costs will be paid by Warbler Labs from some other sources. It is important for the future long-term development and health of the protocol to show that Goldfinch/Warbler Labs can seek justice and a refund.
This issue, it seems to me, is even more important than compensation for creditors’ losses.

If very briefly and very simplified, then how I see the process of getting out of this situation:

0. Mandatory prosecution of unscrupulous borrowers and the allocation of part of the funds for this purpose (sources of financing can be determined later).
and then there are stages, mostly voiced above by other participants, for example @Antoine:
1. Decide whether or not to use GFI tokens to create reserves to cover creditor losses.
in case of a positive vote
2. Determine the size of the pool to compensate for losses. And one-time (this is important, this should be done only once, and then other sources should be identified) to replenish the pool [at the expense of GFI].
3. Determine the share of creditors’ losses that can be covered by the pool.
4. Determine the sources of replenishment of the pool (for example, it could be a small percentage of the profit received on successfully repaid loans).