I would suggest that Gro should be incentivising participants to keep funds on the protocol and reward this by offering a higher restitution of the % of wallet lost, to those that are continuing to ensure the protocols viability.
Thanks for detailing out the proposal. I appreciate the team owning up and being as transparent as you could given what you’ve to work with.
Background on me: I’m a retail investor through Argent and big supporter of zkSync ecosystem projects. I’ve lost more than 50% and still supporting the protocol.
Note to the founder: Treasury should be protected and #Build. You should plan to have a runway of 30 months as we embark on an interesting journey here. You need to watch your burn rate and make sure the protocol doesn’t fail with other decisions. Not saying this lightly, Trim & go lean as you continue to build. If you don’t have conviction and believe you can’t survive then cut your loses and ours.
Making up for the loss will take time and everyone needs to understand and support it. Anyone who doesn’t, should have some incentives as they leave and I presume they left. With that In perspective, my suggestion is to take a hybrid approach of a) use platform token to incentivize users to stay on the platform with vesting b) Debt approach- use the Tx/revenue generated to make the supporters come as close you can to recover their loses. This promotes both reward for staying in the protocol and also gives the protocol time to apply corrective action as you build. I don’t think only one doing part will work.
Logic is sound and agree.
In stepping back to look at the overarching picture, I get the sense, that there’s a lot of distraction here. Investing in crypto assets is highly risky and speculative. The functioning of the cryptocurrency market is outside of anyone´s control. BTC has no intrinsic value.UST didn´t have any intrinsic value either. If it goes to zero, the team takes no blame. Vault users took on the risky portion of a speculative endeavour. Losses for vault users are not outside the intended operation of the protocol. The protocol worked as intended and managed to protect PWRD.
Yes, the protocol worked as intended by protecting PWD users with Vault user funds. The problem is that the protocol took in DAI, USDC, and USDT in Argent and levered up UST with it without a plan to derisk and exit if a very well known potential death spiral happened (which it did). Covering a 3-7% haircut (exiting that position at a 10% depeg with 2x lev) would be much easier to justify as “it worked as it should have” than the 40%+ losses that Vault users are left with. If you are going to lever up something with known risks like UST you have a have an exit plan. Im not going to argue for making Vault users whole as this is defi and there are risks and if every time something bad happens we resort to bailouts we are back right where we are now in our current financial system, but some form of compensation to Vault users as an apology for the poor risk management would seem like a potentially reasonable business decision. It by no means is necessary as “this is how it was designed to work,” but if a 40% haircut on what you thought was USDC,DAI,USDT collateral stables is how it was designed to work, it needs a new design.
Thank you for this post! Looking forward to see the data on the losses. In terms of figuring out a potential compensation plan it would also help to have information on the following:
- DAO treasury and expected yearly expenses of it
- Current GRO team allocation status (already paid out vs. still vested)
- Expected income from the protocol performance fees
I’ll also post here what I’ve already shared on the discord:
Gro Protocol allocates assets across a mix of stablecoins and protocols to maintain a balanced and diversified exposure. This spreads risk so that if any of these fail, only the portion of capital exposed to that particular stablecoin or protocol is at risk.
^This is Gro’s description of the vault. Additionally, there is the hodl tax - so it’s pretty clear that this should be a “deposit and forget” strategy for the user. Given that, it can not be expected that the users regularly check on the chosen strategies.
Originally, I’ve chosen Gro because there was only DAI, USDC and USDT involved. Never would I’ve expected to incurr a 50% loss due to UST failing…
So to me the first mistake of the team was to automatically migrate all vault users to this new, high risk strategy. This should have been an opt-in, so only those comfortable would migrate while inactive users would have stayed in the old vault (and earn less).
Considering my comment above, I do think that this should be considered when thinking about a compensation plan. Users that deposited knowing that they would be exposed to UST accepted the risk for the higher returns.
The second mistake was this:
This is @anon6700304 (from Gro) posting in the forum thread to Vote 9. And then one month later a blog post with the following strategy allocation is published (Gro Renewed — Part 2: Vault & PWRD | by Gro | Gro Protocol | Medium):
So after stating that they will not start with UST or MIM (and for sure not both at the same time), they chose a strategy that consists of ~65% UST and MIM.
The third mistake was not acting faster or having a stop loss strategy for such a scenario.
Given those facts I’d argue that the team should admit their mistakes and participate in a potential compensation by giving up some of their vested GRO.
It’s hard to say what would be a good compensation plan without having all the data. Maybe a mixture of vesting GRO from the treasury and team allocation together with a debt token would not strain the budget of those positions too much?
But more data is definitely needed before making any decision
I can say that I was an Argent user that had looked into GRO and the strategies used for vault prior to the switch and had put a decent amount of my savings into it. I then dropped even more money into after the strategies changed without being aware of the changes in allocations. I’m pretty disappointed to find out how things played out. It seems like a significant change in the risk profile was adopted by the governance process which I could not have even participated in. Other people voted to take on significantly more risk with the money I had put in with potentially very little risk to themselves (it’s unclear to me how much overlap exists between GRO holders and the Vault contributors). The fact that no stop losses existed to prevent this type of failure is doubly disappointing.
Seeing that UST was likely to only continue to decrease in price, I sold as soon as I could. Surprisingly, the price for vault had only gone up since that time, I’m unclear how because I correctly predicted that the price of UST would continue to fall so I’m a bit confused as to how we managed to keep selling and not take on greater losses. I’d appreciate clarity on what’s led to the prices we’ve seen over that period.
Lastly, I’m not currently a vault holder but I believe that if GRO has a long term future it won’t be the expectation that Vault holders are taking on up to 100% losses for 12% returns. It just isn’t sustainable or worth the risk for the reward, I’d rather sit on eth and leave money in a trad bank or money market account. GRO would need to demonstrate that they are learning and improving as they go along and recognize when mistakes are made and take accountability for them when appropriate. In my opinion this is one of those scenarios. A debt token or airdrops or some combination of solutions are merited to demonstrate a serious commitment towards improvement and a sustainable protocol for those that were effected . . . Current vault holders or not (although I think it’s reasonable to say, if you stuck with us we are going to incentive that kind of behavior). I agree that those who were impacted and invested prior to the change in strategies should also be better compensated unless they also voted for the change in strategies. Right now I’ve gota bunch of USDC left over from the sale but I haven’t settled on where to place it, if GRO demonstrates a focus on long term sustainability then I might be convinced to come back but that would probably include some greater degree of assurance that what I invest in isn’t get switched out from under me or some constraints on what kind of strategies could be invested into without me having to take action myself to switch into the new strategies myself (e.g. moving funds into a v2 Vault from the original Vault strategies).
TLDR - DeversiFi donated up to $1000 USDT to all users who were affected by the Anchor collapse, the exchange is continuing to show growth even after this setback.
Similar to Chopper71 I am an argent user via ZkSync, with a 45% loss, I am only a small account in the Vault strategy ~13gvt but this was all of my funds on ZkSync.
All I will add is my experience of one other Ethereum project and there actions following on from the UST depeg.
I had exposure directly to Anchor UST via DeversiFi’s investment directly on their exchange, this was offering near to the 20% return as a solution via the DeversiFi exchange, users could access high yield returns without costly / multiple transactions, bridging etc (similar to the benefits of vault on ZkSync) an ideal DeFi solution for small account.
Most users of aUST on DeversFi were impacted having to sell at a loss, and the protocol was slow to react to the depeg (of their own admission).
The protocol made the decision to donate up to $1000 (capped) to all users affected by the depeg (criteria was holding aUST after a certain date, measured on chain), I personally did not benefit from the donation as I exchanged my aUST to USDT a couple of days before the crash (unfortunately I didn’t do the same for the Vault strategy!).
I appreciate an injection of funds isn’t always the best solution in the long term but I thought the support for small accounts was a nice touch, and it certainly will retain me as a user knowing a protocol will support you if you show them loyalty in the long run.
Echoing the sentiment from vires, we need more data but a mixture from treasury and team allocation seems like a good approach.
GRO would need to demonstrate that they are learning and improving as they go along and recognize when mistakes are made and take accountability for them when appropriate. In my opinion this is one of those scenarios. A debt token or airdrops or some combination of solutions are merited to demonstrate a serious commitment towards improvement and a sustainable protocol for those that were effected . . .
Again, it comes down to the commitment from the core management team at Gro. There are decisions that can be left to the DAO, but ultimately it is the decision of the team on how they will respond to this situation. There were mistakes made before, during, and after the depeg, and those mistakes were made by the executives at Gro, not a smart contract. This was not a code exploit, it was poor communication, risk management, and overall decision making. If the Gro team hopes to earn back the trust of investors, then they need to take accountability for the mistakes made and demonstrate a serious commitment back to those who were impacted.
I experienced a heavy loss in the Vault so thanks for suggesting a compensation plan. I have not withdrawn my vault tokens but I agree with a comment above where the author suggested the current situation seems to be: staying in for a 3.3% return, against the possibility of a heavier loss to protect PWRD in the future.
In the near future, I would like to be able to withdraw the tokens without affecting any compensation entitlements - this may be already the case.
As a user, In terms of compensation, I would like to get a report on how much the compensation would be and how long would take to be paid. I have been involved in previous default scenarios, where I have been receiving small payments for many years, if necessary.
The source of funds for the compensation needs to balance the financial strength of the protocol, the incentives for the team, etc. To make sure the project can be successful and the compensation can be materialised. Probably more information would be required to make a fair assessment here.
A few mention that an automated stop-loss would have been appreciated. In hindsight this would have been a good idea, but it would be good to know whether those would be happy to taking a loss when USDT de-pegged last week; only for USDT to bounce back.
Might be good idea to follow up with a survey that tests the support of taking a haircut when a peg temporarily falls below the tolerance? It might raises a need for being able to vote for asset-specific tolerance levels.
To leverage or not to leverage
Since this is the second time a selected strategy has failed and Vault users have felt unfairly hit, I wonder if: either the risk of entering the leveraged Vault aren’t clear enough PWRD successfully weathered the storm, though I am left wondering if the overall net outcome from the ordeal for Gro Protocol is negative. It raises an existential question: does the leveraged component of PWRD cause more harm than good?
A few Vault users mentioned that they felt powerless to being able to withdraw, for some users this is a feature and others a bug. Either: we could consider introducing a 3rd product that utilises the same strategies, but does not use leverage: this would offer users more control, and at the same time make the decision to use leverage more of a conscious one.
Alternately consider dropping PWRD / considering a different strategy of protection.
Set and forget, manual opt in, or better comms?
A few Vault users have mentioned being surprised that strategies change over time. I personally think the website is clear that funds are deployed in a “continuously optimised portfolio of strategies”. But it does make me wonder whether there should be an explicit opt in to accept changes in strategies.
A point was raised by @keyneom that GVT holders should be able to vote in governance: currently GVT holders can earn GRO by depositing their position into the single-sided GVT pool. The point of GRO token is to govern the protocol, and letting someone join in governance by temporarily depositing a large position could open up the protocol to attack… But maybe there is some room for per-pool tolerances to be set by GVT and PWRD holders; but not overall strategies? To me it seems the real issue here is holders not having sufficient awareness about strategies or strategy changes. I think opting into receiving pool-specific emails could help here, but maybe there are other ways to push out this info?
Insurance and/or increased decentralisation
Random left-field thought, when voting for strategies - I wonder if we should consider allowing users to stake in return for a percentage of revenue from that strategy? And the stake acts as a partial insurance fund if the strategy were to fail Where people stake with their GVT. Whichever direction we go, I think it would benefit GRO to let users have more nuanced control in the form of gauges e.g. to vote to pull from a strategy early. (May need to consider a malicious larger GVT holder doesn’t deliberately trigger an early exit to cause panic a less-liquid market)
I don’t wish to dwell on this matter longer than necessary, but I do feel like it is worth spelling out the grievances incurred as a result of the events and decisions of two weeks ago.
We incurred a $500k+ loss for the following reasons:
- The Gro team chose to increase exposure to UST after we invested, but the front end does not adequately display what strategies are in play, or the total potential exposure to each strategy. The numbers displayed and “safety” assessments are arguably misleading. It wasnt until the week before the depegging that I noticed just how exposed we were, and even then it wasn’t clear by how much.
- Upon realizing that the exposure was in fact quite high, I attempted to withdraw our funds but initially was unsuccessful, as the Gro frontend simply refused to allow it or explain why. After contacting the team, they eventually realized it was a slippage issue due to the size of the withdrawals. This front-end issue which failed to inform the user for the reasons for transaction failure, and failure to offer workarounds without team support, cost us 4 days of withdrawal time that would turn out to be crucial, as by the time I had received the information, it was the 9th of May.
- Just as UST started depegging, with this new found slippage knowledge, I was able to take out 1/3rd of our Vault exposure over a dozen smaller withdrawals, but slippage limits prevented me making any further withdrawals. The protocol design choices made by the Gro team blocked us from being able to make an informed decisions about our liquidity, which locked in ~$1m of Vault exposure whether we wanted it or not.
- Early on in the depeg, Despite incurring only a 6% loss, key decision makers on the Gro team refused to cut losses, despite the fact that the downside risk was much higher than the realised losses. The inaction of the team turned a recoverable situation (If a 6% loss had been wrongly incurred, governance actions could easily correct the loss) into a situation which could now very well kill the protocol. It’s one thing to say that the intention of the protocol is to be decentralised, but then for users to be blocked from making their own decisions and though inaction driven by blind faith, gambling with the money of investors, and no less betting on on the non-failure of the largest ponzi scheme seen in crypto’s history, that was imploding in front of everyone’s eyes in real time, is quite disappointing.
- Eventually the Gro team decided inaction was no longer an option, and ran harvests and started selling UST to help people pull their funds. This proved that not only could the Gro team have taken such actions 2 days prior and prevented a run on the vaults altogether, but that the risk management strategy employed by the team was, as has been identified in this thread, non-existent. I’m thankful that the decision was made to sell UST and pull funds eventually, but it didn’t prevent the $500k of losses actually incurred.
These 5 reasons I believe fully explain what happened. The underlying protocol itself is a solid concept and the PWRD pool worked as intended. Unfortunately it was let down by slow and poor decision making. Decentralised or not, the expectation exists that if you are the steward of a protocol and it’s users, you take matters into your own hands if you have the power to do so, and the situation demands it. The LPs, GRO holders and DAO members have all implicitly entrusted the Gro team to take charge in situations like this, and to make decisions around strategies that look after the long term interests of the protocol.
In terms of the potential ramifications, it is very unlikely that any users would want to pursue recourse outside of this forum if genuine attempts at reimbursing or covering the losses are offered. It is my recommendation that all of these above issues are addressed by the team:
- Rework the frontend to display the strategies and allocations in play with links to further clarification, rather than displaying “exposure” percentage figures with a very ambiguous meaning and safety scores which mean next to nothing.
- Improve the withdrawal flow such that larger withdrawals are not totally blocked from occurring (the option to bypass slippage), with better error handling and messages to users.
- Related to the above, build an option to allow vault holders to pull out their prorated share of the underlying strategies directly to their wallet so that slippage and such can be handled by individual liqudity providers in these situations.
- A commitment to avoiding high-risk yield strategies. Many very smart people were fooled by Anchor, and many very prominent people in crypto shilled the scheme with the full belief that it was a legitimate stablecoin. As stewards of a yield aggregation protocol, having a leveraged pool adds additional risk, and so the strategies used to generate the yield need to be risk adjusted accordingly. A commitment to avoid all ponzi-like strategies where the underlying collateral is at risk should be made to encourage the return of users, and an acknowledgement from the team that UST was a poor decision from the get go would help build back the trust of now-former liquidity providers.
- Design and enact an emergency risk management policy which is pre-approved by the DAO to force the team to take descisive action long before double digit percentage losses have occurred, whether or not the team believes in the strategy.
Further, some form of compensation is nescessary. There are some ideas in this thread already but a legitimate attempt at making vault holders (at least mostly) whole would be strongly advised, given:
- The losses incurred stem from decisions made by the team, not by the protocol
- Users need to know that the Gro team intends to act in good faith in order to stake their funds in the protocol
- Good faith is achieved when despite the hardship and the suffering price of GRO, a commitment is made to repair the damage and support users in spite of the immediate price impacts (which have already been significant), and most importantly, publicly own the mistakes that have been made.
- Long term, if executed correctly, a loss recovery plan would instill faith in myself and other protocol users, as well as new users looking back on this history, that this was simply a moment to battle-test the Gro protocol and uncover issues which, now resolved, make Gro a much stronger offering for long-term yield seeking users.
In regards to compensation, my recommendation is that a combination of strategies be employed to cover the losses over time. No one expects compensation to happen overnight, and rebuilding Gro to where it was just a few weeks ago will also take a while. It will be a difficult 6 months, but I have faith that Gro can come out of this on the other side with a renewed sense of purpose and ultimately a much stronger product.
If the end result of this plan is that despite potentially the worst thing that can happen to a leveraged stable-coin pool actually happening, the team was able to recover the funds, identify and admit to the mistakes, strengthen the protocol, and make users whole again, then that is an extremely compelling case to use Gro again in the future.
This is a good summary. Investors accepted some risk but the practical operation of the vault was not ideal.
The UST depeg was not the same as stress on the USDt peg. You could build in a stop loss threshold. The huge volatility of UST/USD gave ample opportunities to cut the position after a stop loss had been triggered.
If the worst scenario for a leveraged Gro vault following an asset going to zero is - 49% and the lessons are learnt then the position for Gro is recoverable.
Argent will have all the loss data. I have USDc following sale and I don’t expect Gro to cover the loss in USDC upfront. If you make a sensible offer in vested Gro token then you will probably get buy in whilst avoiding the upfront hit to Gro.
one way to alleviate the impact is to distribute GRO tokens through the existing mechanism of the 12-month vesting contract.
Agree this was my back of an envelope calculation
I disagree that we can make assumption based on position size if the loss was meaningful for user or not. We definitely should go with linear approach.
Hi @vires-in-numeris - here’re some data points. I’ve also tried to put in reference links so it’s easier to contextualise the numbers as well. Hope that helps!
DAO treasury and expected yearly expenses
The DAO currently has ~$13.7mn in stablecoins or stablecoin products (and more in other tokens). There’s an additional 8mn GRO tokens that have not been minted that belong to the DAO treasury beyond the 5mn GRO tokens minted for the LBP - see more details in this article and vote. You can view the full treasury position here.
On expenses, the committed ongoing spend is at up to 1.8mn USDC (Vote 12), 96 ETH & 1920 AVAX (Vote 11), and 18,000 vesting GRO (Vote 10). Marketing run-rate has yet to be voted in but you can see the additional figures here. Additionally there was also one-off commitment in Vote 3 on security, Vote 13 on Votium, and Vote 14 on Votium.
Bringing the data above together, the DAO treasury has >4 years of runway after accounting for the above one-off and ongoing spend.
Current GRO team allocation status (already paid out vs. still vested)
None of the allocation has been paid out. Vesting started on 28th September 2021 and is on 1-year cliff with 3-year vesting. Team members joining after 28th September 2021 is also subject to the cliff and vesting; their vesting start date corresponds to when they join.
Expected income from the protocol performance fees
This is the uncertain part as the performance fees is 5% of yield earned on user funds in Vault & PWRD (10% for Labs). Out of this sum, 90% of that goes into the vesting bonus pool and the remaining 10% goes into the DAO treasury. This means performance fees are highly correlated to protocol TVL that had seen a great drop with UST depeg.
Another source of income for the DAO are yields earned on treasury; so far funds have only been deployed to Aave, Compound, and Gro so the yield level ranges from 1-5% on stablecoins. We could also put funds into ETH and stake with Lido or use other protocols / assets subject to the whitelist from Vote 10 or further DAO votes.
What about a linear approach with a max. cap? That would mean it’s linear for vast majority of users but still helping smaller wallets relatively speaking.
I would vote equally distributed compensation so that users get same percentage of their position. So no max cap. In the end loss was also same for everyone percentage wise, so why wouldn’t compensation be? There wasn’t “max cap” for loss either.
Since there is a lot of more small investors, I understand that you get more pressure to make more compensation for small positions (percentage wise) but still you don’t want to provoke pro investors / large positions. So try to make it as fair as possible for everyone.