Bancor Update — June 29, 2022

This post was originally published on this site

Bancor Update — June 29, 2022

It’s been an extremely challenging time for the Bancor community following recent emergency actions. On June 19th, we were faced with a decision that nearly every DeFi builder dreads: Stand by and do nothing as liquidity providers face potential catastrophe, or intervene and protect users. If the protocol collapses, most users get nothing. If the protocol survives, we can work to restore the system and create the best possible outcome for all users.

Members of the Bancor DAO have been working tirelessly to restore the protocol and develop an improved model with a strong focus on BNT fundamentals and the long-term sustainability of the protocol. It has become clear that the risk of unlimited BNT minting must be removed from the system, and the BNT emissions should be curtailed. Contributors are assessing the solutions now and look forward to presenting them in an open forum to the Bancor DAO. While there are no short-term easy fixes to the issues at hand, those who have been with us a while know we have the brainpower, technical talent and grit in the Bancor community to come back from this set-back.

In this update, we’ll address some commonly asked questions, give a high-level overview of what happened, the current state of the protocol and the path to recovery.

What Happened?

  • During the week of June 12th-18th, a handful of large pending withdrawals showed up in the cooldown contract, creating the perception of a “mass exit” from the protocol.
  • The pending withdrawals coincided with extreme sell pressure of 18+ months worth of accumulated BNT rewards, resulting in a rapid decline in the BNT price relative to the most liquid tokens on the network.
  • This created a large deficit in the protocol, whereby the BNT owed to pending withdrawals rose significantly above the protocol’s holdings and its accrued fees.
  • The amount of BNT distribution needed to cover the pending withdrawals could have produced a runaway effect that would leave most liquidity providers with nothing.
  • Per the Bancor DAO’s intervention policy, BNT distribution was paused, effectively preventing the risk of collapse and protecting user funds.
  • These measures are temporary and pending a DAO vote as technical solutions are worked on, firstly to address the current deficit, and secondly to harden the system with protections against similar events in the future.

State of the Protocol

Currently, a majority of the protocol’s largest token pools are in a state of deficit. This means that if the protocol were to pay users the full value of their deposits all at once (i.e., a mass exit), it would require distributing so much BNT that the token could spiral to zero, and the payouts would ultimately be worthless to recipients.

Pool surpluses or deficits in the protocol continuously change based on the price of BNT. There is an inverse correlation between the BNT price and the size of the deficit: As the BNT price has risen relative to the prices of listed tokens on the network, the deficit has decreased, and as the BNT price has declined relative to listed tokens, the deficit has increased. The more liquidity in a given token on Bancor, the more its price divergence from BNT affects the deficit or surplus.

Withdrawals

Since BNT distribution for withdrawals is currently paused, TKN liquidity providers who withdraw from the protocol now will not receive BNT as part of their withdrawal. In other words, you can decide to hold off on withdrawing until the issues are resolved, or withdraw now with TKN only and zero BNT.

It is important to note that keeping BNT issuance live could have resulted in that BNT being immediately traded for TKN, contributing to most users losing everything.

Any compensation for users who withdrew from the protocol while in a state of pause is ultimately up to the Bancor DAO and can be proposed via the governance process.

As an LP, you can check the amount you will receive prior to executing the withdrawal in app.bancor.network. To make the process easier for those wishing to withdraw, there’s a live proposal in governance to cancel the cooldown and withdrawal fee while BNT distribution is paused.

The Path to Recovery

We deeply appreciate the community’s support during this trying time and the flood of ideas to fix the current issues, some of which can already be found in Bancor’s governance forum. We’ll soon be sharing additional proposed solutions with the Bancor DAO, and collectively deciding as a community the best options for reactivating the protocol.

Further analysis is needed on the root causes and proposed solutions moving forward. They are:

1. Excessive BNT liquidity mining rewards

Proposed Solution: Discontinue BNT rewards

2. Unlimited BNT minting

Proposed Solution: Limit BNT minting

3. Cooldown-induced panic

Proposed solution: Disincentivize users from remaining in a perpetual state of pending withdrawal.

4. Insufficient revenue to cover protocol liabilities

Proposed Solution: Ensure organic fee revenue matches liabilities

1. Excessive BNT Liquidity Mining Rewards

In Bancor V2.1, BNT liquidity mining rewards were introduced to attract new liquidity into Bancor pools and incentivize long-term liquidity provision. In spite of trading fees increasingly generating competitive and sustainable returns for liquidity providers, the rewards programs, as voted on by the DAO, persisted in the largest liquidity pools on Bancor.

The BNT rewards structure did not sufficiently protect against the risk of large players claiming and dumping their accumulated BNT rewards while withdrawing large sums of liquidity from the system in a volatile market situation.

In Bancor’s Impermanent Loss Insurance model, the source of payouts is inflationary BNT. However, there is a counteracting deflationary force, which uses a portion of fees collected from across the network to buy and burn BNT. So long as the rate of inflation is equal to the rate of deflation, then the insurance model is stable (i.e. the insurance pays for itself).

The problem is that all BNT liquidity mining emissions are inflationary. Therefore, the rate of minting for incentives acts in direct opposition to the insurance model. For example, if 1 BNT is minted for insurance and 1 BNT is burned from fees, everything is fine. If during that period, we print an additional 100,000 BNT in rewards — then the insurance model is not working. Liquidity mining is insurance liability; they are the same number. The liquidity mining rewards are — and always were — paid for by the insurance budget.

To eliminate this untenable stress on the IL insurance system, discontinuing BNT liquidity mining rewards has been proposed in Bancor V3. In addition, to further improve the deflationary mechanism of BNT, there’s a live proposal in governance to increase Vortex Burn to 100% for V2.1 & 30% for V3.

2. Unlimited BNT minting

As recent events have shown, unlimited issuance of BNT for withdrawals puts undue strain on the protocol and runs the risk of BNT hyperinflation as a result of mass-exit events.

The obvious answer, then, is to limit the amount of BNT minted for withdrawals from the protocol. While the details of this design are still being worked out, what is clear is that the BNT “printer” needs to be curtailed. Any reimbursement to users should be funded entirely by organic fee revenue — and we are confident we can get there.

3. Cooldown-induced panic

With the advent of Bancor V3, a cooldown contract was introduced which gave visibility into requested withdrawals. This created a perception that everyone in that contract would leave the protocol at the end of the 7-day cooldown period. The reality is there had been many holders who remained in a state of cooldown for weeks without leaving, and who might not have had intentions of leaving at all. However, that signaling of a potential intent to leave appears to have contributed to the induced mass panic.

To address that, conditions and requirements will be proposed to disincentivize LPs staying in a perpetual state of cooldown, or the DAO could remove the cooldown entirely.

4. Insufficient revenue to cover protocol liabilities

Protocol fees should always cover protocol liabilities and in an improved model, we propose it be enforced by the code rather than assumed by the DAO.

The “superfluid” nature of Bancor V3 lays the groundwork for additional revenue streams — outside of just trading — to emerge atop Bancor liquidity pools that can support profitable and reliable fee-earning strategies in DeFi.

Bancor is here to stay

DAO community members are now assessing solutions to address the root causes above, and will soon open up a detailed action plan for a broader discussion with the Bancor DAO.

Bancor has been building in DeFi since 2017. We won’t stop pushing the boundaries in pursuit of a better DeFi experience for token projects and their holders. Recent events were a wake-up call on how risk is handled in the protocol, and the need to improve analytics and threat-modeling among contributors and across the Bancor community. Change is needed, and we deeply appreciate core community members for sticking with us, helping to identify vulnerabilities and working together to push forward.

The DAO governance process will determine how to address impacted users' funds, and the resolution may need to be gradual, and be supported by future revenue. This is crucial to avoiding the runaway effect in BNT that created these issues in the first place. Challenges lie ahead, but these challenges also present an incredible opportunity to strengthen the protocol and drive additional value.

We are not perfect and DeFi is still evolving. Situations like these are a necessary part of that evolutionary process. We will survive this battle testing and emerge from the fire stronger and better.


Bancor Update — June 29, 2022 was originally published in Bancor on Medium, where people are continuing the conversation by highlighting and responding to this story.

Leave a Reply

Your email address will not be published.