Recovery Mode removal

Overview

The removal of recovery mode has been discussed extensively throughout the past month, with numerous users voicing their discontent with the potential of their positions being at risk of liquidation by factors outside of their control. We first clarified our intention to remove Recovery Mode in the March Community Forum.

In summary, we are in process of removing Recovery Mode, as we believe that, moving forward, this is the best route in terms of user experience. Removing Recovery Mode will entail changing the Maximum LTV from 90.9% to 75% on all assets.

Hubble is a borrowing protocol first and foremost. Recovery mode leads to a potentially unpredictable borrowing experience, and thus defeats one of the primary goals of the platform: providing a predictable borrowing experience.


Background

Recovery Mode was inspired by Liquity, an Ethereum Protocol that exclusively accepts ETH as collateral. Liquity was the main inspiration behind the creation of Hubble Protocol. Hubble’s Maximum LTV (Liquidation Ratio) of 90.9% and Recovery Mode were both mirrored from the Liquity model. However, Hubble is fundamentally different to Liquity.

Hubble sees itself as a flexible, community-focused project. We want the freedom to react to changes in the market and build continuously to expand the features and functionality of the protocol.

Summary of recovery mode

As mentioned above, Hubble’s Liquidation Ratio is 90.9% across all collateral assets. However, when the system-wide LTV reaches 66.6%, Recovery Mode kicks in. When Recovery Mode activates, liquidation on loans with an LTV above 66.6% can be triggered, starting from those with the highest LTV. Liquidations continue on a loan-to-loan basis until the system LTV is once again below 66.6%.

Recovery Mode ensures that USDH is always 150% backed by crypto-assets, as a system-wide 66.6% LTV ensures that at least $1.50 of crypto is deposited for 1 USDH at all times. For example, at the time of writing, the System LTV is ~50%, which translates to USDH being ~200% crypto-backed.

Recovery Mode was necessary to ensure that system LTV doesn’t reach as high as 90.9%, as this would essentially bring the protocol within 10% of insolvency.


Motivation for Changes

In the Liquity model, Recovery Mode makes sense, as it means that the stability of the protocol is automated with the liquidation mechanism of Recovery Mode. Moreover, because Liquity accepts only ETH as collateral, a single LTV is a sensible approach.

In Hubble’s case, multiple collateral types entail different levels of risk, and having a 90.9% LTV for all assets regardless of risk, with Recovery mode to counteract it, is not in the best interest of borrowers or the protocol.

Furthermore, because Recovery Mode is triggered by the system LTV ratio, positions can be liquidated as result of other positions being opened.

Example

  • Protocol TVL (Collateral) = $50M
  • Total borrowed USDH = 31M
  • System LTV = 62%.
  • Current Borrower has a 10x leverage on a position with an 85% LTV. They closely manage their position, repaying USDH or adding collateral as needed to keep their position safe.
  • New Borrower is a whale. They deposit $15M in collateral, and mint 12.45M USDH. They have an 83% LTV ratio.
  • Protocol TVL is now $65M
  • Total USDH borrowed = 43.45M
  • System LTV = 66.8%

If a whale borrows millions of USDH at an extremely risky LTV, this could have an adverse impact on the system-wide LTV, and users who had managed their position closely, but with a high LTV, can then be liquidated by a factor outside of their control. We believe that this is an undesirable outcome.


Proposed Changes

Recovery Mode has never been activated, as users have generally borrowed in responsible manner. However, according to feedback, Recovery Mode is undesirable. In response to this, we are planning to remove Recovery Mode entirely. To replace Recovery Mode, Hubble will implement a series of changes, starting with:

  1. A protocol-wide 75% LTV ratio.
    This will be implemented in the first phase of the transition away from Recovery Mode. All new loans will have to be taken with an LTV lower than 75%, and liquidations will occur if a loan reaches 75% LTV
    Eventually this will transition into:

  2. Per-asset LTVs, where different assets have different maximum allowed LTV ratio’s. In the case of a multi-asset loan, the LTVs of the different assets will be aggregated to calculate a local liquidation ratio (maximum LTV).

An in-depth look into this second phase is beyond the scope of this discussion, and it will be raised in future discussions once Recovery Mode has been phased out and the protocol-wide 75% LTV has been implemented for a sufficient period.


Impact of Changes

Pros:

  • No risk of other users impacting your risk of liquidation
  • Potential for risk-based LTV ratios
  • Higher system-LTV
  • More predictable loan experience

Cons:

  • Lower maximum-LTV on individual loans

With system LTV now shifting to 75%, new loans can only borrow USDH at a maximum 75% value of deposited collateral. This will reduce the flexibility in terms of borrowing, but will serve for an overall safer experience.

High-risk borrowers, of which there are relatively few on the platform, will be most affected. At the time of writing, there are 561 loans on the platform, and 23 of them have an LTV higher than 70%: View liquidation page

Impact on Current Loans

  • Loans that are active and above 75% LTV when the changes are implemented will not be liquidated.
  • All users will be prompted to acknowledge the change.
  • All actions taken on loans above 75% will have to be directed at bringing LTV below 75%
    For example:
    – Adding collateral
    – Repaying a portion of USDH

Conclusion

To summarise, the changes are as follows:

  • Remove recovery mode
  • No maximum system-LTV of 66.6%
  • No maximum loan LTV of 90.9%
  • Maximum-LTV changed to 75% across all assets & loans
  • System LTV moves fro 66.6% to 75%
  • USDH backed by minimum 133% crypto-assets at all times
  • Current loans above 75% will not be liquidated
  • Users will be clearly notified of the changes when entering/using the app
  • All actions on current loans will have to bring LTV below 75%

From Hubble’s point of view, the removal of Recovery Mode will increase the user experience with more predictable loans.

1 Like

Most lending platforms offer 80% TVL and the lure of being able to go to 91% may be a marketable feature.

I suggest in addition to these changes that the stability USDH be used as a means for a user to hold a higher then 75% TVL. If that user gets liquidated there would be USDH that could be used to pay back the loan. This would encourage more USDH held in the system creating a more stable price.

3 Likes

The plan to remove Recovery Mode - its possible effects, and the article itself is very well thought out. I can sense that this is planned very well and have undergone careful considerations since ultimately, the goal is to have a stable and predictable platform for defi users.

I would agree with Jesse that lowering the TVL kinda sucks. It’s no longer capital efficient for me to do this, especially that I tend to degen with DeFi protocols.

His suggestion about a user having an allotted USDH on the stability pool should allow him higher than 75% LTV, sounds like a good idea.

On top of that, I personally would recommend having an oracle built around identifying current market sentiment so we’ll be flexible with our LTV. If everyone’s bullish (identified thru that oracle) people tend to degen and maximize their capital efficiency. Maybe we could raise LTV above 75% if that’s the case and lower it if everyone’s bearish. At least this way, we’re dynamic and can cater to everyone’s interests.

1 Like

Considering the stability pool deposits for users is a great idea, but still an optimization. Potentially, most users will not participate in the stability pool and so we need to decide on the LTV for them.

Capital efficiency is great but should not endanger the protocol solvency. If we allow 90% for everyone then the protocol is 10% away from being insolvent and, even if you had capital efficiency for a few months, you won’t be able to use USDH for much longer because people will have lost trust in USDH.

The reason you want trust in USDH is for people to confidently LP it such that it allows borrowers to go in and out of USDH if need be (for leverage or for repaying their debt).

As such, liquidation LTV needs to be good enough that it doesn’t endanger the protocol. The best way going forward for this, is to create multiple types of vaults with different debt ceilings, to allow degens to go 89% LTV and while keeping a cap on how much USDH is only backed by 110% collateral. Given that a 90% LTV for SOL is risky, you’d have to pay a premium on this in the form of fees to compensate the people that hold USDH and are taking on your risk.

1 Like

Following a similar vein of allowing a user greater flexibility with their LTV ratio, you could consider a tiered HBB or USDH loyalty system where the amount staked in the SP or HBB staking would determine the users LTV. If you make these tiers so that to have say a 90% LTV the user would have to hold a substantial amount of HBB or USDH staked this creates a skin the game scenario where the users are encouraged to act responsibly to protect their stakes. This could be particularly effective with HBB as the tokens price is highly correlated to the success of the protocol (over the long term) so any misbehaviour to the detriment of the community, should cause more pain than if applied to the stable coin. This also has a secondary consequence of creating significant demand for either HBB and USDH.

This could lead to high value users who in theory should have a better understanding of defi and the risks of leverage to utilise the protocol at high LTV ratios. Whilst allowing normal users who are happy with the 75% LTV to not have to participate in the system.

Although I agree with Marius that any decision made on raising the LTV should not increase the risk of insolvency for the protocol.

1 Like

I think there’s a long-term division (I wouldn’t go so far as saying conflict) of interest to consider here in terms of HBB; that being the introduction of governance in future. If there are Hubble users who both want to utilize the protocol at high LTVs, but also take part in voting in governance, then they would essentially have to divide their HBB holdings between the two positions.

As @y2kappa mentioned, having different vaults and assets with different LTVs is a solid building block moving forward, where vault parameters can be adjusted and different vaults opened as voted by users etc. As opposed to having an ‘exclusive club’ of users with extra benefits, of which the parameters would inevitably have to adjust as time goes on and the protocol grows.

That being said, use cases are significant factors to consider for both USDH and HBB, and this could lead to a larger conversation surrounding staked tokens and how these might be utilized within the protocol/governance.

1 Like

I like this a lot tbh, you could potentially slash their staked HBB if it triggers a liquidation at 90% LTV, and therefore the buffer for liquidation is 10% for the collateral and how much HBB we can sell to cover the loss.

I think 75% or less if a user isn’t contributing to the stability pool makes sense. It also encourages more participation within hubble if the user wants to leverage more.

What I am suggesting is there needs to be sufficient USDH to cover the margin without a risk to the collective TLV.

I actually think if the stability pool was implemented in this way the entry level LTV could be even lower.

1 Like

Here’s a complementary idea to the stability pool being used to maintain LTV:

Allow the option to select a strategy with the stability pool to ‘stabilize’ the individual LTV

  1. Option to buy mSOL (or another asset if you really want) with USDH (Buy the dip)
  2. Option to repay the debt (Set and forget)

This would encourage the individual to maintain a lower LTV ratio then even 75%

4 Likes

@Jesse-N This would be incredible, especially given the issues Solana occasionally faces during periods of congestion. Stability Pool funds auto-stabilizing a user’s loan without any active management is the dream. I really think that has the potential to blow open Hubble’s target market and bring some of our lost TVL back to Solana. Personally I know several people who have drifted to other blockchains due to liquidation scares during periods of congestion. If Hubble can remove those events… sky is the limit.

4 Likes

Can I add on top of this idea? In the future, we may accept tokens more volatile than others. Can I use my USDH staked in the stability pool to buy those liquidated tokens instead? Then, have those tokens convert to USDH, then use that USDH to pay for my loans.
It’s a double whammy benefit for me since I am a trader. I profit off of volatility and having my technical analysis skill be put to good use through this route is really a breath of fresh air.

1 Like