GIP-43: Reevaluate LP and Backer Rewards

Authors: The Warbler Labs team

Summary & Background

Note to the Community

Given GFI’s utility as a governance token and requirement in membership, the community’s input and perspective on how to address the imminent reward depletion is critical; community feedback on this proposal and participation in the discussion of this proposal is invaluable. While this proposal lays out two approaches, alternative proposals from the community are highly encouraged in order for the community to collectively reach the best decision for the protocol. This proposal is based on the facts and circumstances as of March 3, 2023. Various conditions such as the withdrawal queue may materially change the GFI emission rate.

Summary

At the current rate of emissions, the GFI rewards for LPs (the StakingRewards contract) will run out at the end of June 2023 and Backer Rewards will run out in mid-April 2023. Those dates are significantly sooner than had originally been anticipated. Also, given the low liquidity environment, which prevents people from easily exiting the position after those dates, we are proposing a way to gracefully extend the rewards, and split the remaining tokens between Backers and LPs.

To that end, we present two options below on how to achieve these goals, using principles of relative risk and liquidity to guide the proposal.

Background on how we got here.

The introduction of withdrawal mechanics (GIP-25) caused a large unstaking of FIDU from the StakingRewards contract. The remaining staked FIDU was far below the current target of 100m FIDU, causing the contract to emit rewards at the max rate. And because Backer staking rewards simply mirror what the LP staking rewards are (GIP for reference), this max rate was then amplified by also providing it for Backers.

Principles for the proposed options

Before diving into the specifics of the proposed options, it’s worth explicitly mentioning the principles upon which the proposal is based. At a high level, we want to do something that is fair for all stakeholders. We believe what is fair will take into account two main things

  • Liquidity - If the investor can exit their position easily, then there is less need for token rewards, since they are not locked in, and the market can simply decide whether they want to keep participating in the pool
  • Risk - What is the relative risk of the investments? More risk should equal more tokens, and less risk should mean less tokens.

Using those as guides, we can say that LPs have greater liquidity and less risk overall than Backers, therefore remaining tokens should skew towards Backers. Backers are locked into multi-year loans, and sit in the junior position. But to be clear, this is not a lop sided distinction. LPs still have low liquidity as well — the Senior Pool withdrawal pool is substantial right now — and indeed still take on risk. While most deals have 20-25% buffer from Backers, there are some old deals that do not include this. So these principles skew in favor of Backers, but not super heavily.

In addition, we also use the principle that **when it comes to tokenomics, gradual changes are better than sharp cliffs Sharp cliffs can cause highly volatile actions, and have knock on effects that are hard to predict.

Proposed Options

As such, this GIP focuses on two options (see summary chart below for APY and saved GFI implications):

  1. Option 1: No change to the StakingRewards or Backer Rewards (based on the current state, rewards for Backers would run out in mid-April and rewards for LPs would run out at the end of June)

    Pros: This is the “null” option. It is therefore extremely simple. The argument in favor here might say that the community originally allocated a certain percentage to both Backers and to LPs, and those should just run out. The fact that there are cliffs, and that people cannot easily exit right now is unfortunate, but not relevant.

    Cons: The argument against this might say that when those allocations were made, it was assumed that they would last longer, and that especially the LPs would be able to exit at any time. Given that those assumptions have changed, it is reasonable to re-allocate rewards and make them last for a longer duration.

  2. Option 2: Split remaining rewards 60/40 between Backers and LPs.

    Details: This would mean that of the roughly 2M tokens remaining in the LP contract, 1.2M would be allocated to Backers. If 60% feels too low for Backers, remember that they have less than half the capital of LPs. Thus the rewards per dollar is even higher than 60/40.

    In addition, the rate at which remaining LP rewards are emitted would be gradually reduced by 50% each month for 6 months, so that they extend out for roughly 36 months. Backer Staking Rewards (The LP Reward matching part) would be turned off, while the Backer Repayment Rewards would be increased slightly. The net effect for both LPs and Backers would be a reduction from current levels, but those rewards would extend out over roughly 36 months. Full details can be seen below.

    Pros: By splitting the remaining rewards, we are ensuring that both Backers and LPs are continuing to receive rewards throughout the life of their loan, which we think better fulfills the spirit of what the original allocations were meant to do. This approach also ensures a smooth decrease over time, which creates more stability for the protocol. Lastly, if unitranche deals continue (like the new Cauris deal last December), then this means the tokens can be used for new growth of the protocol, which is always good.

    Cons: This approach requires some minimal smart contract changes, which brings risk. One could argue it is less fair to LPs, but based on the principles laid out above, we think it’s a reasonable compromise.

As a reminder, alternative paths or parameters are strongly welcome from the community.

User Implications

Here is an analysis of the impact of each option on LP or Backer APYs as well as the potential GFI savings that would result (see table in Summary), based on the current state.

Option 1 Option 2
Overview No change to the StakingRewards or Backer Rewards Gradually reduce the StakingRewards max rate by 50% each month for 6 months; disable Backer Staking rewards while reallocating 1.2M GFI to Backers and setting the Backer repayment target to $34M.
Date LP rewards would end End of June 2023 March 2026, or later
Date Backer rewards would end Mid-April 2023 March 2026, or later
Base USDC APY Impact None None
LP Total Est. APY (w/ GFI) Next 2 months: 20.7%; Months 3 to 6: 12.9%; Months 7 to 36: 7.8% Next 2 months: 12.6%; Months 3 to 6: 8.6%; Months 7 to 36: 8.0%
Backer Total Est. APY (w/ GFI) Next 2 months: 25.5%; Months 3 to 6: 16.9%; Months 7 to 36: 16.9% Next 2 months: 21.0%; Months 3 to 6: 20.6%; Months 7 to 36: 18.6%

Competitive Comparison

In any scenario in which StakingRewards is reduced or removed, Goldfinch’s base USDC yield opportunities still remain competitive for new LPs and Backers when compared to RWA protocols and well-established lending protocols (see below, further details at rwa.xyz). Most other RWA protocols are not providing any native token rewards currently, with the exception of ClearPool and Compound.

Avg. Base APY (USDC) as of February 24, 2023 Reward APY, as of February 24, 2023 Total
Goldfinch Senior Pool (LP) 7.8% Assume >0 ~7.8%
Goldfinch Borrower Pool (Backers) 17% average Assume >0 ~17% average
Maple 8.32% 0% 8.32%
Centrifuge 9.31% 0% 9.31%
TrueFi 2.08% 0% 2.08%
Credix 14.13% (solana) N/A 14.13%
ClearPool 5.78% 1.15% 6.93%
Aave 2.1% N/A 2.1%
Compound 1.61% 0.77% 2.38%

Source: rwa.xyz

Additional Context

Early on in the lifetime of the protocol, GFI liquidity rewards may have attracted significant TVL (see chart below). There was roughly $75,000,000 USD growth and 122,476 GFI distributed by the StakingRewards contract between the launch of the liquidity mining rewards in January 2022 and the end of April 2022 (source). However, the rate of TVL growth on the protocol has slowed to near zero despite distributing rewards at reasonable levels.

Furthermore, TVL or staked FIDU growth may not be possible until the $30M+ in withdrawal requests have been fulfilled so the current StakingRewards emission rate could remain at max emissions; if that happens, the remaining StakingRewards (~2,000,000 GFI) will run out by the end of June and the remaining Backer rewards (~350,000 GFI) will run out in mid-April. This rate of depletion is much faster than the rewards forecasted when originally designed, and therefore the parameters must be revisited in order to protect the protocol’s most at-risk investors.


source

Jan ‘22 - April ‘22 May ‘22 - Feb ‘23
Overall TVL Growth $75M -$11.8M
New Loans $61.8M $2.1M (unitranche)
GFI Emissions (by token amount) 122,476 6,966,593
GFI Emissions (by cost) $275,571 (at $2.25 / GFI) $4,179,955(at $0.60 / GFI)
Emissions / Overall TVL Growth 0.16 -59

Voting

  1. Option 1: No change to the StakingRewards or Backer Rewards
    1. Backer rewards run out mid-April and LP rewards at the end of June
  2. Option 2: Split remaining rewards 60/40 between Backers and LPs.
    1. Every month for 6 months, reduce max rewards emitted for Senior Pool LP liquidity mining by 50%.
    2. Reallocate 1.2M GFI from the Senior Pool LP liquidity mining rewards allocation to the Backer Rewards allocation.
    3. Disable the Backer Staking rewards system introduced in this GIP.
    4. Increase Backer rewards for interest repaid by decreasing the max number of interest dollars eligible for rewards from $100M to $34M.

Sources

  1. Data for calculating rewards emissions: https://docs.google.com/spreadsheets/d/1Tgn3NI7Thcqp8o2iJfdJ5is5CehWHKd7-BTlI7toe8o/edit?usp=sharing

  2. Data for calculating impact on APYs and GFI savings: https://docs.google.com/spreadsheets/d/1UUWKuuWK_GE7L7FmLFIALdX71i_jmBNChuVG2hPnXCQ/edit#gid=795845629

  3. Senior Pool Liquidity Mining: After depositing into the senior pool, senior pool LPs can stake their Fidu to earn GFI rewards. The reward rate is variable.
    See the Senior Pool Liquidity Mining docs.

  4. Backer Rewards After depositing into the junior tranche of a tranched pool, a backer earns rewards from successful borrower interest repayments. The amount of rewards is directly related to the amount of interest repaid.
    See the Backer Incentives docs.

  5. Backer Staking Rewards: After depositing into the junior tranche of a tranched pool, a backer earns rewards as if they had deposited into the senior pool and staked the FIDU they would have received.
    See the Backer Incentives docs.
    See the Backer Participation in Staking Rewards docs.

Note: protocol experienced no change in capital inflows when the protocol did increase liquidity rewards by removing the vesting schedule (GIP-10), there is evidence to suggest that no reasonable increase in GFI rewards will sufficiently incentivize senior pool capital supply in the current low liquidity, high withdrawal demand environment.

3 Likes

Interesting proposal, for now, I am not a backer of the protocol and only a LP provider so my comment might be a little biased.

But I think because the backer has a lower capital USD value they are already receiving more rewards than the LP’s.
as you mentioned in the details of option 2;
‘‘remember that they have less than half the capital of LPs. Thus the rewards per dollar is even higher than 60/40.’’

Then I have a question

What is the % of backer capital compared to LP capital?


Because In the user implications analysis above I think the Reward ratios of option 1 already give the Backer a fair premium over the safer and more liquid LP.

This is premium % shows in option 1 that in the coming 2 months backers receive a 25% more, months 3 to 6 they receive 31% more and months 7 to 36 they receive 116% more than the LP’s.

This is premium % shows in option 2 that in the coming 2 months backers receive 66% more, months 3 to 6 they receive 140% more, months 7 to 36 they receive 132% more than the LP’s.

My conlusion
I think option 1 is fair to both the Backer and the LP. Backers already receive a higher reward because of the higher risk involved with the borrower and they receive more staking rewards because their capital amount is less the LP’s amount. In months 7 to 36, the backers receive a 132% premium in comparison to the LP for a long amount of time.

Shifting this balance to 60/40 would make it unfair for the LP in my opinion.

I am interested to hear a Backer’s perspective on this!

2 Likes

Nice proposal. I vote for Option 1. In my opinion it is more fair for backers and LP providers.

I vote for proposal number 1. I think that in this situation this is the fairest option.

I think, Option 2 is unfair for the LP in my opinion. My vote is for Option 1! because Option 1 already give the Backer a fair premium over the safer and more liquid LP.

1 Like

I think it’s worth finding a compromise so I will vote for option 2, plus they will continue to receive rewards

I also don’t see the motivation to shift additional rewards from Senior LP to Backers, as Backers are already compensated for the higher risk with 10% higher USDC APY (7.8% vs ~17%). Especially given the current liquidity environment for LP holders, which at current rates are locking users into ~1 year hold time. Finally, any decrease in GFI rewards for Senior LP will only cause more pressure on the withdraw queue.

As written, I would strongly support Option 1.

If, instead, you are more concerned with the pace of GFI token distribution, I would be interested in supporting a proposal that would look to slow the pace of distribution across both LP and Backers (without changing the distribution of how much GFI is allocated to each).

Personally, I don’t stake, but I know that many have faced the problem of leaving the LP and withdrawing their rewards. And since for the project to exist, it is very important for us that those who provide liquidity are satisfied, it is logical to vote for option 2

Agree with those who believe there is no need to shift rewards towards backers and I’m an investor in both types of pools. It seems that senior pool is having a harder time to attract capital at the moment.

Additionally, high gas fees really hurt those with <100k deposits. Reducing the rewards will make it even more expensive to claim relative to reward size. I understand, however, most value is coming from whales.

I may support a proposal with a reward distribution that is more in favor of senior LPs.

I think that reducing emissions makes sense. I don’t agree that re-allocating so that Backers receive a higher proportion makes sense.

The LP/Jr diversified pool is one of the distinguishing features of the protocol compared to others. We’re at a point in time… the liquidity situation will change in the future and it’s important not to set any precedents here.

I voted for option 1. If a percentage of remuneration was prescribed, why change it? I see no reason to change something, everyone knew the conditions in advance

I think it is necessary to find a middle ground so I will vote for option 2, plus they will continue to receive awards and this cannot but rejoice

Option 1 is terrible for the protocol, the emissions need to change, and we are spending GFI that will run out soon and not actually effectively incentivizing the long term health of the protocol. Option 1 is unpalatable and doesn’t accomplish anything for anyone.

I’d be ok with turning off all emissions before I’d be ok with Option 1. I am a large LP holder but Goldfinch is much more important than the returns I get on my LP stake in the next 3 months (which is how long until the emissions run out anyways if we don’t change them…)

Option 2 is fine and at least buys more time.

1 Like

I vote for the 2nd option. Splitting remaining rewards and gradually reducing LP rewards could be seen as more equitable and better fulfilling the spirit of the original allocations.

Wow, how thoughtful. But I don’t see the point in different percentages. I vote for option 1

First of all, I would like to note the high quality of the proposal, including the detailed information provided, which allows, among other things, to make an independent analysis.

In my opinion, one side of the question is whether to take into account the background. For example, after Goldfinch Flight Academy, significant amounts of GFI were allocated by Senior Pool members (those who held funds in the pool at the snapshot date) without a lock. It is clear that not everyone was lucky enough to be among them, but nevertheless, such a “gift” (if I remember correctly) was not made for the Bakers.

If we consider the situation at the moment, then participation in the Senior pool is hindered by both relatively low profitability (relative to Bakers) and low liquidity (including the inability to quickly withdraw funds, only in order of priority).

I think both scenarios described in the proposal are acceptable.

In favor of the first option is the fact that nothing needs to be changed. And as is well known, if something functions satisfactorily, then there is no need to change something, because. there is a risk not only to improve, but also to harm. But the disadvantage of this option is that it is less flexible and less oriented towards the long-term development of the protocol. Indeed, no amount of reasonable GFI payouts will make the Senior Pool attractive enough in a bear market and low liquidity.

Option 2 is undoubtedly a better fit for the current situation. It makes sense to reduce the emission of GFI, since the expected effect of this emission has not been achieved for a long time.
This option assumes a gradual decrease in yields for both liquidity providers and bakers, shifting the incentive towards the backers. At the same time, I don’t think that the bias towards the bakers is very significant.

If choosing between options 1 and 2, then I choose option 2, but only because of the greater flexibility and suitability of the situation.

P.S. As an addition, I would suggest “stretching” the GFI emission time horizon not only for 3 years, but even for 5 years, gradually reducing the GFI emission every month or every 3 months throughout this time.

1 Like

Thank you @mikesall for the detailed proposal.

When you mention the following, are you saying that at the current rate, the 16.2% allocation of GFI tokens assigned to liquidity providers, and the 8.0% allocation of GFI tokens assigned to backers will have been distributed by the end of June and by mid-April respectively?

I wholeheartedly agree with @Noun12. Option 2 is the better option. I’ll add that I’m in talks with investors considering deploying capital via Goldfinch. Letting rewards dry out would make other credit protocols a lot more attractive to capital providers.

When we let rewards dry, I bet that we’ll wake up in a few months realizing that we made a mistake and see that in order to attract capital, we need to move GFIs assigned to other purposes, or mint more GFIs to reward capital providers.

Council has approved the proposal

1 Like

This makes sense.I will vote for option 2, plus they will continue to receive awards and this cannot but rejoice

Hey there. I am backer of Almavest Baskets #6 and #7. It is stated that total outcome ($USDC+$GFI) in Option2 should be 21% after that update. I have noticed significant drop in $GFI rewards and my calculations show that outcome is around 16-17% now. Am I missing something?