GIP-02 | Increase the existing Leverage Ratio from 3x to 4x

Author: Sam Eyob, CIO Warbler Labs

Summary: This builds on the back of a proposal by @Billy Bill which was supported by @SAM226 in Discord to increase the existing Leverage Ratio from 3x to 4x. This proposal is further backed by the scale Goldfinch has achieved deploying $39m across 18 countries with a 0% default rate. This proposal suggests Goldfinch take on incremental risk with the leverage ratio which would result in higher APY’s for Backers (in addition to this existing GFI rewards), and a higher utilization of the Senior Pool.

More concretely, this proposal suggests two things for consideration by the community:

  1. Goldfinch increases the Leverage Ratio from 3x —> 4x resulting in:
    a. On average increasing their top-line APY returns by 2-3% (eg. 15% → 18%)
    b. In exchange for Senior Pool decreasing the first loss portion by a modest 5%, but also being able to increase pool utilization and Senior Pool APY’s given more senior capital is deployed into every pool
    c. Borrowers would be able to close pools more quickly due to smaller Backer pools being required
  2. This would affect any open pool from Almavest Basket #6 onwards until either a new leverage ratio has been approved, or until the automated Leverage Ratio as conceived in the whitepaper goes live.

Motivation: The Goldfinch journey to $39m across 18 countries has been amazingly quick, but has been measured in terms of investing in extremely high-quality real-world borrowers with proven track records. We now have on-chain data to prove these Borrowers are committed to making their payments, and as such should be comfortable taking modestly more risk with higher leverage (i.e. 3x —> 4x) in order to:

  1. Increase Backer APY’s
  2. Increase Senior Pools utilization and APY’s based on these and future trusted Borrowers
  3. Attract more high-quality Borrowers to the Goldfinch

0% Default rate since Goldfinch Protocol Launched

$39m in Active loans across, 18 countries, supplied to high-quality Borrowers like Almavest and Cauris Finance who have supplied capital to 232k real-world end-users


Specification & Requirements :

  • Goldfinch increases the Leverage Ratio from 3x, to 4x: Backer cash APY’s would increase on average by ~2-3% based on the below table, in addition to being supported by variable 30-40% rewards which were approved in this snapshot vote

Assuming the average Borrower pays an APY of 10 - 14%, Backers would receive on average between ~2-3% additional top-line APY as a result of increasing the leverage ratio from 3x to 4x

  • Borrowers able to launch faster pools: See above “Risk Trade-off for Senior Pool.” Smaller Backer Pools would mean they can be filled faster

Benefits

  1. Increasing Backer cash APY returns by 2-3% (in addition to variable GFI rewards)
  2. Increase Senior pool utilization and thus Senior pool APY due to reduced cash drag
  3. Borrowers able to launch faster pools

Downside: First-loss protection provided to the Senior Pool by the Backers reduced from 25% to 20%. This reduction is supported by a clear track record of high-quality Borrowers consistently repaying into the protocol with zero defaults

Voting

  • Yes: Increase the leverage ratio from 3x to 4x affecting pools from Almavest Basket #6 onwards

  • No: Do nothing

Resources:

  1. Original Discord Billy Bil proposal

  2. Discord support from SAM226

  3. Live on-chain Goldfinch portfolio quality

  4. Medium block post highlighting the breadth of countries capital has gone into with no defaults on interest payments

  5. Leverage model summary in Goldfinch Docs

  6. Goldfinch front end highlight 0% default rate

  7. Almavest Basket #6 pool link

  8. Snapshot vote for Backer staking distributions

2 Likes

Disclaimer: this proposal affects us directly (as ALMA, one of the pools currently open). But I wanted to share a datapoint with the community as they consider the proposal from @SEyob.

Besides Goldfinch, we (ALMA) also source capital from large asset managers. The leverage ratios we are subject to from them typically are 7-9x. (Meaning, we have to find first-loss protection of between 10-15% to be able to get their capital). What @SEyob has proposed is reducing Senior LP first-loss protection from 25% to 20%. This would still provides significantly more downside protection to senior LPs than what’s commonly found in the market.

I vote Yes. This benefits all stakeholders of the protocol.

Disclaimer: this proposal affects us directly as Cauris – we have a pool open and also invest in Backer NFTs.

Similar to Alma, historically we’ve raised capital with 10% to 20% junior allocation. We understand that Goldfinch need to first show some historical performance before bringing ratios closer to what we see elsewhere. We’re in favor of switching the junior ratio from 25% to 20%.

We expect pools to fill faster and it will provide a higher return to Junior investors, while providing in-market protection to Senior investors. Filling more pools will result in deploying more Senior capital (FIDU) in borrower pools, which will also benefit Senior investors.

1 Like

Я голосую Нет.

  • 0% Default rate - эта статистика лишь за пол года, нельзя на ее основе повышать риски.
  • В компаниях из реального сектора first-loss protection 10-15% - в пул инвестируют непрофессионал, академия давала возможность улучшить знания, но этого недостаточно, поэтому нельзя увеличивать риски. Напомню, что в случаи дефолта Бекер теряет Usdc и GFI больше не начисляется.
  • Упоминание о доп. доходности пула неактуально, так как для ее прогнозирования надо прогнозировать цена токена GFI через 4 года.

Увеличение плеча снизит диверсификацию Сеньор пула. Альма уже занимает там почти 30м, Стратос хочет занять 20м сейчас и до 100м в будущем. Увеличение плеча лишь поможет им быстрее вычерпать весь пул.


I vote No.

  • 0% Default rate - this statistic is only for half a year, it is insufficient term to increase risks on its basis.
  • In companies from the real sector, first-loss protection 10-15% - a non-professional invests in the pool, the academy made it possible to improve knowledge, but this is not enough, therefore, risks cannot be increased. Let me remind you that in case of default, Becker loses Usdc and GFI is no longer charged.
  • Mention of additional the profitability of the pool is irrelevant, since in order to predict it, it is necessary to predict the price of the GFI token in 4 years.

An increase in leverage will reduce the diversification of the Senior pool. Alma is already covering almost 30m there, Stratos wants to take 20m now and up to 100m in the future. Increasing the leverage will only help them scoop out the entire pool faster.

3 Likes

I’m a fan of this proposal. It can substantially help grow total lending on the protocol, while making the APYs even more compelling for Backers to fill the pools. Also, because these borrowers are expected to have 0% default rates, I think the marginal credit risk for both LPs and Backers of going from 25% → 20% is small.

I support the proposal with the current risk indicators to increase the attractiveness of the APY.

Proposing changes to the leverage ratio without discussing what changed in the historical loss performance of the underlying collateral assets (i.e. the historical and recent performance of the underlying loans) is an incomplete proposal (from a tradfi credit stand point).

Typically when we made these sort of adjustments to the senior advance rates, they were done with extensive analysis of trends seen in the underlying collateral. For instance, a common scenario was this:

  • At the time of origination, a 75% senior advance rate equated to a ~3.5x loss coverage/breakeven, which for that specific type of collateral, came out to be an IG credit (somewhere between ~BBB - single-A)

  • After seeing improvements in origination volumes, and early loan life performance, we see trends indicating that maintaining that 3.5x loss coverage/breakeven on forward collateral ROA performance allows for up to an 82% advance rate, so a bump up to an 80% advance rate was warranted.

  • We’d also use this as more of a negotiation tactic as well - i.e. if the borrowing company asked for a lower interest rate, then we might push back and meet them half-way on the interest rate, but also offer a higher advance rate.

Long-story short - would an increase of Junior leverage from 3x to 4x result in a significant change in the projected breakeven multiple?

I see 0% default rate mentioned a lot in this proposal. I believe this metric is slightly misleading because -

  1. Most of the pools sourced so far are vetted by the goldfinch/warbler team to kick off the borrower pools.
  2. Fully repaid pools. We only have small capital repaid pools (max being 300k USDC pool and total 3 pools of 550k USDC)) so far and yet to see big ones (Millions pools) fully repay yet.

So I would caution a bit on rallying on 0% default rate history.

Nevertheless I will vote yes - because increase in risk is small and also motivates backers to supply capital into pool.

Is that 0% default rate the default rate of all the collateral pools backing the loans, or the default rate on the loans themselves?

I also support this proposal. I think the incremental risk, compared to the benefits for the Senior Pool (increased utilization, and pools filling faster) is substantial.

@JeremyKim That’s a very fair point about wanting to see changes in underlying collateral, though to be clear, all existing loans would stay at the current leverage ratio of 3X. So I’m not sure a change in underlying collateral assets is quite the right metric here. All existing loans are already out the door, and the terms won’t change. This would be for any pools that are still live (or yet to be created) as of the time the change is made.

So the question is more about the overall track record of Goldfinch, and whether it warrants taking on some incremental risk now for the benefit of filling pools, and higher Senior Pool utilization, which increases APY’s, and helps keep capital in and momentum up. Given that 20% protection is still on the high end (or higher) than a lot of industry averages (per what @givatury said), this tradeoff seems good to me.

@Silverfinch. To a few of your points. The 0% is technically for over a year now. Granted the first half of 2021 had much lower volumes than today, so your point is taken. It’s also worth noting that much of the volume today is actually small baskets of loans to multiple companies, so inherently have more diversification than the early loans which were generally to individual companies (so should be a bit safer).
I think the diversification issue of the pool at large is a good one to always keep in mind. We certainly want more Borrowers. Stratos is a new Borrower, so they will help, and there are many more who have expressed interest. But more importantly, these Borrowers are in fact credit funds themselves who use the loans and diversify them usually into a few companies. So the actual diversification is greater than it may seem.
Lastly, the mention of higher yields for the Backers is not dependent on GFI token price. Those higher yields are on the USDC (due to the higher leverage ratio). And it’s also actually higher yields for the Senior Pool too, because utilization would be higher.

To recap, the benefits are very clear (higher APY’s for Backers, pools filling faster, and higher utilization for the Senior Pool), and the risks are incremental, with solid track record thus far backing up the high quality Borrowers who are coming to Goldfinch. So I think the tradeoff is worth it.

Right - are the collateral pools for these warehouse loans revolving? If so, then the composition of the collateral pool at the time of origination could be different than the collateral composition of the collateral pool now.

What you say is that the warehouse loans haven’t been amended, but don’t add any insights into how the ROA of the underlying collateral pools backing these warehouse loans have changed.

In this scenario, assuming that the underlying collateral pool performance hadn’t changed, this would be a slam dunk for the backers - their risk doesn’t change, but they get more leverage. However, since the additional leverage is at the expense of the senior position taking on more risk, I could see how a senior investor might not want this, even with the diversification aspect from FIDU.

If the performance of the underlying collateral pools have improved (i.e. you see a sort of rooster comb of monthly static pool collection curves of each deal trending upward), then it would be a slam dunk for both FIDU and backers.

@JeremyKim Right - are the collateral pools for these warehouse loans revolving? If so, then the composition of the collateral pool at the time of origination could be different than the collateral composition of the collateral pool now.

What you say is that the warehouse loans haven’t been amended, but don’t add any insights into how the ROA of the underlying collateral pools backing these warehouse loans have changed.

These, along with your other points are great, and really speak to the additional work that needs to be done to better mesh the Defi + TradFi worlds when it comes to understanding what you’re investing in, and how collateral/security works both on/off-chain. More specifically, the key Borrower performance indicator that’s observable, and thankfully on-chain is their default rate. That though is very much informed by understanding atomized collateral level performance (e.g. vintage/cohort level, business cycles etc…) which informs/can help predict whether the Borrower will have issues with future repayments.

When it comes to this proposal, all existing pools will receive zero direct benefits if this proposal passes as those pools are locked at 3x leverage ratio, and will not move to 4x.

Though as you rightfully mention (my own interpretation of your intent), future risk-reward trade-off (e.g. changing future leverage ratios from 3x → X) should be informed based on historical performance. I fully agree with this. As the Goldfinch protocol matures and better meshes the Defi and TradFi worlds, there will surely be plenty of additional product/features that could be surfaced to enhance the risk-reward profile for both the Senior and Backer participants (e.g. Warbler and/or 3rd parties could stand up new products/service for the Goldfinch protocol).

When it comes to the existing Borrowers and their future pools, in addition to the new Stratos pool which has launched, I’d strongly urge you to join the Borrower datarooms. There you can speak directly and confidentially with the end Borrowers to better understand the facts/circumstances and nuances of their track records (e.g. the security packages, LTV’s, covenants Borrowers apply to their end-borrowers when they lend), as it relates to the asset classes, steady/improving vintage performance, and geographies they would deploy Goldfinch funds into (here’s a link to the Goldfinch Discord Borrower channel).

Overall, I see the incremental risk-reward being win-win for all as the stance Goldfinch took from the start with a 3x (i.e. 25% first-loss coverage) was exceedingly conservative. Nevertheless, directionally agree that atomized collateral data either coming directly from Borrowers or via 3rd parties would be fantastic tools to help create a robust and transparent on/off-chain ecosystem for all.

1 Like

Actually, tradfi structured credit analysts will take historical performance and build projection models, and the structure of the investments will be informed by those projections, but also the originator’s growth strategies. (For instance, if an asset originator used to originate in regions x, y, and z, but feel that region z has the lowest risk-return profile and made the decision to exit that region, then without much change in performance, an investor would focus on data from only regions x and y, which may be more attractive now that they dropped a region generating assets with lower ROA).

I think a key wrinkle to all of this is the willingness of the borrower and the underlying asset originators to share the details of their data. Completely understand that loan-level data would not be shared, nor extensive details on the borrower’s analysis process.

Having looked at the latest Stratos and Almavest datarooms, these datarooms are definitely good starting points, but if I were to put on my “tradfi structured credit investor” hat, we would usually ask for additional details (perhaps those are covered in the Telegram groups).

This proposal is now live for ~72 hours on Snapshot for voting.

1 Like