FIP-13 A Framework for FEI lending deployments

We should be pursuing more framework based approaches for making decisions about PCV, TRIBE rewards, and FEI deployments so we can streamline governance processes. @arcology brought up a similar point in Discord.

This post focuses on FEI deployments specific to lending platforms including Rari Fuse, CREAM, Sushiswap Kashi, and Aave.

I think we should aim for a few objectives with lending FEI deployments:

  1. Bootstrapping markets - without an original source of FEI to borrow, markets cannot have borrowers and can’t have the APY to generate yield.
  2. Partnership - being able to easily bootstrap USD liquidity for leverage in a market is a strong value prop for potential partners (for example the Olympus pool party integration)
  3. Visibility - by being in many markets with high liquidity, people see and use FEI in diverse places.
  4. Earning yield for PCV - by earning FEI back to the protocol, we improve the collateralization and net PCV equity

Sizing for FEI deployments should be scoped to the maturity of the market.

We currently have 10M FEI lent out on FeiRari. Which we consider a medium maturity market.

I propose the following framework:

Maturity Range Examples
High 25-50m Aave, Compound
Medium 5-10m FeiRari Fuse, C.R.E.A.M
Low 1-2.5m Fuse and Kashi partnerships, e.g. Olympus

This framework is somewhat subjective, but it will allow us to move faster on these integrations as we assess the market demand for FEI liquidity across various platforms and userbases.

We should also aim to cap the size of overall lent out FEI, to prevent a Soros attack (borrowing and shorting FEI to drain protocol reserves). I propose a soft limit of 10% of PCV value to be authorized under this framework.

If approved by the community, we can move to make a single combined DAO vote for:

  • 2.5M Kashi FEI-ETH
  • 2.5M Kashi FEI-TRIBE
  • 5M C.R.E.A.M FEI
  • 1.33M Olympus PoolParty FEI

And make future DAO votes based on this framework without explicitly snapshotting each integration. For example future Kashi pairs and Fuse pools with partner protocols can get 1M FEI to bootstrap liquidity.

  • I support this framework, including other projects not referenced
  • I support this framework, only for projects listed
  • I do not support this framework

0 voters


Establishing this framework is imperative to standardize future deployments across DeFi. Having points of reference, and more information regarding the decision making process. Fei has an unprecedented ability to supply liquidity to markets for utilization, and doing so will increase the circulating supply of Fei and expand our userbase, while also spreading name recognition. I see this as a win for the protocol.

As a sidenote, this ability to bootstrap lending markets through governance also showcases the value of holding Tribe and being an active and engaged voter within the Fei ecosystem.


I think the sizing and framework is adequate. It allows us to be lean without introducing too much red tape.



Frameworks are something I heavily advocate for, because as Fei grows and matures we need a standardized approach for scaling and efficiency yet having enough flexibility to allow us to be a bit more ‘fuzzy’ - both of which points are mentioned by @GrantG and @countvidal .

I’m pleased that this approach is being considered, I have no objections to this approach but I do have a question:

Would there be any circumstances that the ‘maturity’ assessment for a particular project changes? In this scenario, would we adjust the FEI deployment sizes accordingly, and how? Does it need a governance vote for both (increasing and decreasing deployment size) directions, or only upwards?

I would also urge you to consider that with all deployments, one key question to think about is how we generate revenue for each project, in terms of both economic and social benefits.

Finally, would all deployments (current and future) be considered for TRIBE incentives? Or would this be on a case-by-case basis or only for certain deployment sizes?


Good idea to have this more detailed framework, it is interesting to have this comparative analysis and do not look isolated case by case.

Are there any metrics that could be considered for analyzing the maturity of the market or it will be a more subjective analysis? Will we take into account smart contract and counterparty risks?

Would we also consider the potential benefits of each one to evaluate the risk/benefit of each alternative. By the framework, I understand we are focusing more on the risk part.

Is there any guideline about the duration of these deployments?

The sum of all the deployments should respect this limit? How the collateralization ratio of the project could affect this total limit?


per the post: the focus is on FEI deployments, but the proposed pools have ETH and TRIBE

  • what’s the source for those tokens, and should that be a part of the framework (so we avoid that extra step on upcoming deployments)

any reason to exclude Bancor from this approach?

ultimately framework/standardization is the only way to scale out, tunning and ongoing reviews will be required but well worth the effort

Guys, I’d like to reiterate the position I’ve already brought up elsewhere.

DAO should strive to make decisions informed and probably even steered by data, modelling and model-based predictions. Figures and assessments shouldn’t come out of thin air. Specifically two types of data-driven models should be in DAO’s focus: subjective DeFi crowd sentiment and objective onchain dynamics.

I’ll soon publish the first iteration of the tool for sentiment analysis of the DeFi Twitter feed with the fine-tuned BERT language model under the hood. Later it should be augmented with tools for understanding ledger data: deep learning on non-Euclidean domains could come particularly handy here.

Yes, it’s complicated and a lot has to be built to make this happen. But that’s the right avenue to pursue, I’m urging everybody to support me in this bold and important vision.

1 Like

Yes naturally projects can increase in maturity based on time being active (Lindy) and TVL. They can also decrease in the event of some kind of exploit or other reason to downscope activity. It might make sense to revisit our classifications regularly.

The benefits are generally increased liquidity and visibility for both projects, some comarketing etc. As demand for FEI borrowing increases the DAO would also generate revenue and partners would earn their fees and such.

Absolutely we should consider them all, but I think TRIBE incentives should be separate and more case-by-case for now.

Maturity would be based on TVL, time since deployment, and the comparative risk of the assets in the protocol. For example CREAM carries high tail risk due to the number and diversity of assets involved but has a long track record. Kashi has very low tail risk as the pairs are risk isolated but the system is newer.

For now we should assume the deployments are indefinite, until we have more data on how the FEI markets form or a strong need to reduce the circulating supply of FEI.

Yes this was my thinking, we could be a bit looser on the number given the current collateralization ratio of the protocol. At a minimum, we should keep the amount of protocol-lent FEI less than the protocol’s PCV equity.

ETH and TRIBE will be supplied by external participants looking to unlock liquidity in their assets, possibly for leverage for example.

Bancor is an AMM, this would focus on lending deployments.

Obviously data can be used to enhance our decision making as a DAO and I’m in full support of that. The main questions are:

  1. How much better of decisions can we make with data?
  2. What data do we have available and how can we filter which data is relevant?

The idea behind this approach is to get a good-enough system in place to allow us to move quickly and gather the very data we’d need to make better decisions in the future. The benefits of this approach to a high degree will be more on the visibility and partnership side, as other DAOs can see we have a credible commitment to provide liquidity when they add FEI to their Fuse and Kashi pools for example.

I’m not opposed to data at all, I think this is a stepping stone and want to make sure we’re focusing our energy on pareto-optimizing outcomes, where 80% of the benefit can come from simply looking at very high level metrics discussed above.

1 Like

The first DAO proposal for FIP-13 is live. Here are the final numbers:

C.R.E.A.M - 5M
Kashi FEI (asset) / TRIBE (collateral) - 2.5M
Kashi FEI (asset) / ETH (collateral) - 2.5M
Kashi FEI (asset) / xSushi (collateral) - 2.5M
Kashi FEI (asset) / DPI (collateral) - 1M
Fuse Pool 6 (Tetranode’s Locker) - 1M
Fuse Pool 7 (CLG + Tetranode UpOnly) - 1M
Fuse Pool 3,3 (Olympus Pool Party) - 1.333M
Fuse Pool 19 (Index Coop) - 1M
Fuse Pool 24 (Harvest Finance) - 1M

Voting will end at approximately 10am PT / 4pm UTC on Aug 2:

1 Like

Made a PR with code for second round:

Fuse Pool 9 (Reflexer and Frax) - 1M
Fuse Pool 25 (BarnBridge) - 1M
Fuse Pool 26 (Tokemak) - 1M
Fuse Pool 27 (StakeDAO) - 1M