[GIP-15] Engagement with Risk DAO

Summary

RiskDAO is a service DAO spearheaded by B.Protocol and 1kx that focuses on providing a public risk assessment framework and associated audits to DeFi lending and borrowing protocols.

It is proposed that the Gearbox DAO engage with the RiskDAO to:

  1. Develop mathematical simulation models for stress tests of correlated assets depegging (in particular for sUSD and FRAX stable coins).
  2. Maintain a public risk monitor dashboard to feature stress test results for all of Gearbox markets and key risk indicators.

EDIT: Item (2) includes stress tests and LTV recommendation also for existing Gearbox markets. And also for future ones.

Motivation

Setting the right risk-related parameters, most notably loan-to-value (LTV, aka collateral factor), is crucial for lending markets solvency and adoption. Aggressive parameters would encourage borrowers adoption at the expense of higher insolvency risk, and conservative parameters would do the opposite.

Gearbox has an in-house Risk Committee, however it is becoming best practice to engage with external risk assessment firms (e.g., see Aave) similarly to how external code auditors are engaged.

Further, the RiskDAO offers additional monitoring tools and simulation models w.r.t the tools the risk committee is offering.

The Risk DAO

At Risk DAO we’ve developed a novel simulation model. We take real world liquidation data of popular assets from centralised exchanges, along with the price trajectory of the assets. We then extrapolate the liquidation sizes and price trajectory to the asset we wish to analyse, and simulate the outcome based on the asset’s available DeFi liquidity. Our approach eliminates most assumptions over user behaviour during market crashes, and makes it more feasible to analyse the risk of a platform prior to its launch, and for multichain lending platforms, where the data for user behaviour is even more sparse.

The simulation model is used to reason about the safety of collateral factor and debt ceiling values (and in the full report we also use it to reason about additional risk parameters).

With the model we run hundreds of thousands of simulations to estimate the expected system insolvency under each scenario. The figure below shows a single simulation run, with price trajectory (in green), market liquidity (red), liquidation volume (yellow) and stability pool size (orange) change over time.

The online dashboard shows the result of daily simulations w.r.t current market conditions, and (potentially) recommends new values for collateral factors.

In addition, the dashboard also tracks key indicators such as dex liquidity, oracle integrity, and alerts when whales open big positions.

Furthermore, it also lets the dev team understand the root cause of collateral factor recommendations, and let them simulate different scenarios.

Below are some snapshots of the system:

Expected liquidations

Market composition - shows the % of each asset out of the collateral and debt

Collateral factor recommendation

Engagement model

Risk DAO is a revenue driven service DAO.

The pricing model is $50k per quarter. The DAO members are willing to accept up to ⅓ of the payment in vested protocol tokens.

4 Likes

Hey @yaron,

Welcome and thanks for putting together this proposal! I believe it’s a cool to have an independent third party involved in risk monitoring.

A couple of questions on your proposal:

The proposal comprises stress-modelling for these two assets only?

Does that mean you going to advise the optimal LTV values for all Gearbox assets or before-mentioned two only (SUSD and FRAX)?

Could you please elaborate on ETA, i.e. how long it’s going to take to deliver simulation models and the risk monitor dashboard?

3 Likes

No, for all of existing Gearbox markets, and the new ones. From what I understand, in the short term, the new markets are expected to be sUSD and FRAX.

All. Sorry if it wasn’t clear. I will edit the message to clarify it.

We expect to have a first version of the dashboard live after 1 month. Of course it would require more fine tuning and maintenance over time.

2 Likes

Please check our existing assets here AllowedList Policy - Gearbox Protocol
With v2 we’re going to add STETH, CVX, LDO, FRAX, LUSD, SUSD, GUSD, FXS

@ilgiz ser, didn’t I miss any of the v2 assets?^

I see. FXS, CVX and LDO are normal speculative assets.
Anw, the engagement covers all new markets, as they are added to Gearbox, but in terms of timelines, more assets will require more times.

I’d say trickest part is evaluating risks for different LP tokens:

  • Curve LPs (frax3crv, susd3crv, 3pool, lusd3crv, stethCrv etc)
  • Convex LPs risks (pretty close to previous question - the only risk added is contracts security risks)
  • Yearn LPs for Curve pools
1 Like

To some extent frax3crv is less risky than raw frax, as it has lower volatility (albeit higher smart contract risk).
(Likewise for the other curve LPs.)

Gearboxes are getting more and more popular, hope to collaborate with more projects!

2 Likes

Idk shit about finances or risks, but RiskDAO are cool people and have had good community traction so far. Would think this is a great idea, and it also doesn’t seem too costly for at least one quarter. Cool!

Hi @yaron,
Can you please elaborate a bit more on the deliverables and a relative timeline. Like what would be the dashboard main parameters and charts, who should be running front-end and back end, what engagement (if any) you would need from Gearbox, how much time math models would take, etc.

The engagement should be as clear as possible for DAO members and delegates, I believe

Hey, basically this is an example for the deliverable https://demo.riskdao.org/ (make sure you also toggle the Pro View).
The demo is for Compound compatible platform, and of course adjustments for gearbox would be needed.

The estimated timelines are to go live after one month with the standard collaterals, and it will take more time with the stable assets.

The engagement should be seen a perpetual engagement, with perpetual consultation and dashboard maintenance. As gearbox evolves, further monitoring and mathematical modelling will have to be developed.
As significant amount of effort is required to bootstrap the system and to develop mathematical models, we are asking for a minimum of 3 month commitment.

2 Likes

sounds clear to me. RiskDAO has established itself as a reliable provider of solutions and methodologies for assessing the risks of DeFi protocols, so it would be interesting to apply their knowledge to the gearbox’s case, ofc if the DAO approves it.

regarding payment structure:
current contributors get GEAR tokens at 150M FDV, locked for 3 months and then linearly vested over 9 months. so if I understand correctly: 50k is splitted to 33.3k USDC payment and 1,111,111 GEARs according to vestings described above.

1 Like

Snapshot voting is started https://snapshot.org/#/gearbox.eth/proposal/0x9d711e587592e8b4b98f742d7778c876a1284a6452ce8c8eb7529e2e49ca6199

2 Likes

Excited the proposal passed.
The wallet address for our operational expenses, for the USDC payment, is 0x4d1858221250dAC0641D831a4Ff1C9d26f3388a9

The address of the DAO multisig, that will hold the GEAR tokens, is 0xe27B57A99e5CAc0E0f16eE0d5D768126B05422A8

2 Likes