Automated Risk Manager
Risk management is one of the most time-consuming jobs for any lending protocol. For any given asset, there is a wide range of risks and parameters that one needs to observe. In many cases, over 200 risk properties need to be monitored at any point in time for any individual asset.
To scale risk management, there are two approaches a protocol could take.
Delegate risk to a risk manager: This is similar to what is right now happening with Aave, ZeroLend, and other lending protocols, where there is a dedicated party (e.g., Chaos Labs, Gauntlet) that is actively managing risk for the entire protocol. The upside to this is that there is a professional firm that is managing risk for the protocol, and in most cases, we can assure that the risk manager will do their job. But this is not a scalable solution. As more markets and assets are added to the pipeline, risk management cannot be scaled.
Isolate risk and remove the need for a risk manager: This is similar to permissionless protocols like Morpho/Uniswap, where anybody can create a pool, and risk is assumed by the user. The upside is that risk is not at all managed by the protocol, and the protocol can scale to list as many assets as possible. However, the downside is that without a professional risk manager, users who otherwise can’t evaluate risk for themselves will risk themselves. Protocol designers should ideally design protocols so that users don’t end up hurting themselves.
So, the solution ideally is to find something that can:
Actively manage the risk of a pool.
Work at a large scale, to 1000s of assets at the same time.
Introducing the Automated Risk Manager (abbr. ARM)
The ARM is a simple smart contract with a set of rules that make decisions on the various risk parameters of the protocol.
The smart contract is a decentralized strategy with business logic written into the contract itself. Its execution is automated using a keeper service like Chainlink Automations or Gelato Functions.
The ARM can have rules that are conservative and update risk parameters at fixed time intervals.
Some basic rules can be:
Ensure that the supply cap for a given pool and an asset inside is no more than 30% of the total supply of a token.
If liquidity for a token (on Uniswap, Balancer, Curve, etc.) drops below a certain limit, the LTV for that asset is automatically set to 0%.
If there is a lot of volatility for a given asset then gradually reduce the LTV for the asset.
If the demand for an asset reaches the supply cap, the cap should be increased again to allow more supply.
Avoiding Manipulation
Risk management actions can be executed over a period of 3 days using TWAP data that is fed to it daily. This ensures that the strategy takes into account untampered data and makes safer decisions.
In the worst case, a human can override the decisions made by the ARM at any time.
Last updated