๐Ÿš‘
Risks and Emergency Modules
How the protocol behaves in case of distress

๐Ÿ”Ž TL;DR

    Angle is mostly at risk when a sudden collateral price decrease is combined with a severe drop in demand for leverage by Hedging Agents.
    The protocol has the safety modules and incentives to be able to resist and maintain peg even in extreme market conditions.

โšก Protocol's Risks

As with any stablecoin protocol, the main risk is to be unable to maintain convertibility between stablecoins and collateral, which might affect peg. In this case, governance might have to deter exits with high fees, or to freeze transactions. As long as the protocol remains over-collateralized, there is no risk that it happens.
The protocol becomes under-collateralized if is not hedged enough by HAs, and there are not enough SLPs to over-collateralize the protocol.
Here, we explore all the responses the protocol can give in case of distress. It is interesting to note that it is most at risk when both events occur: decrease in demand for leverage longs, and collateral prices drop (or increase in the price of the asset the protocol tries to peg). Contrary to other stablecoin protocols, the first condition is necessary, and a collateral price drop alone is not sufficient to put the protocol at risk of under-collateralization, and hence bank run.

๐Ÿ›ก๏ธ Hedging Agents Insurance

Hedging Agents can enter the protocol with even very small leverage like x1.01, or x1.1. To be over-collateralized, the protocol doesn't need HAs to take on highly leveraged positions.
Even if demand for perpetual futures decreases, the protocol should remain cheaper than most protocols allowing to get on-chain leverage, like Aave or Compound. In normal times, when demand for leverage is high and almost all the collateral is hedged, transaction fees are set so that it is slightly more expensive to open a position. When demand for perpetual futures drop, fees drop (following the hedging curve) and Angle becomes much cheaper than other on-chain protocols.
It should also be cheaper to get leverage with Angle than with centralized exchanges. With Angle, traders only pay entry and exit fees: there is no funding cost. It is more cost efficient for Hedging Agents that have a long-term horizon to open positions on Angle rather than on a centralized exchange, as this funding rate ususally represents a high cost.
Governance has the ability to vote to activate governance token distribution to Hedging Agents to increase incentives: HAs could be paid to get long (even with really small multipliers) using Angle.
These three effects should help the protocol attract the remaining demand for leverage in times of deleveraging in the market, helping it stay solvent and over-collateralized.

๐Ÿ’ฐ Protocol Surplus

In normal times, the protocol accumulates some surplus from transaction fees and lending returns not distributed to SLPs. It can also generate revenue from governance token sales through the bonding curve, and from collateral prices increase when the collateral of the protocol is not fully covered by HAs.
While everything that can be done with this surplus is still to be determined (like auctions where governance tokens are burnt against a portion of the surplus), the surplus mostly serves as the first buffer to better deal with drops in collateral ratio.

๐Ÿ€ Standard Liquidity Providers Insurance of the Insurance

If the incentives put in place do not attract enough HAs to fully insure the protocol against the volatility of the collateral and if there is not enough surplus accumulated, the protocol can rely on the collateral brought by Standard Liquidity Providers to ensure full convertibility of the stablecoins. There is an equilibrium threshold to expect with Standard Liquidity Providers (SLPs). Indeed, the fewer SLPs there are, the more interesting it becomes to be a SLP because of the multiplier effect that is spread among SLPs.
To prevent the cascading effect that could be induced if SLPs all exit the protocol, a slippage is introduced to discourage SLPs from exiting when the collateral ratio is too low: this is a way to make sure that there will always be SLPs in the protocol. Note here that HAs are never affected by exit restrictions: they can find incentives to enter at any point in time thus bringing extra-liquidity, regardless of whether there has been a price drop or not.
Besides, when the protocol starts to be too unhealthy, as explained in this page, there are recollateralization incentives for SLPs: a fraction of the transaction fees and lending yield going to SLPs can be automatically put aside, and will only be recoverable by SLPs when the protocol is collateralized back again. This means that a new SLP giving money to recollateralize a pool may receive transaction fees and yield accumulated before her arrival in the pool.
Governance can also choose to distribute a bigger fraction of the transaction fees and lending returns to SLPs to make them stay, or to offer more governance tokens through the staking contracts

๐Ÿ’ฑ Dynamic Transaction Fees for Users

In the meantime, transaction fees for users minting and burning will adjust automatically based on the hedging ratio of the protocol. This means that it will be more expensive for users to mint stablecoins (thus adding collateral into the protocol) if the collateral that is already there is not hedged enough by Hedging Agents.
Besides, if the collateral ratio keeps decreasing, governance can choose at its discretion to mitigate bank run scenari by increasing burn transaction fees by inducing a collateral ratio dependency.
You can have a look at Angle's analytics page to follow the different hedge and collateral ratios, and the current transaction fees.

๐Ÿท๏ธ Governance Token Sale

In case where the protocol is on its way to becoming under-collateralized and can no longer sustain full convertibility of stablecoins to collateral, then governance has the ability to change the parameters of the bonding curve to make buying governance tokens cheaper.
Like Maker does when auctions are made where governance tokens are sold against stablecoins, here the protocol proposes to issue tokens at a smaller price to burn some stablecoins and hence increase the collateral ratio of the protocol.

โŒ› Collateral Diversification and Settlement

In the above sections, a protocol with a single collateral type was described. It is important to note that the protocol will accept different collateral types, diversifying the overall risk of the protocol. With asset diversification, Angle is less prone to having the issue where one asset drops and affects the stability of the whole system.
If one collateral becomes too risky, the protocol has the ability to incentivize liquidity providers to contribute more to one pool than to another through staking incentives and transaction fees.
It is also possible to revoke a collateral for a given stablecoin if it reveals too risky: we call this process collateral settlement. If there is collateral left in the pool at this time, it is sent to another contract that is specifically designed to handle the settlement.
The goal of settlement is to choose how much each category of stakeholders is going to get based on their claims, and then to reimburse each users, HAs and SLPs which asked it based on that. This process has been thought to heavily favor governance token and owners.
The following page explains how in details this collateral settlement process works:

๐Ÿ“œ Stablecoins Independence

The protocol's stablecoins are independent from one another, meaning that if one stablecoin fails, other stablecoins will not be impacted. The collateral pools related to each stablecoin are different and separated from one another, and the contracts behind have different addresses.
Division of pools and collaterals
In case of a security breach, to trigger the emergency shutdown at the level of the protocol, all collateral for all stablecoins should be settled.
Last modified 4d ago