AIP - 73: Enable Transmuter rebalances through negative fees

Angle Improvement Proposals
1

This is a proposal to enable the protocol governance multisig to adjust Transmuter fee parameters under certain conditions.

Context

Transmuter maximum exposures

Right now the Transmuter is exposed at 67.86% to EUROC and 32.14% to bC3M (you can check the Transmuter’s exposure to its collateral assets in real time here).
Transmuter comes with variable fees that depend on the exposure to each of the assets in its backing. Here is how the initial setup fees are looking like for Transmuter.

In particular, it’s worth noting that fees are such that there are some exposures for which it’s going to be impossible to mint with one token or burn for another one.
For instance, if approximately 3.7m more agEUR are minted from EUROC, the system will reach its maximum exposure to EUROC, and it’d be impossible to mint more agEUR from EUROC.

At this point, it’d be only possible to mint with bC3M or to borrow agEUR (at a rate probably higher than what is given through agEUR staking contract) with the Borrowing module.

The thing is that the type of users who can mint from bC3M is probably not the same as the one who can mint from EUROC. bC3M requires a Backed KYC and people who can already access bC3M might be less interested in agEUR savings contract because they can get what powers its underlying yield.
As such, in this situation even if there is demand for more agEUR to be minted from EUROC, the protocol may be unable to fill it, and might leave aside some potential for growth.

It’s not impossible that people come to mint with bC3M but this is far less likely.

Support for negative fees

The Transmuter mechanism supports negative fees. These can only be set by the governance multisig.
Reason for having negative fees is to give financial incentives to arbitrageurs to mint from one asset.
While some bC3M may be fine with just minting and buying agEUR at a discount some more advanced players may prefer to simply close their arb in a couple of transactions, meaning minting agEUR from bC3M and instantly burning these agEUR for EUROC earning more EUROC than what they used to buy bC3M in the first place.

With Angle savings infrastructure being in the discussion, should this be implemented, we may get to the point where many more people are coming to mint with EUROC thus bringing the protocol to this limit exposure.

Proposal

Proposal is to enable the governance multisig to set negative fees to open arbitrage opportunities of up to 0.5% whenever the system’s exposure to EUROC reaches 74.5%.
Idea is to let this arbitrage opportunity open till exposure is brought back to 66%.

This means that the protocol would set negative fees for the operation to mint agEUR with bC3M and burn these agEUR for EUROC. This could take place by for instance reducing the minting cost from bC3M to as low as -0.5% and setting the burning cost for EUROC to 0%.

At current levels, if EUROC was to reach 74.5%, bringing exposure to 66% would imply doing a 1.4m€ swap in the reserves to balance reserves of both assets.

The arbitrage fee comes from the fact that Backed takes a .2% minting fee and that arbitrageurs/market makers taking the arb may want to be compensated for this.
The 0.5% proposed here is an upper bound on the fee that can effectively be set by the governance multisig.

In the end and depending on the level of competition between market makers this system would emulate, the arbitrage fee paid by the protocol (which is in fact equivalent to a fee to swap EUROC to bC3M) would in most cases end up being far smaller. To achieve the best outcome, the governance multisig could for instance increase the arbitrage fee little by little till reaching the value at which it becomes profitable for a market maker to take the arb. Like start at 0.2% arbitrage level and increase by slots of 5bp till someone takes the opportunity.

Implementation

Having arbitrage loops within the protocol is something which can be potentially dangerous if not well handled. While this could be implemented onchain, we suggest in the first place to rely on this governance controlled approach so the scope of these rebalances and arbitrage remains under control.

It’s notably important that during those updates the fee to mint with EUROC or to burn for bC3M are increased (above the arbitrage level) so that people cannot infinitely trick the system into opening an arbitrage. Like it shouldn’t be possible to mint EUROC for less than 0.5% otherwise it becomes possible to take close to the arbitrage loop, mint with more EUROC so as to put the system again in the loop.

As soon as the arbitrage has been taken and the protocol exposures rebalanced to the desired level (66% +/- 0.5%), governance should set the fees as they were initially set in the schedule.

Down the line, it’s going to be important to automate these fee adjustments (after audits), but as agEUR liquidity is being built and grown, this approach seems like a fair tradeoff between execution efficiency, cost and security.

Value to the protocol

This proposal guarantees that whenever people have minted a lot of agEUR with EUROC, the protocol will have the tools to increase back again its exposure to bC3M and therefore will be able to sustain the yields it is paying through its savings infrastructure.

It is key in ensuring that what is being built works as well with $1m in TVL as with $1bn

2

Hey,

Wanted to update the proposal with a more automated solution.

While the initial proposal was to enable the governance multisig to set negative fees to open arbitrage opportunities of up to 0.5% in the system, after a careful mathematical analysis, we’re now proposing to take advantage of the support for negative fees in the system to let the rebalances happen automatically.

In practice, the fee schedule we’re pushing for here is the following.

How to interpret this: whenever the exposure to bC3M reaches less than 40%, there is a negative fee to come and mint with bC3M.
For instance at 33% exposure the fee to mint from bC3M is -0.35%, which means that to get 1 agEUR you can bring 0.9965€ worth of bC3M. At which point burning this agEUR for EUROC comes with no cost: an arbitrageur can earn 0.35% from the process of rebalancing the protocol’s reserves.

Note that the evolution of the fees between two thresholds in fees is linear, which means that there is a price discovery in what the market makers minting from bC3M will be able to accept. For instance, the fee between 40% exposure to 33% evolves between 0% to -0.35%: this means that there is going to be a form of competition between market makers to be the first one to take the arbitrage: like one may be fine with a 0.35% discount for the whole process while another one may be fine with a 0.3% bonus for this, in which case this last market maker will take the arb first.

With the fee setup proposed and assuming market makers are ready to accept a 0.35% discount, this would result in a convergence point to 67% EUROC - 33% bC3M (as currently).

The 0.35% estimated arbitrage fee here comes from the fact that Backed takes a .2% minting fee and that arbitrageurs/market makers taking the arb may want to be compensated for this.
The upper bound enabled within the system proposed here is an upper bound on the fee that can effectively be set by the governance multisig.

Given the high burning cost for C3M, these negative fees do not open infinite arbitrage loops within the protocol and there is no risk of having the protocol’s reserves depleted: it will always converge.

People could mint agEUR with EUROC to drastically increase the exposure to EUROC, hence trigerring the negative fees for bC3M, but their arbitrage would only aim bring the protocol back in a situation where it’s exposed to around 67% EUROC - 33% bC3M if they were to burn all the agEUR they got.

Note as well that as part of the fee schedule introduced, 0% mint and burn fees are introduced for EUROC. This is meant to enable people to view the upcoming stEUR contract as the place where you can put your capital at rest, and enter with no fees.
Fees will not always be null but the majority of the time the exposures should lie in an area where minting with EUROC comes with no cost.

Implementation

There are fee setters available in the Transmuter system, the protocol governance multisig will just have to set the fees with the schedule proposed in the Google Sheet.

2 Likes
3

I like setting agEUR minting from EUROC at 0bps as that’s an unlock for achieving singleness of money on-chain (all € stablecoin are at par). It is good PR as people will have no friction to test (and stay) with stEUR. I would be a bit more aggressive with burn fees and keep them to 0% until 55%, then linear to 0.2% at 25% then 100% maybe at 0%. The rationale is that we might want people to exit the system frictionlessly at scale. But let’s keep it like that and learn.

The curve for EUROC is maybe a bit steep between 74.5% and 75% (0.10% to 100% fees), but safety against EUROC is probably wise. This could be solved by adding more fiat-backed stablecoin later.

2 Likes
4

Thanks for your reply, just one thing on the burn fees at 0% till 55% exposure to EUROC: it’s doable but it means that technically the convergence point for the reserves of the protocol would be more at 57/58%-42%, as it’ll increase the arbitrage window for market makers.

Also for everyone, here is a summary of the fee schedule as the spreadsheet can be tricky

For EUROC:

  • Mint: 0% fees till 73% exposure, 0.05% at 74% exposure, 0.1% at 74.5% and 100% above 75%. This means that it’s impossible to mint EUROC above a 75% exposure
  • Burn: 100% burn fees below 50% exposure (we cannot have less than 50% exposure to EUROC), 0.1% fees at 50.5%, 0.1% at 59% and 0% at 67% exposure and beyond

For bC3M:

  • Mint: -0.4% fees till 25% exposure, -0.35% at 33%, 0% at 40% exposure, 0.2% at 49.5% exposure and 100% above 50% exposure (exposure to bC3M cannot be greater than 50%)
  • Burn: 100% burn fees till 25% exposure (exposure to bC3M cannot be less than 25%), 0.5% at 26% exposure and beyond
3 Likes