Mutables aims to be a protocol for low-risk investing. We want to reduce market and systemic risk as much as possible. It’s a crypto protocol so we cannot remove all risks. The user will still need to carefully manage private keys and, of course, there is Layer 1/2 protocol risk, risk of code exploits etc. But risk comes in many different guises. Things like protocol complexity and idiosyncratic behaviour can also catch users unaware, so we want Mutables to be a very intuitive way for ordinary investors to earn yield, without the complexity that comes with a lot of DeFi services.
When users holds a position in Mutables, they should know that they will make money when the price of the underlying asset goes up and lose money when the price goes down. This seems like an obvious thing to say but it is not always the case with synthetic asset protocols. They shouldn’t have to worry about liquidations or any other side effects of the protocol. They shouldn’t need to track to manage risks on a daily or hourly basis. They should be able to mint mutables and sit back and wait patiently for returns with a clear understanding of what the outcomes will be as market prices fluctuate.
However, such a system is quite difficult to make work in practice. Like a swan that swims gracefully on a lake, there can be a lot of intense paddling going on below the surface. To provide the smooth user experience we want for the mutable minter requires some additional effort from other actors in the system. Below we dive into the complexity inherent in synthetic asset protocols, in search of a design that meets our needs.
The Complexity of Synthetic Assets
The leader in the synthetic asset space is Synthetix. The debt pool model they pioneered is a great example of protocol engineering and mechanism design. It has been incredibly successful but it also has some minor quirks, in that some of it’s behaviours are a little counter-intuitive.
For example, when you mint a BTC synth you might expect to get full exposure to a BTC price increase, so that when the BTC price doubles, this would double the value of your overall position. But this is usually not the case. While the value of the synthetic BTC (sBTC) behaves as expected, the net profit for the minter depends on what other synths are in the pool. As BTC increases in value, the total debt in the system also increases and so the synth BTC minter is likely to owe more to the protocol, which eats into their BTC profit. Let’s look at an example:
We can see that, while User 1’s BTC goes up $15K, their net position only goes up $6K. This is because User 1 is responsible for 60% of the debt pool, so their debt increases when the BTC price increases causes the debt pool to grow. Much of their $15K gain in cancelled out by an $9K increase in debt.
Next let’s introduce inverse synths, which are synths that behave in the opposite way to a normal synth i.e. their price goes down when the underlying asset price goes up. If BTC goes up in price by $1K, inverse BTC goes down in price by $1K. If there happens to be an equal number of inverse BTC synths and normal BTC synths in the pool, then synth BTC minters would get the full benefit of any price increase. The inverse BTC synth losses would cancel out the synth BTC gains, leaving the total debt unchanged, assuming all other synth prices don’t change. Let’s look at an example with inverse synths.
We can see that when the pool is “balanced” in this way, in that the amount of inverse synths equals the amount of synths, then the profit analysis is much more predictable. One way to strive for this balance might be to provide protocol incentives for users to mint inverse synths (in fact Synthetix already does this), but this doesn’t seem to be effective judging by the inverse synths in circulation. The recent success of cross-asset swaps across Synthetix and Curve is driving phenomenal demand for BTC and ETH synths, which is exacerbating the imbalance between synths and inverse synths. Therefore, in practice, the number of normal synths in circulation far outweighs the number of inverse synths and so the holder of synth BTC would have a somewhat unpredictable gain when the BTC price goes up. With Mutables we would prefer to have a much more intuitive connection between BTC price increases and the value of the user assets. The balanced pool example provides an important insight that might help us find a way forward.
A related issue is that inverse synths (e.g. iBTC) must be frozen if the base asset (e.g. BTC) price moves more than 50% from its initial value. So, in times of high BTC and ETH volatility, iETH and iBTC must be purged and redeployed by the Synthetix core team. This impacts usability as it creates uncertainty around when these assets can be traded, making it harder for minters to manage risk.
Another counter-intuitive outcome is that minters of sUSD are generally shorting an index of ETH and BTC. This is because the vast majority of synths minted are sUSD, sETH and sBTC. So when ETH and BTC increase in value, the total debt increases, and sUSD holders are left owning a share of this bigger debt pool. Let’s work through an example of this:
When we compare this to Example 1, we can see that User 2 has now made a significantly larger loss just by holding a USD-linked synth in a bull market for BTC.
As SNX inflation has been quite generous, minters are well rewarded for their role in managing these risks and complexities. But it does raise the question: could we design an alternative model that simplifies the design, and therefore improves the overall user experience of minters while retaining the ability to reduce DEX swap slippage using synthetic assets?
Does The World Need Yet Another Dollar-Pegged Asset?
Mutables are designed for frictionless exchange. A mutable can be turned into any other mutable, where we can exchange muBTC into muETH with zero slippage. We also want people to be able to trade dollar-pegged mutables to and from any other mutable (e.g. muETH, muBTC). But could we avoid creating another dollar-pegged mutable? After all, we already have Empty Set Dollar. So we’d really like to avoid creating a new financial instrument pegged to the dollar i.e. “muUSD”. We want to make ESD a fully-fledged mutable, so that it can be seamlessly traded with other mutables.
It would be extremely powerful if we could make ESD a mutable. If we could, then ESD would not only be used as collateral to back mutable issuance, it would also be used widely in liquidity pools on the various DEX protocols that would utilise the Mutables protocol. This increase in demand for ESD would increase the returns to ESDS holders. But is it possible to do this?
The Synthetix debt pool model makes this tricky to implement as all synths are over-collateralised. If we were to copy their approach, we would need a theoretical dollar-pegged mutable (let’s call it “muUSD”) to be overcollateralised by ESD. In other words, we couldn’t allow users to swap directly from ESD to muUSD, as the muUSD wouldn’t have enough collateral to back it.
Why is this? It’s because the power of synth swaps in Synthetix comes from the fact that the protocol is basically swapping collateral from one synth to another. A swap from sBTC to sETH is possible because the exchange will be done at market prices as determined by a price oracle. If one sBTC was backed by collateral that is worth 5x the price of BTC, then we could swap the sBTC for 15 sETH (assuming a 15:1 BTC:ETH price ratio) and now the newly minted 15 sETH would be backed by enough collateral to retain the 5x collateral ratio (once the sBTC is burnt). The high collateralisation ratio here is because of three factors:
(a) BTC and ETH are both volatile assets and their price movements can increase the debt owed by users
(b) when a user mints a synth they take on a share of the debt pool and so a user’s debt is not just a function of what they do, it is a function of the behaviour of all the other synth holders, so even if I hold a price stable asset my share of the debt pool is also volatile
(c) the Synthetix token (SNX) that is used as collateral is also volatile.
With dollar-pegged assets like sUSD and ESD, the underlying asset is less volatile but the other sources of volatility remain. With Mutables we remove (c) by using a price stable asset (ESD) as collateral. But we cannot make ESD work as a mutable with a debt pool model because of (b). A debt pool essentially collectivises the risk of all the minters. As we described above, sUSD then becomes a short position on ETH/BTC, as the debt owed by sUSD minters grows when BTC/ETH prices increases. So an ESD mutable would still need to be over-collateralised in the debt pool approach.
So it seems likely that to make ESD work as a mutable, we need to adjust this idea of collectivised risk and maybe move away from the debt pool approach.
Before we outline the proposed solution let’s summarise our design goals:
Simplicity: we would like to build a synthetic asset protocol where the synthetic asset behaves the same way as the underlying asset so that when a user mints a mutable they get the full upside and downside exposure that they would get if they held the underlying asset. ESD as first-class asset: allow ESD to frictionlessly trade with other mutables i.e. to behave as a mutable. No liquidations for minters: shift the risks associated with asset price movements to more experienced investors.
A thought exercise: what if we could have a debt pool modelled on the Synthetix debt pool that is balanced at all times? In other words, what if every time a user minted a mutable, the protocol took the other side of this position? When a user minted 3 muBTC, the protocol would quietly mint 3 inverse muBTC. Or when a user minted 10 muETH, the protocol would quietly mint the same amount of inverse muETH. Let’s call these inverse positions counterweights and work through an example.
A user mints 1 muBTC representing a synthetic position on BTC that they want to hold. If BTC is priced at $30K, they provide 30K ESD collateral to mint 1 muBTC. Note that there is no over-collateralisation here. The protocol mints a counterweight, which is the opposite position to the user.
Scenario 1: BTC goes up 10%
If BTC goes up 10%, the protocol would pay the minter 33K ESD when they burn their muBTC. The 3K ESD is added to the debt of the counterweight. But note that the counterweight was created by the protocol with no collateral. Instead we propose that counterweights are backed by protocol reserves (which we will come back to later). So now the protocol owes an additional 3K ESD. The protocol will need to carefully manage its reserves to ensure it can always pay out gains to minters.
Scenario 2: BTC goes down 10%
If BTC goes down 10%, then the protocol owes 27k ESD to the minter. The minter made a loss. The 3K ESD the minter lost is now used to reduce the debt of counterweight (and therefore the debt of the reserves). Once the minter closes their position, the counterweight profit on the position becomes part of the reserve.
Note that the counterweight has unlimited loss exposure (BTC could go up 10x). The minter has limited downside risk (as BTC can only go to zero). Let’s be clear here: given that the protocol would start with zero reserves, this doesn’t seem to be a credible solution.
To start to address this, let’s add in fees. The Mutables protocol will earn fees from the swaps that it enables. As counterweights make up 50% of the debt pool, they will earn 50% of these swap fees (in fact, given the additional risks they take on it would be acceptable if they earned a higher percentage). So over time the protocol will build up a reserve that it can use to cover losses from counterweight positions. This is a similar approach to how Bancor protocol covers impermanent loss from fees earned. As long as the volume of swaps is high compared to the asset price volatility, fees would cover any losses made by counterweights. The exception here is when an underlying asset appreciates rapidly in a short period of time, where the gains in value to the mutable minters (which equate counterweight losses) are likely to be much larger than the fees received during this time period. This is a big concern. It is possible that sudden price increases could cause a “bank run” if fears grow that the protocol is unable to pay its debts. We come back to this issue below.
Putting bank runs aside for now, it does seem like counterweights have some positive implications:
users that mint muBTC gain full exposure to BTC on the way up AND on the way down, as they don’t need to worry about changes in the overall debt pool. there is no longer any need for freezing / purging as the protocol will be designed to handle large changes in the price of underlying assets (more on this below). as minters of a (theoretical) muUSD are no longer exposed to fluctuations in the debt pool, they don’t need to be over-collateralised, so now we can have one muUSD collateralised by one ESD. In fact, here is no need for a muUSD mutable at all. We can just promote ESD to a fully-fledged mutable. in fact, all mutables are no longer exposed to fluctuations in the debt pool, and now counterweights take on any debt that accumulates in the system, so we do not need over-collaterlisation for any mutable. Every mutable can be initially backed by its market value (at the time of minting) in ESD.
So while we are left with some major issues, we have made significant progress towards our design goals. We have moved from a design where minters have collective responsibility for the debt pool to one where the protocol (plus, as we will see soon, the Mutables token holders) take on that role. In effect, we have removed the debt pool altogether and replaced it with a protocol reserve. It’s also worth noting that counterweights are also not real in any sense. In practice, the protocol is just tracking reserve debts as mutables are minted and burnt. Net payouts to minters increase reserve debts and net losses made by minters decrease reserve debts.
Let’s now go back to the “bank run” issue. It’s clear we need a significant amount of reserves to be a credible solution. We need risk management strategies for those reserves e.g. the ability to hedge against various outcomes. Also at the start, while there are no reserves, the protocol is unlikely to be considered a credible way to get exposure to the underlying assets. We can’t just say that fees are coming soon, and the protocol will pay users back when the protocol starts making a profit. We need a strategy to build up the reserves. Let’s look at what that might look like.
To ensure the protocol has enough reserves to demonstrate credibility at all times, we will provide reserve bonds as incentives. To ensure stability the protocol reserves must be owned by the protocol. This contrasts with a typical liquidity pool, where users own the liquidity provided and can withdraw it any time. With Mutables, minters need to be able to determine at all times that their positions are adequately backed. The protocol-controlled reserves will be acquired by selling bonds, where future Mutables tokens are distributed to users who provide ESD to the reserve. We expect that a large percentage of the Mutables tokens will be distributed in this way. The process will also act as an early price discovery mechanism for Mutables tokens, as we will look to run bond auctions to bootstrap the reserves.
The number of tokens distributed per bond will be determined based on a target amount of reserves. We will set a target for reserves and when we are below target the Mutable token rewards for bond purchasers will be higher. As the amount of reserves committed approaches the target these rewards will be reduced. But the reserve incentives are not just for the bootstrap period. They can also kick in at any time reserves are low. The reserve target will be adjusted over time, so incentives will also fluctuate over time as mutables are minted and reserves start to accumulate from fees.
If the reserve targets are not achieved in the expected timeframe, the Mutables token holders can adjust some parameters through a governance process. Governance can adjust the bond parameters, including the interest rate on the bond and the bond duration. They can also adjust swap fees, adjust inflation or put minting limits on some underlying tokens that are considered high risk.
More details on the targets, incentives, tuning parameters and vesting schedule will be disclosed closer to launch.
Reserve Risk Management Strategies
Careful governance oversight will be required when approving the underlying assets that the Mutables protocol supports, so that the Mutable token holders can ensure that the protocol only supports assets that are not expected to appreciate rapidly. It should be pointed out that while initially we will encourage users to back newly minted mutables with ESD, there is no requirement for the protocol reserves to only be stored in ESD. It would be expected that the reserves will diversify over time and that the reserves will be used to implement various hedging strategies. For example, if there is a sharp increase in amount of BTC that is minted, governance may decide to increase the amount of BTC in the reserves. In this way, if the price of BTC increases rapidly, the protocol reserves would also increase, thus helping the protocol cover additional debts to BTC minters. We also expect that many of the hedging strategies can be automated, so that in the scenario above the protocol will dynamically adjust BTC portion of the reserves as BTC minting and burning fluctuated. This would help limit the burden on governance decision makers.
The protocol will also aim to hold liquidity pool tokens for the associated DEX protocols we work with. In this case, rather than holding BTC in the reserves, the protocol could hold pool tokens for a pool of BTC-pegged assets. This control over liquidity will be key to the success of the protocol. It will give us strategic leverage when working with DEXes.
Future versions of the protocols could also use options and futures to manage reserve risks. This would provide a more capital efficient mechanism to protect against downside risk.
In the case of a black swan event, the strategy of last resort will be to issue additional Mutables tokens to cover debts. Given the likelihood that such an event will coincide with a reduced price for the Mutable tokens, it could entail a large dilution event for Mutables token holders, so Mutables token holders need to be aware of this risk.
We now feel like we have a very simple and easy to understand protocol for investors to earn great returns. Instead of buying and holding a BTC or ETH position, they can now get exposure to these assets along with a nice yield. A minter only needs to provide ESD, so their collateral value will never change. Once they have minted a mutable they will earn rewards paid out in Mutables tokens and as protocol adoption grows, they will also earn swap fees. As the protocol reserves are owned by the Mutables DAO, the minters can be confident that their positions are rock solid under any market conditions. And minters do not need to worry about liquidations, as they do with other DeFi protocols. Overall we have designed a great user experience which we will believe is the key to driving mass adoption of DeFi.