Subscribe to Angle Protocol | Learn About Stablecoins
<100 subscribers
Subscribe to Angle Protocol | Learn About Stablecoins
<100 subscribers
Share Dialog
Share Dialog
Decentralized stablecoin protocols, unlike their centralized counterpart that can accept fiat currency as collateral, need some processes to maintain the tokens they issue at peg. In general, it implies finding some way to absorb the volatility of the collateral used to back the stablecoins and to deal with changes in supply and demand for the stablecoins.
In Part 2, we explored over-collateralized decentralized protocols which, similarly to Maker, rely on collateralized debt positions. The bottom line was that, although robust to bank-runs and hence to massive drops in demand, these protocols tend to be capital inefficient. To create stable tokens more efficiently, there are many different types of protocols, quite often non strictly over-collateralized or even under-collateralized: they are often referred to as algorithmic stablecoins. We will try to explore the biggest trends in this category.
One popular idea amongst algorithmic stablecoin protocols is to directly affect the supply through rebasing. This is done by expanding automatically the supply of coins when price is too high, and contracting it when price too lo.
Rebasing is the idea on which Yam or Ampleforth are built. The way supply is changed in these cases is by directly increasing or decreasing the balance of stablecoins in user wallets based on the price of the tokens. Yet, this is very changed in these cases is by directly increasing or decreasing the balance of stablecoins in user wallets based on the price of the tokens. Yet, this is very hard to achieve in practice in a stable way, and it can hardly resist crisis where faith is lost in the token and everyone wants to get rid of it.
Another decentralized and capital efficient under-collateralized model is the seigniorage share model of projects like Basis Cash or Empty Set Dollar.
Globally, the idea with such designs is that stablecoins are sold by the protocol against collateral. Proceeds from this sale are shared between a pocket of stakeholders owning a volatility token (often a governance token) either through dividends or through token appreciation (in the case where minting is done by buying governance tokens). A fraction of the sale may also go to the protocol’s reserves.
Stablecoins of such designs can then be burnt against volatility tokens (newly issued or auctioned), and sometimes against some collateral too.
The price dynamic of the volatility token is most of the time the inverse of the stablecoin: when stablecoin supply increases (and price is likely to decrease), the price of the volatility token increases, and conversely when stablecoin supply decreases, the price of the volatility token decreases. Some Examples
There are several different design choices with this model: how does redemption works? Is there a reserve of collateral? Is the reserve made up of collateral that keeps a stable value like USDC or not like ETH?
Luna adopts such a mechanism involving a volatility token LUNA for its UST stablecoin.
This video from Terra.money is a very good explainer of this mechanism.
If the price of UST increases above its peg outside the protocol, users could swap LUNA against UST considered to be worth 1$ by the protocol. Those newly minted UST can then be sold on the open market pushing price back down to peg. This mechanism works backwards if the price is below peg (UST are swapped for LUNA).
The FRAX model is a more careful version of this seigniorage share mechanism. Besides the volatility token called FXS, it indeed involves USDC as collateral. A fraction of the value generated when issuing stablecoins is stored in a USDC reserve, and the rest has to be paid in the volatility token (thus increasing its value). In this case not all the proceeds of the stablecoin sale are given to volatility token owners.
In this sense, as Gyroscope team points out in their excellent article about algorithmic stablecoins, “FRAX is equivalent to owning some of USDC and some of a seigniorage shares stablecoin”. This can make the system more robust to some black-swan events.
As a side note, the interest of having an under-collateralized USD stablecoin backed by USDC is that it can act as a multiplier for the amount of USDC in circulation. With 100 USDC you can potentially mint 120 other USD stablecoins.
The general problems with such designs is that what is backing the stablecoin is a new asset (often a governance token) supporting the system in, to use Gyroscope terms, a self-fulfilling way. If the system is in crisis, the value of the governance token backing the stablecoins can decrease which may cause a collapse in the system.
Besides, while this is the case but to a lesser extent for some protocols like FRAX, in the mechanisms from above, the value of the reserve in dollar (or in the asset the stablecoin is pegged to) remains inferior to the number of stablecoins minted. These protocols are therefore naturally under-collateralized and face an inherent bank run risks. There may even be an accelerating effect: having the health of the reserve decreasing entices other stablecoin holders to try exchange it against collateral thus exhausting the reserve even more. This also creates a loss in confidence in the value of the supporting volatility token accelerating the failure of the system.
For some protocols, this approach did not work at all (see Empty Set Dollar price chart below), for others, it simply does not work at all times. On May 24th, after an important drop in the crypto market, UST traded as low as 0.89$ and while the trend for it should normally be different, the price of the governance token also decreased at the same time.
Luna Price in May (source: CoinMarketCap)
UST Price in May (source: CoinMarketCap)
Empty Set Dollar Failure (source: CoinMarketCap)
One solution some algorithmic under-collateralized protocols have already tried to limit bank runs is to simply prevent exits. This is to some extent what the FEI protocol did when it launched and had a stability mechanism called direct incentives. FEI tried to mitigate bank run risk by incurring quadratic redeeming fees from the FEI reserve: this made the system very illiquid especially during supply contractions and created an important liquidity crisis.
Now FEI shifted back to a FRAX like model. Both systems seem to be stable and hold the peg so far, shoutout to them! This is the proof that there is a potential for these designs to work, but given all the risks explained above, this would deserve to be more thoroughly battle tested.
A last remark about projects with a secondary volatility token is that they are hard to bootstrap and rather slow to grow. The volatility token gains in value only if the protocol gains in traction, hence creating a chicken and egg problem.
Algorithmic stablecoins are most of the time capital efficient. Except for rebasing mechanisms, they rely on a secondary volatility tokens which makes them hard to bootstrap
Because of their under-collateralized nature, most are not bank-run resistant
There have been many examples of such designs exhibiting extreme periods of volatility in the past
Angle’s design is inspired by algorithmic stablecoins, while not retaining most of their drawbacks.
Similarly to algorithmic stablecoins:
Angle is capital efficient: minting 1 of stablecoin requires 1 of collateral and no more
Angle is decentralized
Angle has a governance token which can be sold at a discount in some occasions to decrease stablecoin supply
But Angle improves on this design: Angle is bank-run resistant: the protocol is over-collateralized, and well prepared in case of a drop in collateral price
Within Angle, no funds used to mint stablecoins are burnt, and everything stays in the protocol’s reserves. They are even used to accumulate yield by using Yearn-like strategies
Angle is easy to bootstrap and does not require some amount of reflexivity regarding its governance token: it does not need a critical volume or mass to be reached to function effectively
Angle vs. Algorithmic Stablecoins
It is interesting to note that most over-collateralized stablecoin mechanisms like Maker also incorporate some form of seigniorage share design at some point. In the case of Maker, this is used during debt auctions when the first layer of security made up of vaults does not suffice. Angle also has in it a seigniorage share mechanism that can be activated. Yet in this case, it is a last resort when the two additional layers of security made up of HAs and SLPs have failed. The protocol is therefore guaranteed to be more robust than basic seigniorage share mechanisms.
Decentralized stablecoin protocols, unlike their centralized counterpart that can accept fiat currency as collateral, need some processes to maintain the tokens they issue at peg. In general, it implies finding some way to absorb the volatility of the collateral used to back the stablecoins and to deal with changes in supply and demand for the stablecoins.
In Part 2, we explored over-collateralized decentralized protocols which, similarly to Maker, rely on collateralized debt positions. The bottom line was that, although robust to bank-runs and hence to massive drops in demand, these protocols tend to be capital inefficient. To create stable tokens more efficiently, there are many different types of protocols, quite often non strictly over-collateralized or even under-collateralized: they are often referred to as algorithmic stablecoins. We will try to explore the biggest trends in this category.
One popular idea amongst algorithmic stablecoin protocols is to directly affect the supply through rebasing. This is done by expanding automatically the supply of coins when price is too high, and contracting it when price too lo.
Rebasing is the idea on which Yam or Ampleforth are built. The way supply is changed in these cases is by directly increasing or decreasing the balance of stablecoins in user wallets based on the price of the tokens. Yet, this is very changed in these cases is by directly increasing or decreasing the balance of stablecoins in user wallets based on the price of the tokens. Yet, this is very hard to achieve in practice in a stable way, and it can hardly resist crisis where faith is lost in the token and everyone wants to get rid of it.
Another decentralized and capital efficient under-collateralized model is the seigniorage share model of projects like Basis Cash or Empty Set Dollar.
Globally, the idea with such designs is that stablecoins are sold by the protocol against collateral. Proceeds from this sale are shared between a pocket of stakeholders owning a volatility token (often a governance token) either through dividends or through token appreciation (in the case where minting is done by buying governance tokens). A fraction of the sale may also go to the protocol’s reserves.
Stablecoins of such designs can then be burnt against volatility tokens (newly issued or auctioned), and sometimes against some collateral too.
The price dynamic of the volatility token is most of the time the inverse of the stablecoin: when stablecoin supply increases (and price is likely to decrease), the price of the volatility token increases, and conversely when stablecoin supply decreases, the price of the volatility token decreases. Some Examples
There are several different design choices with this model: how does redemption works? Is there a reserve of collateral? Is the reserve made up of collateral that keeps a stable value like USDC or not like ETH?
Luna adopts such a mechanism involving a volatility token LUNA for its UST stablecoin.
This video from Terra.money is a very good explainer of this mechanism.
If the price of UST increases above its peg outside the protocol, users could swap LUNA against UST considered to be worth 1$ by the protocol. Those newly minted UST can then be sold on the open market pushing price back down to peg. This mechanism works backwards if the price is below peg (UST are swapped for LUNA).
The FRAX model is a more careful version of this seigniorage share mechanism. Besides the volatility token called FXS, it indeed involves USDC as collateral. A fraction of the value generated when issuing stablecoins is stored in a USDC reserve, and the rest has to be paid in the volatility token (thus increasing its value). In this case not all the proceeds of the stablecoin sale are given to volatility token owners.
In this sense, as Gyroscope team points out in their excellent article about algorithmic stablecoins, “FRAX is equivalent to owning some of USDC and some of a seigniorage shares stablecoin”. This can make the system more robust to some black-swan events.
As a side note, the interest of having an under-collateralized USD stablecoin backed by USDC is that it can act as a multiplier for the amount of USDC in circulation. With 100 USDC you can potentially mint 120 other USD stablecoins.
The general problems with such designs is that what is backing the stablecoin is a new asset (often a governance token) supporting the system in, to use Gyroscope terms, a self-fulfilling way. If the system is in crisis, the value of the governance token backing the stablecoins can decrease which may cause a collapse in the system.
Besides, while this is the case but to a lesser extent for some protocols like FRAX, in the mechanisms from above, the value of the reserve in dollar (or in the asset the stablecoin is pegged to) remains inferior to the number of stablecoins minted. These protocols are therefore naturally under-collateralized and face an inherent bank run risks. There may even be an accelerating effect: having the health of the reserve decreasing entices other stablecoin holders to try exchange it against collateral thus exhausting the reserve even more. This also creates a loss in confidence in the value of the supporting volatility token accelerating the failure of the system.
For some protocols, this approach did not work at all (see Empty Set Dollar price chart below), for others, it simply does not work at all times. On May 24th, after an important drop in the crypto market, UST traded as low as 0.89$ and while the trend for it should normally be different, the price of the governance token also decreased at the same time.
Luna Price in May (source: CoinMarketCap)
UST Price in May (source: CoinMarketCap)
Empty Set Dollar Failure (source: CoinMarketCap)
One solution some algorithmic under-collateralized protocols have already tried to limit bank runs is to simply prevent exits. This is to some extent what the FEI protocol did when it launched and had a stability mechanism called direct incentives. FEI tried to mitigate bank run risk by incurring quadratic redeeming fees from the FEI reserve: this made the system very illiquid especially during supply contractions and created an important liquidity crisis.
Now FEI shifted back to a FRAX like model. Both systems seem to be stable and hold the peg so far, shoutout to them! This is the proof that there is a potential for these designs to work, but given all the risks explained above, this would deserve to be more thoroughly battle tested.
A last remark about projects with a secondary volatility token is that they are hard to bootstrap and rather slow to grow. The volatility token gains in value only if the protocol gains in traction, hence creating a chicken and egg problem.
Algorithmic stablecoins are most of the time capital efficient. Except for rebasing mechanisms, they rely on a secondary volatility tokens which makes them hard to bootstrap
Because of their under-collateralized nature, most are not bank-run resistant
There have been many examples of such designs exhibiting extreme periods of volatility in the past
Angle’s design is inspired by algorithmic stablecoins, while not retaining most of their drawbacks.
Similarly to algorithmic stablecoins:
Angle is capital efficient: minting 1 of stablecoin requires 1 of collateral and no more
Angle is decentralized
Angle has a governance token which can be sold at a discount in some occasions to decrease stablecoin supply
But Angle improves on this design: Angle is bank-run resistant: the protocol is over-collateralized, and well prepared in case of a drop in collateral price
Within Angle, no funds used to mint stablecoins are burnt, and everything stays in the protocol’s reserves. They are even used to accumulate yield by using Yearn-like strategies
Angle is easy to bootstrap and does not require some amount of reflexivity regarding its governance token: it does not need a critical volume or mass to be reached to function effectively
Angle vs. Algorithmic Stablecoins
It is interesting to note that most over-collateralized stablecoin mechanisms like Maker also incorporate some form of seigniorage share design at some point. In the case of Maker, this is used during debt auctions when the first layer of security made up of vaults does not suffice. Angle also has in it a seigniorage share mechanism that can be activated. Yet in this case, it is a last resort when the two additional layers of security made up of HAs and SLPs have failed. The protocol is therefore guaranteed to be more robust than basic seigniorage share mechanisms.
Angle is not the only non-under-collateralized stablecoin system built to be capital efficient and still robust to bank run: meta-stablecoins like yUSD issued by Yearn or mStable also have these advantages.
Meta-stablecoins are stablecoins that are composed of a basket of other stablecoins. The idea is that having the basket diversifies the risks of the individual stablecoins. This induces a new risk for these stablecoins called composability risk, or the risk that a problem in one token of the basket affects the stablecoin. In yUSD for examples, if one of the coins backing (USDT, Dai, USDC, TUSD) fails, then yUSD would break its peg.
Some protocols, like Gyroscope which is not live yet and that we have already mentioned in this article, are taking action to mitigate composability risk. While such systems are more robust, they just push the problem further. They may be less subject to the failure of a single element, but in case of a systemic crisis, they would still be at risk.
Another issue with these stablecoins is that their market cap will not be able to grow above the market of the stablecoins they are composed of. These stablecoins couldn’t be used to boost alone the entire DeFi ecosystem.
At Angle Labs, we think composability between protocols is great, and we are also committed to mitigating this risk. But contrary to meta-stablecoins, other protocols are not central to the system, and the protocol could work even in a systemic crisis where some major DeFi components fail. Another advantage of Angle is that the size of the protocol is only bounded by the value of the assets accepted as collateral, and not by that of other stablecoins.
The forward-looking statements in this announcement are subject to numerous assumptions, risks and uncertainties which are subject to change over time. Such assumptions, risks and uncertainties could cause actual results or developments to differ materially from the results and developments anticipated by Angle Labs Inc. Even if our anticipated results and developments are realized, such results and developments may nevertheless fail to achieve any or all of the expected benefits anticipated by this announcement. We reserve the right to change the plans, expectations and intentions stated and implied herein at any time and for any reason or no reason, in our sole and absolute discretion, and we undertake no obligation to update publicly or revise any forward-looking statement, whether as a result of new information, future developments or otherwise.
This announcement is not intended to provide legal, financial or investment or other advice and we recommend that you do not rely on, and do not make any financial or other decision based, on this announcement.
Thanks to Guillaume Nervo and Baptiste Debever
Angle is not the only non-under-collateralized stablecoin system built to be capital efficient and still robust to bank run: meta-stablecoins like yUSD issued by Yearn or mStable also have these advantages.
Meta-stablecoins are stablecoins that are composed of a basket of other stablecoins. The idea is that having the basket diversifies the risks of the individual stablecoins. This induces a new risk for these stablecoins called composability risk, or the risk that a problem in one token of the basket affects the stablecoin. In yUSD for examples, if one of the coins backing (USDT, Dai, USDC, TUSD) fails, then yUSD would break its peg.
Some protocols, like Gyroscope which is not live yet and that we have already mentioned in this article, are taking action to mitigate composability risk. While such systems are more robust, they just push the problem further. They may be less subject to the failure of a single element, but in case of a systemic crisis, they would still be at risk.
Another issue with these stablecoins is that their market cap will not be able to grow above the market of the stablecoins they are composed of. These stablecoins couldn’t be used to boost alone the entire DeFi ecosystem.
At Angle Labs, we think composability between protocols is great, and we are also committed to mitigating this risk. But contrary to meta-stablecoins, other protocols are not central to the system, and the protocol could work even in a systemic crisis where some major DeFi components fail. Another advantage of Angle is that the size of the protocol is only bounded by the value of the assets accepted as collateral, and not by that of other stablecoins.
The forward-looking statements in this announcement are subject to numerous assumptions, risks and uncertainties which are subject to change over time. Such assumptions, risks and uncertainties could cause actual results or developments to differ materially from the results and developments anticipated by Angle Labs Inc. Even if our anticipated results and developments are realized, such results and developments may nevertheless fail to achieve any or all of the expected benefits anticipated by this announcement. We reserve the right to change the plans, expectations and intentions stated and implied herein at any time and for any reason or no reason, in our sole and absolute discretion, and we undertake no obligation to update publicly or revise any forward-looking statement, whether as a result of new information, future developments or otherwise.
This announcement is not intended to provide legal, financial or investment or other advice and we recommend that you do not rely on, and do not make any financial or other decision based, on this announcement.
Thanks to Guillaume Nervo and Baptiste Debever
No comments yet