Stability peg management
There are a few mechanisms that can be used for peg management. In this article I will explore some existing solutions;
1:1 Deposit & Redemption (USDT, USDC, TUSD, etc)
Arguably the simplest solution. Deposit 1 USD mint 1 USDT. Burn 1 USDT withdraw 1 USD. If 1 USDT > 1 USD, minters deposit USD and sell USDT earning profit. If 1 USDT < 1 USD, arbitrageurs buy 1 USDT and burn for 1 USD.
Collateral (DAI, sUSD, etc)
Create a debt position off of collateral and mint. Deposit 1 ETH mint 700 DAI. As long as sum of collateral ≥ outstanding debt (minted amount), then the system is table. A qualification here is that the debt itself does not need to be stable, the invariant collateral ≥ debt simple needs to hold true.
If any individual collateral position < debt, liquidators can repay debt to that position and claim an equivalent proportion of the collateral.
The above paragraph implies that debt (DAI) will only be purchased if the profit of collateral seized ≥ cost of debt to be repaid. So if the collateral requirement is 120%, this means that the profit margin of said liquidation is ~20%. So if the whole system was collateralized at 120% it means for every $1 of collateral you have $0.83 debt. If the value of the collateral drops by 10% ($0.9), the collateral to be repaid is $0.1 and the amount of debt $0.083. So for every $0.83 you are earning $0.17. This measures the “band” of value for the underlying token at between $1 and $1.17. The further the collateral value moves from 100% the further the stability band.
The higher a system is over collateralized the further the peg can shift.
Banded Bonds / Seigniorage (Basis, etc)
The following is oversimplified for purposes of illustration. Price < 1, sell bonds (ability to redeem in the future) at discount. Price > 1, allow bonds to be sold. If a coin is current trading at $0.9, you allow purchasers to buy discounted bonds, for example, you can purchase the bond at $0.89. The assets you sold are used to bring back to peg, so in BAC example, buy bonds at 0.89 DAI, the 0.89 DAI is added to trading liquidity to bring the peg > 1. For 0.89 DAI you purchased 1 bonded BAC. Now if the price > 1 you can sell the bonded BAC for BAC. So lets say the requirement is > 1.01, now you can claim 1 BAC from your bonded BAC, and sell to 1.01 DAI. The value of 1 bonded BAC ≥ 0.89 DAI and ≤ 1.01 DAI. Essentially the bond absorbs the volatility.
Volatile Bonds (Frax, etc)
The following is oversimplified for purposes of illustration. You have 2 tokens, 1 with the goal of being stable, the other purely volatile. Price < 1, burn stable token for volatile token. Price > 1, burn volatile token for stable token.
Backstop Bonds (stability insurance)
Provide a secondary backstop token, for example USDC. Backstop providers can stake USDC and earn fees (the scope of the fees are not covered here). If price < 1, sell backstop token to maintain peg. If price > 1, sell stable token to purchase backstop token. Peg is only as strong as the available liquidity in the backstop token. Essentially automated Banded (if backstop is table token) or Volatile (if backstop is volatile) Bonds.
Reflex-Index (Reflexer, etc)
Need more research, initial review is Collateral backed, but with collateral wrapped in less volatile index versions of said collateral. Will expand.
Stable Credit
Stable Credit iteration 3 combines 1:1 Deposit & Redemption, along with Collateral to create a peg band. However, as mentioned in Collateral the invariant collateral ≥ debt must be true. For stable credit, this is not true, as for every $1 of collateral, you have $1 minted and ≤ $1 of debt.
If we consider a stable coin as issued debt, then we need to assume the purpose of said debt. Predominantly, leverage. Using DAI as an example, I want to use ETH as collateral, mint DAI, and then sell DAI for more ETH. So if we consider a clean room experiment, and we have 2 ETH which minted 2000 DAI, and we have 1 AMM which has 1 ETH and 1000 (of the 2000 minted DAI). Then the minter of the 1000 DAI would like to leverage their ETH. They would sell their 1000 DAI for another ≤ 1 ETH. So now their position is ≤ 2 ETH and 1000 DAI debt. The AMM has ≤ 1 ETH and 2000 DAI. So the price of 1 DAI in the AMM is actually 2 DAI per ETH sold. So assuming all minters sell for leverage, the band of value for minted DAI is ≥ 0.5 ≤ 1 (disregarding liquidation calculations for now)
When you start reaching sufficient liquidity then most of these calculation no longer matter, but it is important to build alternative solutions into the system for them. Along with 1:1 and Collateral, for iteration 4, we propose Volatile Backstop Bonds.
