The use of stablecoins has exploded in the last year, yet fewer and fewer people understand how stablecoins actually work.
For some reason, the creators of stablecoins are so obsessed with complex designs that nearly every white paper is mired in equations and newly invented terminology, as if the creators were trying to convince you: Trust me, you’re not smart enough to understand this.
But I don’t agree with this point of view. At the end of the day, all stablecoins are designed to be very simple. Next I will show you a simple visual language to understand how all stablecoins work.
Think of each stablecoin protocol as a bank, each of them owning assets and liabilities, capturing value in some way and distributing that value to “equity” holders.
A normal full-reserve bank model, pictured above, has its real assets on the left, the actual dollar reserves it holds, and its liabilities (called “digital dollars”) on the right, which are claims on reserve assets.
In a full reserve bank, each liability is matched 1:1 with reserve assets. If someone owns digital dollars and asks for cashback, the holder will be given physical dollars and the corresponding digital debt will be destroyed. This is how Tether, USDC, and other fiat-backed stablecoins work.
Equity in the bank belongs to the shareholders (investors in the bank) who profit from the fees charged by the bank. In Tether’s case, the owners of Tether Ltd. are shareholders and its profits come from Tether’s minting and redemption fees.
Every liability of a full reserve bank should be closely tied to the U.S. dollar, as it is always redeemable for $1 in reserves. As long as banks maintain cheap convertibility, arbitrageurs will have no trouble maintaining their peg to the dollar.
So this is a normal full reserve bank. It’s an obvious model, but it will help illustrate how crypto banking is different.
Full Reserve Crypto Stablecoins
How to create a fully encrypted reserve bank whose liabilities are stable dollars?
Given that cryptocurrencies reinvented money, the first thing to do is to exchange USD assets for crypto assets. But cryptocurrencies are volatile, so a 1:1 guarantee won’t work if your debt is in USD. If the value of cryptocurrencies falls, the bank will face under-collateralization.
So just do the obvious: put an extra crypto buffer to give you a buffer if the crypto asset crashes.
Basically this is how MakerDAO works.
Dai’s peg is currently stable.
Note that reserve assets are significantly larger than total liabilities (Dai), which ensures the safety of the entire system.
Now let’s look at Synthetix, which takes a different approach: Instead of holding a diversified basket of crypto assets, Synthetix issues the sUSD stablecoin against its own SNX token. This SNX is also a “stock token”, in other words, the only asset that Synthetix allows as a deposit is its own token. Due to the high volatility of SNX, Synthetix requires a 600% overcollateralization of every sUSD in circulation.
The peg for sUSD is currently stable.
Both MakerDAO and Synthetix are similar to traditional full-reserve banks, except that their assets are overcollateralized in cryptocurrencies. In a way, their pegs are secure because there is some mechanism for exchanging stablecoins into their underlying assets. (In both approaches, there is a system of interest rates that target the desired price.)
There is another stablecoin, often referred to as an “algorithmic stablecoin.”
Algorithmic stablecoins are not redeemable at all and have no depositors in the traditional sense, making them less like traditional banks and more like central banks. (Central banks tend to use methods other than callables to keep prices stable.)
Each algorithmic stablecoin works slightly differently. In order to analyze an algorithmic stablecoin, one needs to try to understand its role in two important situations: when the stablecoin is above the peg, and when the stablecoin is below the peg.
Algorithmic Stablecoins
Structurally, the simplest algorithmic stablecoin is Fei.
Fei shrugged off the peg almost immediately, and has also recently gained notoriety. The following diagram describes how Fei works:
Fei’s current anchor position has been broken.
Fei operates much like a real central bank, defending its own peg directly in the market. Note that Fei is not meaningfully over-collateralized and most of its assets are in cryptocurrencies. This means that in the event of a black swan event, Fei’s assets could be significantly lower than its liabilities, preventing it from maintaining its own peg.
Fei’s algorithmic mechanism is quite complex, all Fei’s trading activities are conducted through Uniswap, and a technique called “reweighting” is used for actual trading, and “direct incentives” (actually a capital control) are used.
But the end result is the same: the protocol participates in the open market, pushing prices towards the peg.
Similar to the algorithmic central bank’s Celo protocol, it issues the Celo Dollar (cUSD) stablecoin. Celo Dollar uses CELO as its reserve collateral (Celo’s native asset), as well as a diversified portfolio of other cryptocurrencies.
Like Fei, the Celo Protocol has continued to use a Uniswap-style model to trade Celo Dollars on the market. The Celo Reserve is initially made up of a large amount of reserve assets that are designed to remain over-collateralized at all times. If Celo’s assets fall below 200% of its liabilities, the system will recapitalize by charging transaction fees for CELO transfers.
Therefore, the main difference between Celo and Fei, apart from their trading mechanics, is the assets they hold and their rules around staking.
The peg of Celo Dollar is currently stable.
A third stablecoin of the same type is Terra’s UST. Its collateral is LUNA, Terra’s native asset, and like FEI and Celo, the Terra protocol acts as a market maker for stablecoins. If the stablecoin system runs out of assets, it replenishes the inventory by increasing the native LUNA supply.
The peg of UST is currently stable.
Fei, Celo and Terra do not allow redemption. Instead, they trade their currency on the open market, which means they are willing to buy and sell spreads.
On the surface, this may appear to be very different from redemption, but it’s actually a closer continuum than it appears. This is because, economically speaking, a credible commitment to market making is the same as allowing coinage and redemption.
Imagine a stablecoin backed by ETH, call it the STBL token, and the protocol launches the ETH/STBL pair to the market. This means that the protocol will be willing to sell 1 STBL for $1.01 ETH and buy 1 STBL for $0.99 ETH. If STBL falls below the peg, it will continue to trade STBL until its ETH is exhausted.
If STBL tokens were minted and redeemed instead, it might allow anyone to mint 1 STBL for $1.01 ETH and redeem 1 STBL for $0.99 ETH. If STBL falls below the peg, it will continue to redeem ETH with STBL until its ETH is depleted.
In traditional central banks, market makers do not allow redemptions, instead allowing the central bank more discretion. But algorithmic market making is different because smart contracts can make unbreakable, self-enforcing promises. So market making and callability are two paths to the same goal: providing liquidity and ensuring tight pegs.
Central bank-style algorithmic stablecoins are studied, and there is another, more stable algorithmic stablecoin: Seigniorage Shares.
Mint Share Model Stablecoin
The classic mint stake model stablecoin is Basis Cash, based on the unissued predecessor Basis. It is probably the most quintessential algorithmic stablecoin, and many other designs have since been derived from it. A video showing the working principle of Basis Cash, video link address: https://youtu.be/bHzI8mECz_w. Currently, Basis Cash’s peg has been broken.
Basis Cash can be thought of as two phases of operation: Basis Cash is in a contraction cycle when there are outstanding bonds and the money supply is not growing fast enough to pay off all system debt. However, if demand continues to increase, eventually all bonds will be paid off, the system will enter an expansion cycle, and shareholders will once again be rewarded with newly minted Basis Cash.
The newly minted Basis Cash is the “seigniorage,” the profit the central bank makes by issuing new currency.
Normal central banks keep seigniorage on their own balance sheets for emergencies. Basis Cash, on the other hand, pays all seigniorage to shareholders the moment it receives seigniorage. This gives Basis a very “efficient guarantee” with no assets on its balance sheet at all. This enables it to support a very large high stablecoin supply on 0 assets. However, it also makes it vulnerable to a “death spiral” or crisis of confidence. In fact, Basis Cash does.
Subsequent algorithmic stablecoins are mostly descendants of Basis designs, including the last stablecoin we will examine.
ESD (Empty Set Dollar) is a stable currency carefully issued by the anonymous founding team. The original version of ESD (ESD v1) was based on Basis Cash.
The peg for ESD v1 has been broken and has since moved to a new design.
The innovation of ESD is the fusion of equity tokens with stablecoins, which means that if the stablecoins are staked, more stablecoins will be generated. As you might guess, this caused the stablecoin to become highly volatile, breaking away from the peg, sometimes as high as $2.00, before eventually falling below $0.20.
So far, pure mint stake model stablecoins have generally failed. Many Basis and ESD imitation disks, such as DSD, met the same fate. This at least tells us that the design of stablecoins does matter. These illustrations should help you reason out why mint stake model stablecoins are so vulnerable to a crisis of confidence.
Epilogue
In the early days of DeFi, decentralized stablecoins were basically impossible. For now, these claims seem premature. Algorithmic stablecoins have a lot of design space, and some designs are indeed stronger than others.
One thing’s for sure: you shouldn’t just because the whitepaper insists that a decentralized stablecoin will be powerful. Instead, consider what a stablecoin needs to be stable.