Ethereum inventor Vitalik Buterin said that algorithmic stablecoins should be scrutinized according to how they fare under extreme market conditions, and whether they can safely wind down when hype falls away.
Despite the recent collapse of UST and LUNA, which knocked UST from its $1 peg and wiped billions from the market, Buterin argued in an essay on May 25 that automated stablecoins can make sense while criticizing the exorbitant returns offered by those “doomed to collapse eventually.”
UST’s meltdown, which took place earlier this month after traders lost confidence in LUNA, the volatile coin that propped UST up instead of the cash reserves that back centralized stablecoins, prompted speculation that algorithmic stablecoins are fundamentally flawed.
But some algo-stables are robust, Buterin said. Buterin pointed to MakerDAO’s stable token DAI, which has already survived extreme market conditions.
DAI is an overcollateralized stablecoin, meaning that each DAI in circulation was minted against a greater value of collateral in ETH, USDC, WBTC, and seven other cryptocurrencies. DAI could lose parity with the dollar if all those coins crashed, which is why it asks for extra collateral.
Still, DAI minters can get wiped out in a sudden market crash. That’s exactly what happened during the “Black Thursday” crash of March 2020, when most leading crypto assets shed half 50% of their value in less than two days.
Facing a crisis of under-collateralization, the Maker protocol printed and sold the platform’s governance token, MKR, for DAI to pay off its debts, and added support for the centralized stablecoin USDC to back the creation of DAI. The auction was successful, but MKR fell by 60%.
Buterin posed two questions that can help answer whether an algorithmic stablecoin is resilient enough to withstand shocks.
First, can the stablecoin safely wind down to zero users?
Buterin said LUNA, which he refers to as a volatile coin or “volcoin,” increased in value before its collapse because speculators expected more network activity, and consequently more transaction fees. When that expectation fell away, the system became extremely fragile.
“Only a small downward shock to demand for the stablecoin could lead to the targeting mechanism printing lots of volcoins, which causes the volcoin to hyperinflate, at which point the stablecoin too loses its value,” Buterin said.
He said it’s unlikely that a slow decrease in the volcoin’s expected future fees would lead to a gentle and gradual reduction in the market cap of a volcoin-backed stablecoin. “What’s more likely is a rapid drop in interest followed by a bang.”
By contrast, Buterin said that Reflexer Labs’ RAI stablecoin, which is the same as DAI except that it only supports ETH as collateral, can safely wind down. That’s because investors hold a leveraged position in ETH rather than RAI’s governance token, FLX, eliminating the risk of a “feedback loop where reduced confidence in RAI causes demand for lending to also decrease.”
However, in the unlikely scenario that RAI became the main use for Ethereum, a drop in expected future demand for RAI could crash the price of ETH, risking cascades of liquidations and “a messy collapse”.
Second, Buterin asked what would happen if a stablecoin were designed to return a 20% profit to its users each year. “There is no genuine investment that can get anywhere close to 20% returns per year… what happens if you try?”
He said there are only two ways to drive 20% annual returns for a stable token. The first is to charge a negative interest rate on holders to cancel out the growth, undermining its utility as a stable token. The other is to turn the project into a Ponzi scheme, “giving holders amazing returns for some time until one day it suddenly collapses with a bang.”
These questions are a start. Even if a stablecoin passes these tests, Buterin cited other risks, such as collateralization, code bugs, and governance vulnerability. But “steady-state and extreme-case soundness should always be one of the first things that we check for,” he said.