Two weeks ago an algorithmic stablecoin called TerraUSD (or UST) blew up, incinerating tens of billions of dollars of market value. The idea of an algorithmic stablecoin is that it should always be worth a dollar because of an arbitrage mechanism in which one stablecoin can always be exchanged for a number of units of some other crypto token — for TerraUSD, it was called Luna — with a market value of $1. If that other token (Luna) is worth $100 or $10 or $1 or $0.10, that works fine; if the stablecoin trades below $1, you buy it for $0.97 or whatever and redeem it for some Luna that you can sell for a dollar, making an arbitrage profit and pushing the price of the stablecoin back to $1.
The problem is twofold. One is that, if people want to redeem a lot of the stablecoin, the algorithm will print a lot of the other token (Luna), which will tend to drive down the price of that token, which might undermine confidence in the stablecoin, which might lead to more redemptions, which will lead to more printing, which will drive down the price, etc., in what is called a “death spiral.” This is a very well-known problem that long predates crypto; companies have for years issued bonds that are convertible into fixed dollar amounts of stock, which are called “death-spiral convertibles” and have the same problem. The other problem that is more specific to crypto is that the other token — Luna — is just made up, and its value is tied to confidence in the stablecoin. If a death spiral starts, there is nothing to underpin the value of that token, so it can go to zero fast. Luna was trading in the $80s in early May; it’s at about $0.0002 today. TerraUSD is below 7 cents.
I want to emphasize here that:
- This problem is extremely, extremely well known.
- Algorithmic stablecoins have death-spiraled in the past in extremely public and predictable ways.
- Lots of people loudly predicted that TerraUSD would death-spiral in exactly this way.
- It did.
But I guess we’re gonna keep going until we get it right. Or until we get it wrong a bunch more times:
Among the handful of algo stablecoins that survive is the USDD token launched by controversial crypto entrepreneur Justin Sun on his Tron network just before the collapse of UST. …
USDD went live on the Tron blockchain in early May, according to Sun. It can be used on the Ethereum and Binance Smart Chain blockchains through so-called bridging software. The stablecoin uses an arbitrage mechanism similar to what Terra used to maintain its peg. For example, when 1 USDD drops below $1, traders can make an instant profit by sending 1 USDD to the Tron blockchain in exchange for $1 worth of Tron, the native token of the network, with the arbitrage incentive designed to bring USDD’s price back to $1.
Terra showed that an integration of such algorithmic programming to keep a token stable is easier said than done, largely because the stablecoin is dependent on investors believing that its sister cryptocurrency will continue to appreciate. …
Sun agreed that Terra’s failure shows the flaw of algorithmic stablecoins, but said it also gives an opportunity for new projects to adjust. USDD aims to raise $10 billion through the Tron DAO Reserve — which it called an “alliance” of market participants including Alameda Research and Amber Group — to defend USDD’s peg. That’s a similar setup to Terra’s main developer Do Kwon’s Singapore-based non-profit Luna Foundation Guard.
Look I don’t actually think this is impossible. When we first talked about TerraUSD last month, I liked the idea of the Luna Foundation Guard and its pool of money. The point is that you build up a valuable algorithmic stablecoin on a wave of investor confidence, and then you use that value to build a sort of foreign-exchange reserve fund. You can print Luna for free, so if people value Luna you should print a bunch of it and exchange it for things that (1) people also value and (2) are uncorrelated to Luna. You buy $10 billion of Bitcoin or Ethereum or Treasury bills or gold or whatever and, if the stablecoin goes down, you spend some of that reserve fund to buy the stablecoin and prop up the price. If you build a big fund and show a willingness to deploy it, then no one will doubt your stablecoin, so you won’t have to spend the fund, so your other token (Luna, Tron, whatever) will appreciate, so it will all work in a self-sustaining way. “The basic structure of the trade is (1) Ponzi, (2) acceptance, (3) diversification, (4) permanence.”
I don’t think this strategy is crazy! I mean, of course it’s crazy, but I do think it could work. (Does it … sort of … describe the history of fiat currency?) It’s just, you know, if you slip up on the way to permanence, you vaporize tens of billions of dollars. And it’s extremely easy to slip up, because in the early going the only thing underpinning the value of your stablecoin is confidence in your system. And people keep slipping up.
Also here’s a wild quote:
“An algo stable will exist in the next five-to-seven years,” said Hassan Bassiri, a portfolio manager at Arca, which was an investor in Terra. “And it has to exist or else what are we even doing in this space?”
Losing a lot of money, for one thing. Is this … is now a good time to launch an algorithmic stablecoin? A month ago, all the people who were like “there needs to be an algorithmic stablecoin” put their money in TerraUSD, so it was hard to launch a competitor. Now that TerraUSD has absolutely incinerated all of their money, the field is wide open! People miss TerraUSD — “wasn’t it nice, having a stablecoin that was truly decentralized and algorithmic?” — and so they are willing to ignore how it ended.