Sign In
June 25, 2024
6 minutes

Algorithmic Stablecoins: What are they and how do they work?

Algorithmic Stablecoins: What are they and how do they work?

Focus on algorithmic stablecoins

Previously a relatively obscure asset class, algorithmic stablecoins came to prominence for all the wrong reasons in 2022 due to the high-profile collapse of the UST currency and its parent ecosystem, Terra. UST may have become the most well-known example, but it certainly is not the only instance of this experimental branch of the stablecoin family. This article will explore the various types of algorithmic stablecoins and how they have been implemented. It will then examine the collapse of UST and the causes of the failure.

This article is Part 4 of a Nuant insights series focusing on stablecoins.

Read Part 1, a general introduction to the asset class and the “stablecoin trilemma.” 

Read Part 2, a focus on fiat-backed stablecoins.

Read Part 3, a focus on commodity-backed stablecoins 

General characteristics

Unlike fiat-backed stablecoins or crypto-backed stablecoins, which are backed by reserves, algorithmic stablecoins are issued based on an algorithm. The algorithm aims to achieve price stability by managing the supply and demand of the stablecoin against a secondary token that is not price-pegged.

Algorithmic stablecoins first emerged in 2014 with the now-defunct Nucoins project. They aim to address the shortcomings of other types of stablecoins in terms of the “stablecoin trilemma.” Algorithmic stablecoins are decentralized and are also more capital efficient than crypto-backed stablecoins that require overcollateralization. However, algorithmic stablecoins have performed less well than the other categories of stable assets when it comes to maintaining peg stability.

As decentralized ventures, most algorithmic stablecoins are managed by decentralized autonomous organizations (DAOs). Thus, changes to how the asset is managed can be passed by means of governance votes, which are typically open to all coin holders to participate in.

The three types of algorithmic stablecoins

Rebasing algorithms

Rebasing algorithms aim to maintain price stability by adjusting the supply of stablecoins according to demand. If the demand for stablecoin is high and its value exceeds $1, the supply will increase to bring the price back to $1. Conversely, if demand is low, the supply will decrease to raise the price to $1. This adjustment is called “rebasing.”

A rebasing algorithmic stablecoin such as Ampleforth (AMPL) exemplifies an elastic, non-dilutive cryptocurrency. The supply can change, but users retain the same proportion of the overall supply, so a user holding AMPL tokens may find the amount of tokens changes based on the weighted average of the token price over the last 24 hours.

Seigniorage algorithms

Seigniorage algorithms work with two or more assets. The stablecoin, which aims to be pegged at $1, and a second token that represents a share or seigniorage ownership. This token typically has a price set by the free market and is used to maintain the stable peg of the first token by using a mint-and-burn algorithm.

If the price of the stable asset rises above the peg, the supply of stablecoin is increased by burning the seigniorage token to create more stablecoins and reduce the value back to $1. Conversely, if the stable asset falls below the peg, it is burned and converted to the seigniorage token to stimulate demand.

Basis Cash is one example of a seigniorage algorithmic stablecoin, which also operates a third asset called bonds, which are redeemable when the price of the stable asset falls below $1 as an additional incentive for buyers. In May 2022, Coindesk reported that Do Kwon, founder of collapsed Terra, was also one of the pseudonymous founders of Basis Cash.

Holding the seigniorage asset of an algorithmic stablecoin may confer governance rights. 

Fractional algorithms

Fractional algorithmic stablecoins use a combination of collateral assets like USDC combined with an algorithmic reserve. The algorithmic reserve is not asset-based; instead, it utilizes a balancing mechanism with mint and burn.

One example is Frax, which previously used a fractional algorithm. Prior to the change, 92% of FRAX was collateralized, leaving an 8% algorithmic reserve. If the price of FRAX went above $1, users could mint the stablecoin by depositing 92% USDC and 8% Frax’s FXS token, earning a profit on the arbitrage and increasing the stablecoin supply.

Conversely, if the price fell below $1, users could buy FRAX and burn it to receive $1 in FXS. This mechanism allows for arbitrage and reduces the supply of stablecoins.

Thus, FXS not only confers governance rights but also has a role to play in the collateralization of the FRAX stable asset.

Frax governance voted to stop using the algorithm in February 2023 citing the risks following Terra’s collapse. 

Risks of algorithmic stablecoins

Peg stability

In the absence of physical reserves, the ability of the algorithm to maintain a stable peg becomes pivotal to maintaining investor confidence and thus liquidity in the asset. The collapse of UST in 2022 was the highest-profile example of an algorithmic stablecoin failing to live up to its promise.

UST collapse

Terraform Labs' UST was based on a dual-token system. The first token was LUNA, which was the native token of the Terra blockchain. The second was UST, the stablecoin. The protocol stated that users could create 1 UST by burning $1 worth of LUNA, and, conversely, generate $1 worth of LUNA by burning 1 UST. The principle was that if the market price of UST fell below $1, arbitrageurs would buy it on the open market because they could redeem it for $1 of LUNA at a premium. If UST went above $1, there was an incentive to burn LUNA to generate UST at a higher value, which increased the supply and would bring the price back toward $1.

The circumstances surrounding its failure are complex. In the months leading up to the collapse in May 2022, the market capitalization of UST had been rising sharply, from around $3 billion in November 2021 to over $18 billion at its peak in May 2022. The key driver for the increase was the promise of 20% yields on UST deposits to a DeFi lending protocol called Anchor. At one point, as much as 72% of UST was thought to be deposited on Anchor, and reasonable speculation was beginning to emerge that the yield was unsustainable.

In early May, a series of large transactions converted tens of millions worth of UST to USDC on Anchor and Curve Finance, resulting in UST losing its peg for the first time. The Luna Foundation Guard, a fund that had supposedly been set up as a kind of reserve to defend the UST peg, later confirmed it had deployed over 80,000 BTC (around $3 billion at the time) to help maintain the peg – efforts which ultimately failed. Investors lost faith and began attempting to exit positions in UST and LUNA in Anchor and all other platforms, a sequence that ended up becoming a “death spiral.”

Following the incident, there was some speculation that UST had been subject to a short attack. However, researchers from Nansen disputed this claim, revealing evidence that showed up to seven wallets belonging to separate large investors had been involved in the initial set of transactions.

The UST incident underscores the general risks associated with a stablecoin losing peg stability and the importance of the role of reserves. 

Regulation and compliance

Any purely algorithmic stablecoin is unlikely to pass institutional compliance checks due to the lack of reserves and uncertain regulatory status of the asset.

Following the collapse of UST, several other cryptocurrency operators went under, including major operators such as centralized lender Celsius and hedge fund Three Arrows Capital. The contagion and subsequent regulatory scrutiny of the stablecoin sector means that algorithmic stablecoins face a highly uncertain future in major jurisdictions such as the US and the EU. The EU Markets in Crypto Assets regulation will require stablecoins to be backed 1:1 by reserves, meaning algorithmic stablecoins that only rely on an algorithm to maintain peg stability without the presence of reserves will become unlawful. As capital efficiency is one of the core appeals of algorithmic stablecoins, it would appear that they will become largely outmoded in the new regulatory environment.

Monitoring the risks associated with your digital asset portfolio is a 24/7 job, which is why you need a robust suite of portfolio management tools that are always working on your behalf. Nuant allows you to aggregate your entire portfolio into a single unified view, drill down into the performance and analytics of individual assets, set up market alerts, and automate your KYC and compliance checks, including wallet intelligence to flag risks. Register your interest using the button on the top right to become one of our early users.  

References and further reading

A “short attack” caused UST de-peg, Luna’s 97% price collapse, so who was responsible?, Kitco News. (n.d.). Retrieved March 16, 2023

EU MiCA Regulation is Coming—But what will it mean for crypto funds? (n.d.). Retrieved March 16, 2023

Frax Ditches Algorithmic Backing, Citing Terra Failure, Blockworks. (2023, February 23). Retrieved March 16, 2023

The Fall of Terra: A Timeline of the Meteoric Rise and Crash of UST and LUNA. Genç, K. S. and E. (2022, June 1)

We Need to Talk About Terra’s Anchor, Decrypt. Kelly, D. / L. J. (2022, April 23)

LFG Reserves Dwindle to Just 313 Bitcoins From 80K After UST Crash, Malwa, S. K. and S. (2022, May 16)

UST’s Do Kwon Was Behind Earlier Failed Stablecoin, Ex-Terra Colleagues Say, Nelson, S. K. and D. (2022, May 11)

On-Chain Forensics: Demystifying TerraUSD De-peg, Nansen. (2022, May 27). Retrieved March 16, 2023

Stablecoins: What Portfolio Managers Need to Know. (n.d.). Retrieved March 16, 2023

The End of a Stablecoin—The Case of NuBits. Reserve, Team, R. R. (2018, July 12).

TerraClassicUSD price today, USTC to USD live, marketcap and chart. (n.d.). CoinMarketCap. Retrieved March 16, 2023

Types of Stablecoins: What are Fiat-Backed Stablecoins ? (n.d.). Retrieved March 16, 2023

Updated on
June 25, 2024