Intro

Following the collapse of Terra USD (UST), the crypto community has plenty of reasons to be skeptical about the future of algorithmic stablecoins. Can algorithmic stablecoins be trusted? 

Algorithmic stablecoins are an interesting innovation as they address over-collateralization, a common concern of crypto-backed stablecoin users. For example, Maker DAI is a crypto-backed decentralized stablecoin whereby users deposit Ether in excess of the amount of stablecoins received, in order to account for depreciation. If they deposit $200 in Eth for $100 in DAI, then if Eth depreciates by 25%, they would still be safely collateralized with $150 in Eth. However, this can be capital intensive and also result in liquidation for those not monitoring prices.

Projects developing algorithmic stablecoins aim to provide the stability of stablecoins while requiring less collateral to get started and removing centralization. Rather than relying on collateralization to address volatility, the value of algorithmic stablecoins depends on their ability to adjust the supply of tokens in reaction to changes in the market. In their purest form, algorithmic stablecoins are completely uncollateralized. In practice, many times often rely to an extent on some form of collateralization to help incentivize the peg.

Given the downfall of UST, why would anyone want to use an algorithmic stablecoin? Apart from being capital efficient, they are also decentralized, meaning that they make it challenging for governments or influential parties to influence their usage. They can also be widely used for DeFi protocols.

Making it more capital efficient to access stablecoins is a great idea. Unfortunately, this is a rather new domain that has not been thoroughly tested, and the UST debacle is an example of how things can go horribly wrong. The challenges faced can at times result in failed projects, in turn hindering wider adoption. For example, institutions can generally not use them due to their risks, and failures can create significant fear in the industry. The crypto community as a whole must explore different ways to approach the construction of an algorithmic stablecoin and assess the known risks associated with each of these approaches.

The Stablecoin Trilemma

Many trilemmas exist in the world of blockchain, such as the Blockchain Trilemma of only being able to choose two of three ideals amongst decentralization, security and scalability. In the world of stablecoins, the tradeoff is between:

• Peg Stability
• Decentralization
• Capital Efficient

For example, USDT and USDC are very strong at maintaining their $1 peg and being capital efficient, but this comes at the cost of centralization. Centralization isn’t necessarily bad but can be susceptible to moral hazard, government intervention, and more. In contrast, Maker DAI detailed above requires over-collateralization, making it capital intensive despite being decentralized and maintaining its peg quite well.

At Stably, we use a 100% USD-backed stablecoin which is more capital efficient than Maker DAI and retains its peg, but we require centralization which is not for everyone.

Algorithmic stablecoins are decentralized and capital-efficient and do their best to maintain peg stability. Several have succeeded thus far, despite being more volatile than over-collateralized or centralized stablecoins. At the same time, there have been numerous examples of failures such as TerraUST and Iron Finance. This is often due to periods of high volatility that harm investor confidence, causing a large sell-off unable that’s unable to re-peg. Algorithmic stablecoins also face challenges such as initial bootstrapping and slower adoption. The result has at times been deep investor losses and a skeptical attitude towards them all.

We believe that skepticism in the world of blockchain is important. We also believe that it’s equally important to build an understanding of how these coins work, prior to making any investment decisions. Therefore, we’ll briefly introduce various types of algorithmic stablecoins and their tradeoffs.

Types of algorithmic stablecoins

There are multiple approaches taken to creating an algorithmic stablecoin, and we will cover the three main types of approaches which include rebasing, seignorage, and fractional-algorithmic.

Rebase (Non-Collateralized)

This was one of the first types of decentralized algorithmic stablecoins, and under this model the price is pegged by rebasing the total token supply. By increasing or decreasing the total supply of the token at regular intervals, the price can be adjusted back to $1. Ampleforth is one example of this, whereby they automatically adjust the supply of their AMPL cryptocurrency every 24 hours in what’s known as a “rebase”.

This can be more easily understood through an example:

• You have 1,000 tokens at $1, which then rise to $1.05
• The total supply of this token is increased by 5%, meaning you now have 1,050 tokens and the total supply has proportionally increased as well.
• Arbitragers go to work on decentralized exchanges, with the supply increasing proportionally and in turn the price coming down instantly.

While this sounds good in theory, Ampleforth has had mixed success in maintaining its peg, and another Rebase token known as Base Protocol has had its fair share of challenges in maintaining its peg with its website no longer being functional.

Seignorage

The word seignorage traditionally refers to the difference between the face value of coins and their production costs. While the government would typically pocket this profit, with seignorage stablecoins the profit is made for those that hold the shares. 

Compared to the rebase model, seignorage stablecoins typically use one or more other tokens that incentivize the peg of the stable value through market mechanics.

Sound familiar? This is what LUNA and UST acted as, prior to the infamous downfall.

While we wrote an article about that here, we’ll recap quickly:

• UST could always be traded 1:1 for LUNA. 
• If the price of UST were to be above $1, someone could swap $1 worth of LUNA to UST, and sell it for more than $1, thereby making a profit and putting downward pressure on the token.
• If the price of UST were to go below $1, UST holders could purchase UST, swap it for $1 of LUNA, and make a profit while putting upward price pressure on UST.
• Profits were retained by arbitrageurs and the LUNA Foundation, while with traditional seignorage it’s the government that pockets any profit.

Despite the holding of Bitcoin as collateral, the steep drop in UST combined with a falling Bitcoin price meant that it was insufficient for maintaining UST’s peg.

Fractional-Algorithmic

Sometimes referred to as “hybrid-algorithmic” stablecoins, this class is a mix of part-seignorage and part-collateral.

FRAX is a good example of this, whereby they use a two-token system (similar to UST) that comprises of their stablecoin FRAX and their governance token FXS. FRAX maintains its USD peg by being partially collateralized by USDC (a stablecoin that’s 100% collateralized) and by periodically buying or selling FXS to maintain its market capitalization. 

Essentially, FRAX can be minted by a user who supplies USDC as collateral, as well as FXS tokens, with the proportion of both varying. If the collateralization ratio is 50%, then by depositing $0.50 of USDC and $0.50 of FXS, one could mint $1 of FRAX. If the collateralization ratio were to lower to 40%, then more FXS would be needed per stablecoin, putting buy pressure on FXS and increasing its value.

This is used in conjunction with deep liquidity pools on Curve Finance in order to maintain price stability of $1. In essence, it’s partially collateralized but partially algorithmic.

The exact details of the process can be more fully examined in the link here. In the meantime, the below graphic is a good representation of how the collateralization works, and you’ll note that the process works when redeeming FRAX as well.

Conclusion

Various types of stablecoins exist, ranging from Stably’s 100% USD-backed stablecoin, all the way to fully decentralized algorithmic stablecoins.

While algorithmic stablecoins are decentralized and capital-efficient, they have at times struggled to maintain their pegs during periods of high volatility, such as during Spring 2022 during the notorious collapse of UST which also saw many other stablecoins lose their pegs for brief periods of time at a minimum.

As a new technology, we can expect some additional algorithmic stablecoins to fail. Fortunately, given the open-source nature of these projects, they have a wide variety of approaches and intelligent minds concerned with their sustainability, and rapid advances can be expected.

Diversifying the stablecoins used in deploying one’s investment strategies is generally a good idea. Understanding the potential risks associated with different types of stablecoins is important prior to using them, and makes it easier to select those most likely to stand the test of time.

Stably Ramp makes it easy to access a variety of different stablecoins. Please visit ramp.stably.io to get started or head to stably.io for more information about Stably’s borderless accounts, Stably USD (USDS), and Stablecoin-as-a-Service.