Born to Fail: The Inherent Fragility of Algorithmic Stablecoins

Article Author: Dr. Ryan Clements

Article Compiler: Block unicorn

Algorithmic stablecoins are inherently fragile; these uncollateralized digital assets attempt to peg the price of a reference asset using financial instruments, algorithms, and market incentives, but they are fundamentally unstable and exist in a state of perpetual fragility. So far, several iterations have struggled to maintain a stable peg, with some ending in catastrophic failure. This article argues that algorithmic stability is fundamentally flawed because it relies on three historically proven uncontrollable factors.

1. First, they require a level of support for the demand for business stability.
2. Second, they depend on independent actors with market incentives to arbitrage stable prices.
3. Finally, they require consistently reliable price information.

All these factors are not guaranteed, and it has been shown that they are historically fragile in the context of financial crises or extreme volatility. Regulatory guidelines need to be established for all types of stable registration forms, including issuer registration requirements, clear classification statements, reviews, collateral custody and transparency safeguards, as well as risk disclosure and containment measures. A strong regulatory framework with risk disclosure and containment safeguards is particularly necessary for algorithmic stable trading, as it currently only serves speculative DeFi trading applications with little (if any) social or financial inclusivity value.

Introduction

Financial product innovation is not always a good thing; certain designs of innovations render them inherently unstable. In 2008, a series of complex securitization-driven, derivative-enhanced financial product innovations triggered by housing loans nearly collapsed the entire financial system. Now, a new, increasingly popular, poorly designed, and inherently fragile financial product has emerged that requires appropriate regulation—algorithmic stability.

Algorithmic stablecoins are a contradiction in terms. So far, the market iterations of algorithmic stability have shown a complete lack of stability. They are unregulated, uncollateralized digital assets operating in a state of perpetual fragility. Algorithmic stability does not have a real peg but merely derives value from what the ECB's Crypto Assets Working Group calls "expectations of its future market value." Thus, it is an incredibly fragile payment mechanism. Algorithmic stablecoins are touted by some as a "capital-efficient" antidote to the daily price volatility of popular cryptocurrencies like Bitcoin and Ethereum, limiting their functionality as a currency alternative for consumer transactions, wages, or deferred debt repayments. Others claim that algorithmic stablecoins are "rebuilding traditional banking," becoming a reserve system in the DeFi space. Both comparisons miss the mark, and the utility claimed by algorithmic stablecoins is grossly exaggerated and misleading, as three historical lessons render them inherently fragile.

First, algorithmic stability arbitrage requires a balance of liquidity across the entire ecosystem. If liquidity falls below a threshold level, the entire system will fail. History shows that the fundamental or minimum support level for financial products cannot be guaranteed—especially in times of crisis.

Second, algorithmic stability arbitrage relies on independent participants with market incentives to engage in price-stabilizing arbitrage to maintain the so-called "stable" ecosystem. History again shows that relying on independent, market-driven participants without legal obligations to execute discretionary arbitrage for stable prices is also fragile. History has demonstrated that during crises, information becomes opaque, noise crowds signal, and prices and counterparties become uncertain, easily leading to herd behavior. Information opacity undermines the symbolic "economy" and incentive structure of algorithmic stability.

If the "symbolic" incentive structure in any algorithmic stable ecosystem collapses, the entire ecosystem will fail without a backstop or deposit insurance safety net. Algorithmic stability systems exist in a framework where the system becomes prone to instability and failure when reality deviates from the assumptions embedded in the incentive structure. Multiple iterations of algorithmic stability have failed catastrophically.

Regulatory safeguards need to be established for all types of stable investments, including issuer registration requirements, clear classifications that clarify the forms of stable investments, review rules, collateral custody safeguards, and reporting transparency, risk disclosure, and containment measures. Risk transparency, disclosure, and containment measures are particularly relevant to algorithmic stable investments, as algorithmic stablecoins currently only power speculative DeFi trading applications.

I. Experiences of Various Stablecoins

Stablecoins are crypto assets that attempt to peg their value to another asset (or a basket of assets including reserve currencies or highly liquid government bonds). So far, there has been no unified definition of stablecoins—perhaps one reason why regulatory frameworks have been slow to materialize. The International Organization of Securities Commissions (IOSCO) suggests that stablecoins come in many different varieties and forms.

The most popular form is "off-chain" custodial stablecoins, such as the widely circulated USD Coin (USDC) from Circle and Coinbase, or Diem proposed by Facebook, both of which use "held fiat currency or high-quality liquid assets as reserves." Or Tether, which claims to be backed by a large amount of commercial paper. Other stablecoins are either fully collateralized or "over-collateralized." Over-collateralization means that more than 100% of the stablecoin's value is "held on-chain," using another crypto asset to provide collateral functionality, such as MakerDAO's DAI stablecoin. Using a designated type of crypto asset for over-collateralization (like WBTC, ETH, or other highly liquid crypto assets), DAI is generated, and the collateralization ratio is adjusted for specific locked tokens.

The most unstable and fragile type of stablecoin is "algorithmic," which is not fully collateralized and attempts to maintain a stable peg using market incentives, arbitrage opportunities, automated smart contracts, and reserve token adjustments. These types of stablecoins are described as the "central bank/Fed" of algorithmic stablecoins. In 2021, the market for stablecoins surged to over $119 billion, with a significant and growing share of that market being algorithmic stablecoins.

Despite the catastrophic failure of Iron Finance in June 2021, algorithmic stablecoins claim the benefits of automated operations and the ability to scale without the need for corresponding reserves. Essentially, the protocols supporting algorithmic stablecoins attempt to operate like a central bank, with "less than one-to-one backing," by manipulating the number of tokens "in circulation" in response to changes in their value.

There are various models of algorithmic stablecoins, and there is controversy over their exact definitions. They typically seek to combine the money supply with embedded economic incentives to artificially control the price of the stablecoin. For example, if the trading price of a stablecoin falls below $1, the algorithmic stablecoin system might offer some other type of digital asset at less than $1, such as digital "bonds," "discounts," or issued "equities," with new capital used to maintain the peg. A common structure for algorithmic stablecoins is a "dual-token" system, where one token is used to maintain the peg, while the other is used to "absorb" market fluctuations. The latter token is often referred to as a "equity" or "balancer" token and is typically traded on other decentralized exchanges like Uniswap; the two-token system is often combined with partial collateral dynamics, as described in the next section using Iron Finance's Iron algorithmic stablecoin.

II. The Failure of Iron Finance is a Major Warning Signal for the Product Category

Iron Finance describes itself as a "multi-chain, decentralized, non-custodial ecosystem for DeFi products, protocols, and use cases." Their initial system was a dual-token structure attempting to create an algorithmic stablecoin called "IRON." IRON was pegged to $1 but had no actual backing of $1. They recently announced a relaunch of "V2" for "over-collateralized and soft-pegged" stablecoins. Before its nearly $2 billion failure, each IRON stablecoin was backed by locking 75% of its value in collateralized USDC (a fully reserved, fiat-backed stablecoin) and 25% of its value through locked "TITAN"—Iron Finance's own governance token with an unlimited supply.

Iron Finance collapsed when the value of its unlimited supply governance token TITAN plummeted in the DeFi secondary market. Iron Finance reported that certain "whale" holders engaged in massive sell-offs. The market for TITAN was already weak, and this large sell-off triggered a cascading sell-off of TITAN and IRON redemptions in a "negative feedback loop." This caused the IRON token to lose its peg, which in turn "triggered" the algorithmic minting mechanism of TITAN, providing arbitrage opportunities in the resulting "death spiral."

The direct impact was a massive supply of TITAN in the secondary market. At one point, the price of TITAN was essentially zero, and Iron Finance halted redemptions for the IRON stablecoin—they initially had only a 75% collateral coverage with USDC. The moment the price of TITAN became unstable in the secondary trading market, the entire house of cards of the IRON stablecoin collapsed, with nothing to support this operation.

The idea that algorithmic stablecoins are an early iteration of reserve banking in the DeFi space has been advanced. Iron Finance—when explaining the failure of its so-called stablecoin IRON—in a "post-mortem analysis" report referred to it as "the world's first large-scale crypto bank run." This analogy is fundamentally flawed, as Iron Finance's operational structure was very fragile from the outset.

It attempted to create $1 from $0.75, assuming that its secondary trading governance token TITAN would not fall below a market-determined price floor. Its design assumption was that TITAN itself would remain stable—or better yet, its price would increase. Banks also create money through fractional reserves and loans. However, banks are backed by government deposit insurance—they pay hefty fees in the form of premiums and are subject to extensive oversight and scrutiny.

III. Three Lessons from Financial Market History

Three lessons from financial market history impact the viability of algorithmic stablecoins.

First, if liquidity dries up, any financial product that requires liquidity support or a fundamental level for the entire product category to operate as expected (and assumed) will be prone to failure. Liquidity is unpredictable and affects the prices of all securities. However, if a product requires a minimum level of liquidity to function, then that product is inherently fragile.

As previously identified, the requirement (but lack of) support levels from major financial institutions is a significant factor leading to the failure of auction rate securities markets. The reliance on a fundamental support level may be the biggest issue with uncollateralized algorithmic stablecoin dual-token structures. The token that absorbs volatility must maintain a certain level of demand support—and cannot fall below a price threshold—otherwise, the entire ecosystem will fail. Non-collateralized tokens that claim to be "stable" require consistent (if not increasing) liquidity levels, and once that stops, the peg will fail.

The second historical lesson that makes algorithmic stablecoins inherently fragile and unstable is that they often rely on independent participants with market incentives to execute the arbitrage function for stable prices. Arbitrageurs must intervene and exploit profit opportunities through minting or redemption activities to maintain price stability. Historically, the performance of discretionary price-stabilizing arbitrage has been fragile during crises, as previously noted in the work on exchange-traded funds, where "market discipline may fail when it is needed most."

During the 1987 portfolio insurance failure, arbitrageurs stopped buying undervalued assets. More recently, when the market rapidly shifted pricing in the aftermath of the COVID-19 pandemic in March 2020, arbitrageurs ceased to arbitrage the price discrepancies between fixed-income exchange-traded funds (ETFs) in the secondary market and their underlying net asset values.

The third historical lesson that raises doubts about the long-term viability of algorithmic stablecoins is the widespread information opacity during periods of heightened volatility, panic, or crisis. Effectively integrating price information is a "challenge" for many algorithmic stablecoins. Price "oracles" (external price feeds) are not always reliable, and when token holders vote on which potential price feeds to adopt (from their pool), there are "misaligned" incentive issues. The uncertainty in the price of the TITAN token due to delays in automated "oracle" information feeds led to the failure of Iron Finance in June 2021.

When information is uncertain, cascades and investor herds form, and assets deemed unsafe are rapidly sold off in a panic—this phenomenon was evident during the 2008 global financial crisis, even for certain financial assets like commercial paper and money market mutual funds, which were considered stable before the crisis. Information opacity also affects market participants' ability to engage in price-stabilizing arbitrage, as seen in the case of the 1987 portfolio insurance failure.

IV. Stablecoins as Dominoes in an Emerging Algorithmic Ecosystem

Perhaps the currently most popular algorithmic stablecoin platform is Terra. The creators of Terra, Terraform Labs, recently received significant venture capital backing and investor interest as a "stablecoin for e-commerce creators." Terra uses a governance balancing token (called LUNA) to mint algorithmic stablecoins pegged to the US dollar and Korean won (among others), with built-in monetary supply and economic incentives, including fees and arbitrage opportunities.

These stablecoins are then used as a payment mechanism in the ever-expanding financial "ecosystem" of Terraform Labs, which also includes a protocol (Mirror) for creating synthetic assets that track the performance of US stocks, futures, and exchange-traded funds; a lending and savings platform (Anchor); and a partner payment platform (Chai). Terra also plans to add DeFi asset management, additional lending protocols, and decentralized leverage insurance protocols to this nascent ecosystem.

Terra stablecoins are the "core" connecting the emerging financial "infrastructure," which includes the aforementioned e-commerce payments, synthetic stocks, exchange-traded funds, derivatives, and other financial assets, savings, lending, and borrowing applications. The operation of Terra, as a protocol, incentivizes independent traders to purchase its stablecoins to exchange for LUNA when the stablecoins fall below their peg. The stability of Terra stablecoins transcends DeFi speculation. Given their many applications in the "Terra economy," these algorithmic stablecoins also directly impact the economic prospects of many businesses and consumers.

For this ecosystem to remain viable, Terra stablecoins and the governance token LUNA must have a permanent baseline level of demand. In other words, there must be sufficient arbitrage activity between the two tokens, as well as enough transaction fees in the Terra ecosystem and mining demand in the network. The founders of Terra assert that mainstream adoption of their stablecoins as transactional currency, along with the ability to "stake" them and earn rewards, creates "network effects" and long-term incentives to hold and maintain the ecosystem.

Thus, Terra bets that using stablecoins (and LUNA) in their financial applications "network" will drive permanent demand. This assumption is not guaranteed; Terra stablecoins have previously deviated from their pegs. In many ways, a developing DeFi financial ecosystem supported by algorithmic stablecoins, without real collateral or government guarantees, relies on the permanent interests of market participants driven by individual motivations for sustainability, resembling standing dominoes—once the first falls, all others may be affected.

Conclusion

While there is a search for sustainable price-stable conceptual models, so far, algorithmic stable currencies have shown a complete lack of stability, making them unsuitable as a currency alternative. Unlike collateralized stablecoins, algorithmic stable investment varieties seem "doomed to fail." Financial writer J.P. Koning argues that they are "prone to permanent collapse" because they precariously rely on "circular relationships" between different participants—those seeking "stability" and those seeking "high-return opportunities." Algorithmic stable investments are unlikely to serve any real long-term investment. Improving consumer welfare, or financial inclusivity functions beyond short-term debt speculation, is unlikely to yield inclusive or systemic benefits. As others have pointed out, they are "unstable, threatening their utility."

Like other varieties of stablecoins, algorithmic forms currently lack transparency, scrutiny safeguards, and oversight. As noted in this article, they are also built on a fragile foundation that relies on uncertain historical variables: they require a level of support for baseline demand, they need willing arbitrageurs to participate, and they require an information-efficient environment. None of these factors are guaranteed, and all have proven to be highly fragile in the context of financial crises or extreme volatility. History suggests that they are likely to be prone to instability and failure, and they should be regulated to provide full transparency and strengthen consumer protection and risk control measures so that they do not become interconnected with the larger financial system.

Currently, the financial regulation surrounding stablecoins in the U.S. is fragmented, inefficient, and in many cases overlapping. The clarity of the "regulatory scope" of stablecoins has yet to be determined. They are subject to oversight by the Financial Crimes Enforcement Network (FinCEN) as well as state money transfer and virtual currency licensing. They also pose "bank-like risks"—especially shadow deposits like money market mutual funds, whose monetary policy impacts are entangled with the Federal Reserve. Their systemic risk considerations have garnered support from the Treasury-led Financial Stability Oversight Council with backing from the President's Working Group on Financial Markets. They are also under the purview of the Consumer Financial Protection Bureau (CFPB), the Office of the Comptroller of the Currency (OCC), the Commodity Futures Trading Commission (CFTC), and the Securities and Exchange Commission (SEC).

A comprehensive approach to regulating stablecoins that transcends institutional divides is needed. Ideally, the regulatory framework for all stablecoins would include issuer registration requirements, prudential measures, collateral custody safeguards, and reporting transparency, clear classifications that clarify the forms of stablecoins (and distinguish algorithmic varieties from other types), as well as risk disclosure and containment measures.

Such a framework may require the specific agency "full authority" over crypto trading described by newly appointed SEC Chair Gary Gensler, albeit adapted and applied to stablecoins. Some fully collateralized stablecoins may offer benefits for financial inclusion, such as faster, cheaper global remittances, real-time payments, the application of fiscal stimulus, and the ability to act as transactional agents for thin credit profiles and banking deserts. Therefore, it is necessary to establish a regulatory framework that supports innovation while still creating transparency, risk control, and consumer protection safeguards.