Risks of Algorithmic Stablecoins: Why They Fail and What Australian Investors Should Know
In short: an algorithmic stablecoin tries to hold a steady value, usually US$1, using supply rules and a companion token instead of full reserves of cash or collateral. That works only while people trust it. When confidence breaks, holders rush to exit, the companion token collapses and the peg can fail completely, as TerraUSD (UST) showed in May 2022, when UST and its sister token LUNA lost about US$40 billion of combined value within days. The key points:
- No purely algorithmic stablecoin has held its peg through a major loss of confidence.
- Very high yields, like Anchor’s 20% on UST, were a warning sign rather than a feature.
- Partly collateralised and other decentralised stablecoins are safer the closer they get to full backing.
- New stablecoin laws generally require real reserves, pushing purely algorithmic designs aside, including for stablecoins in Australia.
Moneysmart’s overview of crypto assets explains the risks of holding crypto, ASIC’s guidance on digital assets covers how stablecoins can be regulated in Australia, and the ATO’s guidance on CGT on crypto sets out how losses are treated.
Algorithmic stablecoins are among the most intellectually ambitious and practically dangerous experiments in DeFi history: they attempt to maintain a stable price peg (almost always to the US dollar) using supply adjustment algorithms and token-based incentive mechanisms rather than holding actual fiat currency or over-collateralised crypto assets in reserve. The appeal is obvious: if an algorithm can maintain a stable price without requiring locked capital as collateral, you get the benefits of a stablecoin (price stability, unit of account, medium of exchange for DeFi) with capital efficiency that far exceeds over-collateralised alternatives like MakerDAO’s DAI. The risk is equally obvious in retrospect: algorithms maintain price stability only as long as market participants believe they will, and belief is self-reinforcing in both directions. The catastrophic collapse of TerraUSD (UST) in May 2022 (which wiped out about US$40 billion of combined UST and LUNA value within days) demonstrated that no algorithmic design yet developed can withstand a sustained loss of confidence at scale.
How Algorithmic Stablecoins Attempt to Maintain Their Peg
Pure algorithmic stablecoins typically use one of two mechanisms (or a combination) to maintain their price peg: supply elasticity (expanding and contracting the total supply of the stablecoin based on price) or a two-token system (where a companion “seigniorage share” token absorbs price volatility so the stablecoin can remain stable). Supply expansion: when the stablecoin trades above its target price (e.g., US$1.02), the protocol mints new stablecoins and distributes them to specific holders, increasing supply and theoretically pushing the price back down. Supply contraction: when the stablecoin trades below its target price (e.g., US$0.98), the protocol encourages users to burn stablecoins (in exchange for discounted claims on future supply or companion token rewards), reducing supply and theoretically pushing the price back up. The “theoretically” in both descriptions is crucial: these mechanisms work through economic incentives that depend on market participants believing the peg will be restored, not through any mechanical guarantee.
The two-token algorithmic stablecoin model (of which TerraUSD/LUNA was the most prominent example) creates a companion token that captures the “seigniorage” (the profit from creating new currency) when the stablecoin is in demand and absorbs the contraction cost when supply needs to shrink. In Terra’s case: when UST demand increased (price above US$1), new UST was minted by burning LUNA at the prevailing ratio, distributing the seigniorage to LUNA holders. When UST demand decreased (price below US$1), UST could be burned to receive US$1 worth of LUNA at the current price, contracting UST supply. This mechanism made LUNA an effectively unbounded liability: any amount of LUNA could be minted to absorb UST selling, with no cap on the dilution that could occur. When confidence collapsed and UST selling accelerated, the LUNA printing required to absorb the selling drove LUNA price down, which reduced the UST-to-LUNA exchange value, which required even more LUNA printing per UST burned, creating the hyperinflationary death spiral that destroyed both assets simultaneously.
Rebase or elastic supply models adjust the stablecoin balance in every holder’s wallet proportionally to maintain price stability without requiring any user action. When the price is above peg, all wallet balances increase proportionally (the total supply expands while each holder’s proportional share stays constant). When the price is below peg, all wallet balances decrease proportionally (the total supply contracts). Ampleforth (AMPL) is the longest-running example of this model. The rebase model does not actually create a stable asset in the way users expect from a stablecoin: your AMPL balance goes up and down daily based on price deviations, which means “holding 100 AMPL” has a different AUD value each day. The rebase mechanism creates equilibrium around the target price in the long run (under certain market conditions) but provides no stability guarantee for any specific time window, making it unsuitable as a stable medium of exchange for DeFi protocols or as a store of value.
Fractional-Algorithmic Models
Fractional-algorithmic models (pioneered by Frax Finance, which has since moved to full collateral backing) combine partial collateral backing with algorithmic mechanisms to achieve a middle ground between the stability of fully collateralised stablecoins and the capital efficiency of pure algorithmic designs. Frax’s model dynamically adjusts the collateral ratio (the proportion of each FRAX stablecoin backed by actual collateral like USDC) based on market conditions: high demand and strong peg performance lead the protocol to reduce the collateral ratio (more algorithmic, more capital efficient); weak demand or peg stress causes the collateral ratio to increase automatically (more collateralised, more robust). The fractional design has survived multiple market stress events better than pure algorithmic alternatives, demonstrating that partial collateral backing provides meaningful resilience even at reduced capital efficiency. For Australian DeFi investors evaluating fractional-algorithmic stablecoins, understanding the specific collateral ratio and the conditions under which it might fall further toward algorithmic backing is important context for assessing stability.
The fundamental instability of algorithmic stablecoins in severe market conditions was identified by economists (notably as the “impossible trinity” of algorithmic monetary policy) before TerraUSD’s collapse validated the theoretical concern empirically. The impossibility argument: a stablecoin cannot simultaneously be (1) pegged to a target price, (2) free to float in its supply, and (3) decentralised (with no central issuer guaranteeing the peg). Any two of these properties can coexist but not all three: if you want a stable peg and decentralisation, you need over-collateralisation (giving up supply efficiency); if you want efficiency and decentralisation, you get an elastic money that is stable only statistically (not guaranteed); if you want a stable peg and efficiency, you need a central issuer (giving up decentralisation). Recognising this framework helps explain why every pure algorithmic stablecoin ultimately faces the stability trade-off, and why the DeFi stablecoin market has progressively favoured collateralised designs over algorithmic ones.
The TerraUSD Collapse: A Case Study in Algorithmic Failure
The collapse of TerraUSD (UST) in May 2022 is the most important case study in algorithmic stablecoin risk, affecting millions of investors globally and serving as the definitive empirical demonstration of the death spiral vulnerability. At its peak, UST had a market capitalisation of about US$18 billion and was the third-largest stablecoin by market cap. The Terra ecosystem had attracted enormous capital through the Anchor Protocol, which offered 20% annualised yield on UST deposits (an unsustainable yield funded by investor subsidies rather than genuine lending demand, which was well-known by sophisticated participants but ignored by retail investors attracted to the headline rate). The 20% Anchor yield was the dominant driver of UST demand: when the subsidy began to decline and alternative yield sources emerged, demand for UST was vulnerable to a reversal.
The initial stress event (believed to have involved a large coordinated withdrawal from Anchor Protocol coupled with a large UST sale on Curve‘s liquidity pools) caused UST to depeg slightly below US$1. In a healthy algorithmic stablecoin with strong market confidence, this small depeg would trigger arbitrage: users burn UST to receive US$1 worth of LUNA, reducing UST supply and restoring the peg. But as the depeg deepened (from US$0.99 to US$0.98 to US$0.95), two things happened simultaneously: the LUNA price fell as the market anticipated the inevitable LUNA minting that would occur at scale; and confidence in UST’s recovery eroded, reducing the demand for UST from new buyers who would have provided natural demand pressure. The erosion of confidence became self-fulfilling: as LUNA fell in value, each UST burned bought less and less dollar value of LUNA, while simultaneously each new unit of LUNA minted to absorb the selling diluted existing LUNA holders further. Within seven days, LUNA had fallen from about US$80 at the start of May 2022 to a tiny fraction of a cent, and UST was trading at pennies on the dollar.
The contagion from the UST/LUNA collapse spread throughout the DeFi ecosystem and into centralised exchanges, affecting protocols and funds with direct exposure. Three Arrows Capital (3AC), a major crypto hedge fund, had substantial exposure to LUNA through its conviction in the Terra ecosystem; the collapse of LUNA was a primary factor in 3AC’s subsequent insolvency, which cascaded to counterparties including Genesis Trading, BlockFi, and Voyager Digital. The systemic nature of the collapse demonstrated that even investors who held no UST or LUNA directly could be affected through counterparty exposure to entities that did. For Australian DeFi investors who held crypto on platforms affected by the 3AC cascade (including some Australian platforms), the indirect exposure to algorithmic stablecoin risk was a painful lesson in systemic risk propagation. The DeFi risks of algorithmic stablecoins extend beyond direct holders to the broader ecosystem.
Post-Mortem Analysis of the UST Collapse
The post-mortem analysis of the UST collapse identified several warning signs that were present but either ignored or rationalised by investors who were comfortable with the 20% yield: the fundamental economic unsustainability of the Anchor yield (funded by investor subsidies, not genuine lending demand); the unbounded LUNA minting liability implicit in the burn-to-receive-LUNA mechanism; the concentration of UST demand in a single protocol (Anchor) that created catastrophic vulnerability to demand reversal from that single source; the rapid growth of UST supply without corresponding growth in genuine economic utility (most UST was in Anchor to earn yield, not in active DeFi use); and the inadequacy of the Luna Foundation Guard’s Bitcoin reserve as a stabilisation mechanism given the scale of the attack. Australian investors who suffered losses in the UST collapse and reflect on these warning signs in retrospect can identify the research framework that would have flagged the risks: the same questions (yield source? collateral backing? death spiral resistance? demand concentration?) that the framework in this article asks.
For Australian DeFi investors, the ATO’s tax treatment of UST/LUNA losses applied the standard capital gains tax framework to an unprecedented scenario. Assets that became effectively worthless in the UST collapse (UST stablecoins, LUNA tokens) created capital losses for Australian holders, calculated as: cost base minus sale proceeds (where “sale” might be the conversion during collapse or the effective abandonment of worthless assets). The ATO rules mean that simply holding a token that has become nearly worthless is not itself a CGT event: a capital loss generally arises when the holder disposes of it, for example by selling or swapping it for a tiny amount, as the ATO’s guidance on CGT on crypto explains. Once realised, UST/LUNA capital losses could offset capital gains in the same or subsequent income years. The complexity of the UST/LUNA collapse (involving multiple transaction types across multiple wallets and platforms) made professional crypto tax advice particularly important for affected Australian investors.
Warning Signs and Risk Assessment for Algorithmic Stablecoins
Identifying algorithmic stablecoin risk before it materialises requires specifically looking for the structural vulnerabilities that precede collapse. The primary red flags: (1) yield derived from protocol subsidies rather than genuine lending demand (any stablecoin yield above market lending rates must be subsidised from somewhere; understanding the subsidy source and its sustainability is essential); (2) unbounded minting liability for the companion token (if there is no cap on how much of the companion token can be minted to defend the peg, the mechanism has no floor in a death spiral); (3) concentrated demand from a single protocol or use case (demand that can reverse simultaneously and completely creates vulnerability to rapid supply contraction); (4) rapid growth in supply without commensurate growth in genuine economic utility; and (5) inadequate collateral reserves relative to the stablecoin supply (for fractional models). The presence of multiple red flags simultaneously should be disqualifying for any serious DeFi investment in an algorithmic stablecoin.
Peg deviation history provides empirical evidence of how a stablecoin has performed under actual market stress. A stablecoin that has maintained its peg within a tight range (0.5% or less) through multiple major crypto market downturns (the March 2020 crash, the May 2021 selloff, the November 2022 FTX collapse, and various 2023-2026 events) has demonstrated more practical resilience than one that has only operated in calm market conditions. Checking peg deviation charts on DeFiLlama’s stablecoin section and CoinGecko reveals the historical price range and any specific periods of stress. For algorithmic stablecoins with limited operational history (less than 12 months or no experience through a major market downturn), the absence of stress-test history is itself a relevant risk factor regardless of how elegant the mechanism design appears. DeFi risk assessment gives significant weight to empirical track record over theoretical design quality.
The dollar-backing ratio (for fractional algorithmic stablecoins) determines how much of any redemption demand can be met with actual collateral rather than requiring the algorithmic mechanism. A fractional stablecoin with 50% collateral backing can withstand 50% of its supply being redeemed for collateral before the algorithmic mechanism must handle the remaining redemptions. As the collateral ratio increases toward 100%, the mechanism approaches the safety of fully collateralised designs; as it approaches 0%, it approaches the pure algorithmic vulnerability. Monitoring the current collateral ratio of any fractional stablecoin you hold (available from the protocol’s own interface or on-chain) and understanding the conditions under which it adjusts (automatically by the protocol, or through governance votes) provides an ongoing risk metric that can inform hold or exit decisions if the ratio deteriorates.
Alternatives and the Regulatory Response
Alternative risk management approaches for Australian DeFi investors who want stable value exposure in DeFi without algorithmic stablecoin risk include: using collateralised stablecoins (DAI, LUSD) for DeFi activity requiring some level of decentralisation; using centralised stablecoins (USDC, USDT) for maximally stable peg performance with accepted counterparty risk; and accepting the yield trade-off by using lower-yield but more stable options rather than chasing algorithmic stablecoin yields. The risk-adjusted return calculation that considers the probability of a death spiral collapse (which can result in 90-100% loss of the stablecoin’s value within days) makes high-yielding algorithmic stablecoins unattractive relative to lower-yielding but more stable alternatives when the collapse probability is properly accounted for. The expected value of holding a stablecoin with 20% nominal yield and 5% annual death spiral probability is lower than the expected value of holding a stablecoin with 5% yield and near-zero death spiral probability.
The regulatory response to algorithmic stablecoins in Australia and globally has been significantly shaped by the UST collapse. Australian Treasury’s work on stablecoin regulation, which proposes treating payment stablecoins as a type of stored value facility, came after collapses such as UST highlighted the risks for retail users. The proposed regulatory frameworks in Australia and internationally (the European MiCA framework and the US GENIUS Act passed in July 2025) generally require stablecoin issuers to hold adequate reserves, which leaves little room for purely algorithmic models in retail-facing products. For Australian DeFi investors evaluating any stablecoin product that claims to be “algorithmic” or “decentralised” without clear collateral backing, the regulatory trajectory suggests these products are likely to face increasing compliance headwinds that could affect their viability as investable assets in the Australian market. Monitoring Australian crypto regulatory developments is part of the ongoing risk assessment for any algorithmic stablecoin position.
Frequently Asked Questions
What are the risks of algorithmic stablecoins?
Algorithmic stablecoins try to hold a steady value, usually US$1, through supply rules and a companion token rather than full reserves. The main risk is a death spiral: if confidence falls, holders rush to exit, the companion token collapses and the peg can fail completely. TerraUSD showed this in May 2022, when UST and LUNA lost about US$40 billion of combined value within days.
How do algorithmic stablecoins try to maintain their peg?
Most use supply changes, a companion token or both. When the coin trades above its target, for example US$1.02, new coins are created to push the price down. When it trades below, for example US$0.98, holders are encouraged to burn coins in exchange for the companion token or future rewards. This depends on traders believing the companion token will keep its value.
How do fractional-algorithmic models work?
Fractional-algorithmic models (pioneered by Frax Finance) combine partial collateral backing with algorithmic mechanisms to achieve a middle ground between the stability of fully collateralised stablecoins and the capital efficiency of pure algorithmic designs. Frax's model dynamically adjusts the collateral ratio (the proportion of each FRAX stablecoin backed by actual collateral like USDC) based on market conditions: high demand and strong peg performance lead the protocol to reduce the collateral ratio (more algorithmic, more capital efficient); weak demand or peg stress causes the collateral ratio to increase automatically (more collateralised, more robust). The fractional design has survived multiple market stress events better than pure algorithmic alternatives, demonstrating that partial collateral backing provides meaningful resilience even at reduced capital efficiency.
What happened in the TerraUSD collapse?
TerraUSD peaked at about US$18 billion, making it the third-largest stablecoin, largely because the Anchor Protocol paid around 20% a year on deposits. In May 2022 large withdrawals and sales pushed UST below US$1. Holders swapped UST for newly minted LUNA, flooding the market, and within about a week LUNA fell from around US$80 to a tiny fraction of a cent.
What warning signs preceded the UST collapse?
Warning signs included Anchor's 20% yield, which was subsidised rather than earned from real borrowing demand, the unlimited amount of LUNA that could be minted to defend the peg, most UST demand coming from a single protocol, and supply growing far faster than real use. The bitcoin reserve built to defend the peg was also far too small for a full run.
How do you assess algorithmic stablecoin risk?
Key red flags include yields well above normal lending rates, a peg defended mainly by a companion token the same team controls, demand concentrated in one protocol, little or no hard collateral, and no history through a major market downturn. A coin that has held its peg tightly through several crashes and holds substantial real backing is lower risk, though never risk-free.
What safer alternatives exist for stable value in DeFi?
Common alternatives are fiat-backed stablecoins such as USDC and USDT, which carry issuer and freezing risk but have held their pegs far better, and over-collateralised crypto-backed coins such as DAI, which rely on smart contracts and collateral. Spreading holdings across more than one design, and avoiding yields that look too good to be true, reduces reliance on any single mechanism.
What are the risks associated with Risks of Algorithmic Stablecoins?
Algorithmic stablecoins can fail completely and quickly: TerraUSD lost its peg and collapsed within days, wiping out holders and taking associated protocols with it. The failure mode is reflexive, because the mechanism that maintains the peg depends on confidence in the peg holding. Advertised yields on these systems are frequently funded by the same reflexivity rather than by revenue. Treating any uncollateralised or partially collateralised stablecoin as safe because it has held its peg so far is the specific error to avoid.