Why Algorithmic Stablecoins Are Inherently Fragile

The core mechanism is a positive feedback loop. A price drop below peg triggers contractionary measures (e.g., burning tokens, minting debt), which destroys confidence, causing further selling. This is the opposite of a dampening, stabilizing system.

• Negative Feedback vs. Positive Feedback: Stable systems (like a thermostat) correct deviations. Algo-stables amplify them.
• Death Spiral Speed: Depegs can happen in <24 hours from a minor shock, as seen with Terra/LUNA and Iron/TITAN.

All contraction/expansion logic is triggered by an external price feed. This creates a single, manipulable point of failure. A brief oracle failure or flash crash can trigger irreversible protocol death.

• Manipulation Surface: Low-liquidity pairs or flash loan attacks can feed false data.
• Irreversible Actions: Once the protocol mints/burns based on bad data, the damage is permanent. This is a fundamental mismatch between oracle latency and on-chain finality.

If the stablecoin is truly backed by exogenous collateral (e.g., USDC, ETH), it's not algorithmic—it's a collateralized stablecoin. The 'algorithmic' label is a marketing term for undercollateralization. The promised efficiency is just leverage.

• Leverage = Risk: UST was marketed as 'decentralized' but was simply a ~140% levered bet on LUNA.
• Inevitable Run Risk: Undercollateralized systems cannot survive a loss of confidence, as there is no asset buffer. All $10B+ TVL in algo-stables to date has evaporated.

Proposals to 'save' a depegging algo-stable (e.g., adjusting parameters, adding collateral) are theater. By the time governance can act, the reflexive death spiral is unstoppable. Governance is a liability, not a safeguard.

• Speed Mismatch: Governance votes take days. Market moves in seconds.
• Adversarial Forks: During a crisis, the community and tokenholders' incentives violently diverge, leading to contentious forks that destroy remaining value.

Algorithmic models like Terra/LUNA create a reflexive feedback loop between the stablecoin (UST) and its governance/volatile asset (LUNA).

• Problem: A drop in stablecoin demand triggers minting of volatile tokens to maintain peg, diluting holders and collapsing confidence.
• Solution: Overcollateralized models (MakerDAO's DAI, Liquity's LUSD) use exogenous, non-reflexive assets as a capital buffer, decoupling stability from native token price.

All algorithmic stablecoins are oracle-dependent, but undercollateralized ones are hypersensitive.

• Problem: A manipulated price feed for the collateral asset (e.g., MIM's reliance on TIME) can mint infinite stablecoins against worthless collateral, as seen in the Wonderland exploit.
• Solution: Robust, decentralized oracle networks (Chainlink) with multi-layer aggregation and circuit breakers are non-negotiable infrastructure, not an afterthought.

Stability mechanisms like seigniorage or bonding curves rely on perpetual, deep liquidity that evaporates during stress.

• Problem: Iron Finance's TITAN and Basis Cash demonstrated that arbitrage incentives fail when the backing 'reserve' is an illiquid, ponzi-like token. Peg defense becomes a liquidity black hole.
• Solution: Direct, on-chain liquidity pools of high-quality assets (e.g., USDC in Curve pools for FRAX's hybrid model) provide a tangible defense fund that doesn't rely on tokenomics.

Algorithmic parameters (minting fees, collateral ratios) are controlled by governance tokens, creating a centralization vector.

• Problem: A malicious or compromised governance vote (e.g., a Beanstalk flash loan attack) can instantly drain the protocol's treasury or mint unlimited stablecoins.
• Solution: Immutable core contracts (Liquity) or time-locked, multi-sig executed upgrades with strong community oversight (MakerDAO) mitigate this existential risk.

Algorithmic systems are hyper-optimized for a specific monetary environment and shatter under regime change.

• Problem: UST's reliance on the Anchor Protocol's ~20% yield created unsustainable demand. When macro conditions shifted and yields collapsed, the entire demand thesis evaporated overnight.
• Solution: Stability must be derived from utility and organic demand (e.g., DAI in DeFi lending), not artificially subsidized yields that attract mercenary capital.

Even partially algorithmic models introduce fragility, trading some safety for capital efficiency.

• Problem: FRAX's fractional-algorithmic design must dynamically adjust its collateral ratio (CR) during volatility. If the CR hits 0% during a bank run, it becomes a purely algorithmic stablecoin with all the attendant risks.
• Solution: A hard floor for collateralization (e.g., MakerDAO's >100% minimum) is a circuit breaker that prevents a full transition into a fragile state, prioritizing survival over optimal capital use.

Algorithmic stablecoins create a positive feedback loop between price and demand. A rising price attracts more users, which expands the supply. A falling price triggers redemptions, forcing supply contraction and further selling pressure. This makes them inherently pro-cyclical and unstable.

• Death Spiral: The 2022 collapse of Terra's UST and LUNA erased ~$40B in market cap in days.
• No Anchor: Value is purely based on the market's belief in the mechanism, not an external asset.

All algorithmic systems require a trusted price feed to trigger mint/burn functions. This creates a single point of failure that is vulnerable to manipulation, latency, and centralization.

• Attack Vector: A manipulated oracle can trigger incorrect supply expansions or contractions.
• Centralized Reliance: Most feeds depend on a small set of CEX APIs (e.g., Binance, Coinbase), reintroducing custodial risk.

Algorithmic stablecoins attempt to perform central banking without a lender of last resort or diversified revenue base. During a crisis, demand for stability is inelastic, but the protocol can only offer more of its own volatile governance token.

• No Real Backstop: Contrast with MakerDAO's DAI, which uses $10B+ in diversified collateral and the PSM.
• Cantillon Effects: Early adopters and insiders benefit from seigniorage, while late users bear the brunt of de-pegs.

To survive, algorithmic systems require constant parameter tuning (e.g., redemption rates, fees). This places immense pressure on often-slow and politically charged DAO governance, making them unable to react at market speed.

• Reaction Lag: Governance votes take days; market crashes happen in minutes.
• Attack Surface: A malicious proposal or voter apathy can cripple the system permanently.