Stablecoins: Three Designs and How Each One Can Break

The Three Stablecoin Architectures

Stablecoins are cryptocurrencies designed to maintain a stable value relative to a reference asset, typically the US dollar. They serve as a bridge between volatile crypto markets and the world of everyday payments and lending. But not all stablecoins are built the same. There are three primary designs: fiat-collateralized, crypto-overcollateralized, and algorithmic. Each uses a different mechanism to maintain its peg, and each has a distinct vulnerability that can lead to a break.

• Fiat-collateralized stablecoins are backed by reserves of traditional currency or cash-equivalents held in bank accounts.
• Crypto-overcollateralized stablecoins are backed by other cryptocurrencies, with a surplus of collateral to absorb price swings.
• Algorithmic stablecoins rely on code to expand and contract the coin supply in response to demand, with no collateral backing.

Understanding these designs is crucial for anyone using or evaluating stablecoins. Historical episodes of de-pegging reveal exactly how each model can fail.

Fiat-Collateralized: Reserves and Trust

Fiat-collateralized stablecoins, such as USDT (Tether) and USDC (USD Coin), maintain their peg through a simple promise: for every token in circulation, the issuer holds an equivalent amount of fiat currency or highly liquid assets. Users can always redeem one token for one dollar from the issuer, creating an arbitrage mechanism. If the market price of the token drops below one dollar, traders can buy tokens cheaply and redeem them for one dollar from the issuer, pocketing the difference. This buying pressure pushes the price back up. If the price rises above one dollar, traders can mint new tokens by depositing dollars, increasing supply and pushing the price down.

The peg holds as long as two conditions are met: the issuer actually holds sufficient reserves, and those reserves remain liquid and accessible. The failure mode is a loss of trust in reserves. If users suspect the reserves are insufficient or frozen, redemption may halt. This triggers a bank run: everyone rushes to redeem at once, but the issuer may not have enough liquid assets. The token then trades far below its peg. In 2022, the collapse of FTX and Alameda Research led to a brief de-peg of USDT when fears spread about Tether's exposure to those firms. The peg eventually recovered, but the episode highlighted the risk: fiat-collateralized stablecoins depend entirely on the honesty and solvency of a central issuer.

Crypto-Overcollateralized: Collateral Volatility and Liquidations

Crypto-overcollateralized stablecoins, like DAI, are backed not by fiat but by other cryptocurrencies. To create one DAI, a user must lock up more than one dollar's worth of Ethereum (or another accepted crypto) in a smart contract. For example, a user might deposit 1.50 dollars worth of ETH to mint 1 DAI. The overcollateralization absorbs fluctuations in the collateral's value. If ETH drops, the system can liquidate the collateral to maintain the peg. The arbitrage mechanism works differently: if DAI trades below one dollar, users can buy DAI and use it to repay their loans, effectively redeeming it at base value. If DAI trades above one dollar, users can mint new DAI by depositing more collateral and sell it for profit, increasing supply.

The failure mode here is extreme volatility in the underlying collateral. If the crypto backing falls in value too quickly, the system may not be able to liquidate positions fast enough, causing a cascade of liquidations and undercollateralization. This can break the peg. During the crash of May 2020 (Black Thursday), the Ethereum network became congested, and liquidations stalled. DAI briefly traded near 0.95 dollars. The system survived due to emergency actions and later improvements. But the risk remains: a sudden, severe market crash can overwhelm the liquidation mechanisms, causing DAI to trade below parity until stability is restored.

Algorithmic: Supply Elasticity and the Death Spiral

Algorithmic stablecoins use no collateral. Instead, they rely on algorithms to adjust the token supply to keep the price at one dollar. The classic example is the now-defunct TerraUSD (UST). In its design, UST was minted and burned using a sister token, LUNA. When UST's price fell below one dollar, users could burn UST to mint LUNA, reducing UST supply and increasing demand. When UST rose above one dollar, users could burn LUNA to mint UST, increasing supply. This arbitrage was meant to maintain the peg automatically.

The failure mode is a devastating one: a loss of confidence triggers a death spiral. When many users try to exit UST simultaneously, they burn UST to mint LUNA, flooding the market with LUNA tokens. LUNA's price plummets, making the arbitrage less attractive. The supply of UST shrinks more slowly than the collapse in demand, so the peg breaks. UST trades far below one dollar, and LUNA collapses. In May 2022, UST de-pegged and fell to near zero, taking LUNA with it. The root cause was a bank run on the algorithmic system: once trust vanished, no mechanism could stop the downward spiral.

Comparing the Vulnerabilities

Each design has a signature weakness.

• Fiat-collateralized suffers from the need for trusted, audited reserves and the risk of illiquidity or fraud.
• Crypto-overcollateralized is vulnerable to rapid declines in the collateral's market value and network congestion.
• Algorithmic is the most fragile, because it lacks any real backing and depends entirely on sustained demand for its own token.

No system is perfect. Even fiat-collateralized stablecoins face regulatory risk: if authorities freeze issuer bank accounts, redemptions stop. Crypto-overcollateralized stablecoins must constantly manage liquidation parameters. Algorithmic stablecoins have proven to be inherently prone to bank runs, as history shows.

Lessons for Users

When choosing to use or hold a stablecoin, one must understand which design is in play. The trade-offs are clear: fiat-collateralized offers simplicity but relies on a central party; crypto-overcollateralized offers decentralization but risk from volatile markets; algorithmic offers pure code but has no safety net. Diversifying across types may reduce the impact of a single failure. But no stablecoin is risk-free. The history of de-pegs shows that each design can break in its own way, and the consequences can be severe. Stay informed, and never invest more than you can afford to lose.