What is a stablecoin: The peg, the exit door, and why it breaks

A stablecoin is a crypto token designed to hold a stable value versus a reference asset, most often the U.S. dollar. It is not “a digital dollar” so much as a dollar claim with a mechanism, and stress tests always come down to whether 1 token can be converted into $1 quickly at size.

Key Takeaways

• A stablecoin’s $1 price is a market outcome supported by an “exit door” like issuer redemption, onchain liquidation, or arbitrage.
• The main types of stablecoins are a fiat backed stablecoin, a crypto backed stablecoin, and an algorithmic stablecoin, and each fails in a different way.
• USDC’s March 10–13, 2023 SVB weekend showed how secondary-market prices can gap when primary redemptions are constrained.
• Stablecoin regulation expanded across the EU (MiCA), Singapore (MAS), the UK (FCA), and the U.S. (GENIUS Act), tightening rules around reserves and redemption.

Stablecoins and the idea of a peg

A stablecoin works because it targets a specific reference value and then recruits markets to enforce it. That target is the peg, usually 1 token ≈ $1. The important detail is that the peg is not a law of nature and not a guarantee. It is a price relationship that holds when the conversion path between the token and the reference asset stays open and credible.

That framing matters because “what is a stablecoin crypto” is often answered with a vibe, not a mechanism. Stablecoins are crypto-assets designed to maintain a stable value against a reference asset, typically the U.S. dollar. The stability goal is achieved by backing assets and or by rules that change supply. When those supports are questioned, the token can trade away from $1 even if the design intends otherwise.

A depeg is the name for that break in the relationship. Sources define depeg as a material deviation above or below the intended reference value. Kraken’s education material also makes a point traders learn quickly on screen: small deviations can happen simply because liquidity differs by venue and time. That is why “stable” should be read as “anchored,” not “fixed.”

The clean mental model is to treat a stablecoin like cash in a prime brokerage account. The number on the screen only matters because there is an exit door. If the exit door is wide, arbitrage keeps the peg tight. If the exit door narrows, secondary markets start pricing the risk that conversion will be slow, limited, or loss-making.

How stablecoins keep their value

Three mechanisms show up again and again, and each one defends the peg with a different exit door. Eco groups them as: (1) fiat reserves held by a custodian, (2) crypto collateral locked in smart contracts at a ratio above 1:1, and (3) algorithmic supply adjustments that mint and burn against a paired asset. The mechanism determines who can convert, what they convert into, and what happens when everyone tries at once.

For a fiat backed stablecoin, the exit door is issuer redemption. The issuer mints tokens when dollars come in and redeems tokens for dollars when tokens come back. The peg holds when authorized holders can reliably redeem at par and when secondary-market traders believe that redemption is real at scale. This is where stablecoin reserves matter. Eco notes that large issuers publish monthly attestations from independent accounting firms, citing BDO for Tether and Deloitte for Circle, to show what backs the supply.

For a crypto backed stablecoin, the exit door is liquidation. Users mint the stablecoin by locking more than $1 of onchain collateral for each $1 issued, which is collateralization by design. If collateral prices fall and a position breaches thresholds, the protocol liquidates collateral to keep the system solvent. MakerDAO, now rebranded as Sky, pioneered this model in 2017 with DAI, using overcollateralized positions and auctions when collateral drops through liquidation levels.

For an algorithmic stablecoin, the exit door is arbitrage against a paired asset and the system’s own mint and burn rules. The design tries to pull price back to $1 by expanding supply when price is above peg and contracting supply when price is below peg, often by swapping between two tokens. It can look smooth until confidence flips, because the mechanism depends on the market value and liquidity of the paired asset staying intact.

Major stablecoin types and examples

The stablecoin market is dominated by custodial dollars, not clever onchain engineering. Eco reports stablecoins as a roughly $319.6B asset class as of April 29, 2026, with fiat-backed tokens about 84% of the market. That dominance is not ideological. It is mechanical. Redemption for dollars is the simplest story to underwrite, even if it concentrates risk in banks, custodians, and policy.

The two names most beginners recognize are tether and usd coin. Eco’s April 2026 figures put USDT around $189.6B in circulation and USDC around $77.6B, with DAI around $4.7B and Sky’s USDS around $8.4B. Those numbers are time-sensitive, but the hierarchy is the point. Liquidity tends to cluster where the market already is, and that liquidity itself helps keep pegs tighter on more venues.

“Types of stablecoins” is not just taxonomy. It is a map of who eats the loss when something goes wrong. With fiat-backed coins, the loss shows up when reserves are impaired, inaccessible, or redemption is constrained. With crypto-collateralized coins, the loss is pushed onto leveraged minters through liquidations when collateral falls. With algorithmic designs, the loss can land on anyone holding the stablecoin when the reflexive loop breaks.

The comparison people actually care about is usdt vs usdc, because both are fiat-backed but they do not share the same banking footprint, regulatory posture, or redemption access. Eco highlights that Circle’s USDC is MiCA-compliant while Tether chose not to seek MiCA authorization and was delisted from regulated EU exchanges. That is not a moral judgment. It is a distribution fact that changes where liquidity can legally sit.

Why people use stablecoins

Stablecoins exist because most crypto activity still wants a stable unit of account. Traders want to park PnL without taking BTC or ETH volatility. Protocols need a dollar-like yardstick to price loans, set liquidation thresholds, and quote liquidity pools. That is why stablecoins are the working capital of what is defi, not a side feature.

Eco lists four core use cases that keep demand persistent: cross-border payments, DeFi collateral, treasury management, and on and off ramps between exchanges and bank accounts. The point is not that stablecoins are always cheaper or always better. The point is that they settle like tokens while behaving like cash most of the time, which is a powerful combination when banking rails are slow or fragmented.

DeFi collateral is where stablecoin design choices turn into second-order effects. The Federal Reserve note on the March 2023 USDC episode describes stablecoins as performing dollar-like functions in decentralized finance and as run-able liabilities for issuers. When a major stablecoin wobbles, protocols that treat it as “cash collateral” can transmit the shock. During the SVB weekend, Dai’s Peg Stability Modules offered one-to-one exchange facilities against USDC and other stablecoins, and those PSMs were rapidly drained, pulling stress into a stablecoin that had no direct SVB exposure.

For beginners, the simplest way to understand demand is to watch where pairs are quoted. Centralized exchanges quote a large share of spot markets against USDT or USDC because stablecoins reduce the need to touch banking rails for every trade and let traders rotate between risk assets and “cash” instantly.

Key risks and how depegs happen

Depegs are not all the same event. They are the market repricing the exit door. When the door is redemption, the failure mode looks like a run. When the door is liquidation, the failure mode looks like cascading liquidations and auction stress. When the door is algorithmic arbitrage, the failure mode looks like reflexivity and a confidence spiral.

USDC’s March 10–13, 2023 SVB episode is the cleanest case study because it separated primary and secondary markets. The Fed documents that Circle disclosed at 10 p.m. on March 10 that it could not access $3.3B of reserves held as uninsured deposits at SVB, about 8% of reserves at the time. Redemption requests surged. When Circle shut down primary market operations over the weekend due to banking-hour constraints, USDC lost its peg on secondary markets. The Fed reports a trough of $0.86, while other sources cite venue-specific prints, which is exactly the point. The secondary market trades the probability-weighted value of future redemption when redemption is not currently available.

Algorithmic failures are the other end of the spectrum. TerraUSD’s May 2022 collapse showed how quickly an algorithmic stablecoin can go from “works fine” to “no bid.” Eco describes UST losing about an $18B market cap in three days, while Changelly describes about $60B wiped across the Terra ecosystem and a LUNA supply explosion. The precise number depends on what is being measured, but the mechanism was the same. Once the paired asset could not credibly absorb redemptions, the arbitrage loop turned into a feedback loop.

Risk also comes from governance and control. Centralized issuers can freeze addresses to comply with sanctions, which Eco and Changelly both treat as a real operational feature. That can reduce some categories of fraud and compliance risk for platforms, but it creates a different risk for holders who rely on permissionless transferability.

Regulation and practical safety checks

Stablecoin regulation stopped being a theoretical debate and became a set of named rulebooks between 2023 and 2026. Eco lists major frameworks: the EU’s MiCA regime, Singapore’s MAS stablecoin framework, the UK FCA’s final stablecoin rules, and the U.S. GENIUS Act passed in February 2026. MiCA took full effect for stablecoins on June 30, 2024 and classifies them as “asset-referenced tokens” and “e-money tokens.”

Regulation matters because it shapes the redemption promise and the reserve constraints that make that promise believable. Eco’s summary of the GENIUS Act draws a bright line between payment stablecoins and yield-bearing tokens, and it specifies reserve asset requirements for payment stablecoin issuers. That is not a guarantee against depeg, but it is an attempt to standardize what stablecoin reserves can be.

A beginner-friendly evaluation process starts with how stablecoins maintain their peg, not with branding. The checks that actually map to failure modes are:

1. Identify the exit door. Is conversion supported by issuer redemption, onchain liquidation, or algorithmic arbitrage. 2. Verify the backing path. For fiat-backed coins, look for published attestations and what assets are listed as reserves. For crypto-backed coins, inspect the collateral model and liquidation design. 3. Map the bottleneck. The Fed’s USDC case shows primary redemption can be constrained by banking hours and operational limits, even when assets exist. 4. Check dependency chains. The Fed highlights how DeFi interlinkages like Dai’s PSMs can import risk from a centralized stablecoin into a “decentralized” one. 5. Price the possibility of a depeg. Kraken’s framing that small deviations can be liquidity-driven is useful, but it should not be confused with solvency. A tight peg is a symptom of confidence and liquidity, not proof of safety.

The Take

I’ve watched too many people treat stablecoins as “cash” and then act surprised when the cash behaves like a claim. The March 10–13, 2023 USDC weekend is the template: Circle disclosed $3.3B stuck at SVB, primary redemptions slowed over the weekend, and secondary markets did what they always do when the exit door narrows. They priced the delay and the uncertainty.

The expensive misconception is thinking the peg is a promise. It is a trade that clears because someone can convert 1 token into $1 fast enough at size. When that conversion is redemption, liquidation, or arbitrage, the only question that matters in stress is whether the door stays open when everyone reaches for it at once.

Sources

• Federal Reserve
• Eco
• Kraken
• Changelly