Stablecoins cluster into three design families: fiat-collateralized, crypto-collateralized, and algorithmic. Fiat-backed coins dominate usage but depend on custodian and disclosure quality; crypto-collateralized coins are transparent and permissionless but must over-collateralize and rely on liquidations and oracles; algorithmic models promise capital efficiency but have repeatedly failed under stress. Regulation is tightening in both the EU (MiCA) and the U.S. (GENIUS Act 2025), raising the bar on reserves, disclosure, and redemptions.
How regulators and economists classify stablecoins in 2025
Most authorities and research bodies bucket designs by how they secure the peg: fiat-backed, crypto-collateralized, and algorithmic or uncollateralized. The design choice shapes use cases, centralization, and failure modes.
Policy backdrop has materially shifted. In the EU, MiCA governs e-money tokens (EMTs) and asset-referenced tokens (ARTs), with the EBA and ESMA issuing standards and guidance. Some platforms that handle EMT payments may also need payment/e-money licenses alongside MiCA authorizations.
In the U.S., the GENIUS Act (signed July 18, 2025) establishes a federal framework for payment stablecoins, including full-reserve requirements, disclosures, and oversight.
Fiat-collateralized stablecoins
How they work
Issuers mint tokens against fiat and cash-equivalent reserves (e.g., T-bills), targeting 1:1 redemption. Stability depends on reserve quality, segregation, transparency, and redemption mechanics.
Examples and disclosures
USDC publishes regular reserve information and monthly third-party assurance; Circle states that a Big Four firm performs monthly attestations. Circle also clarifies redemption access and institutional workflows.
Tether (USDT) releases quarterly assurance reports by BDO Italia; its Q2 2025 opinion states reserves exceeded liabilities by several billion dollars. Tether also lists minimum redemption amounts and fees on its site.
PayPal USD (PYUSD), issued by Paxos, markets 1:1 redemption and reserves in deposits, Treasuries, and cash equivalents; see Paxos and PayPal docs and white paper.
Redemption, fees, and access
Direct redemption varies. Tether requires at least 100,000 USD and charges the greater of $1,000 or 0.1% per redemption, making exchanges the practical route for smaller users. USDC offers institution-grade redemption rails and documents instant and tiered redemption limits.
Pros and cons
Pros: low volatility when reserves are high-quality and liquid; deep liquidity. Cons: issuer and custodian risk; audit/attestation ambiguity; policy changes (e.g., chain support changes) can affect users.
Crypto-collateralized stablecoins
How they work
Users lock volatile crypto (e.g., ETH) in smart contracts and mint a dollar-pegged asset at an overcollateralized ratio. Pegs are maintained via incentives and liquidations; some systems also add a peg stability module (PSM) to allow swaps against other stables.
MakerDAO’s DAI
DAI began as ETH-only collateral and evolved into a multi-collateral design using vaults, a PSM, and real-world assets such as short-dated U.S. Treasuries via partners. The Fed’s note explains both the vault path and PSM path for new DAI.
Recent analyses document the shift toward diversified collateral that includes fiat-backed stables and RWAs, making DAI a hybrid between pure crypto-backed and fiat-linked designs.
Liquity’s LUSD
Liquity requires a minimum 110% collateral ratio for ETH-backed loans; loans are interest-free and liquidations enforce solvency. This offers high capital efficiency compared with many DeFi peers.
Pros and cons
Pros: on-chain transparency; permissionless access; diversified collateral options. Cons: liquidation and oracle risk; need for overcollateralization; PSMs can re-centralize risk by depending on other stablecoins.
Algorithmic models (and why they struggle)
What “algorithmic” means
Algorithmic designs target stability using supply changes or two-token seigniorage mechanisms rather than fully matched reserves. Examples include rebasing (e.g., AMPL) and seigniorage share models (e.g., TerraUSD/LUNA).
Case studies
Academic and policy work has dissected the 2022 TerraUSD collapse as a reflexive bank-run or “death spiral,” where redemptions into a volatile balancing token overwhelmed the system. Other algorithmic experiments such as FEI ultimately wound down after security and governance stress.
Central bank and BIS papers flag persistent weaknesses in unbacked/algorithmic models, noting repeated peg losses under stress.
The FRAX evolution
Frax moved from a fractional-algorithmic approach toward near/full collateralization in v3, retaining algorithmic market operations while targeting a 100% collateral ratio. Ratings and project docs describe this transition.
A fourth bucket emerging: synthetic, delta-hedged “dollars”
Protocols like Ethena’s USDe maintain a dollar target with crypto collateral hedged via short perps (basis trades), plus liquid stables. Official materials and independent coverage call it a “synthetic dollar,” not a classic fiat-backed coin; risk depends on derivatives markets, funding rates, and counterparty exposure.
Side-by-side comparison
| Dimension | Fiat-collateralized (USDC, USDT, PYUSD) | Crypto-collateralized (DAI, LUSD) | Algorithmic / Synthetic (UST legacy, AMPL; USDe) |
|---|---|---|---|
| Peg mechanism | Redeemable against fiat/T-bill reserves | Overcollateralized vaults; liquidations; PSM | Supply rules or delta-hedged strategies; often no 1:1 fiat reserves |
| Transparency | Off-chain attestations/audits and reserve reports | On-chain collateral + governance docs | Varies; algorithmic code is open, but off-chain hedging adds opaqueness |
| Redemption path | Issuer redemption; exchanges for retail | Repay debt/close vault; PSM swaps | Mechanism-specific; may lack direct fiat redemption |
| Typical fees/limits | Issuer fees and minimums (e.g., USDT 100k min redemption; 0.1%/≥$1k) | Protocol fees; liquidation penalties | Trading/hedging costs; funding-rate risk |
| Main risks | Custodian/regulator risk; disclosure quality | Oracle/liquidation risk; collateral drawdowns | Reflexivity, derivatives market stress, funding-rate reversals |
Key data points are documented in issuer pages, protocol docs, and central-bank or policy research linked throughout this article.
What 2025 regulation means for each design
European Union (MiCA)
MiCA brings authorization, disclosure, and supervision for ARTs/EMTs. The EBA and ESMA publish technical standards; firms facilitating EMT payments may also require e-money/payment licenses in addition to MiCA authorization. Expect stricter redemption rights, liquidity, and governance controls.
United States (GENIUS Act)
The GENIUS Act creates a national framework for payment stablecoins, mandating full-reserve backing with high-quality liquid assets, disclosures, and supervisory oversight. The White House announcement confirms enactment on July 18, 2025.
Due-diligence checklist before you hold or integrate a stablecoin
- Reserves and disclosure cadence: look for named custodians, asset breakdowns, and third-party assurance or audits.
- Redemption terms: confirm minimums, fees, settlement rails, and who can redeem (retail vs institutions).
- Market share and depth: dominant fiat-backed coins have the most liquidity; verify up-to-date market stats from reputable trackers or official bodies.
- Smart-contract and oracle design (for crypto-collateralized): review liquidation mechanics, collateral ratios, and oracle sources.
- Regulatory posture: check MiCA status in the EU and GENIUS Act alignment in the U.S. for payment use cases.
- Stress history: study past depegs or failure cases and what changed afterward.
Frequently asked questions
What’s the most reliable design for payments today?
For mainstream payments, fiat-collateralized EMTs under MiCA and payment stablecoins under the GENIUS Act are likely to see the clearest compliance pathways and merchant support, assuming strong reserve disclosures and redemption rails.
Are crypto-collateralized stablecoins truly decentralized?
They remove fiat custodians but still rely on oracles, liquidations, and sometimes PSMs that swap into centralized stables, creating hybrid dependencies. Evaluate each protocol’s governance and collateral mix.
Do algorithmic stablecoins work?
They can function in benign markets but have repeatedly broken under stress. Policymakers and research groups have documented these shortcomings extensively since 2022.
Where do synthetic, delta-hedged dollars fit?
They are not classic fiat-backed stables and depend on derivatives markets and counterparties. They may offer yields via funding spreads but inherit new risks.
Sources and further reading
- Federal Reserve FEDS Note on stablecoin design categories.
- EU MiCA overview and EBA landing page for ARTs/EMTs; news on additional licensing for EMT payment activity.
- White House fact sheet confirming the GENIUS Act signing.
- USDC transparency and redemption resources; Tether assurance and fees; PYUSD reserves.
- MakerDAO PSM and evolution of DAI collateral; Liquity LUSD docs.
- Algorithmic failures and analyses: TerraUSD (CRS/BIS/academic).
- Synthetic dollars: Ethena USDe docs and press coverage.
