What actually keeps a stablecoin at $1?
A peg is only as hard as the redemption path. If a holder can reliably exchange one token for one dollar of reserves, one dollar of overcollateralized crypto value, or one dollar of hedged collateral, arbitrage usually pulls price back toward par. If redemption is slow, discretionary, or dependent on a reflexive governance token, the peg becomes a confidence trade rather than a hard conversion claim. Regulators now frame stablecoin risk around reserve quality, redemption design and governance, not branding.
Stablecoin tokenomics is mostly liability management. The token is the liability. The reserve, collateral, hedge, or governance backstop is the asset side. The key question is not “is it stable,” but “who can redeem, against what asset, through which mechanism, under whose control, and with what delay.” That lens matters more than whether the protocol calls itself fiat-backed, decentralized, or synthetic.
Which peg models matter in practice?
Most live stablecoins fit into four economically distinct models. The table below is the cleanest way to compare them.
| Model | How the peg is held | Structural strength | Structural weakness | Representative examples |
|---|---|---|---|---|
| Fiat-backed custodial | Issuer promises 1:1 redemption against cash and cash-equivalent reserves. | Tight peg, simple mental model, strong secondary-market arbitrage. | Issuer can blocklist, pause, upgrade contracts, or gate redemption. | USDC, USDT |
| Crypto-backed overcollateralized | Users lock volatile collateral and mint less than the collateral value, with liquidation and redemption logic defending solvency. | Lower reliance on banks. Onchain visibility. Redemption can be protocol-native. | Capital inefficient. Oracle and liquidation risk. Governance can re-centralize the design. | DAI, LUSD |
| Synthetic / delta-neutral | Spot collateral is offset with short derivatives to target a synthetic dollar exposure. | Less dependence on banks. Potentially scalable if hedge markets are deep. | Exchange, custody, funding-rate, and governance-committee risk remain. | USDe |
| Algorithmic / reflexive | Peg relies partly or mainly on supply contraction, seigniorage, or conversion into a volatile endogenous token. | High capital efficiency in benign conditions. | Confidence-sensitive and vulnerable to death spirals under stress. | Legacy FRAX V1, TerraUSD |
The decentralization ranking is not the same as the peg-tightness ranking. Fiat-backed coins usually hold the tightest peg because a legal issuer manages reserves and redemptions directly. Governance-minimized crypto-backed systems usually disperse control better, but they pay for that with lower capital efficiency and narrower operating flexibility. That trade-off is structural, not rhetorical.
Why do fully backed stablecoins still depeg?
“Fully backed” does not eliminate redemption-timing risk. A reserve portfolio can be solvent on paper and still trade below par if the market doubts immediate access to that reserve, or doubts which holders can actually redeem. On March 11, 2023, Circle said $3.3 billion of USDC reserves, about 8% of the total reserve at the time, were at Silicon Valley Bank. USDC then traded below $1 until access to those funds was clarified.
Reserve composition and legal control matter as much as headline backing. Circle says USDC is always redeemable 1:1 and backed by highly liquid cash and cash-equivalent assets, with the majority of reserves in the Circle Reserve Fund, an SEC-registered Rule 2a-7 government money market fund. That is materially different from an uninsured bank deposit stack, but it still depends on regulated intermediaries, market plumbing, and issuer operations.
Admin rights are part of tokenomics, not a side note. Circle’s EVM contract stack includes pausable, upgradeable, blacklist, and mint rights, plus ownership roles, and Circle’s USDC terms expressly reserve the right to block transfers to and from certain addresses. That design can improve compliance response. It also means the stablecoin is not censorship-resistant in the strong sense.
USDT has similar peg logic but a different transparency profile. Tether publishes daily circulation data and quarterly reserve reports. In its March 31, 2025 attestation, Tether reported U.S. Treasury exposure approaching $120 billion and $5.6 billion of excess reserves. In its September 30, 2025 attestation, it reported $181.2 billion of reserves against $174.4 billion of liabilities. But on March 3, 2025 Tether also said it was taking a historic step toward a full financial audit, which is an important distinction from saying such an audit had already been completed.
Are crypto-backed and synthetic dollars really decentralized?
Some are more decentralized than fiat-backed coins, but the answer depends on where authority sits. Maker’s original architecture is structurally decentralized on issuance relative to custodial stablecoins, yet DAI is still governed by token voting, parameter changes, oracle design, and emergency controls. Maker’s own documentation states that MKR holders govern key parameters, and the governance module explicitly centers voting, proposal execution, and time-delayed spell execution.
DAI’s peg today is also partly a function of explicit stablecoin routing. Sky’s LitePSM allows fixed-ratio swaps between DAI or USDS and external stablecoins such as USDC, with no slippage in normal operation. That mechanism is operationally effective. It also imports external issuer risk into a system many users still describe as decentralized by default.
Liquity’s LUSD is structurally cleaner from a decentralization purist perspective. Liquity V1 describes itself as non-custodial, immutable, and governance-free. LUSD stability relies on overcollateralized ETH borrowing, a Stability Pool, and hard redemptions at face value, with a 110% minimum collateral ratio. The trade-off is equally clear: less governance means less flexibility to add emergency policy discretion later.
USDe is not bank-centric, but it is not authority-free. Ethena says direct mint and redeem access is limited to approved KYC/KYB-cleared parties, backing assets sit with off-exchange settlement providers, and ENA holders elect a Risk Committee that governs critical decisions. That is a very different centralization surface than a bank reserve stablecoin, yet it is still a managed system with committees, custodians, and exchange dependencies.
What did algorithmic stablecoins teach?
Reflexive peg support fails when the backstop asset falls with confidence. Legacy FRAX V1 made this logic explicit: the collateral ratio moved up or down based on market price, and redemptions below 100% collateral were funded partly with newly minted FXS. Frax later retired that original mint-redeem design, and current docs describe frxUSD, not the old FRAX v1 mechanism, as the current stablecoin line.
TerraUSD showed the terminal version of this failure mode. The SEC’s February 16, 2023 complaint describes UST as an algorithmic stablecoin that supposedly maintained its peg by being interchangeable with LUNA, and says that when UST depegged in May 2022, both UST and its sister tokens collapsed toward zero. Once the market no longer trusted conversion into the endogenous asset, the mechanism accelerated the break instead of absorbing it.
The durable lesson is simple. If the stabilizer is a volatile asset whose value depends on belief in the stablecoin itself, then the system is not diversified enough to survive a real run. “Algorithmic” can improve capital efficiency in expansion. It does not magically create exogenous collateral in contraction.
What should builders, treasuries, and DAOs diligence first?
Start with the control map. Before looking at APY, integrations, or market cap, identify who can do each of the following: mint, redeem, pause, upgrade, blocklist, change collateral rules, change oracle sources, alter fees, and trigger emergency shutdown. If those rights sit with one company, one multisig, or one small delegate set, the stablecoin is operationally coordinated even if it is issued on a public chain.
- Reserve risk: What exactly backs the token, and can those assets be liquidated fast enough in stress?
- Redemption risk: Who can redeem at par directly, and who is forced into secondary markets?
- Governance risk: Are changes rule-bound and delayed, or committee-driven and immediate?
- Collateral quality risk: Is the backstop exogenous, or does it depend on the protocol’s own market cap?
- Infrastructure risk: Does the stablecoin depend on banks, custodians, exchanges, or specific oracle paths that can fail independently of the peg logic?
The most decentralized stablecoin model is usually the one with the fewest discretionary levers, not the one with the loudest decentralization branding. On that criterion, immutable overcollateralized designs like Liquity score better structurally than issuer-administered fiat coins or committee-managed synthetic dollars. But they also give up coordination tools that can improve peg tightness and incident response. That is the real trade-off: operational flexibility versus distributed control.
At FinDaS Tokenomics, that is the starting point for any stablecoin-related token economy design review. The right question is rarely “which peg model is best” in the abstract. It is “which failure mode can your treasury, protocol, or users actually tolerate, and how much central authority are you accepting to reduce it.”
