A stablecoin holds its price with two forces working together. The first is credible backing, a promise that each coin can be swapped for something worth about a dollar. The second is arbitrage, traders who earn a profit whenever the price drifts and, by chasing that profit, drag it back.
The backing sets the anchor, and arbitrage does the pulling. If a coin slips to $0.99, a trader can buy it cheap, swap it for a full dollar, and keep the difference, and all that buying lifts the price back toward $1. The faster and more open that swap is, the tighter the peg holds in normal trading.
What changes from coin to coin is what stands behind that dollar, and that choice decides how much stress a peg can take before it slips.
What a peg really is, and why it needs defending
A peg is a target price a coin is designed to track, not a law of nature. On an exchange, supply and demand set the price: more sellers than buyers push it below a dollar, heavier buying pushes it above. No central bank stands behind a stablecoin, so its peg must be defended by design.
The closest real-world parallel is a currency board. When a country pins its money to the dollar, its central bank defends that rate by buying its own currency when it weakens and selling when it strengthens, drawing on a reserve to do it. A stablecoin does the same job with different tools. A pool of reserves or collateral stands in for the war chest, and arbitrage traders stand in for the trading desk. Researchers who study these coins describe the peg as an equilibrium held by arbitrage between the issuer and the open market, not a fixed feature of the coin (source: Warwick, What keeps stablecoins stable).
Because that defense is a market process and not a legal guarantee, the price can wander a little until the profit motive drags it back. If the coin itself still feels fuzzy, it helps to start with the basics of a stablecoin before the machinery here. It also helps to see how bitcoin compares to fiat money, because a stablecoin gives up bitcoin's wild price swings on purpose in exchange for a steady value. The simplest way to defend a peg is also the oldest one: keep a real dollar in reserve.
Credible backing is the anchor
Fiat-backed coins hold reserves and promise to swap one coin for one dollar, and that promise is the anchor the price is measured against. The issuer keeps assets worth as much as all the coins in circulation, usually cash and short-term government debt, so each coin has a real claim behind it.
Here is an everyday way to picture it. Imagine a coat check that hands you a paper ticket for each coat you drop off. The ticket is not the coat, but you trust it because you can swap it back any time. A cash-backed coin works the same way. The coin is the ticket, the dollar in reserve is the coat, and the promise to swap keeps the ticket worth about a dollar. The Bank of England puts it plainly: whoever issues the coin holds the same value in real money, and the holder has the right to swap it back whenever they want (source: Bank of England stablecoin explainer).
Reserve quality is the whole game. Cash and short-term government bills can be sold fast at close to face value, even in a panic, so they can fund swaps on demand. Riskier holdings are slower to sell and can lose value at the worst moment. Circle reports that USDC is held mostly in cash and short-term US Treasuries (source: Circle transparency disclosures), while Tether reports that USDT is backed mainly by US Treasuries alongside smaller holdings such as gold and bitcoin (source: Tether Q1 2026 reserves attestation). Same one-dollar target, different risk behind it.
There is also a gap between saying reserves exist and proving it. Most issuers publish attestations, which confirm reserves matched coins on a single date, rather than full audits over time, so knowing how reserves are verified tells you how much the promise can bear. One caution: that check is about the issuer's own reserves, a different thing from an exchange showing it holds the coins it owes you.
Which design a coin uses is what sorts the main stablecoin types, and a few dollar tokens now hold short-term Treasuries and pass the interest to holders. Those are close to tokenized real-world assets, not plain payment coins. Reserves set the anchor, but they do not move the price by themselves. That job belongs to arbitrage.
Arbitrage does the actual pulling
Arbitrage is the force that actually moves a coin back to its peg. When the market price sits below the swap value, traders buy the cheap coins and redeem them for a full dollar, which pulls supply out and lifts the price. When it sits above, they create new coins at a dollar and sell until the premium fades.
Here is the discount case as a worked trade, assuming a coin trades at $0.99 and a firm can redeem directly with the issuer.
Discount case: a coin trading at $0.99 Buy 1,000,000 coins on the open market = $990,000 Redeem 1,000,000 coins with the issuer = $1,000,000 Gross gain = $10,000 (about 1%) Effect: 1,000,000 coins leave the market, nudging the price up toward $1
The trade is only worth doing while the price is off the peg, so traders race to do it first, and their combined buying closes the gap fast in normal conditions. The premium case runs in reverse: a coin trading at $1.01 lets a firm create coins at a dollar and sell them higher, adding supply until the price eases down. The Federal Reserve describes this same self-correcting loop, where any gap between the market price and one-to-one redemption offers a profit that traders move to capture (source: Federal Reserve, The stable in stablecoins). Most of this runs between large firms and the issuer, since direct redemption is usually limited to approved partners. Retail still feels the result, because the price you see on an exchange barely leaves a dollar while someone stays ready to profit from any gap. The loop only works while there is a genuine dollar to redeem for, which raises a fair question about coins that never touch a bank.
Over-collateralization and liquidation hold a crypto-backed peg
Crypto-backed coins cannot keep dollars in a bank, so they lock up more crypto than the coins they issue and let code sell that collateral if it falls too far. The holder deposits more than the coin is worth, and that extra cushion absorbs price swings before the system has to step in.
Think of a pawn shop. You pledge something worth more than the cash you borrow, you get it back when you repay, and if you walk away the shop sells it. Here the shop is code on a blockchain. DAI is the long-running example. To create it, a user locks collateral such as ether at a minimum ratio, often around 150%, so roughly $150 of ether backs every $100 created against it. Price feeds track the collateral all the time, and if a vault falls below its ratio the system auctions the collateral to cover the debt and charges a penalty (source: MakerDAO whitepaper). Because 150% is only the floor, holders usually lock up well above it, so a vault opened at around 300% can lose half its collateral value before it reaches the liquidation line.
The risks differ from the cash kind. A fast crash can drop collateral faster than the system can sell it, and forced sales can push prices lower and trigger more sales. The code leans on price feeds, so a bad or slow feed can misfire. And when much of the collateral is itself another stablecoin, the two are linked, which is a quiet way for one coin's trouble to become another's. Because this design lives inside on-chain apps, DeFi apps on Ethereum are useful background for how those contracts fit together, though the deep protocol detail is their subject, not this one. A third design keeps crypto backing while removing the liquidation risk.
Delta-neutral hedging: a peg without a dollar in a bank
A newer model keeps crypto backing but cancels out its price swings with a hedge. The issuer holds a crypto asset and opens a short futures position of the same size, so a fall in the asset is offset by a gain on the short, and the combined value stays near a dollar.
The word "delta" just means price exposure, and "neutral" means it has been canceled out. A worked example makes it clear. Say the system holds $100 of staked ether and also shorts $100 of ether futures. If ether falls 10%, the staked ether loses about $10, but the short gains about $10, so the pair is still worth about $100. If ether rises 10%, the gain and the loss swap places, and the total still lands near $100. The dollar value comes from the hedge, not from cash in a vault. The clearest live example is Ethena's USDe, which holds a mix of crypto assets, including staked ether, plus matching short positions (source: Ethena USDe documentation).
This is a newer design, and its yield comes from two places, the rewards on the staked crypto and the funding that short positions collect when traders are mostly betting long. The risks match the machinery. When traders lean short, funding can flip and the hedge starts to cost money. The crypto and the short often sit with exchanges and custodians, so an exchange failure is a real worry. The key point is that the backing is real, but it sits outside the coin, in the hedge itself. Every design so far leans on real assets. The one design that tried to skip real assets is the one that failed hardest.
Why algorithmic mint and burn is the weakest
Algorithmic coins hold little or no reserve and defend the peg by minting and burning a paired token instead. A holder could swap the stablecoin for a dollar's worth of the paired token, so supply shrank when the price dipped. It worked until confidence broke, because the backstop was part of the failing machine.
Picture an IOU that is backed by a second IOU you also print yourself. That is the trap. TerraUSD, or UST, is the case study. It kept its peg through a swap with LUNA, its own paired token, where one UST could always be exchanged for a dollar's worth of freshly minted LUNA. A lending program called Anchor paid a very high yield, about 19.5%, which pulled in most of the demand and packed the system with yield-seeking money. In May 2022, heavy withdrawals broke the calm, the price slipped, and the swap turned into a trap. As holders fled UST for LUNA, LUNA's supply ballooned about eighty times in two days, from roughly 0.4 billion to 32 billion coins, and its price crashed from about $31 to a cent (source: Richmond Fed brief on the Terra collapse). There were no reserves to recover to, so unlike a fully backed coin, it did not bounce back.
That reflexive loop is the difference people most often miss. A coin with outside backing holds its value whether or not anyone trusts it, while UST was backed by LUNA and LUNA's value came from UST, so the support vanished right when it was needed. Deciding what to do when a coin actually breaks its peg is its own topic, and this guide stays on the normal mechanism rather than the response. For most people the practical takeaway is short: recognize a purely algorithmic design and weigh how risk and return trade off before holding one. Even a coin with honest backing needs somewhere deep enough to trade.
Why liquidity is the peg's quiet backstop
Backing and arbitrage only hold a peg if the market is deep enough to trade in size without moving the price. Arbitrage closes a gap by buying or selling large amounts fast, which is cheap on a deep order book and costly on a thin one. So liquidity is the quiet backstop that keeps the whole mechanism honest.
The reason is depth. On a deep market, arbitrage can buy or sell size without moving the price much, so a gap closes fast; on a thin one, the same trade shifts the price against itself and the gap lingers. From BloFin's operational view, the USDT and USDC order books are deep enough that arbitrage closes small gaps quickly, and sizable orders usually clear with only minor slippage in normal conditions. That steadiness is the practical sign of the liquidity a peg needs to snap back, and when a book thins out, that same arbitrage costs more to run, which is usually where small dislocations show up first.
This is why two coins with similar backing can behave differently under stress. The one on deeper markets gives traders room to work, so its price snaps back faster. The fine detail of how a large order actually fills, and how much market depth a venue carries, is its own subject worth reading on its own. Where you trade matters too, since centralized and decentralized exchanges hold coins and route orders in different ways. Put backing, arbitrage, and liquidity together, and you can start to rank one peg against another.
What makes one peg stronger than another
What separates a strong peg from a fragile one is how liquid and genuine its backing is, how open the swap path is, and how deep the market trades. The model sets the ceiling on how much stress a peg survives, while live liquidity governs its day-to-day steadiness. The table lines up the four main designs.
| Model | Peg anchor | Correcting force | Main failure mode | Strength under stress |
|---|---|---|---|---|
| Fiat-backed (USDC, USDT) | Cash and short-term Treasury reserves | Redemption arbitrage | Reserves frozen or hard to sell | High while reserves stay liquid |
| Crypto over-collateralized (DAI) | Surplus crypto locked in code | Liquidation plus arbitrage | A crash outruns liquidation | Moderate to high |
| Delta-neutral synthetic (USDe) | Crypto plus a matching short | Hedge plus arbitrage | Exchange failure or negative funding | Moderate |
| Algorithmic (past coins like UST) | A paired token, little reserve | Mint and burn a second token | Death spiral once confidence goes | Low |
Read down the table and a pattern shows up. The designs near the top lean on assets that live outside the coin and can be sold for close to a dollar, so their worst case is a delay rather than a collapse. The designs lower down lean more on their own token or on markets that can seize up, so their worst case is permanent. Capital efficiency rises as you move down the list, and so does the cost of being wrong.
That is why real backing is worth caring about. From BloFin's operational view, a design that rests on real, reachable assets behaves very differently under stress than one propped up by a second token. BloFin's own RWUSD shows the contrast in practice: it is a yield product recorded inside BloFin Earn, subscribed with USDT and benchmarked to tokenized Treasury returns, so it sits on real backing rather than on confidence in a paired token, though it is not a coin you send from wallet to wallet. No design is risk-free, and the ranking shifts with reserve quality and conditions. Still, the pattern from real events holds: pegs backed by liquid, reachable assets on deep markets tend to recover from shocks, while pegs that lean on their own token do not. For a fuller sense of BloFin's RWUSD and how a real-backed yield product differs from a plain payment coin, the guide covers the terms.
Frequently asked questions
What does it actually mean when a stablecoin loses its peg?
Losing the peg means the price drifts away from its target and does not snap right back. Small wobbles are normal, since a coin ticks a fraction above or below a dollar all day as buyers and sellers trade. A depeg is bigger and stickier, like a slide to $0.95 that holds for hours instead of seconds. The line is not a fixed number, but the useful test is whether arbitrage is closing the gap or has stopped working. When the gap stays open, it means traders no longer trust that they can redeem for a full dollar.
Who actually keeps the price at a dollar, and can a regular person profit from it?
Most of the peg defense is done by large firms that can redeem straight with the issuer, usually only above a high minimum. They buy a cheap coin, redeem it for a full dollar, and pocket the gap. A regular person almost never redeems directly, so the everyday version is just trading on an exchange with other people. You still benefit, because those professional firms step in whenever a gap opens and keep the price you see close to a dollar. In effect, a small group defends the peg while everyone else rides along.
Is a stablecoin safe just because it has never lost its peg?
No. A clean track record shows the design has not been tested hard yet, not that it cannot break. UST traded near a dollar for months before it collapsed in days. The better question is what has to stay true for the coin to hold. List those conditions, such as reserves being real and liquid, redemption staying open, and markets staying deep, then decide how comfortable you are with each one. A coin with few and easy-to-check conditions is sturdier than one that quietly depends on many things going right.
Does more collateral always make a crypto-backed peg safe?
Not on its own. Extra collateral buys time, which matters, but it does not remove every risk. A sharp crash can drop the collateral faster than the system can sell it, so the buffer runs out mid-fall. The code also leans on price feeds, and a wrong or slow feed can trigger the wrong sales. And when the collateral is itself another stablecoin, a problem in that coin flows straight through. So a high ratio helps, but the quality and reachability of the collateral matter just as much as the size of the cushion.
Can a stablecoin hold its peg if the backing cannot be reached right now?
Not while the money is stuck, even if it truly exists. In March 2023, Circle held about $3.3 billion of USDC reserves, roughly 8% of the total, at Silicon Valley Bank when it failed on a Friday. Banks were shut all weekend, so Circle could not move the cash, and USDC slid to about $0.86 before recovering once the US government guaranteed the bank's depositors that Sunday (source: Federal Reserve note on Silicon Valley Bank and stablecoins). The reserves existed the whole time, just out of reach. A peg needs backing that is real and reachable right then.
Do rules like the GENIUS Act keep a peg from breaking?
No. The US GENIUS Act, signed in 2025, sets standards for what can back a payment stablecoin and how issuers report reserves, and once its rules take effect it will bar those coins from paying holders yield (source: GENIUS Act, Public Law 119-27). Those rules are not in force yet. Stronger reserve and disclosure rules would make a peg easier to defend, because clearer backing makes redemption more credible. But no rule promises that a government will buy the coin at a dollar, so the day-to-day peg still rests on arbitrage and market depth.
Researched and written by the BloFin Academy editorial team with AI-assisted drafting. Updated July 2026. Primary sources: the Bank of England, the US Federal Reserve, the Richmond Fed, academic research from the University of Warwick, Circle and Tether reserve disclosures, the MakerDAO and Ethena protocol documentation, and the GENIUS Act (Public Law 119-27). All facts independently verified against cited sources current as of July 2026.
This article is educational and general in nature, not financial, investment, tax, or legal advice. Stablecoins carry real risks, including loss of the peg, and no stablecoin is guaranteed to hold its value. Nothing here is a recommendation to buy, sell, or hold any asset. Do your own research, and consider a licensed professional before making financial decisions. BloFin does not provide investment advice.