How to Know If a Stablecoin Is Actually Stable

The word 'stable' is doing a lot of work in crypto. It is in the name 'Stablecoin' and used in every pitch deck and product description. But for most of the people who lost money in the 2022 contagion, it turned out to mean something completely different.

TerraUSD was called a stablecoin. It had a ticker, a peg, a whitepaper, and billions in market cap (tokens in circulation, each supposedly worth a dollar). When TerraUSD and its sister currency LUNA collapsed in May 2022, it took roughly $40 billion in value with it. Not slowly. Over just a few days.

USD Coin (USDC) is also called a stablecoin. It briefly lost its peg, meaning it traded below $1, in March 2023 when $3.3 billion of its reserves got caught in the Silicon Valley Bank collapse. It recovered. Nobody lost their holdings. The two events have almost nothing in common except the word 'stablecoin'.

The standard definition of a stablecoin is a cryptocurrency designed to hold a fixed value, usually by pegging its price to a currency such as one dollar or one euro. That is accurate as far as it goes. But what that definition leaves out is everything that matters. The definition describes how a token behaves when things are going well. It says nothing about what holds the peg together, or what happens when it breaks. Those questions have very different answers depending on what is actually holding the peg in place.

Three stablecoin models, three very different risk profiles

There are three main ways to build a stablecoin. Each involves a different mechanism for maintaining the peg and a different failure mode when that mechanism comes under pressure.

Fiat-backed stablecoins hold one unit of real currency in reserve for every token minted. Circle does this for both USDC and EURC, with Deloitte auditing the reserves monthly. What you are trusting is the issuer's operational integrity, not a mechanism.

Tether (USDT) sits in the same category and is the largest stablecoin in circulation by market cap, though its reserve composition has historically been harder to verify than Circle's. That makes counterparty risk harder to price.

Crypto-backed stablecoins like DAI, governed by MakerDAO, require borrowers to lock up more crypto than they mint, typically 150% or more. That overcollateralisation buffer helps absorb price swings, and if necessary the system can liquidate positions to defend the peg. More decentralised than fiat-backed. Not immune to extreme markets.

Then there is algorithmic. No real reserves. The peg is held by code and market incentives. When everything works, it appears to be like the others. When it does not, there is nothing underneath to catch it.

Smart contract risk runs across all three categories. Every stablecoin built on a blockchain lives in code. Bugs, exploits, and bridge vulnerabilities are risks that do not exist with a euro sitting in a bank account. That is a variable worth pricing in, not a reason to avoid them.

The EUR stablecoin market shows the scale of the opportunity. EUR-denominated stablecoins represent around $1.2 billion in onchain market cap against an offchain EUR market of roughly $16.1 trillion, per Token Terminal data. That is approximately 0.007% penetration. The demand is real. The market is early. Understanding the risk differences between these three models is what separates someone positioned to use them from someone positioned to be surprised by them.

How algorithmic stablecoins work and why they break

To understand what happened to TerraUSD, you first need to understand what it actually was.

UST was built by Terraform Labs and launched by its founder Do Kwon. The design was ambitious: a stablecoin with no dollars behind it, no reserves, no vault. Instead, the peg was maintained through a linked token called LUNA and a set of on-chain incentives. If UST traded above $1, the protocol encouraged users to burn LUNA and mint new UST, increasing supply and pushing the price down. If UST traded below $1, users could burn UST to mint LUNA, reducing supply and pushing it back up. In theory, market arbitrage would keep the peg tight.

The death spiral starts when confidence breaks. A large sell-off pushes UST below $1. The protocol responds by minting LUNA to absorb the selling. More LUNA in circulation dilutes its value. A falling LUNA price weakens the defence mechanism, which causes more UST selling, which mints more LUNA, which falls further. The system designed to restore the peg accelerates its collapse. In May 2022, that process played out over roughly 72 hours. UST went from $1 to near zero. LUNA, which had been trading above $80, fell to fractions of a cent.

The bank run dynamic is familiar even to people who have never touched cryptocurrency. When enough depositors believe a bank might fail, they withdraw simultaneously, and that collective action causes the failure they feared. UST worked the same way. The mechanism was not broken by a hack or a bug. It was broken by coordinated disbelief.

Do Kwon had publicly dismissed concerns about the model. Terraform Labs had accumulated Bitcoin reserves as a secondary defence mechanism. But selling the Bitcoin reserves did little to stop it. Kwon was later convicted of fraud related to misrepresentation of the protocol's stability. The design was always the problem.

Not every algorithmic project has disappeared. Frax Finance operates a partial-algorithmic model with real collateral backing a portion of its supply, a more conservative design sitting closer to the crypto-backed category than pure algorithmic. It has survived conditions that destroyed UST. But the purely algorithmic model, where the peg depends entirely on incentive mechanics with no underlying asset, has been discredited as a design.

Terra collapse: the big picture

Forty billion dollars gone in less than a week. Not stolen. Not hacked. Just gone, because the mechanism holding the peg stopped working and there was no collateral underneath it to slow the fall. The UST collapse triggered a cascade of failures.

Three Arrows Capital

Su Zhu and Kyle Davies had built 3AC into a $10 billion crypto hedge fund with heavy investment in LUNA. When LUNA went to near-zero in days, those positions were wiped. Lenders started calling in loans. 3AC could not cover them. The fund filed for bankruptcy in July 2022. Its founders left the country before arrest warrants were issued. The total defaults ran into billions, though the exact figures vary between lenders.

Celsius Network

Celsius was a yield platform offering returns on crypto investments of up to 18% annually, with around $12 billion in customer assets at its peak. Those yields required Celsius to put capital to work using strategies that were heavily dependent on the Terra ecosystem. When the losses hit, Celsius froze withdrawals in June 2022 without warning. The platform filed for bankruptcy shortly after.

Voyager Digital

Voyager lent hundreds of millions to 3AC. When 3AC defaulted, Voyager could not cover its own obligations and filed for bankruptcy in July 2022.

The chain nobody mapped until it broke

Across all three, the structure was the same: everyone owed everyone, and nobody had a buffer. Celsius had direct DeFi exposure. Voyager's problem was simpler. It had just lent 3AC the money. UST did not just hurt people who held UST. It hit customers on Celsius who had never heard of Terraform Labs. Most people framing this as a 'crypto contagion' story focus on the institutions. The actual damage was to retail depositors who had no way of knowing their yield depended, even partially, on the Terra ecosystem staying intact.

Regulators had been pointing at exactly this structure for years. Interconnected debt, opaque risk, no reserve requirements, no depositor protections. The Terra story gave regulators a live case study. MiCA in Europe picked up pace, as did US stablecoin legislation. Every significant policy document written on algorithmic stablecoins since 2022 references Terra as the primary example of how not to design them. The $40 billion figure is what made the headlines. But confidence was shattered when retail depositors lost access to money they thought was safe.

Iron Finance: the earlier warning nobody took seriously

Terra happened in May 2022. Iron Finance happened in June 2021 using an identical mechanism but on a smaller scale. It was a partially algorithmic stablecoin project running on the Polygon network. Its stablecoin, the IRON token, was pegged to $1 and backed roughly 75% by USDC, with the remainder maintained algorithmically through a second token called TITAN. When large holders began selling TITAN in June 2021, the price started dropping. The algorithmic mechanism minted more TITAN to defend the IRON peg. This feedback loop dropped the price from around $65 to nearly zero in hours.

Mark Cuban was publicly invested and had written about the IRON token on his blog, describing his yield farming strategy. After the collapse he wrote that he had been hit and called for stablecoin regulation. He later admitted he had not fully understood the risk of the algorithmic model. Iron Finance could have become the case study that prevented Terra. The mechanism was identical. But the crypto news cycle moved on within days. Instead, the same faulty model was rebuilt at much larger scale and marketed more aggressively.

Fiat-backed is not risk-free: the SVB story

On 10 March 2023, Silicon Valley Bank collapsed. It was the second-largest bank failure in US history. Within hours of the FDIC announcing it had taken control of SVB, Circle confirmed that $3.3 billion of USDC's cash reserves were held there. USDC briefly traded at around $0.87. A stablecoin backed by real dollars, issued by one of the most reputable companies in the space with full regulatory compliance, lost 13 cents of its dollar peg inside a weekend.

A stablecoin can be fully collateralised and still fail, because the problem is liquidity, not solvency. The reserves were real. They just could not move fast enough. Traditional banking has a central bank that acts as lender of last resort. None of those mechanisms exist on-chain, so reserves take the impact directly with nothing to absorb it first. The BIS described this structural vulnerability in their paper "On Par: A Money View of Stablecoins."

By Monday, after the US government announced it would guarantee all SVB deposits beyond the standard FDIC limit, Circle confirmed the funds were accessible and the peg restored. People who held USDC through the weekend lost nothing, though the 48 hours before the government guarantee were uncomfortable. Fiat-backed stablecoins inherit the risks of the traditional banking system and can fail if the backing bank fails. USDC's stability depends on Circle's ability to manage operations and on the regulated financial institutions holding its reserves. SVB was a regulated financial institution and still failed.

What protected USDC holders was the US government's decision to guarantee deposits beyond the standard FDIC limit. A policy decision made in Washington on a Sunday rescued a stablecoin peg.

The EU equivalent of that protection is the EU Deposit Guarantee Scheme, which covers deposits up to 100,000 euros per institution. That is the floor for euros in a bank account. EURC reserves sit in institutions covered by that framework, but EURC itself is a regulated electronic money token, not a bank deposit. Holding EURC tokens is not the same as holding euros in your own account. Large holders face different protections than individual depositors. The custody and compliance frameworks around both tokens are materially stronger than most alternatives. The SVB episode showed that even fully-backed stablecoins can lose their peg.

What the regulatory response actually changed

Terra collapsed in May 2022. MiCA came into full effect for stablecoins in June 2024. It took two years for institutions to respond to something the market already understood was broken.

The EU's Markets in Crypto-Assets regulation, known as MiCA, is the most comprehensive stablecoin legal framework. It defines what stablecoins can be and who can issue them under what conditions. Algorithmic stablecoins are not explicitly banned under MiCA, but their requirements make a purely algorithmic model functionally impossible within the EU.

Circle moved early. Its European subsidiary, Circle Internet Financial Europe SAS, secured an Electronic Money Institution licence from the ACPR (the Autorité de Contrôle Prudentiel et de Résolution, the same authority that supervises French banks) in July 2024. That licence makes EURC and USDC among the first major stablecoins to achieve full MiCA compliance. The European Banking Authority coordinates EMI supervision across the EEA, meaning the compliance framework Circle operates within applies consistently across member states, not just France. Circle CEO Jeremy Allaire had publicly positioned the company toward regulatory engagement for years before MiCA passed. That positioning paid off in competitive terms: when major exchanges began delisting non-compliant stablecoins, Circle's products had a clear lane.

The Deloitte attestations are the practical output of that compliance. Every month, Deloitte independently verifies that EURC reserves match the tokens in circulation. The reports are public and independently verifiable, not a marketing claim. That mechanism was absent from UST, where the reserves were the LUNA token itself. Tether also lacked this verification for years.

The US is moving more slowly. The SEC and CFTC have both staked claims to stablecoin jurisdiction without a unified framework emerging. Congressional stablecoin legislation has moved through committees and stalled more than once. The direction is toward something resembling MiCA, with reserve requirements, audits, and issuer authorisation, but the timeline remains uncertain. For now, the EU framework is the clearest regulatory model in existence.

MiCA changed the legal standing of issuers, not the technology. The tokens work the same way they did before June 2024.

How to read a stablecoin before you use one

Five questions. Work through them in order.

Question 1Who issued it and are they licensed?

The issuer is the entity responsible for the reserves, the redemptions, and the compliance obligations. Check whether the issuer holds an EMI licence from a recognised European authority. Circle's licence is held under ACPR supervision and is publicly verifiable. If you cannot find a named regulator and a specific licence, that is a strong indicator to avoid that stablecoin.

Outside the EU, the picture is less clean. US-issued stablecoins operate under a patchwork of state money transmitter licences with no unified federal framework yet. Paxos, issuer of USDP, operates under a New York Department of Financial Services charter, one of the stricter US state frameworks available, but federal oversight of stablecoin issuers remains unresolved between the SEC and CFTC. That does not make stablecoins issued outside the EU inherently unsafe, but the accountability structure is less defined.

Question 2Are the reserves independently audited, and can you see the reports?

Deloitte audits EURC reserves monthly and the attestation reports are available on Circle's transparency page. That is the standard to compare against. There is a meaningful difference between an issuer saying their reserves are fully backed and an auditor confirming it. If a stablecoin's reserve verification is infrequent, produced by an unknown firm, or not publicly accessible, treat it as a disqualifying red flag rather than an administrative gap.

Chainalysis, the crypto investigation software, can supplement reserve reports for issuers who are less forthcoming. Tracking the flow of funds between issuer wallets and custodian accounts gives you partial visibility into reserve movements when official reports are lacking.

Question 3Where are the reserves held?

Reserves held across multiple regulated financial institutions are more resilient than reserves concentrated at one. When SVB collapsed, Circle temporarily lost access to $3.3 billion in reserves but has since diversified custody. For EUR-denominated stablecoins, check that reserve institutions are operating within the EEA under ACPR or equivalent supervision.

Check what the reserves consist of. Cash and short-term government securities are the cleanest backing. Commercial paper and corporate bonds introduce credit risk that pure cash does not carry. Tether's reserve composition has historically included a higher proportion of non-cash assets than Circle's. This credit risk can cause the stablecoin to lose its peg during market stress.

Question 4Can you redeem it, and on what terms?

A stablecoin that holds its peg in secondary markets and can be redeemed directly with the issuer at par is more useful if markets dislocate. Circle Mint allows qualified businesses to redeem USDC and EURC at 1:1 with no fees. Retail users typically access redemption through exchanges, which adds a layer of dependency. Know which route applies to your use case before you need it.

Smart contract risk sits here too. Every on-chain stablecoin operates through code. Check whether the contracts have been audited, by whom, and how recently. The Block and CoinDesk both maintain coverage of major contract audits and security incidents across the blockchain sector.

Question 5Does the issuer have a compliance kill switch, and under what conditions?

Under MiCA, issuers can freeze specific tokens if instructed by regulators. Circle's contracts include this function. It is a disclosed compliance mechanism, not a hidden risk. But it means your holdings are not entirely beyond reach of the issuer.

A business accepting USDC as payment, or a protocol holding it as treasury float between spending cycles, is not meaningfully exposed to this. The freeze function requires regulatory instruction to trigger and has been used a handful of times in documented enforcement cases. For users where asset control is a design requirement rather than a preference, a DeFi protocol where the trustless property is load-bearing, or a treasury large enough that issuer counterparty risk sits inside the risk framework, understand it before assuming a stablecoin behaves like self-custodied cryptocurrency.

None of these checks require specialist knowledge. The stablecoins that failed badly, UST and the IRON token, did not hide what they were. The whitepapers explained exactly how they would fail. Most people did not read them, or did not understand the failure mode they described. Reading the documents is most of the due diligence.

What stable actually means

The stablecoins that held through 2022 had one thing in common: something real behind the peg. Reserves that could be verified. Issuers that could be held accountable. UST had neither.

UST had eleven billion dollars in market cap at its peak. It had a founder who publicly called out sceptics by name. It had institutional backing, exchange listings, and a yield mechanism that attracted billions in deposits from retail cryptocurrency holders worldwide. What it did not have was anything backing it except the mechanism itself. When that broke, the marketing did not matter.

The word stablecoin obscures key differences and is not useful on its own. A fully reserved, MiCA-compliant electronic money token audited monthly by Deloitte and a purely algorithmic token maintained by arbitrage incentives are not the same instrument. Calling them both stablecoins is technically true but practically misleading. The regulatory shift since 2022 created a clearer line between those two things. MiCA drew it in Europe. The full reserve model with independent audits, licensed issuers, and disclosed custody arrangements is now a defined category with legal standing. EURC and USDC sit inside that category. The instruments that collapsed sit outside it.

The SVB episode showed that fiat-backed stablecoins carry real-world banking dependencies. Smart contract risk is present in every on-chain token regardless of what backs it. The US regulatory picture is still being written. But these are knowable risks that can be assessed and managed.

Terra, Iron Finance, the contagion through Three Arrows Capital and Celsius Network. All of it happened within fourteen months. Most people who lost money had no framework for evaluating what they held before they held it.

The audits are public. The licences are verifiable. Reading them takes twenty minutes.

Stable means something specific. Worth knowing what.