In May 2022, a single redemption mechanism turned $40 billion in combined market value to zero in a matter of days. Not through fraud, not through a hack, but through the precise mathematical logic the system was designed to run. Every stablecoin in circulation today sits somewhere on a spectrum between that reflexive collapse and a genuine dollar claim, and the distance between those two points comes down entirely to collateral quality, queue discipline, and who gets made whole before retail holders do. What follows is exactly where each major structure sits on that spectrum, and why the gap keeps closing at the retail holder's expense.
Stablecoins are not currencies. They are redemption promises, and every redemption promise has a queue, a collateral stack, and a set of actors who get paid out before you do. Academic research on narrow stablecoins draws a sharp line between fully collateralized structures like USDC and the asset-backed-but-stretched model that Tether operated for years. Understanding where that line sits, and why it keeps moving, is the only way to read these instruments honestly.
The Collateral Stack Is the Product
How a Stablecoin Run Propagates: From Pressure to Collapse
How a Stablecoin Run Propagates: From Pressure to Collapse
Source: Article analysis, SVB 2023, Terra/Luna May 2022
Source: Article analysis, SVB episode 2023, Terra/Luna collapse May 2022
Strip away the branding and a stablecoin is a liability. Someone issued it, and on the other side of that issuance sits an asset: a Treasury bill, a commercial paper note, a bank deposit, or in Tether's documented earlier years, a mix of secured loans and corporate bonds that would not survive a serious liquidity stress event. The asset quality determines whether redemption works under pressure or collapses into a discount.
Circle, the issuer behind USDC, restructured its reserve composition substantially after the SVB episode. As of mid-2026, reserves sit predominantly in short-duration US Treasuries and overnight repo agreements, with custodians spread across BlackRock's Circle Reserve Fund and a small number of regulated US banks. That structure is not accidental. It is a direct response to the observable failure mode that almost cost the peg in 2023.
Tether's collateral history is more complicated. At various points between 2018 and 2022, independent reserve attestations showed meaningful allocations to Chinese commercial paper, secured loans to undisclosed counterparties, and other instruments carrying real credit and liquidity risk. Tether has since shifted heavily toward US Treasuries, and its attestations now show a profile closer to a money market fund than to the earlier structure. Whether that transition is complete and independently verifiable at the level regulators now require is a separate question, and one the GENIUS Act's reserve reporting provisions are designed to force into the open. Issuers who cannot answer it cleanly will find the compliance burden doing the work that market discipline failed to do earlier.
Why Fully Collateralized Stablecoins Still Break
Collateral Quality Comparison Across Stablecoin Structures
Collateral Quality Comparison Across Stablecoin Structures
| Feature | USDC (Circle) | Tether (USDT) | Algorithmic (Terra) |
|---|---|---|---|
| Collateral Type | US Treasuries, overnight repo | Treasuries (now), prior: loans, corp bonds | None (algorithmic mechanism) |
| Liquidity Under Stress | High (short-duration) | Medium (improving) | None |
| Depeg Risk (Liquidity) | Low to Medium (SVB event) | Medium (historical) | Extreme |
| Reserve Verification | Audited, regulated | Attestations only | N/A |
| Retail Holder Protection | High (post-SVB reforms) | Partial | None ($40B wiped 2022) |
Source: Article analysis, Circle/USDC disclosures, Tether attestations, Terra post-mortem
Source: Article analysis, Circle/USDC disclosures, Tether attestations, Terra/Luna post-mortem
The SVB episode proved something theoretical models often miss: even a fully collateralized stablecoin can depeg if its collateral is illiquid or inaccessible at the exact moment redemption pressure arrives. USDC was not insolvent. Its assets covered its liabilities. But those specific assets were trapped inside a failed institution over a weekend when secondary markets were pricing uncertainty, not fundamentals.
This is the difference between solvency risk and liquidity risk, and it matters enormously for how runs propagate. A solvent institution with locked assets looks identical to an insolvent one during the hours before resolution. Market participants do not wait to find out which category applies. They sell the token, and the secondary market price becomes the operative reality regardless of what the balance sheet says.
The narrow stablecoin framework in academic work on this topic argues for reserves held in direct claims on central bank money or very short-dated government instruments with guaranteed same-day liquidity. The Federal Reserve's proposed guidelines around stablecoin reserve eligibility, which have moved significantly closer to finalization under the current regulatory environment, effectively encode this logic. Issuers holding overnight repo backed by Treasuries can meet mass redemptions without fire-selling assets. Issuers holding term paper or bank deposits cannot make the same guarantee, and that gap is precisely where the queue forms.
The gap is not theoretical. It is the spread between 87 cents and a dollar, measured in real losses by retail holders who could not wait out a weekend.
How Algorithmic Stablecoins Proved the Point the Hard Way
Tether Reserve Composition: Shift from 2020 to 2026
Tether Reserve Composition: Shift from 2020 to 2026
Approximate share of total reserves (%)
Source: Tether attestations, article analysis (approximate)
Source: Tether reserve attestations, article analysis (approximate composition)
Terra UST did not have a collateral stack. It had a mint-and-burn mechanism tying the stablecoin to LUNA, a floating token whose value depended entirely on demand for the system. In May 2022, large volumes of UST were withdrawn from the Anchor Protocol yield product over roughly 72 hours, with some estimates putting the figure around $2 billion. The redemption mechanism required minting LUNA to absorb the selling pressure. Minting LUNA diluted its price. Lower LUNA price reduced the credibility of the backing, which generated more UST redemptions, which minted more LUNA. The spiral completed within days. Widely cited figures suggest tens of billions in combined market value evaporated in the collapse.
The mechanics matter here because they are not specific to Terra. Any stablecoin whose peg depends on the market value of a correlated asset faces the same reflexive loop under stress. The asset whose value supports the peg is precisely the asset that gets sold when confidence breaks. Basis Protocol ran the same structure at smaller scale and shut down in 2018 before launch after its own team raised similar concerns.
What makes the algorithmic failure instructive rather than simply historical is that the regulatory response after Terra has not eliminated the design. It has driven the design into more opaque structures. Some newer partially collateralized models in DeFi as of mid-2026 blend real collateral with governance token backstops, which is structurally closer to Terra than to USDC. The collateral ratio looks adequate at current prices. It always does, right up until the price the ratio depends on starts moving. Retail holders in those structures are the last to know and the first to absorb losses.
The Redemption Queue Nobody Reads
The Terra/Luna Collapse: Key Numbers at a Glance
The Terra/Luna Collapse: Key Numbers at a Glance
Source: Article reference, Terra/Luna collapse May 2022
Source: Article reference, Terra/Luna collapse May 2022
Stablecoin terms of service contain redemption queues, minimum redemption sizes, and eligibility restrictions that retail participants rarely encounter, because retail participants almost never redeem directly. They sell on secondary markets. But when secondary markets price a discount, the arbitrage that closes the gap depends entirely on institutional redeemers accessing the direct channel. If that channel has friction, the discount persists.
Circle's USDC redemption terms allow direct redemption only for verified business accounts holding minimum thresholds, a pattern common across major issuers. Tether's direct redemption process has had documented minimum sizes, identity verification layers, and processing windows that create meaningful delay between a redemption request and cash settlement. These are operational facts, not accusations. They describe the system as built.
The practical consequence is that retail stablecoin holders in a stress event depend on institutional arbitrageurs to close the peg discount through their own redemption access. Those institutions weigh the spread against their own balance sheet risk in that environment. If the discount is 3 cents and the event looks contained, they step in. If the discount is 13 cents and the cause is unclear, they wait. The retail holder is downstream of that calculation every single time, and no whitepaper describes that dependency in plain language.
How the GENIUS Act Is Redesigning the Reserve Stack
The GENIUS Act, signed into law in 2025 according to public legislative records, establishes federal reserve requirements for payment stablecoins. Permitted reserve assets are reported to be essentially limited to US coins and currency, Federal Reserve deposits, short-term Treasuries, and overnight repo backed by Treasuries. The structure maps almost exactly onto the narrow stablecoin model from academic literature, and it effectively disqualifies the earlier Tether reserve composition as a compliant structure going forward.
Issuers outside the US who serve American users face a parallel compliance path under the Act, which has direct consequences for Tether. Tether operates outside the US regulatory perimeter, and its primary market has historically included significant retail volume from markets where dollar access is otherwise constrained: Turkey, Argentina, Nigeria, Brazil. Those users are not the target of US reserve requirements, but the Act restricts US regulated entities from holding or facilitating stablecoins that fall out of compliance above certain thresholds, creating indirect pressure on Tether's institutional dollar liquidity channels.
The winners from this regulatory redesign are issuers who can meet the reserve requirements at scale without destroying their yield economics. BlackRock benefits directly through its role managing the Circle Reserve Fund. Fidelity has entered stablecoin custody discussions with multiple issuers. The reserve requirement framework does not eliminate stablecoins. It consolidates issuance toward institutions already embedded in Treasury markets, converting what looked like an open protocol layer into an extension of existing financial infrastructure. That is a structural outcome with implications well beyond reserve composition.
What the SVB Discount Revealed About Peg Trust
The 86-cent USDC price in March 2023 lasted less than 72 hours. The FDIC's decision to cover all SVB deposits, including those above the $250,000 insurance cap, resolved the ambiguity before Circle had to mark the loss. The peg recovered to 99.9 cents within days. Most retail holders who sold at 90 cents during the weekend did not benefit from that recovery.
That sequence exposes the asymmetry embedded in stablecoin runs. Sophisticated holders with direct redemption access and institutional relationships get information faster and can act on resolution signals before secondary markets reprice. Retail holders on Coinbase or Binance are trading on the secondary price, which reflects uncertainty rather than fundamentals, and they exit at the panic price rather than the recovery price.
The peg held not because the stablecoin system worked as designed. It held because a political decision to cover uninsured depositors bailed out the reserve structure before the loss became permanent. That is a very different kind of confidence than anything described in a whitepaper. Whether the next stress event receives the same political resolution depends on factors that have nothing to do with the reserve attestation sitting on Circle's website today. The design tells you what happens under normal conditions. The question worth asking is who is actually standing behind the promise when normal conditions disappear. The SVB episode is the clearest evidence available that the answer is not always the issuer.
This article is for informational and educational purposes only and does not constitute financial, investment, or legal advice. The views expressed are analytical observations and should not be relied upon for personal financial decisions. Always consult a qualified financial advisor before making investment decisions.
How FICO 8 Really Works: What the Algorithm Actually Rewards