The word stable describes an ambition, and the machinery underneath decides whether it is true.
Three Designs Wearing One Name
A stablecoin is a token aimed at a fixed value, most often matched to a familiar currency used in everyday trade. The aim sounds simple, and the word stable makes it sound settled. Underneath, though, three very different designs share the one name, and confusing them is the beginner’s first and costliest error.
The first design holds reserves. For every token in circulation, the issuer keeps assets set aside, meant to be worth at least as much, so that holders can in principle hand tokens back and receive real value in return. The second design locks up other crypto as collateral, deliberately locking more value than the tokens issued against it, so a fall in the collateral has room to happen before the promise is threatened. The third design holds no real reserve at all and tries to keep the price steady through rules, incentives, and a companion token that expands and shrinks as the system demands.
These are not variations on a theme so much as different species sharing a label. Judging all three by the same trust is like judging every payment tool in a shop by the card machine alone, when even choosing a point of sale system for a small retail shop teaches that what sits behind the counter matters more than the tidy surface in front of it.
This deep dive takes each design in turn, then turns to the question underneath all of them: the reserves, who vouches for them, and whether an ordinary holder can truly get out. A useful overview of stablecoins and how they are designed will lay out the same families, but the sceptical reading below is where the real education sits.
Keep one distinction in mind throughout. A price that has held steady so far is evidence of something, but it is not the machinery itself. Calm weather is not a hull. The question is always what happens on the bad day, and each design answers that question very differently.
The Reserve-Backed Design and Its Attestations
The reserve-backed design is the easiest to grasp because it resembles older promises. A company issues tokens and holds assets against them: cash in bank accounts and short-term government debt that can be sold quickly for cash. Holders trust the token because they trust the pile behind it and their right to claim against that pile.
Everything depends on the pile being real, complete, and reachable. This is where attestations enter. An issuer may publish reports from outside accountants stating that, on a given date, the reserves matched or exceeded the tokens issued. Those reports are worth reading properly. An attestation is a snapshot vouched for at a moment, which is a different thing from a full audit of everything an issuer does, and the difference matters when trust is the entire product.
Ask, too, what exactly the reserves contain. Cash is one thing. Short-term government debt is another, generally steady but still an asset that must be sold. More exotic holdings, loans, or riskier investments would change the character of the promise entirely, because a reserve that can fall in value at the wrong moment is a reserve with a crack in it.
Redemption is the other half of the design. If ordinary holders can hand tokens back to the issuer and receive their value directly, that right disciplines the whole system. If redemption is open only to large partners, with minimums far beyond a normal holder’s reach, then ordinary people rely on selling to someone else on a market and hoping the price holds. Both designs get called reserve-backed. They do not offer the same safety.
The Crypto-Collateral Design and Its Cushion
The second design replaces the company’s vault with a locked pool of other crypto. A user deposits crypto worth more than the stablecoins they take out, and that deliberate excess is the cushion. If the deposited asset falls in value, the cushion absorbs the fall. If the fall goes too far, the system sells the collateral before the promise behind the stablecoin is broken.
The appeal is transparency of a particular kind. The collateral sits in systems whose rules and holdings can often be inspected directly, rather than resting on a company’s word about its bank accounts. The bargain is that the collateral itself is volatile, sometimes wildly so, and the whole structure is engineered around surviving that volatility rather than avoiding it.
The cushion, then, is the number that matters, understood in plain terms: how much extra value stands behind each token, and how quickly the system acts when that extra value thins. In calm markets the machinery looks elegant. In a fast fall, many positions can approach their limits together, forced sales add pressure to a falling market, and the design is tested all at once rather than one holder at a time.
The Algorithmic Design and Its Failure Mode
The third design is where the word stable has done the most damage. An algorithmic stablecoin tries to hold its value with no real reserve behind it, using a companion token and a set of incentives. When the stablecoin slips below its target, the system offers rewards for actions that should pull it back. When demand grows, supply expands. Balance is meant to emerge from the rules themselves.
The flaw is that the whole arrangement leans on confidence, and confidence is the one asset that can vanish in an afternoon. The companion token has value because people believe in the system. The stablecoin holds its price because the companion token can absorb strain. When belief falters, both legs weaken together, each pulling the other down.
The failure mode, once it starts, feeds itself. A slip below the target prompts the system to create more of the companion token to defend the price. A flood of new companion tokens crushes their value. A crushed companion token defends nothing, so the stablecoin slips further, prompting still more creation. Holders who understood the design as clever engineering discover it was a circle, and circles under pressure become spirals.
This design has failed completely and publicly, wiping out holders who believed the label on the tin. That history is not a detail. It is the central fact about the category, and any new version deserves to be read as a claim that the circle has somehow been squared this time, a claim requiring far more than a confident launch.
There is a simple sorting question that cuts through every pitch in this space: if everyone wanted out this week, what real thing would they be paid from? Reserve-backed designs point to a vault. Collateral designs point to a locked pool. Purely algorithmic designs point to other believers. Only the first two answers name an asset.
Reserves: the Question Underneath Everything
Step back from the three designs and one question sits underneath them all: what, exactly, stands behind the token, and how do you know? This is the creditor’s question, and it applies even where the backing is locked crypto rather than a company vault, because every design is ultimately a claim on something, held somewhere, under someone’s rules.
Break the question into its working parts. What is held: cash, government debt, other crypto, or promises? Where is it held: which banks, which custodians, which systems? Who checks: an auditor, an attesting accountant, public inspection, or nobody outside the issuer at all? How often: daily, monthly, or whenever the issuer feels ready? And who can claim: everyone, or only the large and approved?
The study of a large, long-running example such as tether and its reserve model shows how these questions stay alive for years, debated in public, because the answers shape how much trust the market is willing to extend on any given day.
Regulation is tightening around issuers in many places, which slowly changes what holders can assume about reporting, reserve quality, and redemption rights. That direction is welcome, but it is uneven across borders, and a token available on your phone may be issued under rules far weaker than the ones your own bank lives by. The holder’s reading still has to be done by the holder.
How a Peg Breaks and How It Heals
The peg is the target value the token aims to hold, and it can break in more than one way. The dramatic break belongs to the algorithmic design, where the spiral described earlier simply runs to the end. The quieter break can happen to far better designs, and understanding it will save you from both panic and complacency.
Imagine doubts spreading about the banks where an issuer keeps its reserves. Nothing has been proven. The reserves may be entirely sound. But holders who cannot be sure begin to sell on markets rather than queue for redemption, sellers outnumber buyers, and the traded price dips below the target. The token has not failed. Its price has wobbled because trust in its banking arrangements wobbled, and in markets, wobbling trust is priced immediately.
Healing has a specific mechanism, and it is not hope. While the traded price sits below the target, confident buyers can purchase cheaply and redeem at full value with the issuer, pocketing the difference. That trade, repeated at scale, pulls the price back up. It only works while redemption itself is trusted: while people believe the issuer will pay out fully and promptly. The moment redemption is doubted, the healing trade stops, because nobody buys a discounted claim they fear will not be honoured.
Reading a Stablecoin the Way a Creditor Would
Finish by borrowing a mindset. A creditor does not ask whether a borrower seems nice, modern, or popular. A creditor asks what secures the claim, what the paperwork says, who else stands ahead in the queue, and what happens on the worst day. Read any stablecoin that way and the marketing falls away, leaving the questions that matter.
Start with the claim. When you hold the token, what do you actually own: a right to redeem from an issuer, a share in a collateral system, or merely a token other people currently accept? Then the security: what stands behind it, checked by whom, and how recently? Then the exit: can you, personally, at your size, redeem, and at what cost and delay? Then the record: how has this design behaved in past storms, in its own history rather than its promises?
Hold that line and you will read launch announcements, reserve reports, and market wobbles with the calm of someone who knows which questions decide the matter. What backs it. Who says so. Can I get out. Everything else is decoration.
