Why the backing is external, and what happens when it is not
Every backed coin answers one question: what is in the reserve? The answers divide into two kinds, and only one of them survives a bad week.
The mirror problem
A reserve can hold assets issued by someone else — dollars in a stablecoin, wrapped ether, wrapped bitcoin. Or it can hold the protocol’s own second token, minted by the same contract that mints the first. The second arrangement is cheaper to start: the reserve costs nothing to fill, because the protocol writes it into existence.
It also fails in a particular way. The value of the reserve depends on the price of a token whose main use is backing the coin. If confidence in the coin falls, the reserve token falls with it, so the backing falls, so confidence falls further. Nothing external interrupts the loop, because there is nothing external in it. The arithmetic that looked like backing on a calm day turns out to have been a mirror.
This is not a claim about anyone’s intentions. A mirror reserve can be run honestly by people who believe in it. The property is structural: a reserve that reflects the thing it secures cannot secure it.
What Assetrix accepts
The reserve holds only assets issued outside the protocol, and the contract cannot add itself to the list. At launch the accepted currencies are dollar stablecoins. Wrapped ether and wrapped bitcoin are pre-approved in the sense that the code handles them, but they enter the reserve only if holders vote them in.
The list is not fixed forever, and that is deliberate: currencies get debased, bridged tokens get depegged, issuers get sanctioned. What is fixed is the shape of the list. A vote can add a currency, remove one, or ban one in an emergency. No vote can add a token the protocol itself mints, because the contract has no path to it.
What external collateral costs
Honesty about the reserve has a price, and it is paid in three places.
First, the reserve has to be bought. Every coin in circulation corresponds to real collateral someone deposited, which is why the cost of entry is not free and why the protocol has no way to seed itself.
Second, external assets carry their issuer’s risk. A dollar stablecoin can be frozen, an audit can go badly, a bridge can fail. The protocol cannot make that risk disappear; it can only refuse to concentrate in it. That is what the concentration surcharge is for: a currency that grows past its share becomes progressively more expensive to add to.
Third, volatile collateral moves. A reserve of wrapped ether is worth what wrapped ether is worth. Every level measured against that reserve — including the protection level — moves with it. With stablecoin collateral the protection rule holds exactly as specified; with volatile collateral it holds in units of the collateral, not in dollars. The contract softens this with a spread on volatile deposits, but it does not remove it, and nothing could.
What it buys
A reserve of external assets can be checked by someone who does not trust the protocol at all. The addresses are public, the balances are on chain, the oracles are named. A reader can price the reserve without asking the issuer anything, and can do it during the week when the answer matters.
It also breaks the loop. If the market price of the coin falls, the reserve does not fall with it: the dollars are still dollars. Redemption stays open at backing, which is what makes the lower edge of the corridor real rather than aspirational. A holder who wants out does not need a buyer.
That is the whole trade. A mirror reserve is cheap and reflexive. An external reserve is expensive and inert. For a protocol whose central claim is that entry never gets cheaper, inert is the only reserve that means anything.