Where you are. You have an issuer whose tokens are backed one for one by reserve assets, checked after every posting. That check is arithmetic. This lesson is about the question the arithmetic cannot answer: when the company running the issuer fails, are those assets still yours?
A stablecoin issuer is a company. It has staff, offices, a technology budget and trade creditors. It can be badly run, sued, or simply fail, entirely independently of whether its reserve assets are intact.
So picture the failure. The parent has burned through its capital and owes suppliers more than its operating assets can cover. Somewhere on its books, or somewhere else, sit a billion dollars of bills backing a billion tokens. The liquidator arrives. What happens next depends on a structural decision made years earlier.
The idea in one paragraph
The reserve assets are held in a separate legal entity whose only purpose is to hold them, so that the parent’s creditors have no claim on them: the special purpose vehicle module 5 met in securitisation, doing the same isolation job with different contents. Around that structure sits a licence, because in most large jurisdictions issuing a payment token to the public is now a regulated activity: an e-money token under Europe’s MiCA regime, a permitted payment stablecoin issuer under the United States GENIUS Act. The regimes differ in detail and disagree about plenty, but they share two requirements that matter most to a holder: backing of at least one to one in high-quality assets, and the right to redeem at par. The cage is not paperwork. It is the difference between owning segregated assets and standing in a queue.
Wider than the screen; scroll it sideways.
The same failure, twice
The exercise builds one failing parent twice and asks what a holder recovers.
The parent is genuinely insolvent in both runs: fifty of operating cash against four hundred owed to trade creditors, with its capital gone. In the first run the reserve assets sit in a bankruptcy-remote entity. In the second they sit on the parent’s own books.
Segregated, holders recover the full dollar: the assets were never the parent’s to lose, and the trade creditors have no claim on them. Commingled, the reserve assets join the pool available to everybody the parent owes, and holders recover 0.75 on the dollar, sharing the assets with creditors who have nothing to do with the tokens.
| structure | recovery per token |
|---|---|
| reserve assets in a bankruptcy-remote entity | 1.00 |
| reserve assets on the parent’s own books | 0.75 |
n = 2 trials · parent insolvency, holder recovery per token
The reserve assets are identical in both runs. Nothing about the tokens, the ledger or the backing ratio differs. The entire difference is which legal entity holds the account.
What the regimes actually require
Read past the acronyms and the shared requirements are the interesting part: reserve assets at least equal to tokens in issue, held in high-quality liquid instruments, segregated from the issuer’s own money, with holders entitled to redeem at par. Those are the same four properties whether you arrive at them through European or American drafting, and they are the properties this module’s exercises actually test.
Check yourself
1. Both runs have identical reserve assets and identical tokens. Why do holders recover different amounts?
Because in one run the assets belong to a separate entity the parent’s creditors cannot reach, and in the other they sit in the pool available to everyone the parent owes. The recovery is decided by which legal entity holds the account, not by anything about the tokens or the backing ratio.
2. What do the MiCA and GENIUS regimes agree on, underneath the different drafting?
Reserve assets at least equal to tokens in issue, held in high-quality liquid instruments, segregated from the issuer’s own money, and holders entitled to redeem at par. The detail and supervision differ considerably; those four properties are common ground.
3. An issuer is fully licensed and fully backed. Name something that can still go wrong for a holder.
The token can trade below a dollar in the secondary market, because the licence governs the issuer and its reserve assets rather than the price other people will pay. Redemption can also be slow. Lesson 8 runs the case where sound backing and a shut redemption window produce a price a long way from par.
4. Module 5 introduced the special purpose vehicle for securitisation. Is this the same idea?
Yes, with different contents. There it isolated pooled loans from the originator’s troubles; here it isolates reserve assets from the parent’s. The mechanism and the purpose are identical, which is why the course uses the same name rather than coining a second one.
Do this
Fail the parent twice and compare what holders get.
python3 code/legal_cage.pyThe starter builds the insolvent parent and leaves you recovery_per_token: work out what a holder receives in each structure. In the segregated case the answer is the full dollar; in the commingled case the reserve assets join the parent’s asset pool and are shared across all its liabilities. The assertions check that the caged run returns exactly 1.00 and the commingled run returns 1050/1400.
The completed version is in solutions/legal_cage.py.
What you can now do. You can explain why the reserve assets live in a separate entity, state the four requirements the major regimes share, and demonstrate with a failing parent that the structure rather than the backing ratio decides what a holder recovers. The next lesson asks a question the structure raises immediately: if the reserve assets are earning interest, who is getting it?