I keep a mental list of who actually buys short-dated Treasury bills, and a few years ago the list was boring. Money market funds, banks, foreign central banks, the usual suspects. Then I started seeing the reserve breakdowns from the big stablecoin issuers and had to put them on the same list, because the numbers had gotten to the point where you could not leave them off. Tether and Circle together hold a Treasury bill book that lands somewhere in the neighborhood of the largest foreign official holders. Not a rounding error next to Japan or China. An actual line item you would compare against a mid-sized country.
That surprised me more than it should have, and once I understood the mechanism it stopped feeling like a crypto oddity and started feeling like a plain funding fact. So this is me walking through the loop, and then the part I actually care about, which is what you do with the information in both directions.
Why a stablecoin ends up owning bills
Strip away the branding and a fiat-backed stablecoin is a claim on a dollar. Someone sends the issuer a dollar, the issuer mints a token, and the token trades around as a dollar substitute. The issuer is now sitting on that dollar and has to put it somewhere. It cannot pay depositors interest in most cases, so every bit of yield on the reserve is margin. The obvious place to park it is the safest, most liquid dollar instrument that exists, which is a short-dated Treasury bill. Bills mature in weeks to a year, they are as close to cash as you get, and they pay something.
So the issuer holds mostly bills, some overnight repo backed by Treasuries, a bit of cash at banks, and that is roughly the whole book. The reason it skews so hard to bills rather than longer bonds is redemption risk. If holders want their dollars back, the issuer has to sell reserves fast and at a price it can trust. A ten-year note can drop several points on a bad rate day. A four-week bill barely moves. When your entire product is a promise to redeem at par on demand, you do not want duration. You want the boring end of the curve, and you want a lot of it.
The loop, said plainly
Here is the part that clicks once you see it. Crypto demand goes up, more people want dollar exposure on-chain, so more stablecoins get minted. Every mint hands the issuer new dollars that flow almost entirely into bills. So a wave of crypto risk appetite becomes, a few steps later, a wave of buying at the front of the Treasury curve. And it runs the other way too. When crypto cools and people redeem tokens for actual dollars, issuers sell bills to fund the redemptions. Mint is a buy order, redeem is a sell order, and the underlying instrument is the same one the Treasury sells to fund the government.
I want to be careful not to oversell this. Stablecoin issuers are a marginal buyer, not the marginal buyer. The bill market is enormous and the Treasury issues a staggering amount of it, so on any given day the issuers are a small share of demand. But marginal is exactly the word that matters for pricing. Rates at the very front of the curve get set by supply and demand at the edges, and a large, correlated, fast-moving buyer that swings with crypto sentiment is a real input to that edge. Not the whole story. A term in the equation that was near zero and is now not.
Reading it as a stablecoin risk check
The direction most people should care about first is the token side, because if you touch crypto at all you are probably holding a stablecoin somewhere, even just as a place to sit between trades. The reserve composition is your risk read, and the attestation is where you read it. A monthly or quarterly attestation from the issuer's accountants tells you what is actually backing the thing.
Here is the checklist I run when I bother to open one, and I do bother:
- What is the reserve made of. You want to see it dominated by Treasury bills and Treasury-backed overnight repo. Cash at named banks is fine in small amounts. The further you drift into commercial paper, corporate bonds, secured loans, precious metals, or the issuer's own affiliated entities, the more you are trusting something other than the US government to pay you back.
- How short is the paper. A weighted average maturity measured in weeks is what you want. If reserves are reaching for yield in longer bonds, redemptions during a rate spike get ugly, because they have to sell at a loss to pay you.
- Who signed it and how often. An attestation is a snapshot check by an accounting firm, not a full audit, and the two are not the same thing. Frequent attestations from a real firm beat a glossy annual claim. If the update cadence slips or the firm changes without a good reason, note it.
- Does the total actually match tokens outstanding. Reserves should cover circulating supply with a little slack, not sit suspiciously flat while supply climbs.
The failure mode I always keep in mind is the one that has actually happened to stablecoins, which is a break of the peg driven by reserve doubt. It usually is not the bills that blow up. Bills are the safe part. It is the non-bill slice, the bank deposit stuck at a failing bank, the paper nobody can price in a panic, the affiliate loan that turns out to be a related party holding an IOU. When a token wobbles, look at the part of the reserve that is not Treasuries. That is almost always where the problem lives.
Reading it as a rate input
The other direction is subtler and I hold it loosely, but it is real. If you trade or care about short-end rates, stablecoin flows are one more thing feeding front-end demand, and the useful bit is that the flow is legible. Total stablecoin supply is public and on-chain. You can watch it grow and shrink close to real time, which is not something you can say about most bill buyers. Foreign central bank holdings show up with a lag in official data. Aggregate stablecoin supply you can pull today.
So the rule of thumb I use is simple. Sustained stablecoin supply growth is a quiet tailwind for front-end bill demand, and a sharp contraction is the opposite, a batch of forced sellers hitting the short end while they fund redemptions. It will not move the ten-year and it will not override what the central bank is doing with policy rates. But at the very front, during a stress week when everyone is watching bill yields, a large synchronized redemption wave is the kind of flow that can nudge things at the margin. Treat it as one gauge among several, the way you would treat any single flow signal. Not a trade on its own. A cross-check.
What I actually take from this
The honest summary is that a chunk of crypto risk appetite now expresses itself, a few hops down the chain, as demand for the safest dollar asset there is, and a chunk of the safety of your stablecoin depends on a Treasury bill book you can go read. Those two facts point at the same document. When I want to understand a stablecoin's risk I open the reserve attestation, and when I want a small extra read on front-end demand I look at where aggregate stablecoin supply is heading. Same plumbing, two different questions, and both of them are more knowable than they were a few years ago because the issuers got big enough that their reserves stopped being a footnote.