1 bp ≈ −3 pp
BIS Working Paper 1382, published October 13, ran six and a half years of weekly US data through one question: are stablecoins and bank deposits substitutes? The answer came back with a number attached. A single basis point of deposit-rate increase is associated with roughly three percentage points less stablecoin growth over ten weeks. The blockchain variable is nowhere in the regression. The interest rate is.
BIS Working Paper 1382 · Ahmed & Aldasoro · 13 October 2026

—
One basis point is 0.01%. It is the smallest unit a rate desk will quote you without laughing. It is the difference between a savings account paying 4.00% and one paying 4.01%, which is to say, it is nothing — a rounding error on a rounding error.
According to a working paper the Bank for International Settlements published on October 13, that nothing is enough to move the growth rate of the entire stablecoin market by three percentage points.
I read the paper twice. The first time I was looking for the trick, because a number that size has to be a mistake or a marketing claim. The second time I realized the number is not the interesting part. The interesting part is what the paper implies about what a stablecoin actually is — and it is not the thing the industry says it is.
The Paper, Plainly
The title asks the question directly: “Are stablecoins and bank deposits substitutes?” It is BIS Working Paper 1382, by Rashad Ahmed and Iñaki Aldasoro, and it is the kind of research I wish we got more of in crypto — a narrow, falsifiable question asked against a long dataset instead of a narrative asked against a chart.
| Fact | Detail |
|---|---|
| Paper | BIS Working Paper 1382, published October 13, 2026 |
| Authors | Rashad Ahmed, Iñaki Aldasoro |
| Question | Are stablecoins and bank deposits substitutes? |
| Data | Weekly US data, January 2019 through June 2025 |
| Market over the window | Stablecoin market cap grew from under $5B to over $220B |
| Average growth | ~15% per 10-week interval over the sample |
| Headline estimate | +1bp deposit rate ≈ −3pp stablecoin growth per 10 weeks (OLS) |
| Instrumental-variables estimate | ≈ −9pp, roughly three times the OLS figure |
| Issuer split | USDC ≈ −4.8pp; USDT ≈ −1.9pp |
If you have ever wondered why stablecoin people track the Fed calendar like it is a product roadmap, this table is why. The paper took a market the industry explains with technology — faster settlement, programmability, self-custody — and explained a measurable chunk of its growth with a deposit rate.
15% → 12% — the paper’s illustration: ten-week stablecoin growth after a 1bp deposit-rate rise
The Shape of the Argument
Strip the econometrics and the logic is one line. A non-yielding stablecoin and an interest-bearing deposit are two ways of holding the same dollar. When the deposit pays nothing, the stablecoin’s other advantages — transferability, on-ramp convenience, DeFi access — carry the whole decision, and the market grows. When the deposit starts paying real money, every month you sit in the stablecoin has a visible price tag, and some fraction of marginal holders pay it.
The regression sketch, for those who want it in code rather than prose:
growth_10w(stablecoin mcap) = alpha
+ beta * delta_deposit_rate
+ controls
OLS: beta ≈ -3pp per 1bp
IV: beta ≈ -9pp per 1bp (about 3x the OLS figure)
Two details in the paper do more work than the headline, and neither made the press coverage.
The first is the split between issuers. USDC’s growth is roughly two and a half times more rate-sensitive than USDT’s — about −4.8 percentage points versus −1.9 per basis point. The paper’s explanation is the one I find convincing: USDC’s user base is US institutional and DeFi-heavy, people for whom a bank deposit is a genuine alternative one click away. USDT circulates offshore and in emerging markets, among holders for whom the comparison account may not exist. Read that way, the sensitivity is not a property of the token. It is a property of who is holding it — stablecoin demand is weakest exactly where the holder has somewhere else to put the money.
The second is the control the paper ran on non-stablecoin crypto. Run the same machinery on bitcoin and the broader crypto market and the deposit-rate effect does not show up. That matters more than it looks, because it kills the lazy explanation — “rates rose, all risk assets fell, stablecoins just came along.” If this were crypto beta, bitcoin would respond too. It doesn’t. The substitution is specific to dollar tokens, which is exactly what you would expect if stablecoins compete with deposits rather than with crypto.
The paper also finds the effect extends to stablecoins deployed in DeFi lending protocols — the yield-chasing float shrinks when banks pay more — and that the result survives swapping in certificate-of-deposit rates. It is stubborn.
The Threshold That Makes It a Story Today
Buried in the analysis is a nonlinearity, and it is the reason this paper is news in October 2026 rather than an academic curiosity. When the federal funds rate sat near zero, deposits and stablecoins both paid nothing, and there was no contest to run. The competitive effect switches on once rates climb past roughly 3% — the gap between an interest-bearing account and a zero-yield token becomes large enough to change behavior at the margin.
Which brings us to the environment this paper lands in. The Federal Reserve raised rates 25 basis points in mid-September 2026, to a 3.75–4.00% target range, and the 10-year Treasury has been holding near 5.1%. We are, in other words, running the paper’s headwind live and out of sample — with rates above the threshold where the substitution effect is measured to bite.
The Caveats, Stated Honestly
I said I read the paper looking for the trick. Here is what I found instead of a trick: the paper’s own limitations section, which is more candid than most crypto research of any kind, and which anyone quoting the headline numbers should read first.
- The instruments are weaker than the profession’s comfort level. The authors say so directly — some indicators of identification reliability fall slightly below common standards. The IV estimate is three times the OLS estimate, which is a hint that the true number is somewhere in a wide band, not a precise fact.
- Slows, does not reverse. Higher deposit rates reduce the growth rate of stablecoin market cap. Nothing in the paper suggests they shrink it. Over the sample, stablecoins grew enormously through rate-hike cycles.
- The whole frame assumes stablecoins pay nothing. This is the big one. The paper’s substitution logic rests on the non-yielding stablecoin. If yield-bearing stablecoins or tokenized money-market funds scale up — and every major issuer is working on exactly that — the interest-rate penalty evaporates, and with it most of the paper’s mechanism.
- A policy footnote worth fearing. The paper notes that if issuers were required to hold part of their reserves as bank deposits, the outflow pressure eases — but the deposits concentrate as large, uninsured balances in a few banks. The cure has a concentration risk of its own.
Source: Bank for International Settlements, Working Paper 1382, “Are stablecoins and bank deposits substitutes?”, Rashad Ahmed and Iñaki Aldasoro, published October 13, 2026. Analysis covers weekly US data, January 2019 to June 2025. Secondary coverage consulted for reaction; all numbers above are from the paper as reported.
What to Watch, For Your Reference
I am not telling you stablecoins are doomed or saved. I am telling you the variable that matters changed, and most of the industry’s commentary still argues about token design as if the deposit rate were a detail.
- The spread, not the rate. What kills a non-yielding token is not the level of rates but the gap between what a bank pays and what the token pays — which is the whole gap. Watch deposit pricing at the big US retail banks; that is the marginal competitor now, not other stablecoins.
- Yield-bearing products, as a direct response. The paper’s own assumption is the industry’s escape hatch. The first stablecoin that passes through reserve yield at scale doesn’t just compete better — it invalidates the paper’s premise.
- USDC versus USDT growth during rate spikes. The paper predicts USDC decelerates harder in a high-rate US environment. That is now a testable, ongoing forecast, and you can check it monthly from public circulation data.
- BIS follow-ups. Working papers are not policy, but they are the draft language of future rules. When the Basel committee starts quoting this one, it stops being a paper.
The uncomfortable summary is one sentence, and I have not found a way to soften it: the stablecoin boom, measured honestly, is partly a complaint about bank deposit pricing — and banks can fix a complaint by writing a number on a statement. The technology determined what was possible. The deposit rate determined who showed up.
(The End)


