The Dry Powder Myth

TL;DR Rising stablecoin supply is one of crypto's favourite bullish tells: cash parked on the sidelines, waiting to buy bitcoin. It is testable, so this website tested it against DefiLlama's full history, three ways. As a level it correlates with price at 0.84, a number that looks decisive and is worthless: both series trend up for years, the same trap that makes the power-law chart look like a law. As a flow it carries no usable lead (weekly growth correlates with the next weeks of returns at noise levels, peaking near 0.12). As a valuation ratio, the Stablecoin Supply Ratio, the cheap end leans the right way but the pattern is non-monotonic (returns do not slope cleanly from cheap to expensive) and its forward-return correlation is roughly zero. Stablecoin liquidity is not a tradeable standalone indicator. It is not added to the dashboard.

Every cycle it comes back. Someone posts a chart of total stablecoin supply climbing to a new high, captions it "dry powder," and the replies fill with rockets. The reasoning is clean enough to fit in a screenshot: stablecoins are dollars that already live inside crypto rails, redeemable close to one for one, the nearest thing the market has to cash sitting on the sidelines. If that pile is growing, bitcoin has buyers in waiting.

It is a good story because it is almost mechanical. Money has to enter through the dollar door, and the dollar door is mostly Tether and Circle. Watching the powder, the thesis goes, means watching demand load before it fires.

The appeal of the claim is also what makes it checkable. Total stablecoin supply is public, daily, and long. DefiLlama tracks every issuer back to 2018 and serves the series for free. So this website pulled it and asked the only question that matters for an indicator: not whether the story sounds right, but whether the number would have helped. The answer, tested as a level, as a flow, and as a valuation ratio, is no on all three.

The Number That Looks Convincing

Start with the version that fuels the screenshots. Line up total stablecoin supply against bitcoin's price from 2019 to today and the correlation is 0.84, rising to 0.88 in logs (comparing percentage moves rather than dollar moves). On its own that looks like a settled case.

BTC price and total stablecoin supply, both log scale Two lines that both rise for years. Level correlation 0.84, and it means nothing. 2019 2020 2021 2022 2023 2024 2025 2026 BTC price (log) Stablecoin supply (log)
Total stablecoin supply (amber) and bitcoin price (white), both on a log scale, January 2019 to June 2026. The two climb together through 2019 to 2021, which is where the 0.84 correlation comes from. Then look at 2022: supply holds near its high while price falls by more than half. A relationship that vanishes in the one stretch where a buyer would need it is not a relationship.

It is the oldest trap in statistics. Two things that both rise over a span of years will correlate near one whether or not either has anything to do with the other. Bitcoin's price went up. Stablecoin supply went up. So did global debt, streaming subscriptions, and the number of padel courts in Madrid. The correlation measures a shared direction over time, not a link a buyer could trade. It is exactly the error this website spent three articles taking apart in the bitcoin power law (what the term actually means, the rigorous version of the claim, and the independent verdict on it): a line that fits beautifully and explains nothing.

The chart even shows where the illusion breaks. Through 2022 stablecoin supply sat near its all-time high while bitcoin lost more than half its value. The powder was dry, abundant, and completely uninterested in catching the fall.

Test It as a Flow

The honest version of the dry-powder claim was never about levels anyway. It is about flow and timing: fresh stablecoins are minted, sit briefly, then get deployed into bitcoin a few weeks later. That is a claim about change, not about height, so the test is to strip the trend out of both series and look at growth rates, which removes the shared drift and leaves only the week-to-week movement.

Done that way, the signal evaporates. Weekly growth in stablecoin supply correlates with bitcoin's return in the same week at −0.14, weak and pointing the wrong way. Push supply ahead of price to test the lead, one week through eight, and the strongest reading is about 0.12 around the five-week mark, which sounds like something until it is squared (squaring a correlation gives the share of the outcome it accounts for): a fifth of one percent of the variance, with the sign flipping from week to week on either side. At these levels, stablecoin growth explains a rounding error of bitcoin's next move. Step up to monthly growth and the numbers scatter between −0.18 and roughly zero. There is no horizon at which a rise in the powder reliably comes before a rise in price.

Test It as a Valuation

One framing is left, and it is the one with a real chance. Instead of treating stablecoins as a flow, treat them as a denominator. The Stablecoin Supply Ratio, SSR, divides bitcoin's market cap by total stablecoin supply: it asks how expensive bitcoin is relative to the cash standing next to it. A low ratio means a lot of dry powder per unit of bitcoin, which is the version of the thesis a valuation model would want. Because raw stablecoin supply only grows, the ratio has to be detrended, so the test uses an oscillator: how far today's SSR sits, in standard deviations, from its own trailing year (a standard deviation is the series' typical wobble: the question is whether today's reading is merely high, or unusually high by the last year's own standards).

This is a better question, and it earns a more interesting answer. Sort every day from 2018 on into five buckets by that oscillator, from bitcoin at its cheapest relative to stablecoins to its most expensive, and measure the forward 90-day return of each.

Forward 90-day BTC return by SSR-oscillator quintile A real valuation signal slopes down from cheap to expensive. This one peaks in Q4. +0% +5% +10% +15% +12.7% Q1 cheapest +7.4% Q2 +8.6% Q3 +15.5% Q4 +6.7% Q5 priciest SSR oscillator: BTC cheap relative to stablecoins (left) to expensive (right)
Average forward 90-day return for each SSR-oscillator quintile, daily readings 2018 to 2026. The cheapest bucket (Q1, +12.7%) does beat the most expensive (Q5, +6.7%), which is the direction the thesis predicts. But the best bucket is Q4, slightly above average, at +15.5%. A genuine valuation signal slopes cleanly downward from left to right. This one has its peak in the wrong place.

The cheapest bucket returns 12.7% over the next quarter and the most expensive 6.7%, so the cheap end does lean the way the story wants. But a real valuation signal slopes cleanly from cheap to expensive, and this one does not. The best bucket of all is Q4, bitcoin priced slightly above its own average against stablecoins, at 15.5%. When the strongest forward returns sit one notch from the expensive end, the cheap-versus-expensive gap is mostly luck. The direct correlation, which does not care how the buckets are drawn, confirms it: the oscillator lines up with the next 90 days of returns at −0.01, with the 30-day and 180-day versions no further from zero than 0.06.

There is a deeper reason not to trust even the 6-point gap. Ninety-day forward windows overlap almost completely from one day to the next, so the thousands of readings are not thousands of independent trials. Across seven years there are only about 30 genuinely separate 90-day periods, which means each bucket is a handful of market episodes wearing a large sample size as a disguise. Q4's apparent edge is most likely two or three recoveries that happened to land in the same bin.

The Bar Is the Bar

None of this says stablecoins do not matter. They are the plumbing every dollar of demand flows through, and a sudden contraction in supply would be worth watching for what it says about appetite. What the data says is narrower and firmer: as a timing indicator, total stablecoin liquidity does not clear the bar. The level is a trend artifact, the flow is noise, and the valuation version is a non-monotonic pattern resting on too few real episodes to lean on.

That is the same bar that rejected the return-tail "power law" on this website days ago, applied to a metric that flatters a different bias. A standard that only fails other people's favourite charts is not a standard. So the Stablecoin Supply Ratio does not become a fourteenth indicator, and the dry-powder chart does not get quoted in the dashboard's rationale. The whole point of running the test is to be allowed to say, in public and with the number attached, that this one does not work.