The Engine Was Never the Halving

Four cycles measured from their halving day: the multiple fell from 19.4 to 1.0. Meanwhile one year of US interest outlays equals more than every bitcoin that exists.
The Engine Was Never the Halving

Bitcoin traded at 64,428 USD at 08:59 UTC this morning, 48.9 percent below the high of October 2025. The Fear and Greed index reads 10, deep in its extreme fear band. Fitted against the power law of every daily close since 2010, price sits at the 7.2nd percentile, the lowest weekly reading in the eight years this fit covers.

Three numbers, and one question that decides everything that follows.

The data points at something else. Four halvings, measured the same way each time, show the pattern everybody traded on quietly running out of arithmetic. On the other side of the ledger, the collateral that bitcoin is priced against keeps growing at a rate no supply schedule can answer.

What the tape says

Bitcoin against its own power-law trend

Down 14.4 percent in seven days, 20.9 percent in thirty, 31.6 percent in a year. Over ninety days the net change is 0.4 percent, and that is the number which explains the shape: price rose 25.9 percent from 19 August to 18 October and gave all of it back in the four weeks since. A round trip, not a slow bleed. Market capitalisation 1.30 trillion USD on a circulating supply of 20,112,784 coins.

The fit puts the current trend value at 142,998 USD, with an exponent of 5.593 and an R squared of 96 percent across 5,935 daily points. Price is at the 5th percentile band, which the table gives as 62,363 USD, and the 25th percentile sits at 83,366 USD. In plain terms the model says we are near the cheap tail of its own eight-year record.

Two caveats, and they matter more than the reading. First, the percentile depends on the start year of the fit: 7.2 with 2010, 13.9 with 2011, 17.0 with 2013, 9.8 with 2017, and 0.0 with 2015. “Cheapest ever” is a statement about a model, not a fact about the world. Second, the power law has no mechanism inside it. It describes how price has grown. It cannot tell you why. What follows tries to supply the why.

The halving was never the engine

Four halvings, one decaying multiple

Take every halving date, set the price on that day to one, and plot what happened next. Four curves, same length, same starting point.

At day 940 after the halving, which is where the 2024 cycle stands today: 19.4 times the halving-day price in the 2012 cycle, 5.2 times in 2016, 1.95 times in 2020, and 1.01 times now.

The peak multiples decay the same way: 91 times the halving price in 2013, 29 times in 2017, 7.7 times in 2021, and 2.0 times in October 2025. Both series are monotone, and both come from daily closes rather than memory.

Four observations are not a law, and nobody should pretend otherwise. But there is a mechanical reason to expect this shape, and it is dull: each halving halves a flow that was already irrelevant next to the stock.

Issuance as a share of the coins in existence on each halving day: 12.5 percent in 2012, 4.2 percent in 2016, 1.8 percent in 2020, 0.8 percent today. A halving now removes half of a flow that is smaller than one percent of the stock. Halving a rounding error cannot produce a ninetyfold move on its own. What produced the earlier multiples was demand arriving in a thin market where almost nobody held coins. The supply cut was the story told about the move, not the cause of it.

That has a practical consequence now. Whoever is waiting for the next halving to do what the last three did is waiting on the one variable that has already gone to zero.

What is left is arithmetic

164,250 new coins a year, from 3.125 per block and 144 blocks a day. That is 0.82 percent of supply, worth 10.58 billion USD at today’s price. Hold that number and turn the page.

The other ledger

The collateral: forty trillion, and the interest on it

Total US public debt stood at 40.61 trillion USD on 12 November 2026, of which 32.82 trillion is held by the public. Twelve months earlier the total was 38.12 trillion. The increase is 2.49 trillion in a year, 6.5 percent.

Interest outlays for fiscal year 2026, which ended on 30 September, were 1.36 trillion USD, against 1.22 trillion the year before and 1.13 trillion the year before that. An 11.9 percent increase in the cost of the debt within one year.

None of this is a forecast. It comes from the Treasury’s own daily and monthly tables, and the mechanism behind the increase is in the same source: the average interest rate on the total interest-bearing debt has risen in every monthly reading since April 2026, from 3.340 percent in April to 3.580 percent at the end of October. The stock of debt did not have to grow for the bill to grow. It only had to be refinanced at today’s rates. For confirmation, the two-year Treasury yielded 4.86 percent on 12 November, up from 4.56 percent three days earlier.

Now put the two ledgers side by side in the same unit.

One year of US interest outlays, 1.36 trillion USD, equals 21.1 million coins at 64,428 USD. The entire circulating supply is 20.1 million. The twelve-month increase in the debt, 2.49 trillion, equals 38.6 million coins, or 1.9 times every bitcoin that exists.

These are comparisons of magnitude, not claims that anybody could buy that much. They are still the right scale to hold in mind. The asset is scarce. The claims against it are not, and they are issued by the same entity that sets the price of the unit the asset is quoted in.

The Cantillon side of the trade

M2 grew 5.7 percent in the year to September 2026, an increase of 1.26 trillion USD. Dollar for dollar that is 119 times the value of all bitcoin mined over the same period.

Cantillon’s insight was never about the quantity of money. It was about the path. New money enters at a specific point on a specific balance sheet, and whoever stands there buys before prices adjust. The last receivers buy afterwards. Between the two, relative prices move permanently, and the structure of production moves with them.

Bitcoin’s position on that path is odd: it is not on it. Nobody issues bitcoin to themselves. New coins go to whoever burns energy for them, not to whoever happens to sit at the issuance desk. That is the entire point of a fixed supply, and it is also why bitcoin is last in the queue when the money arrives. First receivers buy bills, then bonds, then equities, then property. Bitcoin gets bought by whoever is left holding a depreciating unit and decides to opt out.

So the price is set at the margin by people with the fewest reasons to sell, and it is measured against the yield they could earn by staying in the unit. That yield just went up.

Why a fixed supply falls when the collateral reprices

Boehm-Bawerk’s old distinction explains the past month better than any chart. Present goods command an agio over future goods, and the size of that agio is the interest rate. Raise the rate and the present value of every future stream falls. Bitcoin has no future stream. What it has is a claim on a future monetary role, which is the longest-dated claim anyone can hold.

A two-year note at 4.86 percent pays you to wait. Bitcoin pays you nothing to wait, and it has cost 48.9 percent of its dollar price since the October 2025 high.

This is not a hole in the theory of sound money. It is the theory working inside an unfree monetary system. A hard asset does not become stable by being hard. It becomes stable when the unit it is priced in stops expanding, and until then its price in that unit absorbs the adjustment for both of them. The same mechanism that will make it a decent unit of account later makes it a volatile asset now. The volatility is the fee paid by everyone who exits early.

Anyone who told you the drawdown could not happen under a fixed supply was selling something. The protocol never promised a price. It promised a schedule, and the schedule is intact. Nine hundred forty days after the fourth halving, block subsidy 3.125, 0.82 percent a year, no committee, no vote, no exception.

Thiers’ Law, without the promise

Gresham observed that bad money drives out good under a fixed exchange rate. The mirror observation, usually named after Adolphe Thiers, holds that good money drives out bad wherever the exchange is free and no legal tender rule fixes the price. Bitcoin is a Thiers asset. It wins where people are allowed to leave, and it wins slowly, at the margin, one balance sheet at a time.

What Thiers does not promise is that the exit is priced kindly. A currency rarely loses its role because a better asset exists. It loses the role when its issuer makes the alternative necessary. In the past six weeks most of the damage came from access rather than from the protocol: a custodian, an index, a lending market, a freeze. None of that was a failure of the cap. All of it is what the exit looks like while the pipes belong to the institutions that benefit from the unit you are leaving.

Three scenarios instead of a target

No price target. Three states of the world, each with the first thing you would see.

The collateral holds. Bills stay near five percent, the debt keeps compounding, the average rate on the stock keeps climbing toward the market rate. Bitcoin trades as a long-duration asset against a high discount rate, below trend, for longer than is comfortable. What you would see first: the average interest rate on the debt still rising while new issuance slows, and drawdowns getting shallower while rallies stay short.

The collateral cracks. An auction fails, or the policy rate is cut into an inflation rate that is not falling. The real return on bills goes negative while the debt still needs rolling over. What you would see first: the two-year falling faster than expected inflation, and gold and bitcoin moving together again after months of moving apart.

The pipes clog. A custody, exchange or policy event with no relation to the protocol. Price falls further while the network keeps producing blocks on schedule. What you would see first: a premium on self-custody and a discount on wrapped claims, exactly as in the past six weeks.

What would make me wrong

If the next twelve months produce a fifth cycle shaped like the third, my claim that the supply cut stopped mattering is wrong, and the measurement above will show it: the day-940 multiple will rise instead of decaying. If the debt stops growing and the interest bill falls, the discount-rate argument loses its subject. And if the percentile in the first chart starts behaving like a forecast instead of a description, then I will have confused a model with a market, which is the very mistake this article warns about.

Every number above comes from a source below or from my own calculation on that data, measured on 16 November 2026. The calculation method is described where the number appears.

Sources

Write a comment