Some Unpleasant Arithmetic

The Interregnum: By Nishan Vense

On September 16, the Federal Reserve raised interest rates for the first time since July 2023. The vote was 12 to 0.

This year, the federal government will spend more servicing its debt than it spends on its military.

Read those two sentences together, because the order is the story.

What’s Actually Happening

Three things, and they belong together.

The Fed hiked. The target range moved from 3.5 to 3.75 percent up to 3.75 to 4 percent. Sixteen of eighteen committee participants expect another before the year ends. Chair Kevin Warsh’s reasoning is orthodox and, on its own terms, correct. Inflation has been persistent, energy prices have kept it elevated, and the policy rate is the instrument a central bank has.

The bond market answered. A week later the ten year Treasury yield hit 5.208 percent, the highest since June 2007. Nineteen years.

And the arithmetic underneath got worse. Annual net interest crossed one trillion dollars. The deficit ran 1.97 trillion through eleven months of fiscal 2026. Total federal debt sits near 40 trillion, and the interest bill on it grew 14 percent year over year.

Higher rates do one thing to a 40 trillion dollar debt stock immediately and with certainty. They make it cost more.

The Fed raised rates to reduce inflation. The first measurable effect was a wider deficit.

The Paper From 1981

Thomas Sargent and Neil Wallace published “Some Unpleasant Monetarist Arithmetic” in the Federal Reserve Bank of Minneapolis Quarterly Review in the fall of 1981, and the core insight is uncomfortable: a central bank fighting inflation alone can make it worse.

Monetarism held that inflation was a monetary phenomenon, so a determined central bank could always defeat it by tightening. Sargent and Wallace asked what happens when the fiscal authority refuses to cooperate.

The answer runs backward from intuition. Tight money today raises the real cost of servicing existing debt. That larger debt has to be paid for eventually. If taxes do not rise and spending does not fall, the only source left is money creation, and there will be more of it later than there would have been had the central bank never tightened at all.

They called it a game of chicken between the monetary and fiscal authorities. Whichever one commits credibly forces the other to accommodate.

Whoever moves second loses.

The term that grew out of this literature is fiscal dominance. It describes a condition, not an announcement. Monetary policy stops being independent in the way that matters, and nobody ever says so out loud.

What the Auction Admits

The most revealing thing on September 23 was not the yield.

Four things moved rates that day. Strong manufacturing surveys. Hawkish Fed commentary. Stubborn oil prices. And a five year Treasury auction met with poor demand.

The first three are the bond market repricing growth and policy. That happens constantly and means nothing in particular.

A weak auction is different. It is the bond market saying it wants to be paid more to hold government paper at all.

When yields rise because investors expect growth, that is a healthy economy. When they rise because investors want compensation for holding one particular borrower’s debt, that is something else wearing the same clothes. In real time the two are nearly impossible to tell apart.

The CBO baseline does not help. Net interest goes from 1.0 trillion in 2026 to 2.1 trillion in 2036. From 3.3 percent of GDP to 4.6. Sixteen point two trillion over the decade. That projection already assumes longer term rates rise. It is not a stress case. It is what happens if nothing goes unusually wrong.

The Compute Bill

I wrote here last month about the 1.5 trillion dollar gap between what the hyperscalers plan to spend on AI infrastructure and what their operating cash flow covers.

Every instrument filling that gap is priced off the Treasury curve.

A nineteen year high in the ten year does not only raise the government’s costs. It raises the hurdle rate for every data center built on borrowed money, at the exact moment the buildout stopped being funded out of profits.

Minsky’s ladder has a second axis I did not emphasize. A borrower can slide from speculative toward Ponzi without doing anything at all, because the price of the refinancing moved underneath them.

The fiscal position and the AI buildout are drawing on the same pool of capital. Only one of them sets the price.

The Question

There is a version of this that is ordinary.

Inflation right now is substantially energy driven. Oil is a supply shock, and rate hikes are a defensible response to one regardless of what the debt stock looks like. Congress is not a law of physics. Deficits have narrowed before, sometimes quickly. The United States borrows in a currency it issues, into the deepest sovereign market on earth, with reserve demand no other borrower enjoys. People have been predicting fiscal dominance in America for forty years and the bond market keeps disagreeing with them.

That reading requires the weak auction to be noise. It might be.

There is another version. Sixteen of eighteen participants expect another hike. If it does not arrive, the interesting question will not be whether the economy changed. It will be whether the arithmetic did.

That is what fiscal dominance looks like from the inside. Not an announcement. A meeting where the reasonable thing to do and the necessary thing to do stop being the same, and nobody in the room says it.

The debt used to be something the government owed. It has become something the government buys, every year, at whatever price the market sets that morning.