The claim was simple enough to check, which is the only kind worth making.
An administration was pulling every lever at once. A sweeping tax cut. Refunds pushed into people’s hands. Tariffs assembled out of emergency powers. And a central bank under open pressure to cut.
The treatise said this cannot work. Not because it is unwise, but because a shortage of money is not a policy setting. You can decide money should be plentiful. You cannot make it so. What you can do is push the shortage somewhere else, and the treatise named the places it would surface.
The long end refused
In August 2026 the Treasury sold ten-year notes at 4.683 percent, the highest auction yield since the financial crisis. The ten-year has been trading near 4.68 percent. The extra return investors demand for lending long is elevated and rising.
This is the part that has aged best, and it is worth being precise about why it counts as a test rather than a description.
The framework did not predict that yields would be high. That is a coin flip. It predicted where. Push the shortage down at the short end, where a central bank has control, and it comes back up at the long end, where it does not. That is a claim about location, and location is the kind of claim that can be wrong.
The manager was not believed
The second prediction was subtler, and to this framework more important. Credibility, the treatise argued, is a reserve built up over decades and spent down every time an institution is told what to do. So the market would not wait for an easing. It would price the instruction the moment it was given.
And it did not wait. Going into the second half of 2026 the Federal Reserve looked likely to stay patient, with inflation sticky, and the market never delivered the boom the appointment was supposed to produce. The chair changed and the easing was already in the price.
They broke in the predicted order
The treatise said the strain would show up first in the things you buy with a loan rather than with money, because for those, wanting the thing and wanting the credit are the same act. Houses and cars first. The weekly shop, paid in cash, last.
Subprime car loans sixty days or more behind reached 6.9 percent in January 2026. That is a thirty-two year record, in a series that starts in 1994, and it is worse than anything during the Great Recession. Meanwhile existing home sales fell 4.2 percent over the first half of 2026, still below where they sat before the pandemic, in a country with plenty of houses and plenty of people who want one.
The finding that forces a correction
Here is where this dispatch stops being a victory lap. The framework’s own instrument, the Credit Counterfeit Index, currently reads minus 1.31. That is deeply below its breach threshold and among the more negative readings in the twenty-year history. On its own account the gauge is saying there is no credit counterfeit at all, that monetary conditions are loose relative to the real bind, and that nothing here should be manufacturing scarcity.
Meanwhile subprime auto delinquency sits at a thirty-two year high. Both of those things are true at the same time, and the framework has to account for it rather than pick whichever supports the argument.
The resolution is in the detail of the delinquency data, and it is the same detail that makes it interesting. Prime segments remain stable, with only modest deterioration in bank and captive portfolios. Subprime portfolios are at cycle highs. The stress is not spread across borrowers. It is concentrated at one end of the credit distribution and largely absent at the other.