Actionable insights — Not The Way It's Supposed To Work

1. Build the base rate first, then treat its violation — not your thesis — as the finding

1. Before interpreting a market move, define the relationship you expect to hold and the population it holds over: which variable, measured from which event, across how many historical instances.
2. Pull every instance in the window and check the direction. A relationship worth calling a base rate should be near-unanimous, not merely a majority — a 60/40 tendency is not violated by one contrary case.
3. Define the trigger event tightly enough to exclude look-alikes. "An actual cutting cycle" is not the same population as "any rate cut," and mixing them destroys the base rate you are trying to build.
4. Measure the current instance against that population and state the magnitude of the deviation, not just its sign — direction alone can be noise, size cannot.
5. Note whether the current instance had conditions that should have made the normal effect stronger. A deviation that occurs despite a larger-than-usual impulse is far more informative than one under a marginal impulse.
6. Only then reach for causes. The deviation is the fact; the explanation is a hypothesis and should be labelled as one.

• Base rates built on too short a window or too loose an event definition — both manufacture "unprecedented" events that are nothing of the sort. And watch the honest failure mode of this method: a relationship can break because the transmission channel changed permanently, in which case the old base rate is retired rather than violated, and the trade is to re-anchor on the new regime rather than to bet on reversion.

2. Close the strongest counterexample on its own terms before offering causes

1. Name the best historical case against your claim yourself, in its strongest form, before a reader can raise it.
2. Concede what is genuinely similar — here, a cut delivered into a still-robust economy — rather than disputing the resemblance.
3. Then find the disqualifying difference in how the episode was classified at the time, not in hindsight. A contemporaneous label ("mid-cycle adjustment") tells you what the market was being told to expect, which is what governed the bond response.
4. Check how the episode resolved. An easing that was reversed within a year was never a cutting cycle and does not belong in the population at all — which turns the counterexample into a validation of the population definition.
5. Only after the objection is closed do you propose causes. Offering explanations while an unanswered counterexample stands is how a base-rate argument gets dismissed.

• The tell that the current easing is itself a 1998-style mid-cycle adjustment: an early reversal to hikes. That would not save the long bond — it would simply mean the deviation was resolved by policy retreating rather than by the term premium falling — but it changes which end of the curve carries the damage.

3. Test a historical analogue on its initial balance sheet, not on its narrative — an inverted starting condition makes it a contrast, not a template

1. When an era "rhymes," write down the two or three state variables that determined the outcome then — fiscal position, debt stock, starting valuation, starting yield — not the story features (a bubble, an easing Fed, a popular technology).
2. Compare each state variable directly. Narrative similarity with inverted state variables is the most misleading pattern in macro, because the story invites the borrowed conclusion while the arithmetic reverses it.
3. Where a variable is inverted, flip the expected sign of the outcome rather than discarding the analogue. The comparison is still useful — as a control, showing what the same monetary posture produces under the opposite fiscal condition.
4. State the inversion in a quantity a reader can check, and preferably in a form that is startling enough to be remembered.
5. Carry the conclusion into a positioning rule: the same policy that was bond-friendly under a surplus is bond-hostile under a deficit, so duration is not a hedge in this regime the way it was in that one.

• Analogue-shopping in the other direction — an era selected because its outcome matches the desired conclusion, with the state variables never checked. The disciplined version always states which variables are matched and which are inverted, and the inverted ones should be the ones doing the work.

4. Read an official intervention as evidence about the diagnosis — you don't manage a market you believe is fine

1. When an authority takes an unusual action in a market, separate two questions: will the action work, and what does the action reveal about what the authority believes?
2. Treat the second as the more reliable signal. Officials have better data on their own funding needs than outside analysts do, and an intervention is a costly, revealed preference rather than a statement.
3. Use it to corroborate an independently derived view — never as the view's only support. The confirmation is worth something precisely because it was reached by a different route.
4. Then ask where the suppressed pressure goes. A price that is administered does not remove the imbalance; it relocates it to whatever remains free to move — typically the currency and the assets priced in it.
5. Position for the relocation rather than for the defended price, and set the falsifier at the point where the defence is abandoned.

• The falsifier that would end the trade: long yields permitted to rise freely with no buyback, no maturity-shortening and no supportive rhetoric — that would say the defence has been abandoned and the pressure has returned to the bond, where it belongs. The opposite tell is escalation — larger buybacks, more bill-heavy issuance, or explicit guidance about a yield level — which confirms the peg and argues for holding the un-printable measuring assets instead of the pegged bond.