The Fed raised rates on September 16, 2026 — the first hike since 2023, reversing a rate-cutting cycle that had only started in September 2024. Its own long-run "neutral rate" estimate sits at 3.0%, a level policy has now sailed well past. Interest-rate rules aren't broken so much as built around an estimate — the neutral rate itself — that keeps moving under everyone's feet.
This is the third and final letter in a short series examining whether the old macroeconomic playbook — inflation, employment, and interest rates, read off their textbook gauges — still works. The first two letters found real nuance: inflation looks broken mainly because the wrong slack proxy was being used, and the employment side is arguably working close to as designed. Interest-rate policy is the odd one out, and the reason is structural: a full year-long reversal, from cutting to hiking, is not a sign the rulebook stopped applying. It is a sign the specific number the rulebook is built around is far less stable than the rulebook assumes.
The Federal Reserve's Open Market Committee voted 12-0 on September 16, 2026 to raise the federal funds rate by 25 basis points, to a range of 3.75%-4.00%, up from 3.50%-3.75%. This was, per the Fed's own characterization, "the first time the benchmark rate has been increased since 2023" — a genuine reversal of the cutting cycle that had begun as recently as September 2024. The Committee's own dot plot released alongside the decision projects year-end rates between 4.1% and 4.4%, with markets pricing in one further 25-basis-point hike in December and additional increases potentially extending into 2027. The Fed's own "longer run" neutral-rate estimate — the rate consistent with neither stimulating nor restricting the economy once inflation is at target — remains anchored at 3.0%, meaning policy, even before this hike, already sat above the level the Fed itself considers neutral, and is now moving further above it still.
| Date | Action | Rate Range |
|---|---|---|
| Sept. 2024 | Cutting cycle begins | — |
| Prior to Sept. 2026 | — | 3.50%–3.75% |
| Sept. 16, 2026 | Hike (first since 2023) | 3.75%–4.00% |
| Year-end 2026 (dot plot) | Projected further hike(s) | 4.1%–4.4% |
| Fed's own long-run neutral estimate | — | 3.0% |
A cutting cycle that starts in September 2024 and fully reverses into a hiking cycle by September 2026 — two years, one full round trip — is the kind of whipsaw that makes any rules-based framework look unreliable to an outside observer. If the Taylor Rule and its descendants were simple, stable dials, a forecaster using them a year ago should have had a reasonable read on where policy would sit today. Instead, the direction of travel itself reversed. That is a legitimate reason for skepticism about treating any current rate-path consensus, including the current dot plot's guidance toward 4.1%-4.4%, as a reliable forecast rather than a snapshot of the Committee's current thinking.
The problem is not that a Taylor-Rule-style framework — set rates based on how far inflation and output sit from target — is conceptually wrong. It is that the framework requires an estimate of the "neutral" or "natural" rate of interest, commonly called r-star, and that estimate is not directly observable; it is inferred, revised, and debated among economists in real time, often after the fact. The Fed's own 3.0% long-run estimate is itself a judgment call, not a measured physical constant, and history is full of episodes where the neutral-rate estimate used to justify a given policy stance was revised substantially within a few years. A rate rule built on a genuinely moving, contested input will produce exactly the kind of whipsaw seen over the last two years — not because the rule is broken, but because the single most important number feeding into it is inherently uncertain in real time.
Calling interest-rate rules "outdated" because policy reversed direction in two years mistakes the instability of one input — the neutral rate — for a flaw in the whole instrument. The more useful question for an investor is not whether the Fed's framework is right, but how much conviction to place in any given dot plot, given how visibly wrong the implied path has been over just the last year.
Read alongside this series' companion letters, the picture sharpens. The inflation letter found this month's CPI print driven substantially by a gasoline price surge — a supply-side event a demand-cooling rate hike is not well-suited to address. The employment letter found a labour market that is genuinely tight by conventional measures, which is a much better textbook justification for tightening. The Fed's September statement leaned on both: "elevated" inflation and a "resilient," strong-productivity economy. Whether this hike is the right tool depends heavily on which of those two justifications is actually driving the decision — and on how much of the "resilience" reflects a genuinely tight labour market versus a shrinking pool of active job-seekers, the open question this series' employment letter raised and could not resolve.
This letter's own framing — that the neutral-rate estimate, not the rules framework, is the unstable element — is itself a judgment call rather than a settled fact; reasonable economists dispute how much weight r-star uncertainty deserves relative to other explanations for policy reversals, including simple forecasting error or genuinely unanticipated shocks. The Fed's own forward guidance has a well-documented history of being revised, sometimes substantially, within a single year, and there is no strong reason to treat the current 4.1%-4.4% year-end projection as more reliable than the guidance that existed a year ago and proved wrong. A continued hiking cycle into a labour market with a quietly falling participation rate carries real recession risk that this letter's other two pieces in this series both flagged without resolving.
Interest-rate policy did not break the old rules this year; it demonstrated, in real time, why those rules are only ever as stable as the neutral-rate estimate feeding them — an estimate that is inferred, contested, and subject to real revision, not a fixed dial. The practical takeaway for positioning is less about picking a side in the academic debate and more about sizing duration risk for continued path uncertainty, watching the same labour-market participation data this series' employment letter flagged as the leading indicator of whether September's hike was the right call, and treating every current dot plot as a snapshot of this month's thinking rather than a forecast to build a portfolio around.