INSIGHTS
What the FED hike settles, and what it does not
Silvan Schriber · September 2026
The FOMC raised the target range to 3.75–4.00% on Wednesday, unanimously, the first increase since July 2023. Sixteen of the 18 participants submitting projections expect at least one more this year; four see two. The statement describes activity expanding at a solid pace, robust capital investment, strong productivity growth, and inflation that remains elevated.
Markets had the move priced at better than 90%, so the decision itself carried little information. What it does not settle is more interesting: how restrictive policy already is, how much further it needs to go, and what the cost of capital looks like on the other side. Those questions turn on the composition of the current inflation, the timing of the investment cycle, and whether the market continues to believe the reaction function.
The inflation picture is three different inflations
Headline CPI held at 3.4% in August (BLS, released 11.09.2026), driven by energy: gasoline up 27.4% year-on-year, fuel oil up 52%. Core CPI eased to 2.4%, the lowest reading since March 2021, with shelter decelerating to 3.0%. On the CPI evidence alone, the domestic inflation process looks close to contained and the problem looks like a relative-price shock originating in the Middle East.
The Fed targets the Personal Consumption Expenditure (PCE) Price Index, and there core stood at 3.3% year-on-year in July (US Bureau of Economic Analysis, 26.08.2026), headline at 3.7%. The gap against core CPI is wide enough to change the policy conclusion, which makes its composition worth examining rather than averaging away.
The Dallas Fed trimmed mean PCE was 2.3% over the same twelve months. When the trimmed mean sits a full percentage point below core, the core reading is being carried by a small number of components rather than by a broad price impulse. In July, financial services and insurance rose 1.2% month-on-month – a category the BEA largely imputes from asset prices rather than observes directly. Powell made the point during his tenure that imputed prices of this kind say little about economic tightness.
Taken together: headline inflation is an energy shock, core PCE is flattered upward by imputation, and the trimmed mean suggests underlying pressure closer to target than the Committee's preferred measure implies. The genuine open question is narrower – whether the energy shock is being absorbed as a one-off price-level adjustment or beginning to propagate through services. August core CPI came in a tick hotter than expected on non-housing services, the component with the closest link to wages. That is the series to watch, and it will not be resolved before December.
Strong investment raises the equilibrium rate – with a lag that runs the wrong way
The capex cycle is the most consequential variable in the medium-term rate outlook and the least settled. The standard framing is that a higher marginal product of capital raises r*, which argues for a higher terminal rate. That is correct as far as it goes.
The part that gets less attention is the sequencing. AI-related capital expenditure lifts aggregate demand immediately and adds productive capacity years later. For the duration of that gap the investment boom is inflationary, and inflationary through channels monetary policy is poorly suited to address, because the demand originates with firms holding strong balance sheets and limited rate sensitivity. The disinflationary payoff arrives afterwards, on a horizon that coincides with no committee's forecast window.
There is a second, more immediate channel. Corporate debt issuance from AI companies has been heavy enough to absorb dealer balance-sheet capacity and push long yields higher independently of policy expectations. That is a crowding-out effect operating through intermediation constraints rather than through saving, and whether it persists is a question about issuance calendars, not about r*.
The implication for planning is narrower than “rates stay higher for longer”. It is that the demand and supply effects of this investment cycle arrive at different times, and the inflation path between the two is not the inflation path after it.
The credibility test has an observable, and it just passed
The 10-year Treasury traded above 5% on 15.09.2026, the highest since July 2007. By Thursday morning it had eased to 4.951%, with the 30-year at 5.297% and the 2-year at 4.692% (CNBC, 17.09.2026). Long yields fell after a hike. That is the textbook response to a central bank confirming its reaction function, and it is worth noting because the opposite outcome is the one that should concern anyone carrying duration.
The distinction is observable rather than theoretical. If long yields rise alongside breakeven inflation, the market is repricing the credibility of the target. If they rise with real yields, it is repricing growth, term premium or supply. The two carry entirely different implications for hedging and for discount rates, and they are routinely conflated in commentary that reports only the nominal level.
Note also the shape: roughly 26 basis points between 2s and 10s. There is very little cushion in that curve.
The differential is now the dominant variable for Swiss and European balance sheets
For institutions in Zurich, Geneva or Frankfurt the domestic question is more consequential than the American one. The ECB raised its deposit facility rate to 2.50% effective 16.09.2026, its second increase this year, with 2026 inflation projected at 3.0%. The SNB policy rate has been 0.00% since 20.06.2025, and the June assessment put conditional inflation at 0.6% for 2026, 0.6% for 2027 and 0.7% for 2028 – comfortably inside the range of price stability, on the assumption of a 0% policy rate throughout. The next assessment follows on 24.09.2026 at 09:30 CET.
The same energy shock is therefore producing a 3.75-4.00% policy rate in the United States, 2.50% in the euro area and zero in Switzerland. That is not inconsistency. It is the franc doing what it does, plus a domestic inflation process that never took hold. But it leaves a spread of roughly 400 basis points at the front end and a comparable gap at ten years – the Confederation 10-year stood at 0.61% on 16.09.2026 against a Treasury near 4.95%.
For Swiss insurers and pension funds still discounting long liabilities against a domestic curve near zero, the appeal of the unhedged USD yield pickup is obvious and the hedging cost is the whole argument. The differential that makes the asset attractive is the same differential that makes the hedge expensive, and the residual after hedging is thinner than the headline gap suggests. The relevant exposure is not the level of US rates but a widening differential combined with franc appreciation – a combination this energy shock makes more rather than less likely, since Switzerland absorbs the terms-of-trade deterioration through a stronger currency rather than through a higher policy rate.
What to test
The scenario worth stressing is not a parallel shift. It is a divergence between the short and long end, in either direction, combined with a widening rate differential against CHF and EUR. Long yields rising faster than short yields would expose refinancing profiles and duration mismatches; long yields falling faster would compress reinvestment assumptions built over the last three years. Both belong in the planning case alongside the more benign path in which productivity-led disinflation does the work and the Fed stops at 4.25%.
Wednesday sharpened the near-term inflation question. It left the longer-term cost of capital, and the sources of its movement, open.
Sources and as-of dates: FOMC statement and Summary of Economic Projections, 16.09.2026. BLS Consumer Price Index, August 2026, released 11.09.2026. US Bureau of Economic Analysis (BEA), Personal Income and Outlays, July 2026, released 26.08.2026. Federal Reserve Bank of Dallas, Trimmed Mean PCE, July 2026. ECB monetary policy decision of 10.09.2026, rates effective 16.09.2026. SNB monetary policy assessment of 18.06.2026; SNB current interest rates as at 16.09.2026; next assessment 24.09.2026, 09:30 CET. Treasury market levels as reported 17.09.2026.