Update, September 16: the Fed raised rates

The committee voted 12 to 0 to raise the target range by a quarter point to 3.75% to 4.00%, the first increase since July 2023 and the fastest cut-to-hike turn since 1999, as this explainer anticipated. The new dot plot shows 16 of 18 submitting participants expecting another hike this year, four of them seeing two more. Chair Warsh again submitted no dot. At the press conference he said inflation "is too high and has been for too long", and named three changes since July: a stronger economy, inflation that did not slow, and rising geopolitical tension. The 10-year real yield closed that day at 2.68%, its highest since 2008, which we explain in our real interest rates explainer.

What is actually being decided

The Fed does not set "interest rates". It sets a target range for the federal funds rate, which is what banks charge each other to borrow overnight. Everything else, from your savings yield to a car loan, is priced off that anchor at some distance.

The range today is 3.50% to 3.75%. On Wednesday there are three realistic outcomes.

OutcomeNew rangeWhat it says
Hike 25 basis points3.75% to 4.00%Inflation is not falling fast enough; the December cut is being partly taken back
Hold3.50% to 3.75%Wait for more data; the argument continues to the October meeting
Cut 25 basis points3.25% to 3.50%Almost nobody is pricing this; the June dot plot had one participant expecting it

A basis point is one hundredth of a percentage point, so 25 basis points is a quarter of one percent. The Fed moves in these increments because the effect is enormous at scale: a quarter point on the overnight rate ripples through trillions of dollars of loans and deposits.

The decision is published at 2:00pm Eastern, which is 18:00 UTC. The Chair takes questions at 2:30pm. September is also one of the four meetings a year that publishes the Summary of Economic Projections, so a new dot plot lands at the same moment. More on that below.

Why a hike is even on the table

Nine months ago the Fed was cutting. That it is now debating the opposite is the whole story, and it happened in four steps.

July 29: the 9 to 3 vote. The committee held rates, but three members dissented in favour of a hike: Beth Hammack of Cleveland, Neel Kashkari of Minneapolis and Lorie Logan of Dallas. Three dissents in the hawkish direction is the most fractured vote in nearly a decade. All three are regional bank presidents holding a vote in this year's rotation, which means they vote again on Wednesday.

August 19: the minutes. The written record of that meeting showed officials "saw a need" for a hike if inflation did not cool. That is unusually direct language for a Fed document.

Late August: Jackson Hole. Chair Kevin Warsh, sworn in on May 22, used the annual symposium to promise price stability in terms markets read as hawkish. Futures pricing for a September hike roughly doubled, from about 35% to 66%, in the days after he spoke.

September 10: the producer price index. Wholesale prices rose 0.4% on the month, more than expected, with oil up about 10% in September and WTI above $94. The consumer price index follows on the 11th, five days before the decision.

The last time this happened was 1999

Central banks cut in a hurry and hike in a hurry, but they very rarely do one and then the other within a year. It looks like an admission that the first move was wrong, and institutions avoid that.

CycleLast cutFirst hikeGap
1998 to 1999Nov 17, 1998Jun 30, 1999about 7 months
2003 to 2004Jun 25, 2003Jun 30, 200412 months
2008 to 2015Dec 16, 2008Dec 16, 20157 years
2020 to 2022Mar 15, 2020Mar 16, 20222 years
2025 to 2026, if it happensDec 10, 2025Sep 16, 2026about 9 months

The 1998 episode is the one everyone in the building remembers. The Asian crisis, the Russian default and the collapse of the hedge fund Long Term Capital Management pushed the Fed to cut three times in ten weeks, from 5.50% to 4.75%. The economy did not slow. By June 1999 the Fed was hiking again, and it did not stop until the rate reached 6.50% in May 2000.

Whether 2026 rhymes with that is not something we can measure. What we can say is that a hike on Wednesday would be the fastest cut to hike turn in 27 years, and that is the reason this meeting is drawing more attention than a routine quarter point ever would.

Who actually votes

The FOMC has nineteen participants but only twelve votes. Every meeting, the vote is cast by:

  • the seven governors of the Federal Reserve Board in Washington, including the Chair
  • the president of the New York Fed, who votes every year
  • four of the other eleven regional presidents, on a rotation that changes each January

The 2026 rotation seats Cleveland, Philadelphia, Dallas and Minneapolis. Three of those four dissented for a hike in July. So of the twelve people voting on Wednesday, three are already on the record, and the outcome depends on whether they persuade four more.

A simple majority decides. Dissents are recorded by name in the statement, which is why they matter: a Fed official does not put their name against the Chair lightly, and three at once is a public signal that the argument is close.

The dot plot, and why the Chair will not be in it

Four times a year, at the March, June, September and December meetings, every participant writes down where they think the rate should end this year and the next few. Each answer becomes an anonymous dot on one chart. Markets treat the median dot as the committee's forecast, which the Fed itself repeatedly says it is not.

The June 2026 dot plot showed a committee split almost down the middle.

Expected end of 2026 rateParticipants
At least one hike9
No change8
A cut1
Median3.75%, the top of the current range

That is eighteen dots, not nineteen. Chair Warsh has said he encourages colleagues to keep submitting projections but declines to submit his own, describing that as consistent with his long held view of the exercise. So Wednesday's chart will again show the committee minus the person running it, which makes the median a slightly odd thing to read as "the Fed's view".

Where "66%" comes from

When you read that markets price a 66% chance of a hike, that number is not a poll. It is derived from the price of federal funds futures, contracts traded at CME that settle on the average fed funds rate in a given month. If the October contract implies an average rate of 3.79% and the current midpoint is 3.625%, the market is pricing roughly two thirds of a quarter point move. The CME FedWatch tool turns that arithmetic into the percentage everyone quotes.

Two things about that number are worth knowing. It moves with every data release, and it has been wrong before. Here is how it has moved in our own coverage over the last three weeks.

MomentPriced chance of a September hike
Before Jackson Holeabout 35%
After Warsh's remarks66%
After the August payrolls report58%, from 49.4% the day before
After Waller said he leaned toward holdingabout 50%
After Thursday's producer prices63%

Five readings in three weeks spanning thirty percentage points. That is not a forecast, it is a live argument, and Friday's CPI will move it again before anyone votes.

What a quarter point changes for you

A hike to 3.75% to 4.00% reaches households at very different speeds depending on the product.

ProductRepricesDirection after a hike
Credit cards, home equity lines, adjustable loansWeeksUp, almost one for one
Savings and money market yieldsWeeksUp, usually by less than the Fed move
Fixed 30 year mortgagesNot directlySet off 10 year Treasury yields, which price years of Fed decisions at once
The dollarSame dayTypically stronger, since dollar deposits pay more

The mortgage row surprises people every cycle. A fixed mortgage is priced off long term bond yields, and those yields reflect what the market expects the Fed to do over the next decade, not on Wednesday. If a hike convinces bond investors that inflation will be controlled, long yields can fall on the day the Fed raises, and mortgage rates with them. It has happened before and it is not a paradox once you know what each rate is tied to.

For anyone earning in dollars and spending elsewhere, or holding assets priced in dollars, the currency row matters more than the rest. A hike makes dollar deposits pay more relative to other currencies, which tends to pull the dollar up and to press on anything priced against it, gold and Bitcoin included.

What Bitcoin actually did on the last eight decision days

This is a macro explainer on a crypto site, so we will close the way we always do: by measuring rather than asserting. We took the last eight Fed decision days and pulled the Binance BTCUSDT session and the exact 2:00pm Eastern hour for each.

Decision dayDecisionSessionDecision hourHour vs average hour
Sep 17, 2025Cut 250.29% lower0.24% lower4.9x
Oct 29, 2025Cut 252.55% lower0.96% lower4.9x
Dec 10, 2025Cut 250.72% lower+0.49%4.0x
Jan 28, 2026Hold+0.06%0.16% lower1.5x
Mar 18, 2026Hold3.60% lower0.70% lower1.5x
Apr 29, 2026Hold0.74% lower0.88% lower2.5x
Jun 17, 2026Hold1.77% lower0.90% lower2.8x
Jul 29, 2026Hold, 9 to 3+0.11%+0.36%2.7x

Three things stand out, and the first one is the strangest.

Bitcoin fell on all three rate cut days. Cuts are supposed to be the friendliest thing a central bank can do for a risk asset, and the sessions read 0.29%, 2.55% and 0.72% lower. We are not going to explain that, because eight observations cannot support an explanation. We are going to point at it.

Six of eight decision days closed lower. Against a baseline where about 48% of all sessions close lower, that is a lean, on a sample too small to call a pattern.

The decision hour is real but modest. It traded 3.1 times an average hour, larger than the 1.9 times we measured for CPI releases, with the three cut days all at 4 times or more and the hold days between 1.5 and 2.8. Yet the mean absolute move on a Fed day was 1.23%, less than the 1.62% baseline. The market shows up for the hour and then does less than usual with the rest of the day.

Set beside the nine CLARITY milestones and the nine CPI hours, this is now the third scheduled event type we have measured, and the same sentence describes all three: the crowd is smaller than the coverage, and the direction is not knowable in advance.