FOMC Minutes Preview (19 August 2026): Three Dissents Meet 31% Hike Odds — Why the Dollar Trades the Conditions, Not the Vote

The minutes of the Federal Reserve's 28-29 July meeting are published on Wednesday 19 August at 2:00 p.m. Eastern time, and they describe a Committee that came closer to raising rates than any in a decade — a 9-3 hold with three regional presidents dissenting for a quarter-point hike, the first unified three-way dissent since September 2016. In the three weeks since that room emptied, the market has cut the odds of a September increase from roughly two-thirds to about 31%, on four data releases the Committee never saw. That gap is the entire trade: this document is a record of a reaction function, not a forecast, and the only part of it that has not already expired is the part that says under what conditions.

There is a second reason these minutes carry more weight than the genre usually does. Chair Kevin Warsh has spent his first months deliberately withdrawing the Fed's forward guidance — the July statement was much shorter than the norm and offered no policy bias, and he has said plainly that it "conveys just the facts. It's steering clear of forecasting." When a central bank stops telling you where it is going, the minutes stop being a footnote to guidance already given and become the primary surviving document about how the Committee thinks. The irony is that this elevation arrives at the exact moment the contents are most out of date.

Why the guidance vacuum makes an old document important

What lands on Wednesday is a narrative account of the July discussion, published three weeks after the decision on the schedule the Fed sets out in its own meeting calendar. It is not a transcript — those appear with a five-year lag — but a summary that has been through a drafting process where every adjective is chosen. That is a weakness if you want colour and a strength if you want structure, because the words that survive editing are the ones the Committee intended to be read. It also arrives into an unusually empty US week, with no American release of consequence before Friday's flash PMIs.

The July statement was, by design, almost content-free about the future. It repeated June's language that "Economic activity is expanding at a solid pace despite elevated uncertainty that owes, in part, to the conflict in the Middle East," noted that job growth "kept pace with the workforce and the unemployment rate has changed little," and ended with the declarative "The Committee will deliver price stability." No bias, no path, and no dots — the July meeting carries no Summary of Economic Projections.

Warsh has been explicit that this is the point rather than an oversight, arguing after the decision that market participants "are learning to play the ball, not the referee." Whatever one makes of the approach, it has a mechanical consequence for anyone pricing the dollar: the volume of official information about the policy path has fallen, so the remaining sources — the minutes, the speeches, and the data itself — carry proportionally more of the load. We traced the same dynamic in the July decision itself, where the front end and the long end of the curve said opposite things precisely because there was no guidance to anchor either.

The four releases the Committee never saw

This is the part most previews get wrong, so it is worth laying out precisely. Every judgement recorded in Wednesday's document was formed on or before 29 July. Here is what has landed since.

The payroll report is the one that reframes the meeting, because the composition was weak in a way that is hard to argue with. Private payrolls did rise 30,000, but government fell 53,000, led by a 50,000 drop in local government education; leisure and hospitality lost 40,000, plausibly the World Cup's end. Healthcare, the economy's most reliable engine, added 22,000 against a twelve-month average of 36,000. The unemployment rate fell to 4.1%, but for the wrong reason: household employment dropped 87,000 while the labour force shrank 264,000, pushing participation to 61.4% — outside the Covid era, its lowest since the middle of 1976. With revisions, the twelve-month average pace of job creation is now just 34,000.

Read the July statement's phrase again with that in hand. Job growth "kept pace with the workforce" was defensible on 29 July. Nine days later the workforce contracted by a quarter of a million people and payrolls went negative. The sentence did not become false so much as it stopped describing anything.

How to read the quantifiers

FOMC minutes are written in a coded dialect, and the code is a headcount. The ladder runs roughly from "a couple" and "a few" at the bottom through "some", "several", "many" and "most" to "almost all" at the top. A separate distinction matters just as much: "participants" means everyone at the table including non-voting regional presidents, while "members" means only the twelve with a vote. A hawkish view held by many participants but few members is a signal about the debate; the same view held by many members is a signal about the next decision.

The June minutes, released 8 July, are the template. They recorded that "almost all of these participants indicated that some policy firming would likely be warranted to return inflation to 2 percent" — near the top of the ladder — while on the appropriate year-end rate, "many participants" saw it within or slightly below the current 3.50%-3.75% range and "many other participants" saw it above. Together those lines are the shape of this Committee: near-consensus that firming is directionally right, an even split on whether it happens this year. June's projections put one quarter-point increase on the board by end-2026.

So the thing to look for on Wednesday is movement along that ladder. If July's minutes upgrade the hike camp from "several" to "many" or beyond, the three dissents understate the hawkishness and the market has repriced too far. If the language holds at June's level while three people vote against, the dissents are the tail of the distribution rather than its leading edge.

That distinction is why the dissent count itself is the most over-read feature of the meeting. It establishes that the Committee is divided; it does not establish direction of travel, and the market's verdict makes the point, since September pricing fell on the day of a hold with three hawkish dissents. Dissenters are by construction the extremes, so the fact that exactly three of twelve voted to hike is as much information about the nine who did not — and some of the loudest inflation worriers were not among them. Governor Christopher Waller has spoken about the possibility that higher rates become necessary yet voted for the hold, while New York's John Williams has said current policy is well positioned to return inflation to target. The dissents tell you about the edge of the committee; the minutes tell you about its middle, and the middle decides.

The scenario map, and what each does to the dollar

The dollar's interest-rate factor keys off the expected policy path, so the question is not whether the minutes sound hawkish — they will, because the meeting was — but whether they contain anything capable of moving a probability that has already fallen from roughly two-thirds to 31%.

A hawkish surprise in breadth. The minutes show more than three participants judging a hike appropriate, or frame the hold as a question of timing rather than of the appropriate level of rates. This reading has the most room to move, because the market has spent three weeks pricing the opposite.

Explicit conditionality. The minutes name the labour or inflation thresholds that would trigger action — the most useful outcome regardless of direction, because a stated condition can be applied to data the Committee never saw. If the condition for holding was a cooling labour market, the payroll report has already satisfied it. If the condition for hiking was inflation failing to return toward target, core CPI at 2.5% is doing more work than the 3.4% headline suggests.

Confirmation. The minutes restate what is known — a divided Committee, no bias, decisions taken meeting by meeting. The most likely outcome, and the one with the least market consequence. The September question then rests on July PCE on 26 August and the August employment report.

The arc of pricing is itself the lesson in why the mechanism matters more than the vote. CME's FedWatch had a September increase at roughly two-thirds in the days after the meeting, 44% once payrolls landed on 7 August, 42% after CPI on 12 August, and near 31% by Friday 14 August. The Fed did nothing at all across that stretch. The hurdle rate for holding dollars moved thirty-odd points on data alone — the same channel we traced through gold and real yields over the identical fortnight.

The dollar through the five factors

Pip Theory scores eight currencies on five fundamental factors, and this week's event touches three of them unevenly.

The interest-rate factor is the direct channel and the one described above: not the level of 3.50%-3.75%, which is not in question, but the market's expectation of the next move, which has softened materially. The growth factor has quietly deteriorated in a way the minutes cannot reflect — negative payrolls, a shrinking labour force, a 0.6% fall in retail sales and a preliminary University of Michigan sentiment reading that dropped to 51.0 from 55.2. The risk-sentiment factor cuts the other way and is the reason a weak-data dollar has not simply fallen: softer rate expectations lifted equities and pulled Treasury yields down, which is a risk-on configuration that typically works against the dollar's haven bid even as it works against its carry.

That is why a single price line is a poor summary of this week. Three factors are moving, two of them in opposite directions, and a document written three weeks ago is about to add information to only one.

What comes after Wednesday

The minutes start a dense three weeks rather than ending anything. July PCE — the Fed's preferred inflation gauge — arrives on 26 August. The Jackson Hole symposium runs 27-29 August, where Warsh is due to deliver his first keynote as chair, and for a Fed that has withdrawn written guidance a set-piece speech carries unusual weight. Then the FOMC meets on 15-16 September, and that meeting does carry a Summary of Economic Projections — the first fresh dot plot since June, when the Committee penciled in one quarter-point increase by year-end. If the data has changed minds, September is where the change becomes visible in numbers rather than adjectives.

Two further dates frame the tail: 27-28 October and 8-9 December. Even after the September repricing, futures have continued to carry higher odds of a move at those later meetings than at September itself — the market's way of saying it thinks the hike was postponed rather than cancelled.

The takeaway

The instinct on minutes day is to score the document hawkish or dovish and trade the label. This one resists that, usefully. The meeting it describes was hawkish on any measure — a fifth hold, but with the most fractured directional dissent in a decade. Yet every level judgement it contains was formed before a negative payroll print, 103,000 in downward revisions, two cool inflation reports and a 0.6% drop in retail sales. Trading the label means trading a description of an economy that the subsequent three weeks substantially revised.

What survives is narrower and more valuable: the conditions. In a regime where the chair has switched off forward guidance on purpose, the Committee's stated triggers are the closest thing to a policy rule anyone outside the building has. Find the conditional sentences, note how many stood behind each, check whether they were participants or members — then apply them to the data the Committee had not yet seen. That exercise tells you something about September. The vote count, three weeks old, tells you only about July.

To learn how Pip Theory builds its fundamental currency-strength scores, see the methodology overview, or track the dollar's factor breakdown on its USD page. Source data: the Federal Reserve's FOMC calendar, the BLS employment situation and producer price index, and Census Bureau retail sales. Market reaction and vote detail reported by CNBC.

Educational macro context only — not investment advice.