Consumer Confidence Falls to 90.8 (July 2026): The Survey Closed 22 July — One Day Before Oil's Round Trip. What the Miss Means for the Dollar

The Conference Board's July Consumer Confidence Index fell 1.4 points to 90.8, missing a consensus near 92.3, with the Present Situation Index down 3.6 points to 114.9 — a third straight monthly decline — and the Expectations Index unchanged at 74.7. But the most consequential number in the release is not an index level at all. It is the cut-off date: responses were collected 1–22 July, and Brent's move above $100 came on 23 July, with the entire war premium then round-tripping to $88.36 by 27 July. The survey missed both halves of the energy story. And June's alarming labour differential, which looked like the narrowest since February 2021 when it was published, was revised from +2.4 to +3.8.

This is a cleaner illustration of the site's argument than any preview could have been. A price-only read of Tuesday morning sees "confidence misses, dollar barely moves, ignore it." Score the underlying factors separately and the release does three distinct things: it marks the growth factor down modestly, it leaves the interest-rate factor essentially untouched, and — critically — it tells you that the commodity factor's July violence has not yet reached household data at all. Knowing what a datapoint cannot possibly contain is as much a part of a fundamental read as knowing what it says.

What actually happened

The release landed at 10:00 a.m. Eastern on Tuesday 28 July, about 28 hours before the FOMC statement, and it was a soft print without being a frightening one.

Chief economist Dana M. Peterson framed it as confidence having "moderated slightly in July, continuing a general downward sloping trajectory since late 2021" — a description of a slow grind, not a break. The full detail is in the Conference Board's release.

Note the oddity in the labour internals: both sides fell. Fewer people said jobs were plentiful, and fewer said jobs were hard to get. That is the signature of a labour market where households are becoming less certain rather than more pessimistic — respondents migrating to the neutral "not so many jobs available" middle. It is consistent with the low-hiring, low-firing configuration that payrolls and claims have been describing from opposite directions.

The revision that erased last month's headline

The preview version of this page led on June's labour differential collapsing to +2.4, with 22.5% of consumers saying jobs were hard to get — at the time the highest share in more than five years. Those were the numbers the Conference Board published on 30 June. They are not the numbers now.

The revised June figures released alongside July's data put the headline at 92.2 rather than 91.2, hard-to-get at 21.7% rather than 22.5%, and the differential at +3.8 rather than +2.4. A 1.4-point swing in the differential is large relative to its monthly variation, and it changes the story from "the household read on hiring just fell off a cliff" to "the household read on hiring is drifting lower."

The survey closed one day before oil's round trip

Here is the detail that reframes the entire release. The preview argued that the Conference Board's late-month window would capture the energy reversal that the University of Michigan's early-July survey missed. That was wrong, and the reason is precise: the July cut-off was 22 July.

Brent settled above $100 for the first time since May on 23 July, with the AAA national average for regular petrol at $4.09 the same day, up 15 cents in a week. Then the war premium unwound completely — Brent closed 8.7% lower at $88.36 on 27 July as the Strait of Hormuz negotiations advanced and Caspian Pipeline Consortium loadings resumed, the move we covered in the Brent round-trip breakdown.

The internals corroborate the timing exactly. Twelve-month inflation expectations were less elevated in July, and write-in mentions of oil and gas prices eased in frequency — the opposite of what a survey that had witnessed a 15-cent weekly pump jump would show. What did increase were mentions of food and grocery prices, which is a slower-moving, stickier complaint and arguably the more durable one.

So the two US consumer surveys are not, in the end, telling contradictory stories about July energy. Neither of them saw it. Michigan's preliminary reading closed before the spike; the Conference Board closed one day before it. The first household data to contain any of this arrives in August — and by then the net move in crude is roughly flat to lower, which means the inflation-expectations shock the market spent a week worrying about may never show up in survey data at all.

Which scenario landed, and what it does to the dollar

The preview mapped three outcomes. The realized print sits in the base case, with one component missing.

The "hope" leg of the base case did not materialise. Expectations were unchanged at 74.7 rather than rising, which matters more than it sounds: the June increase in that sub-index had been attributed to falling energy costs, and a flat July reading in a window with no energy drama is the honest baseline. Households are neither more nor less optimistic about the next six months than they were a month ago. They are simply less positive about the present, for the third month running.

For the dollar the implications are modest and mostly one-directional:

• Growth factor: mildly negative. A third straight fall in the Present Situation Index, with consumption at roughly two-thirds of GDP, is a genuine mark-down of the near-term US growth read — but 90.8 against 92.2 is a drift, not an inflection.
• Interest-rate factor: essentially unmoved. The market carried roughly 40% odds of a July hike and about 95% priced for September into the release, as set out in the July FOMC preview, and a second-tier sentiment miss does not disturb that. The unchanged 61.3% of households expecting higher rates says the same thing from the survey side.
• Commodity factor: untouched by this release, by construction. Any energy pass-through to household expectations is an August story.
• Risk-sentiment factor: neutral here. A soft-but-orderly consumer print does not generate haven demand, so the usual complication — where a bad US number weakens the growth differential while simultaneously bidding the dollar as a refuge — does not really apply to a 1.4-point miss.

That is the value of separating the factors rather than reading one price. The dollar's behaviour on Tuesday is dominated by positioning into a live Fed meeting, not by a consumer survey; attributing the day's move to this release would be the classic error. The release's actual contribution to the fundamental picture is small, specific, and mostly about growth.

What to watch instead

Two things follow from this print, and neither is the headline.

The first is the Present Situation Index streak. Three consecutive declines in the sub-index that measures current conditions, while expectations sit flat, is the shape that usually precedes a softening in actual spending data rather than accompanying it. If August makes it four, the growth-factor mark-down stops being a drift.

The second is the August survey window, which will be the first household dataset to contain the July energy round trip in full. Because the net move was down — Brent finishing lower than where the spike began — the plausible August outcome is that inflation expectations ease further rather than spike, which would remove one of the stagflationary complications the market has been pricing. That is a genuinely useful thing to know in advance, and it is only knowable by reading the cut-off dates rather than the headline.

The Fed announces at 2:00 p.m. Eastern on Wednesday 29 July, per the Federal Reserve's calendar, with no fresh projections at this meeting. Tuesday's survey will not change what the committee does. It marginally strengthens the case that the household economy is cooling — and it conclusively demonstrates that the energy shock everyone spent last week pricing has not yet touched the data the Fed reads.

For the method behind the five-factor framework and how each input is scored, see how the meter works.

Educational macro context only — not investment advice.