PMI in Forex Trading: Two US Surveys, Two Different Formulas

A trader who follows the United States manufacturing sector sees two different PMI figures for the same month, published by two different organisations, sometimes pointing in opposite directions. Neither is a revision of the other and neither is wrong.

Most explanations of the purchasing managers index treat it as one number with one threshold at 50. That description fits neither of the two American releases exactly, and it misses the line the Institute for Supply Management actually publishes for the wider economy.

What follows separates the two surveys, states the arithmetic each one uses, and sets out where the reported thresholds come from.

There Is No Single PMI: Two US Surveys, Two Formulas

The Institute for Supply Management has run a survey of American purchasing and supply executives for decades and publishes the ISM Manufacturing PMI on the first business day of each month. S&P Global runs a separate survey of its own manufacturing panel and publishes the S&P Global US Manufacturing PMI, with data collection for that series beginning in May 2007.

Both carry the same three letters. They are not the same measurement. The panels are recruited separately, the questionnaires reach respondents at different points in the month, and the two headline numbers are assembled from their components under different rules.

On 3 August 2026 both landed within a quarter of an hour. S&P Global released its US manufacturing figure at 13:45 UTC. ISM publishes after 10:00 a.m. Eastern Time on the first business day, which put the second print fifteen minutes behind the first on the same morning.

That sequence matters to anyone watching a currency pair through the release. A trader who reads only the second figure sees a number arriving into a market that has already moved on the first. The economic calendar lists both, and the two rows are easy to read as one event.

The 47.5 Percent Line ISM Publishes, and the 50 Line Everyone Quotes

Explainer pages state a single rule: above 50 the economy is expanding, below 50 it is shrinking. ISM does not say that about the economy. It says that about manufacturing.

In its own monthly report ISM gives two separate thresholds. Anything over 50 percent tells you factories as a group are busier than they were a month earlier. For the wider economy the level ISM names is 47.5 percent, sustained across a period rather than read from one month.

The gap between the two numbers is the whole point. A manufacturing PMI of 48 sits below the sector line and above the economy line at the same time, so the same figure is a contraction in factories and consistent with growth in national output. Manufacturing is a minority of American economic activity, and the threshold reflects that.

The figure has moved before. ISM recalculates the relationship between its index and gross domestic product as the series lengthens, and the level it publishes today is not the level it published a decade ago. Any page quoting a threshold without naming the report it came from should be treated as a page quoting an old one.

The practical instruction is narrow. Read the threshold off the current ISM report, and keep the sector reading and the economy reading apart when interpreting what hawkish and dovish central banks are likely to take from the print.

How a Diffusion Index Is Built From Three Answers

A PMI is not an average of quantities. Nobody reports how much output rose. Respondents choose among three answers about the current month against the previous one: better, the same, or worse.

The index counts the positive share and adds half of the share reporting no change, treating an unchanged answer as half a positive one. That construction is why 50 is the neutral point. If every respondent reports no change, half of one hundred is fifty, and the index sits exactly at the middle.

That construction carries two consequences the headline number hides. The index measures direction and breadth, not magnitude: a reading of 58 says a wide share of firms saw improvement, not that output rose by any particular amount. And a reading of 45 does not mean output fell 5 percent. It means more of the panel reported decline than reported growth.

Both publishers use this construction, which is why the surveys are comparable in shape even where they disagree in level. Diffusion indices are conventionally grouped with the leading and lagging indicators that turn ahead of output data, because the survey is answered before the quarter it describes is measured.

What Each Publisher Weights, and Why the Two Disagree

Each headline PMI is a composite of five component indices, and the two publishers do not combine theirs the same way.

ISM gives each of its five the same share: production and new orders, employment, inventories, and the speed of supplier deliveries. S&P Global applies a graduated weighting instead, with new orders carrying the largest share and stocks of purchases the smallest, and inverts its supplier delivery series so that slower deliveries move the composite in the same direction as the other four.

Collection timing separates them further. S&P Global gathers responses during the second half of the month. ISM receives responses across most of the month, with the bulk arriving late. Two panels answering at different points in a month that turned mid-way can honestly report opposite directions.

Seasonal adjustment is applied by both, though not uniformly across every component. ISM adjusts New Orders, Production, Employment and Inventories, leaving Supplier Deliveries unadjusted in that list. How a seasonally adjusted series behaves when its factors are recalculated is covered on the CPI report page and is not repeated here.

Flash and Final Are Two Releases of the Same Month

S&P Global publishes a flash estimate before the final figure, and both describe the same reference month. The flash for July 2026 was released on 24 July. The final July manufacturing figure followed on 3 August, ten days later.

The flash is not a forecast. It is the same survey counted early, from the responses received by the cut-off, across panels of roughly 650 manufacturers and 500 service providers for the United States. The final figure includes the responses that arrived after the flash was compiled.

Because the flash arrives first, it is usually the release that moves a price. The final print lands into a market that already holds the flash, and the gap between the two is the only new information it carries. A trader treating both as full events is trading the same survey twice.

Flash coverage is not universal. S&P Global publishes flash estimates for a limited set of economies, and for the rest only a single monthly figure exists. ISM publishes no flash at all.

The One Major Release That Is Never Revised

Headline economic data is routinely restated. The non-farm payrolls report is revised in the two months that follow it, and a trade placed on the first print is frequently placed on a number that no longer exists.

Survey responses behind the PMI are handled differently. ISM says the answers it receives are kept exactly as submitted and never altered afterwards. S&P Global makes the same commitment for the survey data beneath its indices once a figure has gone out, while noting separately that the factors it uses to adjust for seasonality are recalculated periodically.

The distinction is worth holding precisely. What is fixed is the raw answer count. What can move is the adjusted series, because a recalculated seasonal factor changes the published history without any respondent changing an answer.

For a trader the effect is that a PMI level from an earlier month can be quoted with more confidence than a payrolls figure from the same month, while a chart of the adjusted series may still not match the one printed at the time.

Composite PMI Is Not the Manufacturing PMI

Alongside the sector indices S&P Global publishes a composite figure, which is often written as the composite PMI. It combines two series only, the output measure taken from manufacturing and the business activity measure taken from services, weighted by the annual value added of each sector.

S&P Global warns in its own releases against reading that composite against the manufacturing headline, because the two are not built on a common basis. The manufacturing headline averages five components under fixed weights, and output is only one of them, so a gap between the two indices is a difference of construction rather than a signal.

The practical error is a chart placing a composite series and a manufacturing series on one axis and reading the crossing points. Nothing is wrong with either series. They are simply two different constructions, and a difference between them carries no meaning.

Who This Page Is Not For

A reader who wants a rule turning a PMI print into a trade direction will leave without one. The index moves a currency through what it implies for policy, and that link runs through expectations rather than through the level printed on the day.

Anyone trading a currency outside the United States needs the publisher and the release schedule for that economy, not this one. Most national PMIs are compiled by S&P Global under a sponsor name, and the panel size, the release time and the availability of a flash estimate all differ by country.

Frequently Asked Questions

What does PMI mean for a forex trader?

PMI stands for purchasing managers index. It is a monthly survey of business conditions answered by purchasing executives, reported on a scale where 50 marks no change from the previous month. For a currency trader it matters mainly through what it implies for interest rate expectations rather than through the index level on its own.

Does a PMI above 50 mean the whole economy is growing?

No. For the ISM manufacturing series, 50 is the line for the manufacturing sector. The level ISM publishes for the overall economy is 47.5 percent, held over a period rather than read from a single month, so a reading between those two numbers indicates a shrinking factory sector alongside a growing economy.

Which release should a trader watch, ISM or S and P Global?

They are separate surveys and both are published. For United States manufacturing the S and P Global figure arrives first and the ISM figure follows minutes later on the same morning, so the second print is read against a market that has already absorbed the first.

Is a rising PMI bullish for the currency?

Not by itself. What moves a price is the difference between the released figure and what the market expected, not the direction of the index. A rising number that falls short of expectations can weaken a currency, and the reverse also happens.

Why do two PMI figures for the same month disagree?

Because they come from different surveys. The two panels are recruited separately, they are asked at different points in the month, and the five components are combined under different weights, so both figures can be accurate and still point in opposite directions.