A plain-English guide to flash PMIs vs ISM.

Wednesday's analysis piece is the live example. S&P Global's US flash PMIs printed at 57.0 manufacturing and 58.7 services, well above consensus (53.7 and 56.0). The 10-year yield tore 14.6bp higher through 5.10 percent on the print; DXY broke 101; gold cracked 58 dollars. The tape moved because both indices printed at cycle highs. This reference piece is the framework for reading a flash PMI print, why it differs from ISM, and where the diffusion-index methodology carries information that a headline growth print does not.

What a PMI actually measures

Purchasing Managers' Index (PMI) surveys ask a panel of purchasing managers at manufacturing or services firms whether specific business conditions are better, the same, or worse than the previous month. The diffusion index is calculated as: percent-better plus half-of-percent-same, with 50 as the neutral threshold. Above 50 means expansion; below 50 means contraction. The specific mechanic is that PMIs measure the direction of change, not the level of activity. A print of 58 does not mean the economy is 8 percent above trend; it means 58 percent of respondents are seeing better conditions this month than last month (roughly).

Because PMIs measure direction of change, they are noisier around inflection points and more informative in the middle of a regime. When the economy is decisively expanding or contracting, PMIs give a clean signal. When it is bouncing along a soft-vs-firm boundary, PMIs can pop back and forth on marginal shifts. That is a specific reason why cross-checking the sub-indices (new orders, employment, input prices, output prices) matters more than reading the headline alone.

Flash vs final vs ISM

S&P Global (formerly Markit) publishes two versions of its US PMI each month: a flash release around the third or fourth Wednesday, based on roughly 85 percent of responses, and a final release around the first business day of the following month. The flash lands earlier and is the market-moving version; the final rarely surprises. The Institute for Supply Management (ISM) publishes its own monthly PMI on the first business day of the following month, alongside a services PMI two days later.

The three specific differences that matter:

1. Panel composition. ISM surveys larger firms drawn from its member roster; S&P Global surveys a broader mix including small and mid-sized firms. The two panels can diverge, and the divergence itself carries information about firm-size differences in the economy.
2. Sub-index construction. ISM's PMI is a weighted average of five sub-indices (new orders, production, employment, supplier deliveries, inventories) with fixed weights. S&P Global's diffusion index weights components differently and includes output prices and input prices as separate series rather than folded into the headline. The result is that ISM PMI tends to be more employment-sensitive and S&P Global tends to be more orders-sensitive.
3. Release timing. S&P Global's flash lands two-to-three business days before the ISM manufacturing print for the same month. When the flash prints materially above consensus, the market often pre-positions for the ISM print; the flash is therefore the specific earlier read that carries the first-mover advantage.

The three sub-indices that carry the macro read

1. Employment. Both series carry an employment diffusion sub-index. A print above 55 on the employment component is the specific signal that hiring is accelerating faster than the headline growth-of-orders would predict. Wednesday's release had manufacturing employment at its fastest pace in more than four years, which is why the tape read the print as a labour-market signal in addition to a growth signal.
2. New orders. Orders lead output by roughly one to two months. A jump in the orders sub-index typically leads a jump in the output sub-index by one release. When orders and output move together (as they did in Wednesday's release), the signal is coincident rather than lagged, which the market reads as evidence of a sustained regime shift.
3. Input prices. The specific pass-through indicator from PPI to CPI. When input prices rise faster than output prices, margins compress and the pipeline pressure builds; when output prices rise faster, firms are passing costs through and PPI is transmitting to CPI. Wednesday's release had both rising, with input rising faster, which reads as pipeline-pressure-building for the coming three-to-six months.

How to read a flash PMI print inside a hawkish-Fed regime

In a normal (data-dependent) Fed regime, a strong PMI print is a "growth-good, inflation-mixed" signal. Under a Warsh-led hawkish regime, the same PMI print reads as unambiguously hawkish because the reaction function has been specified as tolerating slower growth if it delivers inflation control. The specific reason Wednesday's tape moved so violently is not that the growth read was surprising in isolation, but that the specific print combination (strong growth plus firm employment plus building input prices) was the exact pattern that undermines the September SEP's inflation-forecast trajectory.

The framework rule is: in a Warsh-led regime, a PMI composite above 56 is a hawkish signal, a print between 52 and 56 is neutral, and below 52 is dovish. Wednesday's 58.4 sits comfortably in the hawkish bucket and is the specific reason CME FedWatch's October hike odds moved from 55 to 66 percent on the print.

The three specific errors to avoid when reading a PMI

• Reading the headline without the sub-indices. The headline is a weighted average; the story lives in the components. Employment and orders carry the macro read; supplier deliveries carry the supply-chain read; inventories carry the inventory-cycle read.
• Confusing flash with final. The flash is the moving release; the final rarely surprises. When they diverge, that is a signal in itself.
• Ignoring the transatlantic split. The eurozone, UK, and Japan flash PMIs release on similar Wednesday cadence. Transatlantic divergence (US strong, UK weak, as on Wednesday) is the specific quantitative anchor for a "right side of the dollar smile" regime.

Related reading

• Wednesday's analysis piece: the live example this framework describes.
• The dollar smile: the regime framework that Wednesday's transatlantic PMI split anchors.
• The week after an FOMC hike: why Wednesday's PMI print specifically undermines the September SEP's inflation forecast.
• Final-demand PPI and CPI transmission: how the input-prices sub-index feeds through to the following month's PPI print.