Skip to main content

ISM Manufacturing & Services (PMI)

Updated Jun 24, 2026 at 2:35pm

Research Draft Medium 1,248 words

The ISM Manufacturing PMI and ISM Services PMI are monthly diffusion indices published by the Institute for Supply Management, built from surveys of purchasing and supply executives who report whether key business conditions (orders, production, employment) are better, same, or worse than the prior month. Because purchasing managers commit capital ahead of production, the surveys are treated as timely, forward-leaning reads on the business cycle — released within the first three business days of the month, well before the GDP and most official "hard" data they anticipate. Their core tension: they are fast and high-signal at the macro turning points, but they are soft (sentiment-based), volatile, and — for manufacturing especially — measure a sector that is now under ~10% of US employment, so a single month can mislead about the whole economy.

How it's calculated / formed

Each component is a diffusion index scaled 0–100: % reporting "higher" + ½ × % reporting "same". A reading of 50 is the dividing line — above 50 signals the sector is expanding versus the prior month, below 50 contraction; the distance from 50 indicates the breadth (not magnitude) of the change. Note this measures month-over-month direction across firms, not the level or dollar size of activity.

  • Manufacturing PMI is a composite of five seasonally adjusted subindices — New Orders, Production, Employment, Supplier Deliveries, and Inventories — which ISM combines with equal weight since a 2008 methodology change (older descriptions citing 30/25/20/15/10 weights are outdated). Survey panel: 300–400+ manufacturing firms across NAICS sectors. (Institute for Supply Management; Investopedia.)
  • Services PMI (formerly "Non-Manufacturing NMI") is a composite of four equally weighted subindices — Business Activity, New Orders, Employment, and Supplier Deliveries. (Institute for Supply Management.)

ISM publishes calibration thresholds versus GDP that it recalibrates over time. As of the 2026 reports, ISM states a Manufacturing PMI above 47.5%, sustained over a period, generally indicates overall economic (GDP) expansion; for Services the comparable threshold ISM cites is 48.1% (older write-ups citing ~42 / ~48.6 reflect earlier calibrations). ISM also maps each headline to an annualized real-GDP growth rate — e.g. ISM's May 2026 manufacturing release stated the 54.0 reading "corresponds to a 2.2-percent increase in real gross domestic product (GDP) on an annualized basis." The key takeaway holds regardless of the exact figure: manufacturing can print below 50 — technically "contracting" month-over-month — while the overall economy still grows.

Release schedule: Manufacturing on the first business day of the month at 10:00 a.m. ET (second business day in January); Services on the third business day, also 10:00 a.m. ET — both referencing the prior month. (Institute for Supply Management ROB calendar.)

How it's used in practice

Traders and economists read the PMIs in three layers:

1. The headline vs. consensus. The market trades the surprise relative to the economist consensus, and the 50 line as a regime marker (expansion/contraction). A cross of 50 — in either direction — gets outsized attention. 2. The internals. The New Orders subindex is the most forward-looking (tomorrow's production); Employment previews payroll trends; Prices Paid (a tracked component, not in the composite) is a real-time inflation gauge that bond and rates desks watch closely; Supplier Deliveries rising can signal either strong demand or supply bottlenecks (it must be read with context). The New Orders minus Inventories spread is a common leading-of-the-leading heuristic. 3. Manufacturing + Services together. Services is far larger as a share of the US economy, so a divergence — weak manufacturing, firm services — is the normal late-cycle pattern and was the dominant read through much of 2022–2024. Analysts increasingly treat Services (and the broader composite) as the better whole-economy gauge, with Manufacturing as the more cyclical, more volatile early-warning sensor.

For markets: a strong beat is typically read bullishly for cyclical equities (industrials, energy, financials) but can be bearish in a tightening regime by pushing rate expectations up — context (the Fed reaction function) flips the sign. Rates and FX desks react sharply to Prices Paid and Employment.

Adoption, debate & evidence

The ISM surveys are among the most-watched US releases, embedded in the Conference Board's Leading Economic Index and routinely cited by the Fed. That status is well earned at turning points — sustained sub-50 manufacturing readings have preceded or coincided with every post-war US recession. But the honest caveats are real and well documented:

  • False signals are common. Manufacturing PMI has dipped below 50 many times without a recession following (e.g. 2012–13, 2015–16, and 2022–2024 all featured extended sub-50 manufacturing prints with no NBER recession). Gavekal Research argues US manufacturing is better understood as a coincident, volatile indicator that merely looks leading.
  • Duration matters more than a single cross. ISM's own thresholds are explicitly framed "over a period of time," not for a single print — a sustained run of sub-50 manufacturing readings carries far more weight than a one-month dip below 50.
  • ISM vs. S&P Global PMI divergence. The competing S&P Global (formerly Markit) US PMI frequently disagrees with ISM. S&P Global surveys a larger panel (the firm cites 1,300+ US companies vs. ISM's roughly 600–700 across both surveys combined) and uses "live" X13-ARIMA seasonal factors plus an in-house smoothing system, whereas ISM applies seasonal factors forecast in advance for the year. S&P Global says this yields lower month-to-month volatility — by its own published comparison, since 2010 the services index varies from its 3-month average by about 1.5 index points for ISM vs. ~0.8 for S&P Global (S&P Global's comparison; manufacturing-specific point figures not separately verified here). Neither is definitively "right"; they measure overlapping but different panels, so persistent divergences (as in early 2024) are a known source of confusion rather than a resolved question.
  • Flash vs. final. S&P Global publishes a "flash" estimate (based on roughly 85% of responses) about 12 days before its final figure; ISM has no flash — its number is a true monthly surprise, part of why the ISM release itself moves markets.

Strengths & limitations

Works best as: an early, timely, regime-level read on direction — especially the 50-line crosses, the New Orders internal, and Prices Paid for inflation. Strong at confirming cyclical turns and at flagging building stress before hard data.

Fails when: used as a precise recession trigger off a single month (high false-positive rate), or when manufacturing's small employment share is forgotten — a manufacturing slump can coexist with a healthy services-led economy. It is sentiment/breadth, not magnitude: a 55 print does not mean activity grew "more" than at 53 in dollar terms.

#1 misuse: treating any sub-50 manufacturing print as a recession call, or trading ISM and S&P Global PMI as interchangeable. Always read internals (New Orders, Prices, Employment) and pair Manufacturing with Services.

Sources

Disputes flagged: (1) ISM vs. S&P Global PMI frequently diverge — neither is authoritative; (2) whether manufacturing PMI "leads" or merely "coincides with" the cycle is genuinely contested (Gavekal vs. conventional LEI usage); (3) the GDP-expansion threshold figures (currently 47.5 for Manufacturing, 48.1 for Services) are ISM's own calibrations and are revised over time — treat as approximate and check the latest report.