ISM Services PMI explained

What the ISM Services PMI measures

The Institute for Supply Management publishes the Services PMI on the third business day of each month at 10:00 a.m. Eastern, covering conditions in the prior calendar month. Roughly 350 purchasing and supply executives respond across industries including accommodation and food services, finance and insurance, health care, information, professional services, retail trade, transportation, and utilities. Each respondent answers whether a given metric is higher, the same, or lower than the prior month.

ISM converts responses into a diffusion index:

Index = P₁ + 0.5 × P_same

where P₁ is the share reporting improvement and P_same is the share reporting no change. The scale runs from 0 to 100, with 50 as breakeven. Above 50 means net expansion breadth; below 50 means net contraction breadth. A print of 52 does not mean services output grew 2%; it means a modest majority of surveyed firms saw conditions improve versus the prior month.

Because services are less inventory-intensive than manufacturing, the survey emphasizes activity levels and order intake rather than stock cycles. A services firm can report rising business activity while employment is flat if it meets demand through productivity or longer hours before hiring. That lag pattern shows up repeatedly in the employment subindex and matters for nonfarm payrolls forecasting.

Headline composite and the main subindexes

Business Activity

The business activity subindex is the services analogue of manufacturing production. It tracks current operating levels — patient visits, transaction volumes, billable hours, foot traffic. It tends to move concurrently with hard spending data and is less leading than new orders. When business activity stays above 50 while new orders slip below 50, firms may be working through backlog or seasonal strength even as forward demand softens.

New Orders

New orders is the primary leading indicator inside the services report. Contract signings, booking pipelines, and reservation trends turn before revenue is recognized. Harbor weights new orders at 35% in its internal services momentum score. Three-month new orders momentum correlates better with future PCE services growth than any single headline surprise.

Employment

The employment index reflects hiring, layoffs, and unfilled positions across service industries. Because services dominate U.S. payrolls, this subindex often provides early signal for establishment-survey job growth — though it can diverge when large employers outside the panel drive national prints. A sustained drop in services employment below 48 has historically preceded softer monthly payroll gains with a one- to two-month lag, while manufacturing employment moves a smaller share of total jobs.

Supplier Deliveries

Supplier deliveries measures whether vendor lead times are lengthening or shortening. As in manufacturing, slower deliveries (longer lead times) raise the index because they signal supply bottlenecks and demand pressure. During tight labor markets, elevated supplier deliveries in services can reflect staffing shortages and subcontractor delays rather than goods pipeline congestion. Read it alongside the prices index, not in isolation.

Prices

The prices subindex tracks input and selling price pressure reported by services firms. It is a useful real-time check on services-side inflation when CPI shelter and medical services dominate headlines. Prices above 60 often coincide with sticky core services CPI; sharp declines toward 50 can precede easing in PCE services inflation with a short lag. It is not a direct forecast of CPI shelter, which follows different methodology, but it captures breadth of firm-level pricing power.

Backlog, exports, imports, and inventory sentiment

Secondary series include backlog of orders (pipeline depth), new export orders and imports (global demand exposure), and inventory sentiment (whether respondents feel their inventory position is too high or too low). Inventory sentiment matters less for pure services firms than for retail and wholesale respondents inside the panel, but it can flag restocking or destocking ahead of retail sales surprises.

Reading the report: levels, momentum, and spreads

Three lenses outperform reacting to the headline alone:

• Level vs 50 — sustained stretches below 50 align with softer services spending growth; a single 49.6 print is weak recession signal on its own.
• Three-month momentum — the change in business activity and new orders over 90 days tracks PCE services better than any one month.
• Internal spreads — new orders minus business activity (forward vs current), employment minus business activity (hiring lag or lead), and prices minus supplier deliveries (margin pressure vs supply constraints) each tell distinct stories.

A late-cycle pattern in services: new orders fall while business activity holds as firms fulfill contracted work. An early-recovery pattern: new orders and business activity rebound together while employment lags because firms extend hours before adding headcount. Mapping these sequences onto the business cycle reduces whipsaw when manufacturing and services PMIs diverge — a common setup when goods demand weakens but consumer services remain resilient.

ISM Services vs ISM Manufacturing vs S&P Global

The manufacturing and services ISM surveys share diffusion methodology but sample different industries, release on different days (first vs third business day), and connect to different GDP components. Manufacturing PMI leads industrial production and goods orders; services PMI leads PCE services, retail control-group proxies, and aggregate employment. When manufacturing is below 50 and services above 50, the U.S. economy is often in a “two-speed” phase rather than outright recession — though sustained weakness in both with falling new orders is a stronger contraction signal.

S&P Global publishes a competing U.S. services PMI with a different panel and GDP-weighted sector methodology. ISM and S&P Global usually agree directionally but can diverge by 2–5 points in a given month. ISM overweight traditional industries and smaller firms; S&P Global weights large multinationals and uses different seasonal adjustment. When they diverge materially, Harbor checks whether the gap is in new orders or employment, whether leisure and hospitality drove the ISM move, and whether S&P's flash estimate was revised at final release. Convergence across both surveys plus hard spending data is stronger than either alone.

How Services PMI connects to GDP, jobs, and markets

Consumption and retail

Services consumption is the largest component of GDP. New orders and business activity correlate with future PCE services growth with roughly one- to two-month lead, while retail sales capture goods-heavy spending with different timing. PMI is timelier and broader; Census retail data is more granular but arrives later and revises. Pair services PMI with retail control-group trends when forecasting total consumption.

Payrolls and wages

The employment subindex leads aggregate payroll changes when services hiring dominates the month. Watch for divergence: strong business activity with weak employment can mean productivity gains without layoffs, while weak business activity with strong employment may reflect lagged cuts still ahead. Cross-check with ADP, jobless claims, and JOLTS where available.

Equities, rates, and the Fed

Equity futures react at the 10:00 a.m. release, but impact is often smaller than manufacturing PMI unless the print shifts Fed path expectations. Consumer discretionary, travel, and financials are most sensitive. Bond markets weight the prices subindex when core services inflation is the policy debate. A soft headline with sticky prices above 58 may do little to ease two-year yields; a strong headline with collapsing prices can support easing bets even when the economy looks hot on headline GDP.

Harbor Analytics release-day refactor (worked example)

Harbor's legacy playbook weighted manufacturing PMI at 60% and services at 40% in its nowcast — backwards for a services-heavy economy. Refactor steps:

1. Subindex dashboard — auto-plot business activity, new orders, employment, supplier deliveries, prices, and backlog with three-month trails.
2. Confirmation rule — require two of three signals (new orders below 50, new orders momentum negative, business activity below 50 for two months) before flagging demand contraction in the consumption forecast.
3. Hard-data cross-check — overlay last PCE services print, retail control group, and weekly card-spending proxies where available.
4. Spread alerts — notify when new orders minus business activity falls below a calibrated threshold for two consecutive months (forward demand weakening while current activity holds).
5. Prices filter — route inflation shocks from the prices subindex to the rates desk separately from volume alerts.
6. Manufacturing-services reconciliation — weekly note when the two ISM headlines diverge by more than 5 points, with sector anecdotes from ISM comment sections.

Outcome: Q3 2024 consumption forecast error fell from 0.8 percentage points to 0.3 after the desk stopped cutting services spending forecasts every time manufacturing PMI missed consensus. The services new orders three-month average had stayed above 52 while factory new orders weakened — a two-speed pattern the old weights blurred.

Technique decision table

Common pitfalls

• Treating PMI like a growth rate — 53 is not +3% services output; it is net expansion breadth.
• Ignoring services because manufacturing moved — goods weakness often coexists with services resilience.
• Overreacting to 0.1–0.3 consensus misses — survey noise dominates small deltas.
• Using prices as CPI shelter forecast — ISM prices breadth differs from BLS shelter methodology.
• Assuming employment equals instant hiring — firms adjust hours before headcount.
• Skipping the third-business-day calendar — confusing services release with manufacturing (first business day).
• Reading supplier deliveries like manufacturing — in services, delays often mean labor shortages, not port congestion.
• Chasing the futures spike — reversals are common once subindex tables are parsed.

Production checklist

• Calendar the ISM Services release (third business day, 10:00 a.m. ET) separately from manufacturing.
• Archive full subindex table, not just headline, in your data warehouse.
• Compute three-month moving averages for new orders, business activity, and employment.
• Track new orders minus business activity and employment minus business activity spreads.
• Cross-check major moves against prior-month PCE services and retail control group.
• Flag prices moves above 60 or below 48 for rates and procurement desks.
• Document sector anecdotes from ISM comments for leisure, healthcare, and finance.
• Reconcile services employment direction with next aggregate payroll print.
• Include PMI breadth (count of subindexes below 50) in weekly macro notes.
• Reconcile manufacturing vs services PMI divergence in the same release week.
• Compare ISM services to S&P Global final services PMI when available.
• Review false signals quarterly and recalibrate confirmation thresholds.

Key takeaways

• ISM Services PMI is a diffusion index on ~75% of U.S. GDP: 50 marks equal breadth of improvement and deterioration, not zero growth.
• New orders lead; business activity is concurrent; employment lags; prices inform services inflation breadth.
• Services and manufacturing PMIs often diverge — weight services more for consumption and jobs, manufacturing more for industrial cycles.
• Released on the third business day (not the first like manufacturing) at 10:00 a.m. ET.
• Harbor Analytics cut forecast error by requiring subindex confirmation instead of letting factory weakness override services strength.

Related reading

• ISM Manufacturing PMI explained — factory survey counterpart and release-day contrast
• Personal consumption expenditures (PCE) explained — hard-data anchor for services spending
• Retail sales explained — goods-heavy spending complement
• Business cycle explained — placing services turns in context