What the ISM Surveys Say About Business Activity

The ISM manufacturing and services surveys provide timely evidence about business activity before many comprehensive economic statistics are available. Their usefulness comes from asking supply-management professionals whether conditions improved, deteriorated or remained unchanged. Their limits come from the same design: they measure the breadth of reported change, not the precise size of economic growth.

A PMI of 52 does not mean production increased 2%, and a move from 48 to 49 does not by itself mean the sector returned to growth. Understanding the calculation makes both the headline and its component readings more informative.

How a diffusion index works

For a conventional positively oriented component, the unadjusted diffusion index adds the share reporting improvement to half the share reporting no change. If 40% of hypothetical respondents report an increase, 20% report no change and 40% report a decrease, the result is 50: 40 plus half of 20.

Now consider 50% reporting an increase, 20% no change and 30% a decrease. The index becomes 60. That is a broader balance of improvement, not evidence that each company expanded 10% or that national output grew by a particular amount. Company size and the magnitude of individual changes cannot be recovered from that number alone.

Seasonal adjustment can alter the published series relative to the raw response calculation. ISM publishes information on its seasonal factors and methodology. The basic example explains what a diffusion index represents, rather than reconstructing a particular month's adjusted headline.

The manufacturing and services composites differ

The manufacturing PMI combines new orders, production, employment, supplier deliveries and inventories with equal weights. The services PMI uses business activity, new orders, employment and supplier deliveries, also with equal weights. A component's role therefore depends on which composite is being discussed.

An index above 50 generally indicates expansion in the relevant surveyed sector compared with the preceding month; below 50 indicates contraction. It is not a universal threshold for the entire economy. Manufacturing can contract while services expand, and neither survey should be casually substituted for measured real GDP.

Changes in the index level describe changes in the reported balance. A move from 54 to 52 remains on the expansion side but indicates less broad improvement. A move from 46 to 49 remains on the contraction side while showing that the reported decline has become less widespread. Direction and level should both appear in the interpretation.

Supplier deliveries have an unusual sign

The supplier-deliveries index rises when more respondents report slower deliveries. That can accompany strong demand, but it can also reflect shortages or transport disruptions. It is therefore possible for a supply problem to lift the composite even when other measures are weak.

Take a simplified hypothetical manufacturing reading with four components at 45 and supplier deliveries at 70. The average is 50: the sum of 45, 45, 45, 45 and 70 divided by five. Reporting that result as unqualified evidence of healthy demand would miss how the high deliveries component offset the others.

Conversely, faster deliveries can reduce that index while representing an operational improvement for buyers. The sign is a feature of the survey's historical construction, not an instruction to consider every delay economically desirable. Read the report's explanation and other components before deciding what caused the movement.

New orders, employment and prices answer separate questions

New orders concern incoming demand, while production or business activity concerns current operations. Employment addresses reported staffing change. They can diverge when firms complete earlier orders, manage backlogs or adjust productivity. One composite number does not eliminate those differences.

The prices index is also a diffusion measure. A reading above 50 indicates a balance toward higher reported input prices; it does not specify a national inflation rate or the average percentage increase paid by buyers. The Producer Price Index uses a different statistical framework to measure price changes.

Similarly, a survey employment component cannot be read as the number of payroll jobs created. A small firm and a large firm reporting the same direction can have very different effects on total employment. Monthly payroll estimates and jobless claims provide separate evidence about the labor market.

Build a reading from the components

The strongest interpretation explains which components drove the headline, whether demand and output agree, and whether delivery times reflect demand or disruption. It also compares several releases rather than assigning a durable trend to one observation.

Survey responses offer a timely account of conditions reported by participating businesses. Later official statistics may confirm, qualify or contradict that account because they measure different things. The PMI is valuable as early evidence when its construction remains visible, rather than as a shortcut to an exact GDP or employment forecast.