The purchasing managers index asks the people placing orders whether conditions are better or worse than last month. They commit before the output those orders produce exists, which gets an answer weeks or months ahead of the hard data.

What comes back is not a measured quantity. It is the balance of opinion across orders, output, employment and inventories, which is a very different kind of number from a count of anything.

What it actually measures

The midpoint is built to be the boundary. It is the level at which as many respondents reported improvement as reported deterioration, so above it activity is spreading and below it activity is shrinking.

Breadth is what the survey captures. A reading well above the line means improvement is widespread rather than that the improvement is large, and those are genuinely different facts about an economy.

It leads the hard data by months. Orders placed today become output some time later, which is what buys the survey its head start over anything measured after the fact. That head start is the entire reason to accept a number built out of opinion.

Breadth and magnitude are different questions

A diffusion index counts how many respondents improved, never by how much. An economy where every firm improved slightly and one where half the firms improved enormously can produce very different readings from very similar outcomes.

This is worth holding onto when the survey and the hard data disagree. The survey is not wrong in those moments; it is answering a question about how widely conditions are shared, which is not the question output figures answer.

What it does not tell you

It is opinion, collected from a sample. Response rates vary, the panel changes over time, and sentiment can drift with headlines rather than with the orders actually being placed.

Manufacturing is a small share of a service economy. The most-watched version of this survey covers a sector that employs a fraction of the workforce, which is why it can be in contraction while the wider economy grows.

It has produced false alarms. Dips below the line have been followed by recovery without a downturn arriving, and treating the crossing as an event has cost people money. The crossing is worth noting and it settles nothing on its own.

How to read it

Strong expansion. Activity is growing firmly across respondents.

Expansion. Growth, if a modest sort.

Contraction. More managers report worsening conditions than improving ones.

Deep contraction. A pronounced downturn in what is being reported.

You will find PMI Business Cycle on the Macro dashboard, next to GDP Growth, Labor Market and Yield Curve.

Common questions

What does the survey ask?

Whether things improved, held steady or deteriorated since last month, across new orders, production, hiring and stock levels. What comes back is a tally of opinions rather than a count of anything.

Why is the midpoint the line that matters?

Because it is constructed to sit exactly where improvers and decliners balance out. Anything above it means gains outnumber losses, and anything below it means the reverse.

Why is a survey useful at all?

Because the people answering it have already committed money before anything gets made, so what they say arrives months ahead of any measured figure. Speed is bought at the cost of precision.

What does the reading not capture?

Size. It counts how many respondents improved and never by how much, so widespread small gains and concentrated large ones can look identical.

Does crossing the line signal recession?

Not reliably. Dips below it have been followed by recovery with no downturn at all, so the crossing is worth noticing and settles nothing by itself. It is one reading among several.