Euro area PMI and the flash release sequence

One methodology across many countries 

The euro area purchasing managers' indices are produced by a single commercial data provider rather than by any statistical agency, and that is the source of their principal analytical advantage. Because one organisation runs the survey in every country to the same questionnaire, the same response scale and the same weighting scheme, a reading for one member state is directly comparable with a reading for another in a way that national official statistics, compiled by separate offices, generally are not.

The same fact is also the principal caveat. These are not official statistics. The panel composition, the sampling frame and the seasonal adjustment procedure are the provider's commercial property, subscribers receive the detail before the headline reaches the public wires, and there is no statutory obligation to publish a methodology change. The indices are used as if they were official indicators and they are governed as a commercial product.

What is published, and at what level 

The family has three layers. At the bottom are sector indices: manufacturing, which covers goods producers, and services, which covers the far larger part of the economy. Above them is the composite, a weighted combination of the manufacturing output component and the services activity component, which is the aggregate most commonly quoted because it is the one that spans the whole private sector. Above that again are the geographic aggregations: national indices for the individual member states and one euro area index.

An important detail is often missed. The composite does not use the headline manufacturing index. It uses the manufacturing output sub-index, because the headline manufacturing figure is itself a weighted blend of five components including supplier delivery times and inventories, which do not belong in a measure of activity. A reader who tries to reconstruct the composite from the two headline numbers will not reproduce it.

The flash morning, in order 

The flash release is unusual among economic indicators because it is a sequence within a single morning rather than a single moment. On flash day, which falls in the final week of the reference month, the national flash readings for the largest member states are published in succession, each roughly half an hour apart, and the euro area aggregate follows them. The United Kingdom and United States flash readings are published later the same morning, so a single session carries an ordered series of surveys covering most of the developed world.

Because the aggregate is built from national data that has already been published in part, each national reading is conventionally treated as partial information about the aggregate that follows it, weighted by the country's size. The aggregate is therefore the least surprising number of the morning in a mechanical sense, and the smaller member states, whose national indices are not released separately in the flash, are the residual that the aggregate reveals.

The flash is based on the majority of responses received by the cut-off date, and the final reading, published in the first days of the following month, incorporates the remainder along with the full sub-component detail. Revisions between flash and final are usually small, and they are not always small.

The components that carry the information 

• New orders and new export orders. The euro area is an unusually open economy, so the gap between total new orders and export orders is one of the few timely readings available on whether demand weakness is domestic or external.
• Employment. A breadth measure of hiring intentions across the panel, published for each sector and for the composite.
• Input prices and output prices. Two separate series, and the distance between them is conventionally read as an indication of margin pressure, since it describes how much of a cost change respondents reported passing on.
• Backlogs of work and suppliers' delivery times. Read together as capacity indicators, with the delivery times component inverted so that slower deliveries raise the index.
• Future output expectations. A twelve-month-ahead question, which is the only genuinely forward looking item in a survey that otherwise asks about the month just past.

The seasonal adjustment applied to these series deserves a note of its own, because it is applied by the provider rather than by a statistical office and is revised as the history lengthens. A survey question that asks whether this month was better than last month already removes some seasonality by construction, since a respondent comparing consecutive months carries their own sense of what is normal for the time of year. Adjusting the resulting index a second time is standard practice and it is not obviously neutral, and the provider publishes both adjusted and unadjusted series so the difference can be inspected.

How the release is conventionally read 

The composite is read first, against the neutral level and against its own recent trajectory, since a diffusion index describes direction rather than magnitude. The divergence between manufacturing and services is read second, because the euro area has spent long stretches with the two sectors on opposite sides of neutral. The national divergence is read third, since one large member state in contraction and another in expansion is a materially different situation from both being at the same level.

A fourth convention maps the composite onto an implied quarterly growth rate. The provider publishes such a correspondence and central bank staff use survey data in short-term forecasting models. The mapping is a regression fitted to history rather than an identity, and its accuracy has varied considerably across periods, particularly during episodes when the relationship between reported sentiment and measured output broke down.