What is an aggregation in NACE statistics?

An aggregation groups NACE categories into a smaller number of headings for publication or analysis, defined in the metadata for the output, and it is not a NACE level. A section, division, group or class is a level in the classification. An aggregation is a grouping built from those levels, for a particular purpose, and its definition states which categories it includes. National Accounts groupings based on NACE, labelled A10, A21, A64 and A88, are noted by the Central Statistics Office for its publishing of data.

Eurostat notes that correspondence tables are one of the ways in which special aggregations of NACE, such as the Main Industrial Groupings, can be represented. An aggregation should always be cited by its name and its definition, not by a NACE code alone.

Why are aggregations used?

Aggregations are used because a full class level table, with 651 classes, is too detailed for many publications, and because some class level cells would be too small to publish reliably, or would disclose information about a single unit. An aggregation combines classes into a smaller number of headings that can be published, compared, and charted, while preserving the broad structure of the economy. The A10 grouping, for example, presents the economy in ten headings, and the A64 grouping presents it in sixty four, in the National Accounts presentation described by the Central Statistics Office. The choice of aggregation depends on the purpose of the output, the reliability of the detailed cells, and the need for comparability over time. A reader should therefore expect the same underlying data to appear under different aggregations in different outputs, and should not treat a difference in aggregation as a difference in the data.

How is an aggregation defined?

An aggregation is defined by a list of the categories it includes, at a stated level, in the metadata or methodology for the output. The definition should state the NACE version used, the level at which the categories are grouped, and any change in the definition at a classification changeover. Where a class split at the changeover, the definition must state whether the split parts stay inside the same aggregate or move to different aggregates, because that decides whether the aggregate total is comparable across the changeover. Where a class moved between divisions, the definition must state whether the movement crosses an aggregate boundary. A reader who compares an aggregate total across the changeover without checking the definition will compare two totals that may not be defined in the same way, and the comparison will be unreliable. The correspondence table shows the movement of the class, and the aggregate definition shows the consequence for the total. Both are needed.

How are Main Industrial Groupings defined?

Main Industrial Groupings, known as MIGs, are special aggregations used for industrial statistics, grouping activities by their main end use. Eurostat states that correspondence tables between Main Industrial Groupings and NACE Rev. 2.1 are published through ShowVoc, and that correspondence tables are one of the ways in which MIGs can be represented. A MIG is therefore defined by a correspondence, not by a NACE code, and it should be cited as a MIG, with the correspondence version stated. No MIG composition is asserted in this guide, because no MIG correspondence table was consulted in detail for it, and a composition from a secondary source would be unverified. A reader who needs the composition of a MIG should open the MIG to NACE correspondence table for the version in use, and should record the mapping for the categories in question, with the explanatory notes for those categories consulted alongside it.

Aggregation	What it is	How it is cited
A10, A21, A64, A88	National Accounts groupings based on NACE, as noted by the Central Statistics Office	By grouping name, definition, version and reference period
Main Industrial Groupings	Special aggregations for industrial statistics, represented by correspondence tables, as stated by Eurostat	By MIG name, correspondence version, and NACE version
Section or division total	Total for a NACE level, not an aggregation in the defined sense	By level, code, version and reference period

The table distinguishes a defined aggregation from a level total. A level total is a sum at a NACE level, and it is cited by its level and code. An aggregation is a defined grouping, and it is cited by its name and definition. The distinction matters, because a level total can be reproduced from the class table, while an aggregation can only be reproduced from its definition.

What does an aggregation not prove?

An aggregation does not prove anything about a single category inside it, or about a single unit inside a category. A section total that rises can hide a fall in one division, and an aggregate total that stays the same can hide a split that moved activity between its parts. An aggregation is a summary, and a summary should never be quoted as a statement about its parts. A specific business entry is confirmed in the register holding the record, and the explanatory note for a class confirms the meaning of that class inside an aggregate. The aggregate total is the context for those checks, not a substitute for them. To know what happened inside an aggregate, open the table at the next level down and check the correspondence for the classes that moved.

Where is an aggregation confirmed?

An aggregation is confirmed in the metadata and the correspondence table published by the institute that produced the output, for the version in use. The meaning of the categories inside it is confirmed in the Eurostat explanatory notes. The register holding the record confirms the entry for a specific business. With every aggregate figure, record the aggregate name, its definition, the version, and the reference period, and open the next level down before concluding anything about a part of the aggregate.