How are aggregated codes read in tables?

Aggregated codes show totals for a section, a division, a group, or a named grouping, and they hide the movement inside the aggregate. A table that shows a section total shows the sum of all the divisions inside that section, and a table that shows a division total shows the sum of all the groups inside that division. A change in the aggregate can therefore reflect a change in one part of the aggregate, a change in the classification of a part, or a change in the definition of the aggregate itself.

A reader who compares aggregates across the changeover without checking the composition of the aggregate will mistake a reclassification inside it for a change in the total, or will miss a reclassification that leaves the total unchanged.

What is an aggregation?

An aggregation is a grouping of categories for publication or analysis, defined in the metadata, that is broader than a single category. The Central Statistics Office notes particular NACE based groupings, labelled A10, A21, A64 and A88, used by National Accounts when publishing data. Those groupings are aggregations: they combine NACE categories into a smaller number of headings for presentation, and the definition of each grouping is part of the metadata for the output. Eurostat notes, on its correspondence tables page, 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 is therefore a defined grouping, not a NACE level, and it should be cited by its name and its definition, not by a NACE code alone. A table that shows A10, for example, shows the A10 aggregation, not a NACE section, and a reader who treats it as a section will compare it with the wrong total.

How is an aggregate table checked?

An aggregate table is checked by opening the definition of the aggregate, the version used, and the reference period, in the metadata. The definition states which categories are included in the aggregate, and whether the definition changed at the changeover. The version states whether the categories inside the aggregate are on NACE Rev. 2 or NACE Rev. 2.1. The reference period states which period the total describes. A reader should record all three, together with the aggregate name, before comparing the total with another total. Where a class split, the aggregate may stay the same while the classes inside it move, because the split stays inside the aggregate. Where a class moved between aggregates, the aggregate totals will change in opposite directions, even though the underlying activity did not change. The correspondence table shows the movement of the class, and the aggregate definition shows whether the movement crosses an aggregate boundary. Both checks are needed, and neither alone is sufficient.

How are Main Industrial Groupings read?

Main Industrial Groupings, known as MIGs, are special aggregations of NACE used for industrial statistics, grouping manufacturing and related activities by their main end use, such as intermediate goods, capital goods, durable consumer goods, and non durable consumer goods, in the general sense in which Eurostat presents them. Eurostat states that correspondence tables are one of the ways in which MIGs can be represented, and that correspondence tables between MIGs and NACE Rev. 2.1 are published through ShowVoc. A MIG total is therefore an aggregation defined by a correspondence, not a NACE category, and it should be cited as a MIG, with the correspondence version stated. A reader who compares a MIG total across the changeover should check the MIG to NACE correspondence for the version in use, and should check whether any class moved into or out of the MIG at the changeover. No MIG category title or 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. The definition should be read in the correspondence table for the version in use.

Element	Check
Aggregate name	A10, A21, A64, A88, MIG, or other named grouping, as stated
Definition	Which categories are included, as stated in the metadata
Version	NACE Rev. 2 or NACE Rev. 2.1, as stated for the categories inside
Reference period	Period the total describes
Movement inside	Correspondence table, for classes that split or moved

The table is a checklist for an aggregate citation. A citation that names only the aggregate and the figure, without the definition and the version, cannot be repeated, because the same aggregate name can have a different definition in a different version, and the figure will differ accordingly.

What does an aggregate not prove?

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

Where is an aggregate confirmed?

An aggregate 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. A specific business entry is confirmed by the register holding the record. Every aggregate figure should carry the aggregate name, definition, version and reference period, and the next level down should be opened before any conclusion about a part of the aggregate is drawn.