What is a correspondence table?

A correspondence table is a table that shows the relationship between the positions of two classifications, or between two versions of the same classification, so that statistics collected under one can be compared with statistics collected under the other. 2 and NACE Rev. 2.1, and vice versa, are available online. A correspondence table does not assign a code to a business. It records how categories relate, so that a comparison can be made and a move can be traced.

What are the two types?

Eurostat states that two types of correspondence table exist. Correspondence tables between different versions of the same classification describe the detailed changes that have taken place in the revision process. Correspondence tables between two different classifications show the relationship between all positions of the two classifications involved. The manual repeats the two types in the same terms. The first type is the type used for the Rev. 2 to Rev. 2.1 change, because the two versions are versions of the same classification. The second type is used for relationships between NACE and other classifications, such as the International Standard Industrial Classification of All Economic Activities, the Classification of Products by Activity, and other related classifications.

The distinction matters for reading. A table between versions describes change: which categories were created, removed, split, merged, or restructured, and how the scope of each category relates to the scope of the categories in the other version. A table between different classifications describes correspondence: which positions in one classification correspond to which positions in the other, where the classifications have different structures, different levels of detail, and different purposes. A version table with a one to many row means that an old scope was divided. A table between classifications with a one to many row means that a position in one classification corresponds to several positions in the other, because the other classification is more detailed at that point. The same row form has a different meaning in the two types, and the type of the table must be stated before a row is read.

Type stated by Eurostat	What it compares	Example named in the sources
Different versions of the same classification	NACE Rev. 2 and NACE Rev. 2.1	Correspondence tables between NACE Rev. 2 and NACE Rev. 2.1
Two different classifications	NACE and another classification	Correspondence tables between Main Industrial Groupings and NACE Rev. 2.1

Where are the NACE correspondence tables published?

The Eurostat correspondence tables page lists two sets of tables: correspondence tables between NACE Rev. 2 and NACE Rev. 2.1, and correspondence tables between Main Industrial Groupings and NACE Rev. 2.1. The links on that page point to ShowVoc, the web based semantic platform described in the manual as enabling visualisation of the detailed structures of statistical classifications and their correspondence tables. The manual states that ShowVoc covers the structure, the explanatory notes including index entries whenever defined, correspondence tables between classifications, links to legal acts, and methodological documents. The Central Statistics Office in Ireland states that correspondence tables showing the relationship between NACE Rev. 2 and NACE Rev. 2.1 are available on ShowVoc, and that Eurostat provided detailed documentation including correspondence tables defining into which NACE Rev. 2.1 class an activity classified in a particular NACE Rev. 2 class goes.

The manual also states that correspondence tables are one of the ways in which special aggregations of NACE, such as the Main Industrial Groupings, can be represented. A special aggregation groups NACE categories into a different set of categories for a particular analytical purpose, and the correspondence table shows which NACE categories fall in each aggregation category. That use is distinct from the version comparison use, and the table for an aggregation must be read as a grouping, not as a record of change. The aggregation table shows membership. The version table shows movement.

How is a correspondence table used for comparison over time?

That limit is the reason the table must be read with its type. A one to one row can be converted directly, because the scope is identical in both versions. A one to many row cannot be converted directly from the old total alone, because the old total must be divided between the new classes, and the division requires additional information. A many to one row can be converted by adding the old totals, because the joint scope is the new scope. A many to many row requires the joint scope to be treated as a whole, because the individual old and new scopes do not correspond one by one. The type of the row therefore decides what conversion is possible, and a reader who ignores the type will produce a converted series that the table does not support.

What does a correspondence table not do?

A correspondence table does not determine a code for a reader, a business, or an activity. It shows the relationship between categories, as stated in the sources, and it must be read with the explanatory notes, which state what each category includes and excludes. A row that shows an old class corresponding to several new classes does not pick the new class for a particular activity. The new class must be picked by reading the notes for each new class, and the entry must be confirmed with the institute or register that holds the record. A table that is used as if it were a finder tool, by following a row to a single new code, misreads the table where the row is one to many or many to many, and it misreads the table even where the row is one to one, because the table states a category relationship, not an assignment for a unit.

Find the table type on the Eurostat correspondence tables page, read the row type for the class, and confirm any entry with the national statistical institute or the business register that holds the record.