04 / 20 How to read a correspondence table
What are the four correspondence types?
Published Updated
Statistical classification only
Statistical classification only. This guide does not determine any reader's code. Confirm a code with the national statistical institute or the business register that holds the record. This is general information, not advice on any reader's registration.
The correspondences between NACE Rev. 2 and NACE Rev. The manual gives a count for each type, counted in NACE Rev. 2.1, 11 classes split into two or more classes in NACE Rev. 2.1, and 225 classes in a many to many correspondence. The counts sum to 615, the number of classes in NACE Rev. 2, which checks the table as a complete record for the old version. A row must be read with its type stated, because the type decides what the row shows and what it does not show.
What does each type mean?
The manual wording for the many to many type is precise, and it must be read in full. It does not state that each old class in the group corresponds to each new class in the group. It states that the joint scopes are identical. The group as a whole covers the same activities in both versions, while the division of the group into individual categories differs. A reader who picks one old class and one new class from a many to many group, and treats them as corresponding, will misread the table. The group must be treated as a whole, and the individual classes inside it must be compared through their notes, not through the table row alone.
| Type in the manual | Count in NACE Rev. 2 classes | Meaning stated in the manual |
|---|---|---|
| 1 to 1 | 354 | One old class corresponds exactly to one new class, and vice versa |
| n to 1 | 25 | Old classes merged into one new class |
| 1 to m | 11 | Old class split into two or more new classes |
| n to m | 225 | Joint scope of old categories identical to joint scope of new categories |
| Total | 615 | All classes in NACE Rev. 2 |
Why is the many to many group the largest after one to one?
The many to many group is large because many changes in the update were partial transfers and restructurings, rather than clean splits or clean mergers. The manual defines minor changes as including partial transfers of content between classes, and it defines major changes as including restructurings within groups. A restructuring that moves part of the content of several classes into several new classes, while keeping the total scope of the group the same, produces a many to many relationship. The group scope is preserved, while the internal division changes. That pattern is common where a group is reorganised around a new distinction, such as a distinction by source, by channel, or by a new concept, without adding or removing scope at group level.
The size of the many to many group also explains why the net count change is small while the detailed change is large. The net increase of 36 classes is the result of creations minus removals across all types. The many to many group alone involves 225 old classes, more than a third of the classification, whose internal division changed even though the joint scope stayed the same. A reader who looks only at the net count, or only at the split and merger counts, will underestimate the scale of the restructuring. The four type counts together give the scale: 354 classes unchanged in scope, and 261 classes involved in a merger, a split, or a many to many restructuring.
How is the type checked for one class?
The type for one class is checked in the correspondence table itself, in the row or rows for that class. The manual passage gives the counts and the definitions, and it cites the table version, Version 1.05, as the source for the types. A reader must therefore check the version of the table consulted, because a later version of the table could revise a row, and the counts in the manual passage belong to the version cited. The table on ShowVoc, linked from the Eurostat correspondence tables page, is the place where the current table is published, and the version label on the table consulted must be noted beside any type quoted from it.
The type must also be checked in both directions where the table provides both directions. The manual states that complete detailed correspondences between NACE Rev. 2 and NACE Rev. 2.1, and vice versa, are available online. A one to one correspondence is symmetric, by definition. A many to one correspondence in one direction is a one to many correspondence in the other direction. A many to many correspondence is symmetric as a group, but not as individual rows. A reader who checks only one direction may therefore misstate the type, by quoting the reverse direction type for the direction needed. The direction, old to new or new to old, must be stated beside the type.
What does a type not decide?
A type does not decide the new class for a specific activity, and it does not decide an entry. A one to one type states that the scope is identical, but the note for the class must still be read to confirm the scope, and the entry must still be confirmed with the institute or register. A one to many type states that the old scope was divided, but it does not state which new class covers a particular activity. The notes for the new classes state that division. A many to one type states that old scopes were combined, but it does not state that every activity in the old classes has the same treatment in the new class for every purpose. The note for the new class states the combined scope. A many to many type states that joint scopes are identical, but it does not state an individual correspondence, and it must not be used as if it did.
Check the table version, state the direction, read the type for the class, and confirm any entry with the national statistical institute or the business register that holds the record.