How is an old class traced to its new place?

An old class is traced to its new place in three steps: find the correspondence row for the old class, read the explanatory note for each new class the row points to, and confirm the entry with the national statistical institute or the business register that holds the record. No step can be skipped. The table shows the link. The note shows the scope. The institute or register shows the entry. A trace that stops at the table has found a category link, not a confirmed entry.

A trace that stops at the note has found a scope, not a link. A trace that starts from a title has found neither, because titles can stay the same across a restructure and can change across a continuation.

What is needed before the trace starts?

The trace needs the old code with its version, the old title as published, and the date or period of the source that cites the code. The version is essential. A bare four digit code, without a statement that it is a Rev. 2 code or a Rev. 2.1 code, is ambiguous after the revision, because the same digits can sit in different places and different digits can describe the same activities. The CSO in Ireland notes that correspondence tables are available through Eurostat’s ShowVoc platform, and that the tables define into which Rev. 2.1 class an activity classified in a particular Rev. 2 class goes. That definition is the starting point. A trace that begins with a code stripped of its version, or with a title copied from a listing that states no version, should pause and ask the publisher for the version before it proceeds.

Trace input	Why it is needed	Where it comes from
Old code with version stated	Prevents confusion between versions	Source dataset or register entry
Old title as published	Checks the row against the heading	Published structure for the old version
Period or date of the source	Places the code in the right version period	Dataset documentation or register entry
New code, if already shown	Cross checks the row in reverse	Correspondence table, reverse reading

The table above lists the inputs for a trace. It is a checklist, not a set of codes. The specific rows are found in the published correspondence table for the headings concerned.

How is the correspondence row read?

The row is read forward first, from the old class to the new class or classes, counting the destinations. One destination means continuation, rename or merge into a broader heading. Several destinations mean a split, and each destination note must be read. No destination, because the old class has no row, means no activity moved, and the heading continues under its new identifier or title. The row is then read in reverse, from each new class back to its old sources, to see whether the new class draws on other old headings as well. A one to one row that looks simple in the forward direction can hide a mixed origin in the reverse direction. Eurostat states that the correspondence table describes, for each Rev. 2 heading, how activities currently classified therein should be classified in Rev. 2.1. That description is written for headings. It is not a 1:1 guarantee for any specific unit where the table says one to many, many to one, or many to many.

How are the notes used in the trace?

The notes are used to test the row. For each destination, the reader checks the inclusions and exclusions in the explanatory note for the new class, and compares them with the note for the old class. If the activity described in the source sits inside the inclusions of one destination and inside the exclusions of the others, the row and the notes together point to that destination in general terms. If the activity sits inside the inclusions of more than one destination, or inside none of them, the trace is incomplete, and neighbouring groups and divisions must be searched. The manual published by Eurostat, product code KS-GQ-24-007, brings the introductory guidelines and the explanatory notes together, and the guidelines state the general rules for classification. The notes state the content of each category. Both are needed. The guidelines alone cannot place an activity. The note alone cannot show the movement.

Where does the trace end?

The trace ends with the institute or register that holds the record. The meaning of a category is in the Eurostat explanatory notes. The movement between versions is in the correspondence tables. The entry for a specific business is held by the national statistical institute or the business register of the country concerned. This site points to those sources. It holds no register and decides no entry. A code does not decide tax, social security, licensing, funding or collective agreement matters. A completed trace therefore ends with a confirmed entry, stated with its version, or with a clear statement that the entry could not be confirmed from the published sources and must be asked of the institute or register. An unconfirmed trace is a useful result. It prevents a wrong entry from being copied forward as if it were settled.

Start with the version stated beside the old code, follow the row forward and back, test every destination in its note, and let the institute or register confirm the entry that the trace points to.