27.29.    Correspondence tables. The correspondence tables (also known as correlation tables) define the scope of headings of one classification in terms of the scope of headings of another classification. There can be a strict one-to-one correspondence between headings, when the scope of a heading of one classification is equal to the scope of a heading of another classification, or various kinds of split correspondence. A split correspondence exists when the scope of a heading of one classification partially overlaps with the scope of several headings of another classification. Whenever successive versions of the same classification are produced, a correspondence table between the headings of the revised and original versions is issued.  A reverse table, showing the correspondence between headings of the original and revised versions, is also frequently produced. 

27.30.    Conversion tables. The correspondence tables allow a continuous time series to be maintained when various versions of a given classification are used data are to be expressed in terms of another classification. However, if the scope of a heading of one classification partially overlaps with the scope of several headings of another, an exact correspondence becomes impossible and there is a discontinuity in some data series.  For data-processing purposes it is frequently desirable to substitute an approximate, but one-to-one correspondence for a split correspondence. When this is done, the resulting table is called a conversion table. It should be noted that if there are no split correspondences, the data conversion can be carried out using the correspondence table. Examples of such a straightforward conversion are the conversion of HS88 data into SITC, Rev.3, data and the conversion of HS07 data into SITC, Rev.4, data, as those SITC versions used the respective HS version’s subheadings as building blocks. In contrast, the conversion of HS88 and HS96 data into SITC, Rev.4, data would require approximations, as there exist a number of split correspondences between those versions of HS and SITC, Rev.4.  

27.31.    The use of conversion tables containing such approximations is warranted if the scope of the headings involved is quite similar. However, differences in scope between certain basic headings may be so great, as to eliminate the possibility of a meaningful one-to-one correspondence at that level.  In such a case, a correspondence can be established only between basic headings of one version and the higher-level headings of the other. The responsible agency should study which approach should be adopted in order to achieve a reasonable balance between the requirement of data series continuity and data comparability. 

27.32.    Uses of the correlations and conversions tables.  The main uses of correlation and conversion tables in trade statistics include:

(a)    Maintenance of comparable data series when the classification used in compilation is revised;

(b)   Reconciliation of data obtained from various sources (and expressed in different classifications);

(c)    Recompilation of trade data for another purpose (e.g., to analyse trade in terms of broad categories of goods or by various economic activities). 

27.33.    Informing users. To make users aware of the methodology adopted for data conversion, it is good practice to ensure that all correlation and conversion tables are documented and made publicly available as part of the metadata. It is advised that the agency responsible alerts users regarding the data conversions that might diminish data comparability, so they can make their own assessment of whether such data are suitable for their purposes. 

27.34.    Correlation tables provided by the World Customs Organization (WCO). WCO produced the first version of the HS in 1988 (HS88). In 1986, it had issued a publication entitled “Correlation Tables between the Harmonized System and the 1978 version of the CCCN”, to link HS with the Customs Cooperation Council Nomenclature (CCCN).[16]  This was a two-way correlation, that is, from HS to CCCN, and from CCCN to HS. When HS is revised, the WCO issues correlation tables between the new and preceding versions of HS and makes them publicly available.[17] 

27.35.    Correlation and conversion tables available from UNSD. UNSD has created correlation and conversion tables between various versions of the HS, SITC and BEC, so that it could maintain its time-series data on trade.[18] UNSD also maintains correlations and conversions between HS, CPC and ISIC.[19] If compilers need to convert their data from one classification to another, it is good practice for them to take into considerations correlation and conversion tables available from international and/or regional organizations. 

27.36.    Correspondence tables with non-HS national commodity classifications.  If a country compiles data in terms of a non-HS classification (provided that said classification is highly detailed, with criteria similar to the ones applied in HS), compilers are advised to develop a correspondence table between the non-HS classification and HS, and to make it available to interested users.


[16] This text was initially known as the Brussels (Tariff) Nomenclature (BTN), but in 1976 it was renamed the Customs Co-operation Council Nomenclature (CCCN), to prevent any confusion regarding the international organization responsible.

[17] See http://www.wcoomd.org/en/topics/nomenclature/instrument-and-tools.aspx.

[18] UNSD converts data based on the most detailed level of the applicable classifications.

[19] See http://unstats.un.org/unsd/cr/registry/regot.asp.