How does the top down method work?

The top down method identifies a principal activity in stages, from the highest level of the classification down to the class. Paragraph 87 of the introductory guidelines states the hierarchical principle: the classification at the lowest level must be consistent with the classification at the higher levels. The method is statistical only, described here in the abstract, and it does not determine a code for a specific reader.

What are the four steps?

Paragraph 87 lists four steps. First, identify the section which has the highest share of value added. Second, within that section, identify the division which has the highest share of value added. Third, within that division, identify the group which has the highest share of value added. Fourth, within that group, identify the class which has the highest share of value added. Each step narrows the field to the categories inside the category chosen at the step above. A class in a different section, however large its individual share, cannot be chosen if its section does not win the first step.

The hierarchical principle is the point of the method. Without it, a unit could be assigned to the single largest class wherever it sits, and the resulting section, division and group totals would not match the class totals. With the method, the class chosen is the largest class inside the largest group inside the largest division inside the largest section, so every level of the hierarchy tells the same story about the unit. The Eurostat glossary states the same requirement: the classification at each level must be compatible with the previous level.

What does the manual worked example show?

Paragraph 88 gives a worked example with value added shares for eight classes. Section C, manufacturing, holds 52 percent of value added. Section G, wholesale and retail trade, holds 35 percent. Section N, professional, scientific and technical activities, holds 13 percent. Within section C, division 28, manufacture of machinery and equipment not elsewhere classified, holds 42 percent against 10 percent for division 25. Within division 28, group 28.9, manufacture of other special purpose machinery, holds 31 percent. Within group 28.9, class 28.93, manufacture of machinery for food, beverage and tobacco processing, holds 23 percent, against 8 percent for class 28.95.

The result follows the steps, not the largest single class. The largest single class in the example is class 46.61, wholesale of agricultural machinery, equipment and supplies, with 28 percent, which is larger than the 23 percent for class 28.93. The manual nevertheless concludes that the correct class for the principal activity is 28.93, because section G as a whole holds less value added than section C, and the method requires consistency down the hierarchy. The example is the clearest statement in the manual of why the top down method exists: it prevents a single large class in a smaller section from overriding the section level result.

How are the example shares laid out?

| Level | Winning category in the example | Share of value added |

|---|---|

| Section | C, manufacturing | 52 percent |

| Division | 28, manufacture of machinery and equipment not elsewhere classified | 42 percent |

| Group | 28.9, manufacture of other special purpose machinery | 31 percent |

| Class | 28.93, manufacture of machinery for food, beverage and tobacco processing | 23 percent |

The table follows Table 1 in paragraph 88. The shares at each level are the sums of the shares of the categories inside the winning category, not the share of the winning class alone. Section C at 52 percent is the sum of its five classes in the example. Division 28 at 42 percent is the sum of its five classes. The table is a reproduction of the manual example, with the percentages as the manual prints them. It is an illustration of the method, not a finding about any real unit, and no reader should apply the percentages to a real case.

What does this classification not decide?

No part of the method described in this guide picks a code for a specific business. The guidelines describe how statistical units are classified in principle, using value added and the steps set out below. A particular unit application needs its own figures and the judgement of the statistical office that holds the register entry. The guide states the method in the abstract, as the task for this cluster requires. A real business example is not worked through, and applying the steps to a real business without the institute is not invited. That restraint is deliberate, and it is part of the method’s own presentation in the manual.

Where is the entry confirmed?

Confirmation follows the same sequence in every guide in this cluster. The introductory guidelines give the rule, the explanatory note gives the category, and the national statistical institute or business register holding the record confirms the entry for the unit concerned.

Work down the hierarchy in the order paragraph 87 sets out, keep each level consistent with the level above, and confirm the entry with the institute or register that holds it.