Special treatments adopted in compiling I‑O tables

22.46    The symmetric I-O tables are sourced from the S-U tables, and the concepts and definitions used are the same as elsewhere in the ASNA. Issues of particular importance to the I-O tables include statistical units and the distinction between primary and secondary production.

22.47    The content and meaning of the I-O tables produced depend on some particular aspects of compilation including:

• the treatment of intra-industry transactions; and
• the allocation of imports.

22.48    Depending on the treatment of intra-industry transactions, the output of an industry can be defined in three different ways according to whether, and to what extent, these transactions are counted as part of the output.

22.49    Firstly, the output of an industry can be defined as the total value of all flows of products produced by the units classified to the industry. All intra-industry flows are included as output when it is defined in this way. Under this definition, for example, the output of the Motor vehicles and parts; Other transport equipment industry consists not only of fully assembled vehicles but also of motor bodies, engines and other components despatched from (or added to inventories by) any unit recognised as a unit for statistical purposes. This definition of output disregards the fact that many of these components will have been incorporated in finished motor vehicles and will have therefore been counted twice. Output calculated according to this definition could be as much as two or three times the value of finished products of the industry.

22.50    A second definition of the output of an industry confines output to products produced by units within the industry and sold outside the enterprise. This definition also results in some duplication because the components manufactured and sold by one enterprise are all counted separately, although they may have been used in a finished product of another enterprise in the same industry and counted again in the value of this product. Moreover, the components despatched from one unit could be omitted entirely or counted either partly or wholly depending on whether they were used by another unit of the same enterprise or by a different enterprise.

22.51    Thirdly, the output of an industry can be defined as net of all intra-industry transactions; that is, excluding not only the transfers between the unit in industry belonging to the same enterprise, but also all flows between units in industry belonging to different enterprises.

22.52    If the third definition of output is used, the I-O table is said to be net and the main diagonal of an industry-by-industry table is empty. If either the first or second definition of output is used the I-O table is said to be gross and there are entries on the main diagonal.

22.53    For 1974–75 and subsequent years, the ABS I-O tables generally include intra-industry flows and can be described as 'gross', as outlined above. This means that the estimates of output can be directly compared with other information about an industry.

22.54    A further consequence of recording intra-industry transactions is that the level of output is unaffected by the number of industries used (i.e. by different levels of industry aggregation). An important exception is the construction industry, in which output was measured on a net basis prior to the 2001-02 tables.

22.55    Information regarding the use of imports in the economy is not generally available because it is impractical to collect data on how imported products are used. For analytical purposes, the ABS models the use of imports in the intermediate and final use categories using a number of assumptions. In an indirect allocation of imports approach, imports are not distinguished from domestically produced products and their use is therefore based on their contribution to the total supply. Specific rules also determine the disposition of imports which, by definition cannot be allocated to domestic exports but must be allocated to re-exports.

22.56    Various ways are available to record imports in Input-Output tables. The main ones are:

• direct allocation of imports – involves allocating all imports directly to the industries which use them. In this case, all flows recorded in quadrants 1 and 2 refer only to the use of domestic products, and consequently quadrant 1 does not reflect the technological input structure of the industry;
• indirect allocation of imports – involves first recording all imports as adding to the supply of the industry to which they are primary and then allocating this supply along the corresponding row of the table to using industries. The result is that flows in quadrants 1 and 2 contain imported and domestically produced products without distinction. Quadrant 1 then better reflects the technological input structure of the industry and quadrant 2 better reflects the product composition of final demand; and
• direct allocation of complementary imports and indirect allocation of competing imports – this method involves first distinguishing between complementary and competing imports and then allocating the first group directly and the latter indirectly. Complementary imports are defined as those for which no suitable substitute is produced domestically, but determining what is a suitable substitute is largely a matter of judgement. As complementary imports ceased to be separately distinguished from the 2001-02 tables onwards this method is not available in the published ABS I-O tables.

22.57    Each of these methods has advantages from an analytical point-of-view but each also can lead to conceptual and compilation problems.

22.58    Direct allocation of imports is appropriate for many analytical purposes. However, it would be necessary to adjust the imports table and to recalculate the industry-by-industry tables if substitution between imports and domestic production is known to occur, in order to allow for the probable effects of specified import replacement or substitution. In addition, the application of this method requires identification of the destination of each imported product. Although the proportion of imports in total supply (and therefore in total usage) for each product can be established, it may not be known for individual using industries. Of course, it is possible to proceed if one assumes that each using industry draws on imports and domestic production in the average proportions established for the total supply of each product. In the I-O publication, tables with direct allocation of competing imports have been prepared using this assumption. The assumption was applied to detailed working tables which were subsequently aggregated for publication.

22.59    Indirect allocation of imports is appropriate, in the sense that it will result in stable input-output coefficients, where the inputs to the domestic industry to which each imported product is primary are representative of the inputs required to produce the import domestically. Where this is not so, the method will give misleading results. For instance, if coffee (which could be treated as a complementary import) were distributed with the 'other agriculture' product group, an increase in the demand for coffee would necessitate an increase in the output of the 'other agriculture' industry. This, in turn, would require an increase in the inputs to that industry as specified in the published tables unless a specific adjustment is made to the tables. It is easy to compile tables using the indirect allocation method. The initial problem which has to be overcome is matching each imported product with the domestic industry to which the product is primary or would have been primary if it were produced domestically.