The structure of the I-O tables

22.6    The 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 activities.

22.7    The ABS uses an economic statistics model to describe the characteristics of units, and the structural relationships between businesses. Businesses with a simple structure are classified by their Australian Business Number (ABN) on the Australian Business Register (ABR), maintained by the Australian Taxation Office. Businesses with a more complex structure (i.e. where the ABN is not suitable for ABS statistical requirements) are maintained on the ABS Maintained Population register (ABSMP), through direct contact with the business. These units comprise the Enterprise Group, the Enterprise and the Type of Activity Unit (TAU). The TAU represents a grouping of one or more business entities for which a basic set of financial production or employment data can be reported.

22.8    When a unit engages in more than one type of production, the primary production is the activity for which gross value added is the greatest for that unit. The production reported by a unit may include both primary and secondary production. The output of an industry may be a number of products that are jointly produced (e.g. natural gas linked to crude oil). In this case primary products may be distinguished by the principal product with the smaller output treated as secondary production.

22.9    I-O tables can be compiled for industries or products but they are both similar in theory. The distinguishing characteristics of analytical I-O tables are that they are square and symmetric, and they differ from the S-U tables in that the transactions are valued at basic prices rather than purchasers’ prices. The I-O tables provide detailed information about the supply and use of products in the Australian economy and about the structure and inter-relationship between Australian industries.

22.10    Table 22.1 provides a summary of the different dimensions and values shown in the published I-O tables. Detailed information on the content of each published table is provided below the summary table.

22.11    The basic tables of I-O are aggregations of the various components of GDP. The most significant feature of these tables is that they are not symmetrical in that the dimension of the columns differs from dimension of the rows.

22.12    There are four main basic tables used to compile the I-O tables:

• Supply table – shows the output of domestic industries and imports classified across columns, and products classified across rows. The largest values are found on the leading diagonal as industries specialise in their primary products. The columns in the supply table show the products each industry produces, and the extent to which industry specialises in the production of its primary products, as well as the product composition of imports.
• Use table – shows the product groups and primary inputs in the rows, and industries and final use categories in the columns. The rows show the total supply of products, whether locally produced or imported, and show how these products are used by industries as intermediate inputs to production or consumed as final demand by category. At the bottom of the table, the rows show the primary inputs purchased by industries, and by final demand. Reading down the columns shows that you can read the inputs (intermediate and primary) into each industry, and the composition of each final demand category. Therefore, all flows of goods and services in the economy are covered.
• Imports table – shows in the columns the industries to which imported products would have been primary if they had been produced in Australia, and in rows the usage of these products by industry and final demand category. This breakdown is only shown for competing imports, or those products which are produced domestically and imported, so that substitution between domestically produced products and imports is possible. The disposition is not shown for complementary imports, which by definition are products that are not domestically produced. Since the 2001-02 I-O tables, ABS has not measured complementary imports, and assigns all imports as competing.
• Margins table – shows the difference between the basic price and the purchaser’s price of all flows in the use table. Table 4 shows the decomposition of flows at purchaser prices into basic prices, net taxes on products and the sum of all trade and transport margins. Tables 23 to 39 show the detailed disposition of each type of margin, product taxes by type, and product subsidies, to intermediate use and final use categories.

22.13    These four main basic tables make up a record of the estimated flows which occur in the production process. However, the use table is not symmetric which makes it unsuitable for some forms of analysis. This problem is solved by converting the use table to an industry-by-industry flow table by adjusting the rows to show industry use of industry output, rather than products. The ABS does not produce product-by-product flow tables.

22.14    Table 22.2 provides a summary of the basic I-O tables published by the ABS.

22.15    Derived tables differ from the basic tables in I-O in that they are symmetric so that the dimensions of the columns and rows are the same. The dimension is either product by product or industry by industry. In Australia the derived I-O tables are industry by industry.

22.16    Another feature of the derived table is that they are not simply aggregations of the components. Some further calculations are applied in order to produce the tables namely the derivation of coefficients.

22.17    Table 22.3 depicts the industry-by-industry table. A row in the table shows the disposition of the output of an industry group and a column shows the origin of inputs into an industry and final use category. The output of an industry equals the sum of its inputs including its primary inputs so the column total must equal the row total.

22.18    Table 22.3 shows the basic structure of an industry-by-industry table with direct allocation of imports (as is published in Table 5 of the I-O tables) where imports are allocated to the using industries. The flows between the domestic industries are:

• Quadrant 1 – this is referred to as the inter-industry quadrant where each column shows the intermediate inputs into an industry in the form of products produced by other industries and itself. Each row shows how the output of an industry has been used by itself and other industries as part of their production process;
• Quadrant 2 – shows the disposition of output to final use categories by industry group;
• Quadrant 3 – shows the primary inputs to production (compensation of employees, gross operating surplus and gross mixed income, imports and net taxes on production); and
• Quadrant 4 – shows the disposition of primary inputs to final demand categories.

22.19    The sum of quadrants 1 and 2 shows the total usage of goods and services produced by each industry. Total usage equals total supply, with final demand including change in inventories, which may be positive or negative.

22.20    The sum of quadrants 1 and 3 shows the total inputs required to produce the outputs of each industry group. Total inputs equals total supply or outputs, with primary inputs including gross operating surplus and gross mixed income, which can conceptually be positive or negative.

22.21    Table 8 of the I-O tables is an industry-by-industry flow table with indirect allocation of imports. This table shows:

• supply by industry group, including Australian production and products which are imported; and
• the inputs into an industry’s production, reflecting the technological relationships between all inputs into the industry, whether domestically produced or imported.

22.22    In order to balance the table, the row for competing imports is shown below the Australian production; that is, showing total supply (row total) for each industry as being equal to the corresponding total uses (column total). For each column, this row shows the value of imports competing with the output of each industry. This presentation results in the double entry for imports in the table to reconcile total supply and total uses. In a table with direct allocation of imports, the competing imports row is shown above the Australian production row, and shows, for each industry, the total intermediate use of imports by the industry.

22.23    The difference between the direct and indirect allocation of imports is discussed in the allocation of imports section (paras.22.55-22.61).

22.24    The following table provides a summary of the derived I-O tables published by the ABS:

There are four additional tables that are published which are not basic or derived I-O tables. The following table provides a summary of them:

22.26    In Quadrant 1, a row or column is said to refer to an industry; however, a row or column can refer to a product (or group of products) rather than an industry. The structure of products or industries is important in the use of the I-O tables. It is desirable that each product or industry changes as little as possible over time, and that each industry produces a single output, with a single input structure. This approach implies that all products produced by an industry are perfect substitutes for each other or are produced in fixed proportions. It also implies that the input structure does not vary in response to changes in the product mix, and that there is no substitution between the products of different groups of products or industries. This is known as the homogeneity assumption; however, it is not fully supported in the ABS I-O tables.

22.27    The stability of coefficients is affected by the interaction of two factors: (a) the aggregation of products with different input structures; and (b) changes in the product group mix over time. This becomes important when the data sources for the I-O coefficients are infrequent, such that it is necessary to assume that observed coefficients apply in the following years, at least as a starting point. This problem arises in industries producing a range of products that have different input structures.

22.28    There is significant aggregation even in large I-O tables, leading to a departure from these objectives, and affecting the homogeneity of products or industries. There are two ways the aggregation can be made: (a) grouping by industries to create an industry-by-industry table (the ABS approach); or (b) grouping by products to create a product-by-product table. The two methods result in differing impacts on homogeneity, with different implications for the analytical use of the tables. There is no complete solution for the aggregation problem, but appropriate grouping can keep errors to acceptable limits. The groups used are partly dependent on industry classifications, and the practical process of compiling the I-O tables.

22.29    At first sight, the solution to the grouping problem is to narrowly define product groupings. However, this could result in the tables becoming too complicated for users, and would take too long to compile, particularly as the ABS is now producing I-O tables every year. Even with narrower product groupings, there would be instances where a TAU produced products classified to different groups of products, and it would not be practical to split details to different groups. Confidentiality would also become a problem in some industries, as the products covered in a group became more specific.

22.30    For industries, the homogeneity assumption will not be fully met as some industry groups produce a wide range of products at the industry-group level. Similar to above, the classification of industries as establishments or TAUs would make the tables too complicated; the tables would take too long to compile; and there would be confidentiality issues. Grouping industries will still result in secondary production, where the products have different input structures. For example, if the basic iron and steel industry also produces non-ferrous castings, the input structure for this column will show the use of non-ferrous metals, and the corresponding row will show sales of products to industries using non-ferrous castings. These results may not be suitable for users interested only in iron and steel products. The requirements calculated from this table could be misleading, unless the production of secondary products forms a fixed proportion of the industry's output. The proportion of product mix should remain constant where secondary products are jointly produced, or the secondary product is a by-product of the primary production; there is often no correlation between primary and secondary products.

22.31    The extent of secondary production by an industry depends on the range of products produced by individual establishments, and whether the industries are grouped into large numbers of narrowly defined industries, or a small number of broadly defined industries. Where industries are narrowly defined, a large proportion of the products will be produced by industries to which the products are not primary. This conflicts with both the homogeneity requirement and the non-substitution requirement. Where significant proportions of a product can be substituted by products produced by a different industry, there is a weak link between the demand for a product and the output of a single industry. Combining some of these industries could improve homogeneity in one respect, at the expense of creating a more heterogeneous product mix.

22.32    The availability of source data will ultimately affect the grouping of products or industries. Detailed information of sales or output of products is normally available, but information on costs may not be available. In some cases, only input structure detail may be available. A rolling program of case studies is used to gather detailed data on companies’ input and output structures, by direct interview with companies, in order to assist with this problem. In the past, economic activity by some industries was redefined to more appropriate industries to limit the impact of secondary production on the tables, but this is no longer done in order to reflect how production occurs in the economy.

22.33    Regardless of whether products or industries are used in quadrant 1, the same processes are followed to assemble the data. It is necessary to record the product flows in a way that is suitable to compile I-O tables. The same information is required for each product or product group:

• the origin or source of supply, domestic supply by industry, and imports;
• the use of the product, intermediate usage by industry and final demand by category; and
• the difference (margins, taxes and subsidies on products) between the basic price and purchaser's price for each product.

22.34    The supply of imports must be classified in the same way as Australian production. Imports data is sourced from Customs data. These data are initially classified according to the Harmonised Tariff Item Statistical Code (HTISC) which is then concorded to the Input-Output Product Classification. The data enters the I-O tables as a vector and is allocated to the industry to which the imported product is primary.