• Summary
• Downloads
• Explanatory Notes
• Related Information
• Past Releases

Introduction

12.1 Chapter 9 provides an extensive description of input-output and its importance within the overall ASNA. This chapter describes how input-output tables are put together in Australia and how they relate to the rest of the accounts.

12.2 Various tables are included under the broad heading of input-output tables. Essentially each of these tables provides further detail that underlies the aggregates recorded in the gross domestic product account. These summary accounts are focused on the end result of economic activity, whereas the input-output tables provide detailed dissections of that activity and, by showing intermediate transactions, they enhance the description of productive activity within an economy.


The I-O approach to compiling the national accounts

12.3 The input-output tables have been and continue to be an integral part of the Australian System of National Accounts. Input-output tables present a comprehensive picture of the supply and use of goods and services (referred to collectively as ‘products’) in the economy and the incomes generated from production. They also provide an accounting framework within which the commodity flow method of compiling national accounts - in which the total supplies and uses of individual types of commodities have to be balanced with each other - can be systematically exploited, resulting in improvements in the overall accuracy of the national accounts.

12.4 SNA93 recommends use of the I-O framework for compiling basic data, integration of the I-O tables within the national accounts, and compilation of I-O tables at constant prices as well as at current prices. It also recommends that commodity flows data (by-products of the goods and services account) should be compiled at least annually, and that these data should be fully consistent with other parts of the national accounts.

12.5 Commencing with 1994-95, the annual GDP account has been compiled using the commodity flow method (the so called ‘I-O approach to compiling the GDP account’, or simply the ‘I-O approach’). In other words, the compilation of the GDP account is fully integrated with the compilation of the I-O tables.

12.6 Conceptually, the GDP account and the I-O tables are fully integrated and consistent. The GDP account provides three approaches to measuring GDP: summing the incomes generated by production; summing final expenditures on commodities produced; and summing the value added at each stage of production. I-O tables are essentially a further disaggregation of the same three approaches. Whereas intermediate consumption is netted out from the GDP account, I-O tables bring these inter-industry flows of commodities back into focus, thereby providing a more developed articulation of the process of economic production, and the structure and interrelationships of industries. An important feature of the I-O tables is that they are fully balanced matrices which allow for the confrontation of data and the resolution of differences at a detailed level.

12.7 Chapter 9 provides an extensive description of the full input-output framework for compiling a set of national accounts. In that chapter a distinction is drawn between supply and use (S-U) tables and analytical, or symmetrical, input-output tables. The usefulness of S-U tables as a statistical tool is emphasised. The strategy adopted by the ABS in relation to the compilation of input-output tables involves a two stage process whereby a series of S-U tables, in both current prices and in the prices of the previous year, are compiled annually. These tables constitute benchmarks for the annual and quarterly GDP accounts. The process of benchmarking the GDP account to balanced S-U tables is referred to as the I-O approach. The analytic input-output tables are compiled as the second stage of this process when the S-U tables for a particular year are deemed to be final.

12.8 The I-O approach to compiling the GDP account allows for the annual and quarterly current price GDP accounts to be benchmarked to balanced S-U tables. The S-U tables for each year are effectively compiled three times: first preliminary tables; second preliminary tables; and final tables. The GDP account is benchmarked at each of these three stages. The rebenchmarked GDP account is published first in the June quarter issues of the ASNA. This strategy means that the quarterly accounts will never be projected more than seven quarters from a balanced set of annual accounts. Apart from the most recent year (for which a balanced estimate is not available) there will be only one measure of annual GDP, and consequently no statistical discrepancies in annual terms.

12.9 As explained previously, the compilation of balanced S-U tables requires three iterations. The sequence of S-U and I-O tables is scheduled for completion according to the following timetable:

	1st preliminary	end of yr t + 12 months
	2nd preliminary	end of yr t + 24 months
	Final	end of yr t + 36 months
	Input-output tables	end of yr t + 40 months

Classification of industries and products

12.13 The industrial classification used for the 1994-95 and subsequent input-output tables, the 1994-95 Australian Input-Output Industrial Classification (IOIC), is based on the 1993 edition of the Australian and New Zealand Standard Industrial Classification (ANZSIC). However, in some respects, it departs from the usual application of that classification. ANZSIC is mostly applied to establishment units, and this is the starting point for most of the input-output industries. However, for input-output tables it is desirable that an industry corresponds as closely as possible to the production of products primary to that industry. This applies especially where establishments classified to an industry produce significant amounts of products primary to another industry which has quite a different pattern of inputs. In these cases, where practical, secondary or subsidiary production is treated as output of the industry to which such production is primary; this process is called 'redefinition of production'.

12.14 The major redefinitions which are made in the process of compiling the input-output tables for Australia are as follows. Trading activity of miners and manufacturers is redefined to wholesale or retail trade, and any significant manufacturing activity of wholesalers is redefined to appropriate manufacturing industries. The input-output industries: mechanical repairs and other repairs have no counterpart in ANZSIC. They are formed by redefinition of the repair activity undertaken by wholesalers and retailers. Significant capital work on own account, where not primary to the industry of the establishment undertaking the activity, is redefined to construction or an appropriate manufacturing industry. The imputed component of the ownership of dwellings industry is specified in the 1993 ANZSIC as being part of the residential property operators industry (ANZSIC Class 7711). In addition, there are some areas of the economy - for example construction and agriculture - where the nature of available data or other considerations may preclude a strict adherence to ANZSIC sectoring. In these areas, the input-output industries are formed by the redefinition of products from an ANZSIC sector.

12.15 Some of the input-output industries correspond to single ANZSIC classes, but it is not practicable to have an input-output industry for each ANZSIC class. The aim is to provide both a balanced picture of the structure of the economy and comparability between the latest input-output tables and earlier ones. Details of the Australian Input-Output Product Classification (IOPC), and concordances between the items in the classification and data source items of production, imports and exports, are available on floppy disk.

Industrial output

12.16 Generally, the output of industries is defined as the production of goods and services for use as inputs into industries or as final demand. Own account production and transportation not separately invoiced is not shown separately, but rather is included indistinguishably with the rest of producers' output. The actual estimation procedures used for each industry are given below.

12.17 In agriculture, the value of output is the ex-farm or local value, which is estimated by valuing quantities of the various products. (The data are obtained mainly from the annual Agricultural Survey and the Agricultural Finance Survey.) The value of containers, such as bags and cases, is added where appropriate to the local value of output to bring the treatment of rural industries into line with that of manufacturing industries. In services to agriculture, hunting and trapping, the value of output is estimated by reference to measures of the usage by industries of, and in final demand for, products primary to services to agriculture, hunting and trapping. The output of forestry and logging includes the value of forest products, and government and private current expenditure on management of forest resources. The information comes mainly from Australian Bureau of Agricultural and Resource Economics (ABARE) collections and partly from the reports of public authorities. The output of commercial fishing is the local value of the fish obtained from ABARE collections.

12.18 In mining, manufacturing, electricity supply and gas supply, the estimates of output are based on information obtained from ABS economic statistics collections.

12.19 The starting point for these estimates of output is the sum of the following items derived from data obtained from the economic statistics collections: sales and transfers-out of goods produced on the premises; changes in inventories of finished goods and work-in-progress; margin on merchanted goods; capital work on own account; and other specified operating income items. The initial estimates for some industries have a product mixture which is not satisfactory for input-output purposes, and in these cases some secondary output and associated inputs are shifted to the industry to which the output is primary.

12.20 In particular, it is important not to mix processing activity with trading activity because trading activity is treated in a special way in input-output tables. Accordingly, the trading activity of mining, manufacturing and electricity establishments is redefined, partly to wholesale trade and partly to retail trade. In doing this it is assumed that the mining and manufacturing industries, in undertaking this wholesale and retail activity, use the same inputs in the same proportions as the Wholesale trade and Retail trade industries. It is also assumed that the percentage margin earned on these goods by manufacturers is the same as that earned by wholesalers or retailers dealing mainly in similar goods. For the purpose of redefinition, it is also necessary to estimate the inputs into gross margin earned by manufacturers. These estimates are based on the input structure of wholesalers or retailers.

12.21 Apart from trading activity, it is necessary to examine other components of the output of the mining, manufacturing, electricity supply and gas supply industries. Capital work on own account is the subject of special investigations undertaken to determine whether or not the activity is typically primary to the industry reporting it. In those cases where capital work on own account is large, where the activity is primary to another industry and where it is possible to identify the relevant inputs, it is desirable to redefine the capital work on own account to the industry to which it is primary. In other cases, where the capital work on own account is not primary to the producing industry, it is treated as output secondary to that industry.

12.22 Information on a range of other specified operating income items is also collected in economic statistics questionnaires. Investigations are undertaken for industries where revenue from manufacturing on commission is significant. Where commission work is carried out for an establishment coded to a non-manufacturing ANZSIC class, estimates are made of the value of the products produced and materials used, and these are transferred to the industry to which the products are primary (usually the industry of the establishment carrying out the commission work). The commission is absorbed by inclusion in the value of production. Where the commission work is carried out for establishments coded to a manufacturing ANZSIC class, the value of products produced is recorded in the output of the industry commissioning the work. It is assumed that the commissioning establishment does not further process the goods. The commission is recorded as an intra-industry flow.

12.23 Data on rent and leasing revenue are collected in economic statistics questionnaires, and investigations have been conducted into the components of this revenue. Where the value of rent and leasing revenue is large, it is redefined. In other cases it is treated as secondary production of the industry receiving it. Payments for rent and leasing are treated as purchases of products by the industry making the payment.

12.24 The estimates of the output and input for the services to mining industry are derived using information obtained from the ABS surveys of mineral and petroleum exploration and the Economic Activity Survey.

12.25 For tables published for reference years prior to 1983-84, the major component of the output estimate for the services to mining industry consisted of the revenue of establishments that provided services to the mining industries. The main service provided is drilling (on a contract or fee basis). For the tables in respect of reference year 1983-84 and subsequent years, the estimate of output includes the value of exploration activity undertaken on own account. This refers to the exploration undertaken by establishments as a service to their parent enterprises and for which these establishments receive no revenue. The output value attributed to this activity is the current cost of providing the service, i.e. wages paid plus cost of goods and services used by these establishments. For the tables from 1993-94, the estimate of output includes mine operations on a contract or fee basis. Services to mining such as assaying and geological surveying are excluded from this industry and are included in the scientific research, technical and computer services industry.

12.26 With effect from 1994-95, output in the form of exploration activity is shown as gross fixed capital formation. Prior to the 1994-95 tables this output was mainly recorded as intermediate usage by the mining industry. Under this treatment the GOS recorded for the mining industry is now substantially higher as exploration expenditure is not treated as intermediate consumption.

12.27 The output of the residential building construction industry represents the value of work done on new dwellings and the value of repairs to dwellings. The output of other construction comprises new construction other than on dwellings and all repairs to non-dwelling buildings and structures carried out by the construction industry. The estimates of the output of the residential building construction and other construction industries are based on information from a number of different sources, including ABS building statistics, surveys of construction and engineering construction, income tax data, public accounts, annual reports of public enterprises, data from other economic statistics collections (for own-account construction and sales of construction materials), and the Household Expenditure Survey (for expenditure on construction and construction materials).

12.28 The output of wholesale trade and retail trade is defined as gross margin on goods traded (difference between sales and cost of goods sold) plus commission and other non-margin receipts. The value of goods handled is not included in the output of wholesalers and retailers, who are considered to sell only distribution services. Estimates for the output of the wholesale trade and retail trade industries are based on information from ABS collections of these industries. Estimates of output for these industries includes wholesale and retail activity redefined from other industries such as manufacturing. The activities of wholesalers and retailers other than distribution, such as manufacturing, repairing and leasing, are either redefined to an appropriate industry or are shown separately as non-margin output. These non-distribution activities are treated as follows: manufacturing activity of wholesalers is redefined to an appropriate manufacturing industry; repair activity of wholesalers and retailers is redefined to one of the industries: mechanical repairs or other repairs as appropriate; rent and leasing revenue, and service revenue, are treated as non-margin output of wholesale trade or retail trade; and takeaway food selling is treated as non-margin output of retail trade.

12.29 After the separation of non-distribution activities, the gross margin on goods traded is estimated in total and by product. The margin on individual products is estimated using the results of periodic ABS collections of the wholesale trade and retail trade industries, the results of supplementary investigations, and the difference between estimates of the supply of products at basic prices and usage at purchasers' prices.

12.30 In the accommodation, cafes and restaurants industry, restaurants, cafes, licensed hotels and licensed clubs are covered by the economic statistics collections, and considerable data from these surveys are available regarding the output of this part of the industry. However, some adjustments are made for input-output purposes. First, the output of the industry includes a margin on sales of goods (e.g. cigarettes) that is treated in the same way as retail margin. Second, takings from meals served in cafes, hotels, etc. and from the consumption of beverages associated with a meal is treated as meal preparation and presentation, a service activity of this industry that forms part of its output. This output is not treated as a margin. In tables prior to 1994-95 the consumption of beverages in association with a meal was treated as a margin activity.

12.31 In the transport and storage industries the output is defined as revenue derived from the activity of carrying goods and passengers for hire and reward, the storage of goods, and the provision of other transport services. As with wholesale and retail trade, it is necessary to distinguish between the carriage of goods (freight), which forms part of the margin, and other services provided by transport industries, which do not (e.g. furniture removal and passenger transport). Total rail, pipeline, air and water transport revenues are extracted from the reports of the enterprises concerned as far as possible. The estimate of output of the road transport industry is built up from information about items such as the supply of goods potentially carried, primary inputs, motor vehicle running costs, miscellaneous other costs, revenue for services provided, and data collected in ABS surveys. The estimates of transport margin by product are approximate because of the limited information about the types of goods carried by each mode of transport.

12.32 Estimates of output of the communication services industry are based on the published accounts of enterprises in the industry and unpublished information made available by these enterprises. The output of the industry consists of postal, telegraph, telephone and telex revenue; commission earned on agency services (such as conduct of savings bank agencies); technical services provided to national broadcasting and television stations; international communication services and telecommunications satellites; and the coastal radio services.

12.33 In the finance industries, the definition of the output of banks and similar financial enterprises (other than insurance companies) differs from that of other enterprises. The activities of these enterprises are financed to a large extent, or even predominantly, by the excess of the interest they receive over the interest they pay out. If they were treated similarly to trading enterprises (i.e. if their output were taken as equal to their explicit charges for services provided) their output would not be sufficient to cover all their costs and to produce a positive operating surplus. Because of this characteristic of financial enterprises it is necessary to adopt a treatment different from that for non-financial enterprises. The convention adopted in Australian input-output tables from 1968-69 is that part of the interest received by financial enterprises is considered to be a service charge and, therefore, part of their output. Users of these services are deemed to be paying a service fee on funds both lent to and borrowed from these financial institutions. Since the 1977-78 tables these charges related to financial intermediation services indirectly measured (FISIM) have been allocated to final uses and intermediate uses on the basis of various sets of information mainly relating to interest flows. FISIM is discussed in greater detail in Chapter 20. The tables from 1994-95 on also include estimates of imports and exports of these services. In addition to FISIM charges on the output of financial enterprises includes explicit charges made to customers.

12.34 Output of those financial enterprises that provide insurance services is also defined in a manner that is different to all other sectors. For non-life insurance enterprises (such as motor vehicle and property insurers) no explicit charge is made for their services. A service charge is therefore imputed, for national accounts purposes, as direct premiums earned plus inward reinsurance minus outward reinsurance and associated statutory charges plus premium supplements minus expected claims. Premium supplements refer to property income (e.g. interest and dividends) earned on: (a) that part of premiums paid in advance and, (b) claims incurred but not yet paid (as a consequence of delays in finalising payments from the time claims were first lodged). Expected claims are imputed as a centred five year moving averages of actual claims incurred. The value of the service charges (output) for life insurance and pension funds are derived as the explicit charges made by non-mutual insurance enterprises for the services they provide. In the case of mutual funds output is deemed to be equal to the sum of the administrative costs incurred by the fund (including labour costs). This approach, in respect of mutual funds, implies that no net operating surplus accrues to the insurance enterprise itself but rather the surplus is deemed to accrue to policy holders. Refer also to Chapter 20.

12.35 Information on the output of banks, non-bank financial institutions, insurance companies and superannuation funds is available from ABS collections, the Australian Prudential Regulation Authority (previously from the Insurance and Superannuation Commission) and the Reserve Bank. There is very little information for some enterprises in the finance and insurance industries, and the output estimates for these are largely built up from the input side by adding together estimates of primary inputs and of the usage of products primary to these industries.

12.36 Estimates of the output and input for property and business services industries are derived mainly using information from ABS Service Industries Surveys supplemented by information from the Economic Activity Survey. For those activities primary to industries not covered by the surveys, estimates of output and input are derived from taxation statistics. The output of the ownership of dwellings industry is gross rent of dwellings (actual rent paid in the case of tenanted dwellings and an imputed rent for owner-occupied dwellings). Benchmark information for these items is obtained from the Census of Population and Housing, and various indicators are used to extrapolate and interpolate benchmark information.

12.37 The output of the government administration and defence industries is the production of goods and services by general government units coded to these industries. The output is conventionally valued as the sum of intermediate inputs (including the charge for financial intermediation services indirectly measured); compensation of employees; general government consumption of fixed capital; and taxes on production. Ideally, the estimates would be on an accrual basis for consistency with estimates for other industries. Prior to 1998-99, the data required for this purpose were not available as most estimates based on public accounts were on a cash basis rather than an accrual basis. However, from 1998-99 onwards these estimates are on an accrual basis. The output of the defence industry prior to the 1994-95 tables was defined such that expenditures on items that would be considered capital formation in other industries were treated as current expenditures, thereby contributing to the value of defence industry output. From 1994-95 capital-type items acquired by the defence industry are treated as capital formation except where the items are for strictly military purposes. While the acquisition of these assets no longer contributes to the value of the defence industry's output, the consumption of fixed capital on these assets still does.

12.38 The output of the education; health services; community services; libraries, museums and the arts; and other services industries consists of three components. The first of these is production of goods and services by general government units, which represents the sum of intermediate inputs; compensation of employees; consumption of fixed capital and taxes on production (net). The second is production by trading enterprises, such as doctors in private practice and private garbage disposal contractors, which is valued as the revenue from services provided. The estimates for this component are derived using either data on inputs (e.g. compensation of employees) or business income from Taxation Statistics (Australian Taxation Office). The last is production by the producers of private non-profit services to households, such as schools and churches, which is measured as the sum of intermediate inputs; compensation of employees; and taxes on production (net). By convention, there is no operating surplus or mixed income. The estimates for this component are based mainly on public accounts.

12.39 The remaining industries, namely motion picture, radio and television services; sport, gambling and recreational services; and personal services, cover a range of services rendered to persons and, to a lesser extent, to businesses. For the motion picture, radio and television services and sport, gambling and recreational services industries, estimates are made using information from economic surveys where this is available (e.g. for motion picture theatres). Data of reasonable quality are available from public accounts and published reports for public broadcasting and television stations. Total revenues are available for commercial broadcasting and television from the report of the Australian Broadcasting Control Board. For the remainder of this industry, the output estimates are built up from primary inputs, estimates of other costs and demand for the services provided.

12.40 For the personal services industry, the economic surveys provide information about the output of men's and women's hairdressers, and laundry and dry cleaning establishments. For the remainder of the industry, output is estimated from the input side and from measures of the level of demand for the services provided.

Primary inputs

12.41 Primary inputs consist of compensation of employees, gross operating surplus and gross mixed income, taxes on products (net), other taxes on production (net), complementary imports and (in direct allocation tables) competing imports.

12.42 Estimates of the compensation of employees for each industry are the end result of a complex estimation process. Compensation of employees consists of wages and salaries paid in cash, wages and salaries paid in kind and employers' social contributions. Each component is first estimated in aggregate for the economy as a whole, then each aggregate is apportioned to the 107 industries. Compensation of employees for each individual industry is derived by summing these separate components. For wages and salaries paid in cash, an aggregate is derived by applying average earnings from the Survey of Employment and Earnings to an estimate of total employee-status jobs from the Labour Force Survey. This aggregate is allocated to each input-output industry in accordance with proportions established through analysis of the ABS economic statistics collections. The second component of compensation of employees is wages and salaries in kind. This item refers to the situation where employees are remunerated through provision of goods or services rather than in cash. Estimates of the aggregate value of this form of remuneration are obtained by applying expansion factors to Fringe Benefits Tax payments. The third component in compensation of employees is employers' social contributions. These consist of workers' compensation insurance premiums paid by employers and employers' actual and imputed contribution to superannuation funds and separately constituted long-service leave funds.

12.43 The gross operating surplus and gross mixed income of industries is the residual obtained by subtracting from the value of output all intermediate inputs; taxes on production (net); and compensation of employees.

12.44 The gross operating surplus of the producers of government services is defined to be equal to the producers' consumption of fixed capital. This is the value, at current replacement cost, of the reproducible fixed assets used up during a period of account as a result of normal wear and tear, foreseen obsolescence and the normal rate of accidental damage. Unforeseen obsolescence, major catastrophes and the depletion of natural resources are not taken into account.

12.45 Reliable estimates of total other taxes on production (net) are available, but there is little information on their distribution by industry. Fringe benefits tax and payroll tax are separately reported in the ABS economic surveys and this information is used to allocate these taxes. Motor vehicle taxes are first apportioned between persons and businesses, using Government Finance information, and the business part is allocated to industries in proportion to motor vehicle running expenses. Other taxes on production in this group are allocated to industries on various bases. Sometimes the nature of the tax determines the allocation, e.g. road maintenance tax is allocated entirely to the road transport industry, and rates on residential property are allocated to the ownership of dwellings industry. In some other cases, information concerning taxes on production (net) is obtained in the course of investigating industry inputs.

Imports

12.46 Imports represent the value of goods and services purchased from foreign residents. Imports of goods are based on international merchandise trade statistics, but with balance of payments scope, coverage, timing and valuation adjustments. For details of these balance of payments adjustments, see Balance of Payments and International Investment Position, Australia: Concepts, Sources and Methods (Cat. no. 5331.0). Imports of services are also based on balance of payments sources.

12.47 Although the estimates of imports of individual products are recorded on a c.i.f. basis, aggregate imports are shown on a f.o.b. basis to maintain consistency with the total for imports of goods and services shown in the balance of payments. This is achieved by netting off the value of the freight provided by Australian carriers from the estimates of imports c.i.f.

Intermediate inputs

12.48 Intermediate inputs into an industry consist of goods and services used in the process of production. A detailed discussion of the estimation of intermediate inputs using the Australian input-output methodology is provided in the ABS Occasional Paper: The RAS Method for Compiling Input-Output Tables: ABS Experience, which outlines the residual allocation system (RAS). In essence, RAS is a procedure which requires the successive scaling of column and then row elements by successively prorating to column totals and row totals until the matrix elements are fully reconciled with both row and column control totals. The number of iterations required will depend on the consistency of the initial data matrix with the row and column control totals. The accuracy of the final matrix derived will depend on the suitability of the initial matrix as an indicator.

12.49 Ideally, information on the value of the goods and services which comprise the intermediate inputs of industries would be collected from businesses each year. However, it is difficult to obtain this information directly from business, and the ABS also considers that collecting this type of data every year is an unacceptable impost on those businesses responding to ABS surveys. Therefore, the ABS only collects information on intermediate inputs for a selected portion of the intermediate use quadrant in respect of each year. The RAS procedure is used to enable this new information to be combined with the intermediate inputs matrix projected forward from the previous table in order to estimate the intermediate inputs matrix for the current year. The incorporation of this new information is referred to as applying 'modifications' to the intermediate usage matrix.

12.50 Selected data from the Manufacturing Survey and the Economic Activity Survey are the basis of extensive modifications to the input structures of the manufacturing industries. The input structure of the defence industry is substantially derived from data in the Defence Annual Report and ABS Government Finance data. In addition, the input structures of the electricity and gas supply industries have been modified from basic data, and the input structures for the mining industries have been modified to account for the use of contract mining services. In most years approximately 25 per cent of the intermediate inputs (other than the FISIM charge) to all industries are derived from basic data.

12.51 The charge for FISIM is allocated to using industries on the basis of information on interest paid and interest received by the various industries. These interest data are compiled from a number of sources including Government Finance data and the ABS economic statistics collections.

12.52 Rows relating to the margin item: taxes on products (net); and to the following margin products: wholesale trade; retail trade; accommodation, cafes and restaurants; road transport; rail transport; pipeline transport; water transport; air and space transport; port handling; and marine insurance, are wholly modified by margin estimates derived in the compilation of the eleven margin tables. The column totals of the margin tables, which show the total usage of each margin by each industry, are then incorporated as values in the eleven margins rows.

Final demand

12.53 Household final consumption expenditure consists of current expenditure (including household expenditure on motor vehicles and durables) by persons, and the output of goods and services by non-profit institutions serving households. Estimates of household final consumption expenditure are based largely on retail sales, supply of consumer goods, outputs of services, information about the revenues of public enterprises and general government, and estimates of the actual and imputed rent from dwellings. Estimates by industry of origin are based partly on the product dissection of retail sales provided by retail activity surveys and partly on data about the supply of consumer goods. Household Expenditure Survey results are used mainly to validate estimates of household final consumption expenditure provided by ABS economic statistics collections, especially for products produced by industries other than agriculture, mining and manufacturing.

12.54 Government final consumption expenditure is mainly the output of producers of general government goods and services less revenue from any sales, or charges in respect of that output. The classification of government final consumption expenditure by product differs from a classification by purpose in that, for example, departmental administration is classified to public administration in the product classification and not to the activity administered (e.g. education). Estimates are based on public accounts.

12.55 Private gross fixed capital formation includes outlays on, additions to and replacement of fixed assets of all kinds by private enterprises. Commencing with the 1994-95 tables, fixed assets include various intangible assets such as computer software; artistic originals; and mineral exploration. Furthermore, the natural growth of breeding livestock is treated, for the first time, as capital formation in the 1994-95 tables. Purchases of dwellings by persons and all capital expenditure by private non-profit organisations serving households are included. Expenditure on major repairs which significantly extend the life of assets is included. Estimates of private gross fixed capital formation are based on a variety of sources including building statistics, a regular ABS survey of private capital formation and Taxation Statistics. To some extent both the total estimates for this category and the product estimates depend on the output of goods and services not absorbed by other final demand categories or intermediate usage. Taxes on production (net) applying to transactions in land and fixed capital are included in private gross fixed capital formation. These include: stamp duties on all capital transfers; the cost of conveyancing services, real estate services and surveying services associated with the purchase or sale of dwellings; agricultural or pastoral properties, and other non-dwelling real estate; and miscellaneous government charges.

12.56 Public gross fixed capital formation includes all outlays on, additions to and replacement of fixed assets by public corporations and by general government. However, defence expenditure on assets with strictly military use is treated as current and included in government final consumption expenditure. Estimates of gross fixed capital formation of public corporations and general government are based on the accounts of public authorities and additional information supplied by these bodies. The classification of gross fixed capital formation of public corporations and general government relies on information on the nature of the assets purchased and goods and services produced from the assets.

12.57 Changes in inventories are calculated by deducting an adjustment for holding gains from the change in book value of inventories. The major inventory-owning industries are covered by the economic collections, and information for other industries is available from Taxation Statistics, the accounts of public authorities and a regular ABS survey of private inventory holdings. However, these sources do not provide a sufficient dissection of inventories by product. The allocation of the change in book value of inventories to industry of origin is made by inference (but not by automatic apportionment) from the categories of goods sold by wholesalers and retailers and the finished goods of producers, and from the categories of material inputs. Estimating the adjustment for holding gains by product group requires similar inferences about the level of inventories, assumptions about valuation practices, and reliance on partial information about product prices.

12.58 Statistics relating to exports are obtained from the international merchandise trade and balance of payments statistics. The category includes re-exports, for which the corresponding import is included in the row for competing imports. The product detail provided in international merchandise trade and balance of payments statistics is used to allocate exports to input-output products.

Employment

12.59 The employment estimates that accompany the ABS's input-output tables are measured on a full-time equivalent basis, i.e. the full-time equivalent of part-time employment is added to full-time employment. For these estimates the full-time equivalent of part-time employment is assumed to be 50 per cent of the part-time employment. Employment estimates have been derived from the ABS Labour Force Survey. The annual estimates are the average of data for the middle month of each of the four quarters of the reference year.