Latest release

# The I-O compilation process

Australian System of National Accounts: Concepts, Sources and Methods
Reference period
2020-21 financial year

22.95    The I-O compilation begins with the finalisation of the S-U tables, when the balanced S-U levels are disaggregated to I-O levels. This results in unbalanced I-O tables which are then balanced using the product flow method. The product flow method involves a number of steps followed by a quality assessment process at strategic points to check data quality and consistency.

22.96    The figure below shows the steps undertaken in the I-O compilation process. The steps are then described in more detail.

### Figure 22.1 the I-O compilation process

This image is a flow-chart of the I-O compilation process
Step 1: Data - Balanced S-U tables & previous I-O tables. Process - collect data
Step 2: Data - reconciliation and disaggregation. Process - Disaggregate data to IOPC level
Step 3: Data - Initial tables at IOPC level. Process - Implement adjustments identified in initial analysis of raw data - run quality gate checks
Step 4: Data - Industry level analysis of data. Process - Implement adjustments resulting from industry analysis - run quality gate checks
Step 5: Data - Row balancing at Purchaser's Prices. Process - Implement adjustments as a result of row balancing - Run Quality Gate checks.
Step 6: Data - Run compilation process to allocate taxes, margins and subsidies to Use categories, rebalance rows at Basic prices using a proportional iterative fitting process to modify Intermediate use data. Process - Implement adjustments to resolve balancing issues at Basic prices - Run Quality Gate checks.
Step 7: Data - Produce basic and derived I-O tables. Process - Run quality gate checks on released tables
Step 8: Data - Disseminate I-O release, Alternate view tables and Product details release. Process - Release tables via website

Figure 22.1 the I-O compilation process

Step Data      I-O Process
1 Balanced S-U tables  Previous I-O tables Collect data

2  Reconcilation and Disaggregation   Disaggregate data to IPOC level

3  Initial tables at IOPC level   Implement adjustments identified in initial analysis of raw data - Run Quality Gate checks.

4  Industry level analysis of data   Implement adjustments resulting from industry analysis - Run Quality Gate checks.

5  Row balancing at Purchaser's Prices   Implement adjustments as a result of row balancing - Run Quality Gate checks.

6Run compilation process to allocate taxes, margins and subsidies to Use categories, rebalance rows at Basic prices using a proportional iterative fitting process to modify Intermediate use of data. Implement adjustments to resolve balancing issues at Basis prices - Run Quality Gate checks.

7  Produce basic and derived I-O tables  Run quality gate checks on released tables

8 Disseminate I-O release, Alternative view tables and Product details release. Release tables via website

### Step 1: Collect source data

22.97    Obtain the S-U tables and previous I-O tables.

### Step 2: Reconciliation and disaggregation

22.98    Balanced S-U tables are disaggregated to populate the I-O tables ensuring consistency with the rest of national accounts published aggregates.

22.99    Much of the initial data used to populate the S-U tables is classified to the IOPC and IOIG level, and then aggregated to the Supply-Use Product Classification (SUPC) and Supply-Use Industry Classification (SUIC) level. It is at this level that the balancing process of the S-U table is undertaken in order to produce balanced GDP estimates for the national accounts. It is necessary to then disaggregate the S-U products and industries to the I-O level in order to begin production of the I-O tables.

22.100    The disaggregation of the S-U level to the I-O level does not necessarily result in the same I-O distribution, compared with the initial I-O level data. This is due to the balancing being undertaken at the S-U level, and any adjustments made at that level may result in a different S-U level total than the initial total. This difference has to be distributed back to the I-O level and is done in several ways, depending on the component.

22.101    The following table outlines how the S-U level components are disaggregated to the I-O level.

Table 22.7 Disaggregation of S-U level to I-O level—GDP(P)
IndustrySupplyIntermediate consumption
Agriculture, forestry and fishing
Agriculture

Economic Activity Survey (EAS) ANZSIC class-level data are mapped to Input-Output Product Classification (IOPC) using weights derived from Value of Agricultural Commodities Produced, Australia product data and previously published Input-Output (I-O) weights. Any adjustments made during the Supply-Use (S-U) balancing process are split back to IOPCs based on the original IOPC proportions.Intermediate consumption is disaggregated to the IOIG and IOPC level based on the final Total Intermediate Use (TIU) matrix from the previous year's I-O tables.
Forestry, aquaculture, fishing, hunting, trapping and services to agriculture, forestry and fishing

EAS ANZSIC class-level data are mapped directly to primary Input-Output Product Group (IOPG). IOPG data are then mapped to IOPC using weights derived from previously published I-O data. Any adjustments made during the S-U balancing process are split back to IOPCs based on the original IOPC proportions.Intermediate consumption is disaggregated to the IOIG and IOPC level based on the final Total Intermediate Use (TIU) matrix from the previous year's I-O tables.
Agriculture, forestry and fishing support services

EAS ANZSIC class-level data are mapped to primary IOPG. IOPG data are then mapped to IOPC using weights derived from previously published I-O data except for data supplied directly from source data at IOPC level (e.g. ABARES ginned cotton). Any adjustments made during the SU balancing process are split back to IOPC based on the original IOPC proportions.Intermediate consumption is disaggregated to the IOIG and IOPC level based on the final Total Intermediate Use (TIU) matrix from the previous year's I-O tables.
Mining
All except exploration and mining support services

State Mines product detail is mapped directly to IOPC. Any adjustments made during the S-U balancing process are split back to IOPCs based on the original IOPC proportions.Intermediate consumption is disaggregated to the IOIG and IOPC level based on the final Total Intermediate Use (TIU) matrix from the previous year's I-O tables.
Exploration and mining support services

IOPC detail is a combination of EAS, Mineral Exploration and Petroleum Exploration data. Any adjustments made during the S-U balancing process are split back to IOPCs based on the original IOPC proportions.Intermediate consumption is disaggregated to the IOIG and IOPC level based on the final Total Intermediate Use (TIU) matrix from the previous year's I-O tables.
Manufacturing

EAS ANZSIC class-level data are mapped to IOIGs and then split into primary IOPCs according to historical weights and the EAS data items. Any adjustments made during the S-U balancing process are split back to IOPCs based on the original IOPC proportions.Intermediate consumption is disaggregated to the IOIG and IOPC level based on the final Total Intermediate Use (TIU) matrix from the previous year's I-O tables.
Electricity, gas, water and water services

EAS ANZSIC class-level data are mapped to IOIGs and then split into primary IOPCs according to historical weights and the EAS data items. Any adjustments made during the S-U balancing process are split back to IOPCs based on the original IOPC proportions.Intermediate consumption is disaggregated to the IOIG and IOPC level based on the final Total Intermediate Use (TIU) matrix from the previous year's I-O tables.
Construction

EAS ANZSIC class-level data are mapped to IOIGs and then split into primary IOPCs according to historical weights adjusted for the movement in the Building Activity and Engineering Construction Surveys. As the Supply-Use Product Classification (SUPC) and IOPC are a one-to-one match, any adjustments made during the S-U balancing process are applied directly back to the IOPCs.Intermediate consumption is disaggregated to the IOIG and IOPC level based on the final Total Intermediate Use (TIU) matrix from the previous year's I-O tables.

EAS ANZSIC class-level data are used to derive output which is mapped to IOPC based on historical proportions from previous retail and wholesale industry surveys and weights derived from previously balanced I-O tables. Any adjustments made during the S-U balancing process are split back to IOPCs based on the original IOPC proportions.Intermediate consumption is disaggregated to the IOIG and IOPC level based on the final Total Intermediate Use (TIU) matrix from the previous year's I-O tables.

EAS ANZSIC class-level data are used to derive output which is mapped to IOPC based on historical proportions from previous retail and wholesale industry surveys and weights derived from previously balanced I-O tables. Any adjustments made during the S-U balancing are split back to the IOPCs based on the original IOPC proportions.Intermediate consumption is disaggregated to the IOIG and IOPC level based on the final Total Intermediate Use (TIU) matrix from the previous year's I-O tables.
Accommodation and food services

EAS ANZSIC class-level data are mapped to primary IOPCs. Proportions are derived from ABS publications: Accommodation Services, Australia, Clubs, Pubs, Taverns and Bars, Australia and Cafes, Restaurants and Catering Services, Australia. Any adjustments made during the S-U balancing process are split back to IOPCs based on the original IOPC proportions.Intermediate consumption is disaggregated to the IOIG and IOPC level based on the final Total Intermediate Use (TIU) matrix from the previous year's I-O tables.
Transport, postal and warehousing

EAS ANZSIC class-level data are mapped to IOIGs and then split into primary IOPCs according to historical weights. Any adjustments made during the S-U balancing process are split back to IOPCs based on the original IOPC proportions.Intermediate consumption is disaggregated to the IOIG and IOPC level based on the final Total Intermediate Use (TIU) matrix from the previous year's I-O tables.
Information media and telecommunications

EAS ANZSIC class-level data are mapped to IOIGs and then split into primary IOPCs using previously balanced I-O Tables weights. Any adjustments made during the S-U balancing process are split back to IOPCs based on the original IOPC proportions.Intermediate consumption is disaggregated to the IOIG and IOPC level based on the final Total Intermediate Use (TIU) matrix from the previous year's I-O tables.
Financial and insurance services
Finance and Insurance and superannuation funds

Data are compiled at the IOPC level. For details on the sources and methods used to compile finance and insurance and superannuation funds output data see Table 9.21 ANNUAL GROSS VALUE ADDED BY INDUSTRY—Finance services, subdivision 62 and Table 9.22 ANNUAL GROSS VALUE ADDED BY INDUSTRY—Insurance and superannuation funds, subdivision 63.

Any adjustments made during the S-U balancing process are split back to IOPCs based on the original IOPC proportions.

Intermediate consumption is disaggregated to the IOIG and IOPC level based on the final Total Intermediate Use (TIU) matrix from the previous year's I-O tables.
Finance and insurance services

EAS ANZSIC class-level data are mapped to IOIGs and then split into primary IOPCs using previously balanced I-O Tables weights. Any adjustments made during the S-U balancing process are split back to IOPCs based on the original IOPC proportions.

Intermediate consumption is disaggregated to the IOIG and IOPC level based on the final Total Intermediate Use (TIU) matrix from the previous year's I-O tables.
Rental, hiring and real estate services, except ownership of dwellings

EAS ANZSIC class-level data are mapped to primary IOPCs according to historical weights. Any adjustments made during the S-U balancing process are split back to IOPCs based on the original IOPC proportions.Intermediate consumption is disaggregated to the IOIG and IOPC level based on the final Total Intermediate Use (TIU) matrix from the previous year's I-O tables.
Professional, scientific and technical services

EAS ANZSIC class-level data are mapped to IOIGs and then split into primary IOPCs according to previous balanced I-O table weights. Any adjustments made during the S-U balancing process are split back to IOPCs based on the original IOPC proportions.Intermediate consumption is disaggregated to the IOIG and IOPC level based on the final Total Intermediate Use (TIU) matrix from the previous year's I-O tables.

EAS ANZSIC class-level data are mapped to IOIGs and then split into primary IOPCs according to previous balanced I-O table weights. Any adjustments made during the S-U balancing process are split back to IOPCs based on the original IOPC proportions.Intermediate consumption is disaggregated to the IOIG and IOPC level based on the final Total Intermediate Use (TIU) matrix from the previous year's I-O tables.

Government Finance Statistics (GFS) gross expenditure data is classified according to the Government Purpose Classification (GPC). The GPC is mapped to the IOPC based on historical I-O weights and aggregated to the SUPC level. Any adjustments made during the S-U balancing process are split back to IOPCs based on the original IOPC proportions.Intermediate consumption is disaggregated to the IOIG and IOPC level based on the final Total Intermediate Use (TIU) matrix from the previous year's I-O tables.
Education and training

Government Finance Statistics (GFS) gross expenditure data is classified according to the Government Purpose Classification (GPC). The GPC is mapped to the IOPC based on historical I-O weights and aggregated to the SUPC level. EAS ANZSIC class-level data are mapped to IOIGs and then split directly into primary IOPCs except for ANZSIC class 8023 Combined primary and secondary education, which is split evenly between primary and secondary education services. Any adjustments made during the S-U balancing process are split back to IOPCs based on the original IOPC proportions.Intermediate consumption is disaggregated to the IOIG and IOPC level based on the final Total Intermediate Use (TIU) matrix from the previous year's I-O tables.
Health care and social assistance

Supply is derived using the demand-side approach which sums the intermediate consumption of health and social assistance related products and final demand (i.e. final consumption expenditure, and exports less imports). These data are compiled at the IOPC level. Any adjustments made during the S-U balancing process are split back to IOPCs based on the original IOPC proportions.Intermediate consumption is disaggregated to the IOIG and IOPC level based on the final Total Intermediate Use (TIU) matrix from the previous year's I-O tables.
Arts and recreation services
Arts and recreations services, except gambling

EAS data by ANZSIC Class level and institutional sector are mapped to primary IOPCs according to historical weights. Any adjustments made during the S-U balancing process are split back to IOPCs based on the original IOPC proportions.Intermediate consumption is disaggregated to the IOIG and IOPC level based on the final Total Intermediate Use (TIU) matrix from the previous year's I-O tables.
Gambling

Supply is derived using the demand-side approach which sums the intermediate consumption of gambling products and final demand (i.e. final consumption expenditure, and exports less imports). These data are compiled at the IOPC level. Any adjustments made during the S-U balancing process are split back to IOPCs based on the original IOPC proportions.Intermediate consumption is disaggregated to the IOIG and IOPC level based on the final Total Intermediate Use (TIU) matrix from the previous year's I-O tables.
Other services
Repairs and maintenance

EAS ANZSIC class-level data are mapped to IOIGs and then split into primary IOPCs according to historical weights and the EAS data items. Any adjustments made during the S-U balancing process are split back to IOPCs based on the original IOPC proportions.Intermediate consumption is disaggregated to the IOIG and IOPC level based on the final Total Intermediate Use (TIU) matrix from the previous year's I-O tables.
Personal and other services

Supply is derived using the demand-side approach which sums the intermediate consumption of personal and other services products and final demand (i.e. final consumption expenditure, and exports less imports). These data are compiled at the IOPC level. Any adjustments made during the S-U balancing process are split back to IOPCs based on the original IOPC proportions.Intermediate consumption is disaggregated to the IOIG and IOPC level based on the final Total Intermediate Use (TIU) matrix from the previous year's I-O tables.
Ownership of dwellings

Estimates for actual and imputed rental, consumption of fixed capital (COFC) and trade data are all compiled at the IOPC level. The output of residential caravan parks is derived using the movement in service income from the EAS data for this ANZSIC class and applied to the estimate of the previous year. Any adjustments made during the S-U balancing process are split back to IOPCs based on the original IOPC proportions.Intermediate consumption is disaggregated to the IOIG and IOPC level based on the final Total Intermediate Use (TIU) matrix from the previous year's I-O tables.
Taxes less subsidies on products

Government Finance Statistics data classified by purpose and tax type are mapped to the I-O level using historical weights and other observed taxation statistics data. Data are compiled at the IOPC level and aggregated to SUPC for SU tables.

Taxes are not adjusted during the S-U balancing process.

The most recent I-O Table’s distribution of margin by IOPC for each margin type is used to allocate each margin to a product.

Table 22.8 Disaggregation of S-U level to I-O level — GDP(E)
ItemDisaggregation method
Household final consumption expenditure (HFCE)

HFCE data are compiled according to the Classification of Individual Consumption by Purpose (COICOP) and is disaggregated to IOPC level based on balanced S-U table and previous year’s I-O table.

Any adjustments made through the S-U balancing process are applied to the appropriate IOPC based on the intelligence for the decision. If detailed information is not known, it is allocated based on previously balanced I-O tables.

Government final consumption expenditure (GFCE)

GFCE data are compiled according to the Government Purpose Classification (GPC) which is mapped to IOPC using historical input-output weights and aggregated for the S-U tables.

Any adjustments made during the S-U balancing process are applied to the appropriate IOPC based on the intelligence for the decision. If detailed information is not known, it is allocated based on previous I-O tables.

Gross fixed capital formation (GFCF)

GFCF data are compiled according to asset type by sector. With the exception of machinery and equipment, all asset types align with the IOPC. Machinery and equipment is disaggregated to the IOPC level based on previous years’ I-O tables, and converted into current-year prices using price indexes.

Any adjustments applied during the balancing process are based on an informed assessment of how and why the imbalance has occurred. In cases where the required level of detail cannot be sourced, the allocation is based on previously-balanced I-O tables or the original proportions of the correspondence.

Changes in inventories

Changes in inventories is distributed across IOPCs based on how changes in inventories is distributed across IOPCs in the latest set of balanced I-O tables.
Exports of goods and services

International trade data is mapped to the IOPC using the Australian Harmonised Export Commodity Classification (AHECC) to IOPC correspondence. Further adjustments based on Balance of Payment (BOP) data are made to the relevant IOPCs where appropriate.

Further adjustments at the IOPC level may be required for product and industry balancing purposes. All such adjustments are offset to ensure that total exports remain unchanged.

Imports of goods and services

International Trade data is mapped to the IOPC using the Harmonised Tariff Item Statistical Classification (HTISC to IOPC correspondence). Further adjustments based on Balance of Payment (BOP) data are made to the relevant IOPCs where appropriate.

Further adjustments at the IOPC level may be required for product and industry balancing purposes. All such adjustments are offset to ensure that total imports remain unchanged.

Table 22.9 Disaggregation of S-U level to I-O level — GDP(I)
Industry/ItemCompensation of employeesGross operating surplus/Gross mixed incomeOther taxes less subsidies on production
Agriculture, forestry and fishing

Data from the EAS are received at ANZSIC class level and then aggregated to IOIG. Data received at the ANZSIC division level (e.g. government data, payment in kind and fringe benefits tax) are disaggregated to IOIG level using proportions based on either general government output or EAS wages data as appropriate. Data are combined at the IOIG level initially and subsequently aggregated to SUIC.

Any adjustments made during the S-U balancing process are disaggregated to the relevant IOIG level based on industry evidence.

Gross operating surplus/Gross mixed income (GOS/GMI) is derived differently depending on the sector. For the private non-financial corporations sector and unincorporated enterprises GOS/GMI is derived from the EAS data as a residual. GOS for NPISH units and general government units is equivalent to consumption of fixed capital.

For more detail on the derivation of GOS/GMI refer to tables 11.5 to 11.10.

Any adjustments made to GOS/GMI for the private non-financial corporations sector during S-U Table balancing process are allocated to IOIGs using IOIG proportions when originally compiled.

Data are compiled at the IOIG level and aggregated to the SUIC level for the S-U tables. Government Finance Statistics data is classified by purpose and tax type, which are then mapped to the IOPC level using historical weights and other observed taxation statistics data.

Mining

Data from the EAS are received at ANZSIC class level and then aggregated to IOIG. Data received at the ANZSIC division level (e.g. government data, payment in kind and fringe benefits tax) are disaggregated to IOIG level using proportions based on either general government output or EAS wages data as appropriate. Data are combined at the IOIG level initially and subsequently aggregated to SUIC.

Any adjustments made during the S-U balancing process are disaggregated to the relevant IOIG based on industry evidence.

Gross operating surplus/Gross mixed income (GOS/GMI) is derived differently depending on the sector. For the private non-financial corporations sector and unincorporated enterprises GOS/GMI is derived from the EAS data as a residual. GOS for NPISH units and general government units is equivalent to consumption of fixed capital.

For more detail on the derivation of GOS/GMI refer to tables 11.5 to 11.10.

Any adjustments made to GOS/GMI for the private non-financial corporations sector during S-U Table balancing process are allocated to IOIGs using the output proportions.

Data are compiled at the IOIG level and aggregated to the SUIC level for the S-U tables. Government Finance Statistics data is classified by purpose and tax type which are then mapped to the IOPC level using historical weights and other observed taxation statistics data.

Manufacturing

Data from the EAS are received at ANZSIC class level and then aggregated to IOIG. Data received at the ANZSIC division level (e.g. government data, payment in kind and fringe benefits tax) are disaggregated to IOIG level using proportions based on either general government output or EAS wages data as appropriate. Data are combined at the IOIG level initially and subsequently aggregated to SUIC.

Any adjustments made during the S-U balancing process are disaggregated to the relevant IOIG based on industry evidence.

Gross operating surplus/Gross mixed income (GOS/GMI) is derived differently depending on the sector. For the private non-financial corporations sector and unincorporated enterprises GOS/GMI is derived from the EAS data as a residual. GOS for NPISH units and general government units is equivalent to consumption of fixed capital.

For more detail on the derivation of GOS/GMI refer to tables 11.5 to 11.10.

Any adjustments made to GOS/GMI for the private non-financial corporations sector during the S-U Table balancing process are allocated to IOIGs using IOIG proportions from when originally compiled.

Data are compiled at the IOIG level and aggregated to the SUIC level for the S-U tables. Government Finance Statistics data is classified by purpose and tax type which are then mapped to the IOPC level using historical weights and other observed taxation statistics data.

Electricity, gas, water and water services

Data from the EAS are received at ANZSIC class level and then aggregated to IOIG. Data received at the ANZSIC division level (e.g. government data, payment in kind and fringe benefits tax) are disaggregated to IOIG level using proportions based on either general government output or EAS wages data as appropriate. Data are combined at the IOIG level initially and subsequently aggregated to SUIC.

Any adjustments made during the S-U balancing process are disaggregated to the relevant IOIG based on industry evidence.

Gross operating surplus/Gross mixed income (GOS/GMI) is derived differently depending on the sector. For the private non-financial corporations sector and unincorporated enterprises GOS/GMI is derived from the EAS data as a residual. GOS for NPISH units and general government units is equivalent to consumption of fixed capital.

For more detail on the derivation of GOS/GMI refer to tables 11.5 to 11.10.

Any adjustments made to GOS/GMI for the private non-financial corporations sector during the S-U Table balancing process are allocated to IOIGs using IOIG proportions from when originally compiled.

Data are compiled at the IOIG level and aggregated to the SUIC level for the S-U tables. Government Finance Statistics data is classified by purpose and tax type which are then mapped to the IOPC level using historical weights and other observed taxation statistics data.

Construction

Data from the EAS are received at ANZSIC class level and then aggregated to IOIG. Data received at the ANZSIC division level (e.g. government data, payment in kind and fringe benefits tax) are disaggregated to IOIG level using proportions based on either general government output or EAS wages data as appropriate. Data are combined at the IOIG level initially and subsequently aggregated to SUIC.

Any adjustments made during the S-U balancing process are disaggregated to the relevant IOIG based on industry evidence.

Gross operating surplus/Gross mixed income (GOS/GMI) is derived differently depending on the sector. For the private non-financial corporations sector and unincorporated enterprises GOS/GMI is derived from the EAS data as a residual. GOS for NPISH units and general government units is equivalent to consumption of fixed capital.

For more detail on the derivation of GOS/GMI refer to tables 11.5 to 11.10.

Any adjustments made to GOS/GMI for the private non-financial corporations sector during the S-U Table balancing process are allocated to IOIGs using IOIG proportions from when originally compiled.

Data are compiled at the IOIG level and aggregated to the SUIC level for the S-U tables. Government Finance Statistics data is classified by purpose and tax type which are then mapped to the IOPC level using historical weights and other observed taxation statistics data.

Data from the EAS are received at ANZSIC class level and then aggregated to IOIG. Data received at the ANZSIC division level (e.g. government data, payment in kind and fringe benefits tax) are disaggregated to IOIG level using proportions based on either general government output or EAS wages data as appropriate. Data are combined at the IOIG level initially and subsequently aggregated to SUIC.

Any adjustments made during the S-U balancing process are disaggregated to the relevant IOIG based on industry evidence.

Gross operating surplus/Gross mixed income (GOS/GMI) is derived differently depending on the sector. For the private non-financial corporations sector and unincorporated enterprises GOS/GMI is derived from the EAS data as a residual. GOS for NPISH units and general government units is equivalent to consumption of fixed capital.

For more detail on the derivation of GOS/GMI refer to tables 11.5 to 11.10.

Any adjustments made to GOS/GMI for the private non-financial corporations sector during S-U Table balancing process are allocated to IOIGs directly as SUICs and IOIGs have a one-to-one relationship.

Data are compiled at the IOIG level and aggregated to the SUIC level for the S-U tables. Government Finance Statistics data is classified by purpose and tax type which are then mapped to the IOPC level using historical weights and other observed taxation statistics data.

Data from the EAS are received at ANZSIC class level and then aggregated to IOIG. Data received at the ANZSIC division level (e.g. government data, payment in kind and fringe benefits tax) are disaggregated to IOIG level using proportions based on either general government output or EAS wages data as appropriate. Data are combined at the IOIG level initially and subsequently aggregated to SUIC.

Any adjustments made during the S-U balancing process are disaggregated to the relevant IOIG based on industry evidence.

Gross operating surplus/Gross mixed income (GOS/GMI) is derived differently depending on the sector. For the private non-financial corporations sector and unincorporated enterprises GOS/GMI is derived from the EAS data as a residual. GOS for NPISH units and general government units is equivalent to consumption of fixed capital.

For more detail on the derivation of GOS/GMI refer to tables 11.5 to 11.10.

Any adjustments made to GOS/GMI for the private non-financial corporations sector during S-U Table balancing process are allocated to IOIGs directly as SUICs and IOIGs have a one-to-one relationship.

Data are compiled at the IOIG level and aggregated to the SUIC level for the S-U tables. Government Finance Statistics data is classified by purpose and tax type which are then mapped to the IOPC level using historical weights and other observed taxation statistics data.

Accommodation and food services

Data from the EAS are received at ANZSIC class level and then aggregated to IOIG. Data received at the ANZSIC division level (e.g. government data, payment in kind and fringe benefits tax) are disaggregated to IOIG level using proportions based on either general government output or EAS wages data as appropriate. Data are combined at the IOIG level initially and subsequently aggregated to SUIC.

Any adjustments made during the S-U balancing process are disaggregated to the relevant IOIG based on industry evidence.

Gross operating surplus/Gross mixed income (GOS/GMI) is derived differently depending on the sector. For the private non-financial corporations sector and unincorporated enterprises GOS/GMI is derived from the EAS data as a residual. GOS for NPISH units and general government units is equivalent to consumption of fixed capital.

For more detail on the derivation of GOS/GMI refer to tables 11.5 to 11.10.

Any adjustments made to GOS/GMI for the private non-financial corporations sector during S-U Table balancing process are allocated to IOIGs directly as SUICs and IOIGs have a one-to-one relationship.

Data are compiled at the IOIG level and aggregated to the SUIC level for the S-U tables. Government Finance Statistics data is classified by purpose and tax type which are then mapped to the IOPC level using historical weights and other observed taxation statistics data.

Transport, postal and warehousing

Data from the EAS are received at ANZSIC class level and then aggregated to IOIG. Data received at the ANZSIC division level (e.g. government data, payment in kind and fringe benefits tax) are disaggregated to IOIG level using proportions based on either general government output or EAS wages data as appropriate. Data are combined at the IOIG level initially and subsequently aggregated to SUIC.

Any adjustments made during the S-U balancing process are disaggregated to the relevant IOIG based on industry evidence.

Gross operating surplus/Gross mixed income (GOS/GMI) is derived differently depending on the sector. For the private non-financial corporations sector and unincorporated enterprises GOS/GMI is derived from the EAS data as a residual. GOS for NPISH units and general government units is equivalent to consumption of fixed capital.

For more detail on the derivation of GOS/GMI refer to tables 11.5 to 11.10.

Any adjustments made to GOS/GMI for the private non-financial corporations sector during the S-U Table balancing process are allocated to IOIGs using IOIG proportions from when originally compiled.

Data are compiled at the IOIG level and aggregated to the SUIC level for the S-U tables. Government Finance Statistics data is classified by purpose and tax type which are then mapped to the IOPC level using historical weights and other observed taxation statistics data.

Information media and telecommunications

Data from the EAS are received at ANZSIC class level and then aggregated to IOIG. Data received at the ANZSIC division level (e.g. government data, payment in kind and fringe benefits tax) are disaggregated to IOIG level using proportions based on either general government output or EAS wages data as appropriate. Data are combined at the IOIG level initially and subsequently aggregated to SUIC.

Any adjustments made during the S-U balancing process are disaggregated to the relevant IOIG based on industry evidence.

Gross operating surplus/Gross mixed income (GOS/GMI) is derived differently depending on the sector. For the private non-financial corporations sector and unincorporated enterprises GOS/GMI is derived from the EAS data as a residual. GOS for NPISH units and general government units is equivalent to consumption of fixed capital.

For more detail on the derivation of GOS/GMI refer to tables 11.5 to 11.10.

Any adjustments made to GOS/GMI for the private non-financial corporations sector during S-U Table balancing process are allocated to IOIGs directly as SUICs and IOIGs have a one-to-one relationship.

Data are compiled at the IOIG level and aggregated to the SUIC level for the S-U tables. Government Finance Statistics data is classified by purpose and tax type which are then mapped to the IOPC level using historical weights and other observed taxation statistics data.

Financial and insurance Services

There is a one-to-one mapping between the IOIG and SUIC for this industry. This means that movements at the IOIG level are consistent with those in the S-U table benchmarks.

Tables 11.1 and 11.2 provide details on the sources and methods used to compile COE for financial corporations.

For SUIC 620 Finance and SUIC 630 Insurance and Superannuation Funds, GOS is compiled from source data at IOIG level.

For SUIC 640 Auxiliary Finance and Insurance Services, GOS is derived as a residual item from EAS data, also at the IOIG level.

For more detail on the derivation of GOS/GMI refer to tables 11.5 to 11.10.

Data are compiled at the IOIG level and aggregated to the SUIC level for the S-U tables. Government Finance Statistics data is classified by purpose and tax type which are then mapped to the IOPC level using historical weights and other observed taxation statistics data.

Rental, hiring and real estate services, except ownership of dwellings

Data from the EAS are received at ANZSIC class level and then aggregated to IOIG. Data received at the ANZSIC division level (e.g. government data, payment in kind and fringe benefits tax) are disaggregated to IOIG level using proportions based on either general government output or EAS wages data as appropriate. Data are combined at the IOIG level initially and subsequently aggregated to SUIC.

Any adjustments made during the S-U balancing process are disaggregated to the relevant IOIG based on industry evidence.

Gross operating surplus/Gross mixed income (GOS/GMI) is derived differently depending on the sector. For the private non-financial corporations sector and unincorporated enterprises GOS/GMI is derived from the EAS data as a residual. GOS for NPISH units and general government units is equivalent to consumption of fixed capital.

For more detail on the derivation of GOS/GMI refer to tables 11.5 to 11.10.

Any adjustments made to GOS/GMI for the private non-financial corporations sector during S-U Table balancing process are allocated to IOIGs directly as SUICs and IOIGs have a one-to-one relationship.

Data are compiled at the IOIG level and aggregated to the SUIC level for the S-U tables. Government Finance Statistics data is classified by purpose and tax type which are then mapped to the IOPC level using historical weights and other observed taxation statistics data.

Professional, scientific and technical services

Data from the EAS are received at ANZSIC class level and then aggregated to IOIG. Data received at the ANZSIC division level (e.g. government data, payment in kind and fringe benefits tax) are disaggregated to IOIG level using proportions based on either general government output or EAS wages data as appropriate. Data are combined at the IOIG level initially and subsequently aggregated to SUIC.

Any adjustments made during the S-U balancing process are disaggregated to the relevant IOIG based on industry evidence.

Gross operating surplus/Gross mixed income (GOS/GMI) is derived differently depending on the sector. For the private non-financial corporations sector and unincorporated enterprises GOS/GMI is derived from the EAS data as a residual. GOS for NPISH units and general government units is equivalent to consumption of fixed capital.

For more detail on the derivation of GOS/GMI refer to tables 11.5 to 11.10.

Any adjustments made to GOS/GMI for the private non-financial corporations sector during S-U Table balancing process are allocated to IOIGs directly as SUICs and IOIGs have a one-to-one relationship.

Data are compiled at the IOIG level and aggregated to the SUIC level for the S-U tables. Government Finance Statistics data is classified by purpose and tax type which are then mapped to the IOPC level using historical weights and other observed taxation statistics data.

Data from the EAS are received at ANZSIC class level and then aggregated to IOIG. Data received at the ANZSIC Division level (e.g. government data, payment in kind and fringe benefits tax) are disaggregated to IOIG level using proportions based on either general government output or EAS wages data as appropriate. Data are combined at the IOIG level initially and subsequently aggregated to SUIC.

Any adjustments made during the S-U balancing process are disaggregated to the relevant IOIG based on industry evidence.

Gross operating surplus/Gross mixed income (GOS/GMI) is derived differently depending on the sector. For the private non-financial corporations sector and unincorporated enterprises GOS/GMI is derived from the EAS data as a residual. GOS for NPISH units and general government units is equivalent to consumption of fixed capital.

For more detail on the derivation of GOS/GMI refer to tables 11.5 to 11.10.

Any adjustments made to GOS/GMI for the private non-financial corporations and unincorporated enterprises during S-U Table balancing process are allocated to IOIGs using IOIG proportions when originally compiled.

Data are compiled at the IOIG level and aggregated to the SUIC level for the S-U tables. Government Finance Statistics data is classified by purpose and tax type which are then mapped to the IOPC level using historical weights and other observed taxation statistics data.

For SUIC 750 Government administration and regulatory services and 770 Public Order and Safety, data from the EAS are received at ANZSIC class level and then aggregated to IOIG. Government data are received at the ANZSIC division level and are disaggregated between these two IOIGs using their proportions of general government output.

For SUIC 760 Defence, defence wages and salaries are mapped directly to this IOIG.

Other data received or calculated at an ANZSIC division level (e.g. payment in kind and Fringe Benefits Tax) are disaggregated to IOIG level using either general government output or EAS wages data as appropriate. Data are combined at the IOIG level initially and subsequently aggregated to SUIC.

General government GOS is equivalent to the value of consumption of fixed capital on general government assets as derived in the Perpetual Inventory Method.

This division contains minimal private non-financial corporations GOS. It is derived directly from EAS ANZSIC class-level data.

For more detail on the derivation of GOS/GMI refer to tables 11.5 to 11.10.

Any adjustments made to GOS/GMI for the private non-financial corporations sector during S-U Table balancing process are allocated to IOIGs directly as SUICs and IOIGs have a one-to-one relationship.

Data are compiled at the IOIG level and aggregated to the SUIC level for the S-U tables. Government Finance Statistics data is classified by purpose and tax type which are then mapped to the IOPC level using historical weights and other observed taxation statistics data.

Education and Training

Data from the EAS are received at ANZSIC class level and then aggregated to IOIG. Data received at the ANZSIC division level (e.g. government data, payment in kind and fringe benefits tax) are disaggregated to IOIG level using proportions based on either general government output or EAS wages data as appropriate. Data are combined at the IOIG level initially and subsequently aggregated to SUIC.

Any adjustments made during the S-U balancing process are disaggregated to the relevant IOIG based on industry evidence.

Gross operating surplus/Gross mixed income (GOS/GMI) is derived differently depending on the sector. For the private non-financial corporations sector and unincorporated enterprises GOS/GMI is derived from the EAS data as a residual. GOS for NPISH units and general government units is equivalent to consumption of fixed capital.

For more detail on the derivation of GOS/GMI refer to tables 11.5 to 11.10.

Any adjustments made to GOS/GMI for the private non-financial corporations sector during S-U Table balancing process are allocated to IOIGs directly as SUICs and IOIGs have a one-to-one relationship.

Data are compiled at the IOIG level and aggregated to the SUIC level for the S-U tables. Government Finance Statistics data is classified by purpose and tax type which are then mapped to the IOPC level using historical weights and other observed taxation statistics data.

Health Care and Social Assistance

Data from the EAS are received at ANZSIC class level and then aggregated to IOIG. Data received at the ANZSIC division level (e.g. government data, payment in kind and fringe benefits tax) are disaggregated to IOIG level using proportions based on either general government output or EAS wages data as appropriate. Data are combined at the IOIG level initially and subsequently aggregated to SUIC.

Any adjustments made during the S-U balancing process are disaggregated to the relevant IOIG based on industry evidence.

Gross operating surplus (GOS) and gross mixed income is taken from 5204.0 (Total Factor Income by Industry and principal components – Current prices) at division level. General Government GOS is subtracted from the division level GOS/GMI to derive private GOS/GMI for the division. Private GOS/GMI for the division is then split by ratios based on the General Government GOS at IOIG level

Any adjustments made to GOS private during S-U Table balancing are allocated to IOIGs using IOIG proportions when originally compiled.

Data are compiled at the IOIG level and aggregated to the SUIC level for the S-U tables. Government Finance Statistics data is classified by purpose and tax type which are then mapped to the IOPC level using historical weights and other observed taxation statistics data.

Arts and Recreation Services

Data from the EAS are received at ANZSIC class level and then aggregated to IOIG. Data received at the ANZSIC division level (e.g. government data, payment in kind and fringe benefits tax) are disaggregated to IOIG level using proportions based on either general government output or EAS wages data as appropriate. Data are combined at the IOIG level initially and subsequently aggregated to SUIC.

Any adjustments made during the S-U balancing process are disaggregated to the relevant IOIG based on industry evidence.

For the Arts and recreation services division (except Gambling), gross operating surplus/gross mixed income (GOS/GMI) is derived differently depending on the sector. For the private non-financial corporations sector and unincorporated enterprises GOS/GMI is derived from the EAS data as a residual. GOS for NPISH units and general government units is equivalent to consumption of fixed capital.

For more detail on the derivation of GOS/GMI refer to tables 11.5 to 11.10.

For IOIG 9201 Gambling, gross operating surplus (GOS) and gross mixed income (GMI) is taken from 5204.0 (Total Factor Income by industry and principal components - Current prices) at division level. General Government GOS is subtracted from this to derive private GOS/GMI for the division. Private GOS/GMI by IOIG is then calculated from the division total using supply ratios.

Any adjustments made to GOS/GMI for private non-financial corporations and unincorporated enterprises during S-U Table balancing process are allocated to IOIGs automatically as the SUICs and IOIGs have a one-to-one relationship.

Data are compiled at the IOIG level and aggregated to the SUIC level for the S-U tables. Government Finance Statistics data is classified by purpose and tax type which are then mapped to the IOPC level using historical weights and other observed taxation statistics data.

Other Services

Data from the EAS are received at ANZSIC class level and then aggregated to IOIG. Data received at the ANZSIC division level (e.g. government data, payment in kind and fringe benefits tax) are disaggregated to IOIG level using proportions based on either general government output or EAS wages data as appropriate. Data are combined at the IOIG level initially and subsequently aggregated to SUIC.

Any adjustments made during the S-U balancing process are disaggregated to the relevant IOIG based on industry evidence.

For SUIC 940 Repair and Maintenance, gross operating surplus/gross mixed income (GOS/GMI) is derived differently depending on the sector. For the private non-financial corporations sector and unincorporated enterprises GOS/GMI is derived from the EAS data as a residual. GOS for NPISH units and general government units is equivalent to consumption of fixed capital.

For more detail on the derivation of GOS/GMI refer to tables 11.5 to 11.10.

Any adjustments made to GOS/GMI for the private non-financial corporations sector during S-U Table balancing process are allocated to IOIGs using IOIG proportions when originally compiled.

For SUIC 950 Personal and Other Services, GOS for private non-financial corporations is compiled at SUIC level from EAS and COFC data, and disaggregated to IOIGs using historic weights.

For more detail on the derivation of GOS/GMI refer to tables 11.5 to 11.10.

Any adjustments made to GOS/GMI for private non-financial corporations and unincorporated enterprises during S-U Table balancing process are allocated to IOIGs using IOIG proportions when originally compiled.

Data are compiled at the IOIG level and aggregated to the SUIC level for the S-U tables. Government Finance Statistics data is classified by purpose and tax type which are then mapped to the IOPC level using historical weights and other observed taxation statistics data.

Ownership of dwellings

GOS is derived as a residual at the IOIG level.

For more detail on the derivation of GOS/GMI refer to tables 11.5 to 11.10.

This S-U Industry and IOIG have a one-to-one relationship therefore any adjustments made to GOS during S-U Table balancing process are allocated to the IOIG automatically.

22.102    Once the disaggregation is completed it is necessary to reconcile (or confront) these results with additional information to ensure the distribution is coherent. The data sources, in addition to the main data sources used to compile the S-U tables, used in this reconciliation process are obtained from the following:

• ABS data:
• Australian Industry Statistics (AIS) case study data;
• Value of Agricultural Commodities Produced data;
• Mineral and Petroleum Exploration data;
• Building Approvals data;
• Engineering Construction data;
• Private Sector Construction data;
• Water Account data;
• Motor Vehicle Use data;
• Tourist Accommodation data;
• Household Expenditure data; and
• occasional publications;
• Bureau of Resources and Energy Economics;
• Australian Bureau of Agricultural Resource Economics and Sciences:
• Agricultural Commodities; and
• Australian Forest and Wood Products Statistics;
• Department of Resources, Energy and Tourism:
• Australian Petroleum Statistics;
• State Departments of Primary Industries;
• Australian Prudential Regulatory Authority;
• Reserve Bank of Australia;
• Australian Taxation Office;
• Medicare statistics;
• annual reports;
• press and internet articles;
• private industry analysis; and
• industry body commentary.

22.103    The level of detail used in the disaggregation and reconciliation processes will vary depending on the component and its data source. The aim is to use the most detailed information where available. For example, detailed information is available at the IOPC level for international trade in goods statistics. However, IOPC level detail is not available for international trade in services or Government Finance Statistics. In this case, weights from the previous year’s I-O tables are applied to derive I-O estimates.

22.104    The disaggregation (and reconciliation) process will result in imbalances at the I-O levels. These imbalances are resolved during the balancing process described in the steps that follow.

### Step 3: Alignment with the ASNA

22.105    The initial data at IOPC level is reviewed against the quality measures (see section on Quality Measures). This initial data quality assessment is designed to ensure that I-O estimates are fully aligned with the Australian System of National Accounts available at the time that the I-O tables are compiled.

22.106    Given the importance of margins in the I-O framework, detailed ratio analysis are undertaken to review the product allocation of trade margins.

22.107    Adjustments are implemented where required.

### Step 4: Industry level analysis

22.108    The I-O tables are assessed against the quality measures via an industry view and adjustments made where necessary. The industry view assesses Australian production of products by each industry and the use of each product by other industries as part of the production process.

### Step 5: Row balancing at purchasers’ prices

22.109    The supply of each product is balanced against the intermediate and final use of the product by a manual process using the data sources listed in step 2 above.

22.110    A principal edit undertaken in the compilation of I-O tables is to reconcile the flow of products through the productive process. Backward linkages are reconciled to ensure the process of products through the productive process is logical. For example, the Australian production of raw milk is compared to the intermediate use of products produced by the dairy manufacturing industry in milk and dairy product production. As a product is transformed or processed, the value of output should increase due to the increase in value added.

22.111    When all products are balanced at purchasers’ prices (PP), the data are assessed against the quality measures and adjustments made where required.

### Step 6: Allocation of margins, taxes and subsidies and derivation of the Basic Price

22.112    Once the I-O tables are manually balanced at purchasers' prices (PP), the use categories at basic prices (BP) are modelled using the basic relationship PP = BP + MTS where MTS represents Margins, Taxes and Subsidies on products.

22.113    The disposition of each margin, tax or subsidy across the use categories is driven by a number of general rules. These rules are applied to all products except those with specific characteristics (see below) in order to ensure the resulting disposition of MTS is logical and coherent. In the use categories, for example, the retail margins can only be allocated to HFCE while wholesale margins can be allocated to all use categories (including intermediate use) except GFCE and inventories.

22.114    In a number of instances, the standard disposition of MTS is overwritten by special rules. These special rules are applied to products with specific use characteristics and ensure a better allocation of MTS for these products. An example of a special rule relates to goods consumed by tourists which can attract retail margins in exports.

22.115    Margins, taxes and subsidies are independently allocated across the use categories on a proportional basis according to the general and special rules. The total MTS by product is therefore allocated to all use categories for each product, first using the PP and then on BP once derived.

22.116    The BP is derived from the PP by subtracting MTS in intermediate use and final use. Because MTS are allocated independently, the resulting BP tables are unbalanced. A constrained optimisation process is used, once large imbalances have been resolved manually, to balance products and industries in the intermediate use matrix.

22.117    The different margins, taxes and subsidies are added to the BP in a cumulative manner. In most cases, transport margins will be applied before wholesale margins which will themselves be applied before retail margins. For practical reasons the incremental disposition of MTS is not applied but the basic price is assumed to be a better allocator for MTS. This is particularly true for products attracting a lot of retail margins and for which the inclusion of these margins would distort the disposition of other MTS.

22.118    Once derived, the BP is used to allocate MTS across the use categories. Manual adjustments are made where required. This process is repeated twice in order to gradually refine the disposition of MTS and the resulting BP.

### Step 7: Produce basic and derived I-O tables

22.120    The next step is to produce basic and derived I-O tables. Quality checks are again run to assess data quality, and to check that publication standards have been applied to the tables for release.

22.121    The SNA68 transport margin adjustment is made (see section on the SNA68 transport margin adjustment at beginning of this chapter).

22.122    The alternate view tables are then prepared. The treatment applied to compile the alternate view removes the negative effect of the c.i.f./f.o.b. adjustment in imports (see section on c.i.f./f.o.b. adjustment for details).

22.123    The product details tables are then prepared.

22.124    All tables for release are assessed against quality measures and any corrections made as required.

### Step 8: Disseminate I-O tables, alternate view tables and product details publication

22.125    Publication tables are disseminated on the ABS website. The last quality check is run to review the publication production cycle, and to identify and document improvements for the following cycle

## Quality measures

22.126    Twenty quality measures are applied at the various steps in the compilation process described above. The quality measures include internal consistency checks and checks for consistency with published national account aggregates. Most of the quality measures are assessed at every gate with a few exceptions. The complete list of quality measures include:

1. all three measures of GDP are equal when aggregated from the I-O tables;
2. supply matches use in all rows (after product balancing);
3. application of the SNA68 TMA does not affect PP overall estimates;
4. the latest appropriate data for the reference period has been used;
5. there are no negative values in supply;
6. estimates match published estimates in the Australian National Accounts for the relevant reference period;
7. known problems from previous I-O cycles have been resolved;
8. c.i.f./f.o.b. adjustments have been applied correctly;
9. there are no inventories of services and no margins attributed to services products;
10. GST and retail margins are consistent with HFCE;
11. domestic exports do not exceed production unless stock is drawn from Inventories;
12. re-exports do not exceed imports unless stock is drawn from Inventories;
13. coverage and specialisation ratios are sensible and consistent with previous I-O tables;
14. supply is consistent with use for top 50 IOPCs;
15. the COE/TIU and GOS/TIU ratios are plausible;
16. inventory levels are sensible in relation to supply at basic prices;
17. distribution of supply confronted with previous cycle is plausible;
18. distribution of retail and wholesale margins is coherent;
19. supply and final use column totals have not been changed; and
20. issues identified at previous steps have been addressed.