Understanding the Different Approaches to Reporting Health Expenditure in Australia

Health care plays a central role in the lives and well-being of Australians. The Health care industry has grown as a proportion of the Australian economy over the past fifty years, at the same time as wealth and overall economic activity have increased. The industry has doubled its contribution to Australia's gross domestic product (GDP) during this period, from 5.1 per cent in 1969-70 to 10.5 per cent in 2020-21.¹ It is also one of the largest employers in Australia. The Health care industry employed 1.1 million people as of August 2021.²

Reliable statistics are increasingly important to inform the decisions of governments, health care providers and households. The Australian Bureau of Statistics (ABS) produces a variety of economic and social measures on the Health care industry. This includes health expenditure estimates in the Australian System of National Accounts (ASNA) used for economic analysis and forecasting. The Australian Institute of Health and Welfare (AIHW) produces Australia's annual national health accounts (NHA), which are used primarily for health expenditure policy analysis.

Consultation with users found some confusion as to the most appropriate statistics to use, due to an apparent lack of coherence between the statistics, and different health care and economic concepts and definitions. The ABS has partnered with AIHW to analyse the similarities and differences between health expenditure reported through the ASNA and the national health accounts reported in Health expenditure Australia (HEA). This collaboration builds upon previous exploratory work regarding the definitions and coherence of health care statistics, including a reconciliation of the two datasets at an aggregate level.³

The estimates used in this analysis are not the most current estimates. Data from the national accounts is consistent with the 2019-20 issue of the Australian System of National Accounts (ASNA). HEA data refers to the 2018-19 release of Health expenditure Australia. Both publications were released at the end of 2020.

The analysis presented in this paper provides a detailed reconciliation, following the approach developed by the American Bureau of Economic Analysis (BEA).⁴ The reconciliation focuses on individual areas of health expenditure in Australia’s health accounts, and corresponding products in the national accounts. In the context of this paper, 'reconciliation' refers to the alignment of expenditure statistics published by the ABS and AIHW from a macroeconomic perspective, and is not aimed at eliminating all discrepancies.

The purpose of the reconciliation is to transpose the ABS' classification structure onto the AIHW dataset, and to make adjustments to that dataset where appropriate. At the aggregate level, this transposition aligns the ABS and AIHW expenditure series quite closely, as well as the expenditure series for some products.

This paper investigates methods to enable comparisons to be made on a similar basis. It explains the differences and similarities between the statistical and economic concepts, methods and data sources used by both agencies. The paper then demonstrates how these differences can be resolved to enhance statistical comparability and coherence. This results in a consistent statistical narrative of expenditure on health care goods and services, enabling informed decisions in this important policy space. It also adds strength to Australia's suite of macroeconomic statistics as they relate to health output and expenditure.

Further queries on the content of this paper should be directed to:

Director, Non-market Research Section

economic.research@abs.gov.au

The ABS Privacy Policy outlines how the ABS will handle any personal information that you provide to us.

This paper provides a comparison of ABS and AIHW estimates for health expenditure, and discusses the two frameworks used to compile those estimates. The purpose of this paper is to highlight the different nature and purpose of the respective estimates, and to explain key conceptual differences between the two measures. The comparison shows how adjustments based on scope and conceptual differences can be made to ABS estimates to improve their comparability with AIHW estimates. The paper also highlights the distinction between transactions within the general government sector and final consumption expenditure in the ASNA. The remainder of the paper explains the differences between health accounts and national accounts, including scope and statistical concepts derived from the use of different frameworks.

Both agencies produce valuable statistics on health from different perspectives. They make use of different statistical frameworks, classifications and data sources in order to meet the needs of their respective users. Table 1 shows the differences in the production of health statistics between the two agencies.

HEA provides estimates of government and non-government expenditure on health, and the types of health care goods and services purchased. These estimates are presented in an expenditure matrix, in which columns show sources of funding for each area of expenditure reported in rows. This presentation means that services funded by government are shown as government purchases (e.g. residential care services for the elderly), regardless of whether the service is provided by the government  or private sector.

The ASNA focuses on the production of health care goods and services, and the consumption of those products by households, government and industry. In this view, the purchase of a product by government does not necessarily involve a non-market transaction, because governments often purchase health care goods and services from private providers on behalf of households. The two information sets are represented in Figure 1, and overlap around areas of expenditure, which correspond broadly to government and household expenditure reported against products in the ASNA.

Some economic information can be represented using ASNA methodology only. It is part of the information set on the right-hand side of the diagram. This information relates to components of production in the national accounts. An example is intermediate consumption, which refers to expenditure on products used up in the production of health care goods and services. Other examples consist of balancing items used in national accounting. A balancing item ensures that the supply of goods and services can be fully reconciled with the use of those products. It can also be used as a key indicator to assess the performance of an individual industry like Health care based on levels of production.

Section 2 of this paper explains the statistical frameworks underpinning the ASNA and HEA. Section 3 focuses on comparing health-related expenditure estimates published by the ABS and AIHW at the aggregate level, and outlines the various adjustments made by the ABS to source data. Section 4 compares the ABS and AIHW estimates for specific areas of health expenditure, and Section 5 concludes. Appendix A provides additional information about the statistical frameworks underpinning the ASNA and HEA. Appendix B provides additional information about the ABS approach to reporting health expenditure.

The ASNA and HEA both measure government and non-government expenditure on health care services, but each dataset is designed to be used for different purposes. They are compiled from a variety of data sources using separate frameworks.

The ABS uses the supply-use framework to compile estimates of health expenditure in the ASNA, consistent with the 2008 System of National Accounts adopted by the United Nations Statistical Commission (UNSC).⁵ This economy-wide framework balances detailed information on the supply of goods and services with the use of those products throughout the economy. This framework ensures the consistency of data drawn from different sources, detects weaknesses in the quality of data drawn from those sources, and reconciles differences with input data. Data sources and methods are discussed in more detail in Australian System of National Accounts: Concepts, Sources and Methods (see Chapters 9 to 11).

The supply-use tables are balanced once both sides are equal for all products (i.e. supply = use). This process involves applying a series of balancing adjustments that are informed by knowledge of the industry and an understanding of input data quality. Equating the supply and use of each product can mean that expenditure estimates in the national accounts can differ from the original source data. Other adjustments are designed to improve the scope and coverage of source data, and to align the estimates to concepts underpinning the ASNA. An example of the latter adjustment is the replacement of estimates for depreciation collected through business surveys and reported by government with modelled estimates for consumption of fixed capital (COFC). These are discussed in more detail throughout the paper.

AIHW uses an Australian framework to compile estimates of health expenditure in the national health accounts, based on the expenditure matrix. The matrix approach is used to present the same estimates in Health expenditure Australia, as discussed in Section 2. The definitions used in the NHA reflect the definitions and concepts of the System of Health Accounts framework (SHA 2011), which is the international standard for measuring health expenditure, developed by the OECD, World Health Organisation (WHO) and the European Commission. The methods used in the NHA are overseen by the Health Expenditure Advisory Committee (HEAC), which includes subject matter experts and representatives from different levels of Australian government. The NHA differs from other health system reporting in terms of scope, methods and classification systems; other systems tend to focus on specific funding programs, jurisdictions or time periods in the health system. Data sources and methods used in the NHA are discussed in more detail in Australia's national health and welfare accounts:  concepts and data sources.

The NHA have developed in response to domestic policy imperatives around how much is spent and who is spending it, rather than who is producing and supplying health care goods and services. They use a unique classification system based around area of expenditure and source of funding, rather than the core classifications of health care functions, provision and financing recommended by SHA 2011. A key feature of this classification system are intra-governmental flows, which are an important component of funding in the Australian health system. One reason for differences between the HEA and SHA frameworks is that intra-governmental flows would not be visible in a supply-use framework.

AIHW also works within the general framework of the System of Health Accounts to produce internationally comparable estimates of health expenditure. A derivation of the NHA is provided each year to the OECD and World Health Organisation (WHO) in accordance with the SHA classification used for international reporting. The use of differing classification systems produces variations in estimates for particular components of the Australian health system, although derived from the same source data. These estimates are available through the OECD Health Database, which provides comparable statistics on health and health systems across OECD countries based on the System of Health Accounts.

The purpose of this section is to compare aggregate estimates of health expenditure compiled by both agencies. Data used to compile the national accounts and HEA ultimately come from different sources, but some government agencies report similar data through different channels. Different treatments can also be applied to each dataset depending on relevant concepts and standards. Data used in this analysis are taken from the 2019-20 issue of the ASNA and the 2018-19 release of HEA.

The ABS compiles total final consumption expenditure (FCE) in the national accounts, consisting of government and household spending (which can be both imported and exported). AIHW reports total health expenditure financed by government and other sources in Health expenditure Australia, including recurrent expenditure, the medical services tax rebate and capital expenditure.⁶ Recurrent expenditure is the largest component of total health expenditure in the national health accounts. It consists mainly of expenditure on wages and salaries, purchases of goods and services, and depreciation. Recurrent expenditure is comparable to final consumption expenditure measured in the national accounts, as shown in Figure 2.

Total health spending in both datasets has increased at similar rates over time, at an annual average of seven per cent between 1998-99 and 2018-19. Differences between the two datasets remain relatively stable over time, ranging from $281 million in 1998-99 (or 0.6% of ASNA estimates) to $5.1 billion (or 2.7%) in 2018-19. Differences were less than one billion dollars in nine individual years.

The similarity of values for total health spending masks differences between lower-level estimates for government and non-government expenditure in both datasets. In 2018-19, for example, differences between government final consumption expenditure (GFCE) and recurrent health expenditure financed by government reached $10.5 billion. The HEA estimates for government expenditure on health exceed the value of GFCE estimates that year. Differences between household final consumption expenditure (HFCE) and recurrent health expenditure financed by non-government sources reached $15.6 billion in the same year. The HFCE estimates exceed the value of HEA estimates in that year. The total value of these differences nets to $5.1 billion.

Differences between product-level expenditure categories are explored in Section Four.

The ABS compiles health-related government final consumption expenditure in the national accounts using data from Government Finance Statistics, Australia (GFS). GFS estimates are based on information provided by governments and their authorities. AIHW reports total health expenditure financed by government. HEA estimates are compiled from administrative data provided directly by government departments and agencies across all jurisdictions. GFS aggregates correspond broadly to HEA estimates for total government spending on health, as shown in Figure 3.

GFS covers all government activities, including the operations of government-controlled corporations and authorities; however, the data in Figure 3 excludes intra-sector transactions between general government units. Transactions between units within the general government sector are generally removed from aggregates in GFS and ASNA; for example, funds transferred from one government department to another, or between levels of government. This process is known as consolidation, and is performed to avoid double-counting of government transactions. The data in Figure 3 therefore only reflects final health-related expenditures because intra-government transfers have been removed. Section 3.3 discusses consolidation as it applies to the general government sector in more detail.

Several adjustments to GFS data are required to arrive at GFCE. Distinctions are made between final consumption and total expenses in the national accounts. GFCE represents current expenditure incurred by general government on both individual and collective consumption goods and services. The main components of GFCE are a subset of total expenses, and include employee expenses (e.g. compensation of employees), non-employee expenses and consumption of fixed capital, less the value of goods and services sold by general government to other sectors. Non-employee expenses include intermediate consumption of goods and services, as well as government purchases of health services from private providers on behalf of households, known as social transfers in kind (STiK). The key SNA-based adjustment made to GFS data in deriving GFCE is the removal of depreciation from the GFS source data, which is replaced with estimates of consumption of fixed capital. HEA estimates incorporate depreciation instead as part of recurrent expenditure, in accordance with international standards for health accounting.

The remaining expenses of government, such as transfers other than STiK, interest payments and capital transactions, are excluded from the derivation of GFCE in the national accounts. This process is explained in Section 3.3. Transfers other than STiK are a key source of government finance reported in Health expenditure Australia, and explain some of the differences between GFCE and recurrent expenditure in the Health Expenditure Database (HED). They include current and capital transfers from the Commonwealth to other levels of government, such as grants used to fund state and territory health care services.

Some health-related expenditures are also classified to products that are not primary to the Health care industry in the derivation of GFCE. These expenditures mainly relate to the Professional, scientific and technical services industry including research (ANZSIC Division M) and the Public administration and safety industry (Division O). They will generate notable differences between ABS and HEA estimates for health. Figure 4 compares recurrent expenditure by all levels of government on health and GFCE estimates which include the SNA adjustment identified above (substituting estimates of consumption of fixed capital for depreciation).

The HEA series was higher in value than the GFCE estimates throughout the time series. Differences between the two datasets have increased over time, ranging from $1.7 billion in 1998-99 (or 5.8 % of GFCE estimates) to $10.5 billion (or 8.8%) in 2018-19. The bulk of these differences can be attributed to government spending on health products that are not produced by the Health care industry, such as public expenditure on pharmaceutical goods.

Table 2 summarises the adjustments made to derive GFCE estimates from the GFS dataset.

Some GFS data for health are classified to products covering research and public health administration, which are not consistent with the Health care industry definition. For this reason, expenses for research and public health administration are deducted from GFS source data in order to derive GFCE for health, as shown in this table.

Realignment of GFCE and HEA estimates

This component involves removing areas of expenditure that are not consistent with the Health care industry definition specified by ANZSIC. The first adjustment is to deduct the value of the Pharmaceutical Benefits Scheme (PBS) from GFCE and the HEA data. Pharmaceuticals are goods, and are therefore are excluded from health care services in final consumption expenditure. They are produced within the Manufacturing industry division, and the scheme is administered by the Public administration and safety industry. Under the PBS, the government subsidises the cost of listed medicines by making payments to pharmacies rather than manufacturers, because pharmacies are engaged primarily in dispensing and retailing pharmaceuticals to households.⁷ These payments represent another type of social transfer in kind, and their value is recorded as GFCE against two products which are primary to Manufacturing and the Public administration and safety industry.

The second adjustment is to remove aids, appliances and all other medications from the HEA series, because these goods are also manufacturing products. The third adjustment is to deduct the value of administration and public health from GFCE and the HEA series, because these activities are classified to the Public administration and safety industry in the national accounts.⁸ The ABS attempts to distinguish between policy-related health activities by general government (allocated to the Public administration and safety industry) and program monitoring plus management by general government (allocated to the Health care and social assistance industry division). The fourth adjustment is to remove research from the HEA data, also outside the scope of this definition. Expenditure on research is capitalised in the national accounts, and is therefore excluded from FCE in concept.

The last adjustment relates to private not-for-profit institutions serving households (NPISHs). An NPISH unit is a non-market organisation that provides goods and services to households free, or at prices that are not economically significant. NPISHs, in turn, receive a large proportion of their income from households through current transfers, including membership fees, subscriptions and donations. This relationship partly explains why NPISHs have been consolidated with households in the ASNA as one sector.⁹ NPISHs also receive current transfers from general government.

The ABS includes current transfers received by NPISHs in household final consumption expenditure, including transfers from general government units for health-related purposes. The assumption is that funds received by NPISHs from government will be spent on the provision of health-related goods and services to households. Because these same transfers are allocated to government health expenses in AIHW's Health Expenditure Database, government transfers to NPISHs have been reallocated to the GFCE dataset for the purposes of this analysis. It is important to note that all other government transfers in the GFS dataset are recorded as income by other sectors in the national accounts, and are outside the scope of government final consumption expenditure. Transfers from other sectors remain in scope of household final consumption expenditure, and are also allocated to non-government health expenses in the Health Expenditure Database.

The impact of these adjustments is shown in Figure 5, and summarised in Table 3:

Government spending in the adjusted datasets increases at similar rates over time, growing by an average of seven per cent per year between 1998-99 and 2018-19. Differences between them remain relatively stable over time, ranging from $2.9 billion (or 10.1 % of realigned GFCE estimates) in 1998-99 to $6.8 billion (or 6.1%) in 2018-19.

The ABS compiles household final consumption expenditure in the national accounts, and AIHW reports recurrent expenditure financed by individuals, private health insurance funds and injury compensation insurers in the HEA. HFCE consists of expenditure by resident households on goods and services, whether spending is undertaken within Australia or abroad. It excludes expenditure by overseas visitors in Australia. ABS estimates for household spending are compared with recurrent health expenditure financed by non-government entities in the Health Expenditure Database in Figure 6:

The HEA dataset is lower in value than the HFCE estimates throughout the time series. Differences between the two datasets have increased over time, ranging from $2 billion (or 12% of HFCE estimates) in 1998-99 to $15.5 billion (or 21.6%) in 2018-19. These estimates offset differences in government spending shown in Figure 4, where the HEA dataset is higher in value than the GFCE estimates throughout the time series. Considering this point, the value of differences is reduced for total spending in both datasets, ranging from $281 million (or 0.6% of FCE estimates) in 1998-99 to $5.5 billion (or 21.6%) in 2018-19.

HFCE is compiled from survey data. The three main data sources are: the Household Expenditure Survey (HES), Retail and Wholesale Industry Survey (RISWIS), and the Economic Activity Survey (EAS). This profile contrasts with recurrent expenditure funded by non-government sources in the HEA, which is based on a mixture of survey, administrative and estimated data.¹⁰ Several adjustments are made to source data to derive HFCE. The three main adjustments are to include estimates of net expenditure overseas; expenditures made from remote and non-private dwellings which are out of scope of the source data; and to account for "low value" imports not reflected in merchandise trade statistics. See Table 4 below for more details:

Realignment of HFCE and HEA estimates

This component involves removing areas of expenditure that are not consistent with the Health care industry definition specified by ANZSIC. The first adjustment is to deduct the value of aids, appliances, medications and pharmaceuticals from HFCE and recurrent expenditure financed by non-government sources, because these goods are produced by the Manufacturing industry. The second adjustment is to remove other health services, public health and research from the HEA series, because these services are produced by the Public administration and safety industry and Professional, scientific and technical services industry respectively. A separate adjustment is to deduct government transfers to NPISHs from household spending on health in the national accounts, offsetting the adjustment made to government spending described earlier.

Figure 7 compares realigned HFCE and HEA estimates for non-government health expenditure, and the adjustments made to produce the realigned datasets are summarised in Table 5. The HFCE estimates represent health-related expenses incurred by the household sector. The HEA estimates represent recurrent expenditure by non-government sources on health.

The chart shows that the two realigned series are much closer in level terms than the unadjusted series. Adjustments have reduced the differences between the two datasets by several billion dollars during this period, with the difference in 2018-19 falling from $15.5 to $6.1 billion. Differences now range from $263 million (or 2.7% of realigned HFCE estimates) in 1998-99 to $6.1 billion (or 13.9%) in 2018-19.

A key reason for the difference between the two datasets is the distinction between total spending on health and final consumption expenditure, as discussed in Section 3.1. The HEA estimates are based on financial flows that account for the flow of money around the Australian health care system, including Commonwealth transfers to the states and territories. The ASNA estimates do not include current and capital transfers in the derivation of GFCE, and eliminate transactions between levels of government like the GFS system. In other words, the national accounts remove expenditure that is not relevant to the derivation of GFCE. The AIHW applies a similar statistical process to offset specific intra-governmental transactions to avoid double counting of expenditures.¹¹ The expenditure estimates in these datasets are high level macroeconomic aggregates, and while these statistical processes aim to eliminate all relevant intra-governmental transactions, some transactions cannot be completely eliminated or offset owing to source data imperfections.

By illustrative example, assume that the Commonwealth government contributes $20 billion to the National Health Funding Pool, and the states and territories collectively contribute another $25 billion to the pool. There is $45 billion in the pool which can be spent on the provision of health care goods and services by the states and territories. We would overestimate the total amount spent if the Commonwealth transfers into the pool were counted in addition to when those funds were eventually spent on delivering goods and services. Removing transactions that take place within the same institutional sector, such as general government, is known as consolidation in both the ASNA and GFS, as discussed in Section 3.1. This distinction between total spending on health and final consumption expenditure is crucial for national accounting purposes, and is demonstrated in Table 6, using health related expenditure data from GFS and HEA.

In 2018-19, for example, total health related government expenditure from GFS was $164 billion, including $22.2 billion of current and capital transfers between levels governments (or grants). This amount is not part of final consumption expenditure as shown in Table 6. AIHW reported the $22 billion of Commonwealth transfers that were used to fund different areas of health expenditure, but subtracted that funding from the spending reported by the states and territories in order to derive the amount that these jurisdictions spent from their own resources. The impact of this process contributed to the lower value of state and local expenditure in the HEA compared to GFS estimates. Table 6 shows that the impact of consolidation greatly reduces the discrepancy between ABS and AIHW expenditure aggregates.

A significant portion of government spending shown in Table 6 is also recorded in other industries. An example is the final consumption expenditure on the Pharmaceutical Benefits Scheme. In 2018-19, $13.2 billion of GFCE on health was recorded against pharmaceutical manufacturing, which is undertaken by private non-financial corporations in the national accounts. The actual expenditures are made by general government to pharmaceutical companies, even though households are the beneficiaries of the PBS.

The ABS emphasis on final consumption expenditure, which excludes government transfers, is a key difference between HEA data and the GFS estimates which feed into national accounts compilation. Differences are also attributable to final consumption expenditure being reflected in industries other than Health care. Minor differences are attributable to timing differences, as well as differences in methodologies which reflect the different purposes of the respective datasets. Differences between the two datasets at the product level are explored in the next section.

The purpose of this section is to realign lower-level estimates of health expenditure compiled by both agencies.  It complements the realignment of the two datasets at the aggregate level in previous sections of the paper. This detailed realignment reflects the classification and composition of health care spending in those datasets. The lower-level estimates are recorded as areas of expenditure in the HED, and as input-output products in the national accounts. A one-to-one relationship does not exist between most products and areas of expenditure in the ABS and AIHW datasets.

This section does not discuss individual reconciliations for all areas of expenditure and their corresponding products. For example, some areas like dental and patient transport services are similar in scope and definition in both datasets. As a result, differences between the two datasets for these categories are minimal, and do not require explanations. Other areas of expenditure in the HED do not have corresponding products that are in scope of final consumption expenditure for health in the national accounts. They include administration, research and public health. The ASNA has products for the first two categories; however, those products do not exclusively measure administration and research for health care.

Another example is depreciation, which is an expense incurred by government. Depreciation is allocated to selected areas of expenditure financed by government in the Health Expenditure Database. AIHW does not publish estimates for this expense in the HEA. Depreciation corresponds to ASNA estimates for consumption of fixed capital, which is in scope of GFCE for health in the national accounts. There are two health products for COFC in the ASNA. Differences are minimal for these categories in both datasets, even though depreciation is an accounting concept, and COFC is an SNA adjustment, as discussed in Section 2. Depreciation is not discussed further in this section because this expense is not a separate area of expenditure in the HED.

It is important to note that products are classified in the input-output tables according to the industry of origin. Many goods and services are produced by more than industry, and the input-output tables record output produced by primary and secondary industries separately. The primary industry is the industry which has the largest share of output for an individual product. Health care goods and services are produced exclusively by the Health care services industry (IOIG 8401) according to the input-output tables. There is no secondary production of these products.

Health-related products are also supplied by other industries. For example, pharmaceutical goods are produced primarily by the Human pharmaceutical and medicinal product manufacturing industry (1801). A smaller share of pharmaceutical output is attributable to other parts of manufacturing. They include Veterinary pharmaceutical and medicinal product manufacturing (1802), Basic chemical manufacturing (1803) and Cleaning compounds and toiletry preparation manufacturing (1804). Pharmaceuticals are also supplied by the Wholesale trade (3301) and Retail trade (3901) industries, covering wholesalers, supermarkets and pharmacies.

An Excel spreadsheet has been released with this paper, and is available from the Data downloads section. The spreadsheet contains the classification of health expenditure in the ASNA and HED. This includes a correspondence between areas of expenditure and input-output products by industry in Tables 20.1 to 20.3.

The spreadsheet also provides the realignment of the two datasets at the aggregate expenditure level, and the detailed realignment of all products and areas of expenditure, inclusive of spending on health care goods and services. It shows source data and total adjustments for all categories in the GFCE and HFCE datasets, and published totals for the HED. There is no breakdown available for source data and associated adjustments in the HEA dataset. The reason is that AIHW treats all adjustments as edited source data, even modelled estimates.

A list of the categories used in this realignment are shown in Table 7. It is ordered by areas of expenditure, and the categories that are not in scope of the ANZSIC Health care industry definition are marked with asterisks.

Table 8 shows the top five contributors to recurrent health expenditure by source of funding according to the HED. Hospitals represent the largest single contributor throughout the time series. In 2018-19, this area of expenditure contributed 48% to total expenditure funded by government sources, followed by referred medical services (11.6%) and pharmaceutical goods (8.5%). Hospitals contributed 31% to expenditure funded by non-government sources the same year, followed by pharmaceutical goods (22%) and dental services (14.6%).

Table 9 shows the top five contributors to government and household final consumption expenditure as derived from the national accounts. Hospitals, residential care services for the elderly and pharmaceutical goods are the largest contributors to final consumption expenditure throughout the time series. In 2018-19, residential care services for the elderly contributed 28% to government final consumption expenditure, followed by hospitals (26.8%) and referred medical services (13.5%). Hospitals had represented the largest single contributor between 2009-10 and 2017-18. Pharmaceutical goods contributed 23% to household final consumption expenditure the same year, followed by hospitals (16.8%) and dental services (12.3%).

There are two series in both accounts. The HEA series includes private and public hospitals, which make up a single area of expenditure. The ASNA series includes IOPCs 84010010 Hospital services (excl psychiatric hospitals) and 84020010 Psychiatric hospital services. These products correspond to two ANZSIC classes (8401-8402), and are primary to the Health care industry.

The HEA series covers services provided by all hospitals other than defence force hospitals up to 2018-19.¹² Private hospitals are predominantly funded by the non-government sector. They are also contracted by state and territory governments to provide services for public patients, and sell services to individual public hospitals. Public hospitals are predominantly funded by the general government sector.

The HEA data includes some hospital services that are provided offsite, such as dialysis in the home. It excludes expenses incurred by public hospitals through the provision of secondary services, such as patient transport services and health research. These expenses are captured under their respective categories, which are inclusive of primary health care or other services. Health spending by the Department of Defence is excluded from estimates reported in the HEA.

The ASNA series includes services provided by psychiatric and non-psychiatric hospitals as primary activities. Psychiatric hospitals provide services to patients with psychiatric, mental or behavioural disorders. Non-psychiatric hospitals provide diagnostic, medical or surgical services, as well as continuous in-patient medical care in specialised accommodation. Hospitals also provide training to medical and nursing staff as a secondary activity. Dental hospitals are in scope of IOPC 85310010 Dental services.

Data sources

Data for the HEA series comes from several sources. The Commonwealth makes a significant contribution to the delivery of public hospital services through the National Health Funding Pool, while the states and territories are system managers of public hospitals.¹³ AIHW includes Commonwealth transfers as a source of government funding for public hospitals, but subtracts that funding from the hospital spending reported by the states and territories in order to derive the amount that these jurisdictions spend from their own resources.¹⁴ This treatment was discussed in Section 3.3. Transfers are not included in final consumption expenditure. State and territory health authorities provide estimates of spending on public hospital services directly through the Government Health Expenditure National Minimum Data Set (GHE NMDS). These estimates reflect expenses incurred by public hospitals through the provision of hospital services.

HEA data for private and public hospitals is sourced from the Medicare Benefits Schedule (MBS). Other data sources used in HEA for private hospitals are:

• the ABS Private Health Establishments Collection (PHEC) for the period ending 2016-17. This annual collection has ceased, with the results from the final survey published in Private Hospitals, Australia, 2016–17.
• Estimates for 2017–18 are modelled on historical data.
• Private health insurance data are provided by the Australian Prudential Regulation Authority (APRA).
• From 2018–19, the Private Hospital Data Bureau has been used to estimate the patient revenue component of private hospitals.
• The other revenue component continues to be modelled on historical data. Care should therefore be exercised when comparing private hospital spending between 2018–19 and previous financial years.

Data for the ASNA series comes from three sources.

• GFCE estimates are based predominantly on GFS data for hospital services and mental health institutions.
• HFCE estimates are based on household spending reported in the Household Expenditure Survey.
• Service income data are taken from the Economic Activity Survey.

Both GFCE and HFCE are derived from source data by applying the adjustments described in Section 3.

Total spending on hospitals has increased at similar rates in both datasets, at an annual average of six per cent between 2006-07 and 2018-19. However, annual rates of growth differ by several percentage points in individual years. Government spending is the largest component in both datasets, averaging three-quarters of total spending. The HEA dataset is higher in value than ASNA estimates throughout the time series for both government and non-government spending. Differences range from a total of $20.8 billion in 2006-07 (or 40.9% of HEA estimates) to $35.3 billion (or 44.6%) in 2018-19. A key reason for these differences is that ABS allocates a significant proportion of government expenditure on hospitals to residential care services for the elderly, which is discussed further in Section 4.6.¹⁵

Adjusting final consumption expenditure for these differences is shown in Figure 8. The key adjustment is to reallocate a significant proportion of expenditure from residential care services for the elderly to the ASNA series for hospitals. This part of the realignment is based on the original set of weights used to allocate GFS data from various purpose categories to residential care services for the elderly. The other adjustment is to reallocate current transfers to NPISHs from household spending to the realigned GFCE dataset. This adjustment nets to zero for total final consumption expenditure.

Following this reallocation, differences between the Health Expenditure Database and realigned ASNA estimates are significantly lower, ranging from a total of $902 million in 2006-07 (or 2.6% of HEA estimates) to $718 million (or 0.9%) in 2018-19.

There is a one-to-one relationship between both series. The HEA series is unreferred medical services, which is part of primary health care. The ASNA series is IOPC 85110010 General practice medical services. This product corresponds to one ANZSIC class (8511), which is primary to the Health care industry.

The HEA series covers medical services provided to a person by, or under the supervision of, a medical practitioner, but that person has not been referred to that practitioner by another medical practitioner or person with referring rights. The ASNA series covers services provided by registered medical practitioners who generally operate private or group practices in medical clinics or centres.

Some data for the HEA series is sourced from the MBS and the Department of Veterans' Affairs (DVA). Private health insurance data are provided by APRA and other insurers. ASNA GFCE estimates are based on GFS data for general medical services and community health services. Both HFCE and GFCE are derived by applying the adjustments described in Section 3. The key adjustment for HFCE consists of current transfers to NPISHs from government and private sources (see Table 2).

Total spending on unreferred medical services increased at the same rate in both datasets, growing by an average of five per cent per year between 2006-07 and 2018-19. Annual rates of growth differ by several percentage points each year; the exception is 2016-17. Government spending is the largest component in both datasets. It averages four-fifths of total spending in the HEA series, but only represents one-third of expenditure on unreferred medical services in the ASNA dataset.

The ASNA dataset is higher in value than HEA estimates throughout the time series. Differences range from a total of $3.3 billion in 2006-07 (or 32.5% of ASNA estimates) to $7.3 billion (or 37.2%) in 2018-19.

To close the difference in scope between the two datasets, a couple of adjustments can be made to FCE. The key adjustment is to reallocate a proportion of expenditure from residential care services for the elderly to the ASNA series for unreferred medical services. The other adjustment is to reallocate current transfers to NPISHs from household spending to the realigned GFCE dataset.

These adjustments, however, increase the divergence between the two series (ranging from a total of $5 billion in 2006-07 (or 43.1% of ASNA estimates) to $11.8 billion (or 49.1%) in 2018-19), as shown in Figure 9. One reason is that a large proportion of spending on community health services in the HEA series has been attributed to unreferred medical services in the ASNA dataset. This is reflected in the higher value of community health services in the HEA series, and the differences between both datasets for this area of expenditure offset the higher value of unreferred medical services in the ASNA dataset. Another reason is the size of SNA-based adjustments, which reached $1.4 billion in the latest year.

There is a one-to-many relationship between both series. The HEA series is other health practitioners, which is part of primary health care. The ASNA series includes four products:

IOPC 85320010 Optometry and optical dispensing

IOPC 85330010 Physiotherapy services

IOPC 85340010 Chiropractic and osteopathic services

IOPC 85390011 Other allied health services n.e.c.

These products correspond to four ANZSIC classes (8532-8539), and are primary to the Health care industry.

The HEA series covers medical services provided by practice nurses, chiropractors, optometrists, dieticians, physiotherapists, occupational therapists, speech therapists, audiologists, podiatrists, homeopaths, naturopaths, practitioners of Chinese medicine and other forms of traditional medicine. The ASNA series covers medical services provided by registered optometrists, physiotherapists, chiropractors and osteopaths. It also covers allied health care services not elsewhere classified.

Data sources for the HEA series include:

• MBS
• DVA
• Private health insurance data based on surveys conducted by the ABS and APRA.
• Estimates for out-of-pocket expenses are modelled.

Data sources in the ASNA are:

• GFS data, for government expenditure for outpatient services, community health, hospital services and public health.
• HFCE estimates are based on the Household Expenditure Survey.

The only adjustment made to GFCE relates to balancing. The largest adjustment made to HFCE consists of current transfers to NPISHs from government and private sources.

Total spending on other health practitioners has increased at different rates in both datasets. The HEA series grew by an average of 4.5 per cent per year between 2006-07 and 2018-19, compared to 6.2 per cent for the ASNA series. Annual rates of growth generally differ by several percentage points each year. Non-government spending is the largest component in both datasets. It averages two-thirds of total spending in the HEA series, and 55 per cent of expenditure on other health practitioners in the ASNA dataset.

The ASNA dataset is higher in value than HEA estimates throughout the time series, even though recurrent expenditure financed by non-government entities exceeded HFCE until 2015-16. Differences range from a total of $196 million in 2006-07 (or 5.7% of ASNA estimates) to $1.8 billion (or 23.4%) in 2018-19. These differences can be largely explained by the total value of balancing and national accounting adjustments.

The only adjustment able to be applied is to reallocate current transfers to NPISHs from household spending to the GFCE dataset. However, this adjustment nets to zero for total final consumption expenditure, such that there are no differences between realigned and other estimates. Both series for other health practitioners are shown in Figure 10:

There is a one-to-one relationship between both series. The HEA series is community health and other, which is part of primary health care. The ASNA series is IOPC 85990010 Other health services n.e.c. This product corresponds to one ANZSIC class (8599), and is primary to the Health care industry.

The HEA series covers non-residential health services usually provided to the community by state and territory governments. Some of these services are delivered through community health centres and outpatient clinics attached to hospitals, although the latter is captured as expenditure on hospitals in the national health accounts. This series also covers the coordination of health services elsewhere in the community. The term 'other' refers to recurrent health spending that could not be allocated to a specific category, such as spending by providers of regional health services not further defined. The ASNA series covers health care services not elsewhere classified, such as a blood bank operation or health assessment service. It does not exclusively measure spending on community health services.

Data for the HEA series is sourced from the Commonwealth Department of Health (DOH), other insurers and various state/territory departments. GFCE estimates for other health practitioners are based exclusively on GFS data for community health and public services. Household expenditure is estimated using data from the Household Expenditure Survey. Adjustments to source data are made as described earlier.

Total spending on this category has increased at different rates in both datasets. The HEA series grew by an average of 6.7 per cent per year between 2006-07 and 2018-19, compared to 8.4 per cent for the ASNA series. Government spending is the largest component in the HEA series, averaging 95 per cent of total spending on community services.

The HEA dataset is higher in value than ASNA estimates throughout the time series, even though HFCE exceeded recurrent expenditure financed by non-government entities between 2006-07 and 2018-19. Differences range from a total of $3.2 billion in 2006-07 (or 70.6% of HEA estimates) to $6.8 billion (or 64.2%) in 2018-19.

One reason for this difference is that a large proportion of spending on unreferred medical services in the ASNA dataset has been attributed to community health services in the HEA dataset. This is reflected in the higher value of unreferred medical services in the ASNA dataset, and the differences between both datasets for this area of expenditure offset the higher value of community health services in the HEA dataset. Another reason is that final consumption expenditure on community health is recorded against products other than unreferred medical services and other health services n.e.c. For example, the majority of GFS data for community health and public services is classified to general practice and specialised medical services, as well as dental and ambulance services. Balancing adjustments in the ASNA also contribute to the difference between the series.

The only adjustment that can be made is to reallocate current transfers to NPISHs from household spending to the GFCE dataset. This adjustment nets to zero for total final consumption expenditure, such that there are no differences between realigned and other estimates. Both series for community health and other expenditure are shown in Figure 11, but there is no realigned series shown in the graph because the process of adjustment nets out.

There is a one-to-many relationship between both referred medical services series. The HEA series is referred medical services, which constitutes a single area of expenditure. The ASNA series includes IOPCs 85120010 Specialist medical services and 85200010 Pathology and diagnostic imaging services. These products correspond to two ANZSIC classes (8512 & 8520), and are primary to the Health care industry.

The HEA series covers medical services where a person has been referred by a general practitioner or medical specialist. Typically, a general practitioner refers patients to specialists, allied health professionals, and pathology or radiology providers. The ASNA series covers two types of medical services. The first type includes services provided by specialist medical practitioners who generally operate private or group practices in medical clinics or centres. The second type includes pathology laboratory or diagnostic imaging services, such as body fluid analysis, ultrasound and x-ray services.

Data for the HEA series is sourced from the Medicare Benefits Schedule, and private health insurance data provided by the Australia Prudential Regulation Authority. ASNA government expenditure is estimated using GFS data for specialised medical services, community health, paramedical services and public health. HFCE estimates are based on the Household Expenditure Survey. The same adjustments are made to source data as for hospitals, referred medical services, other health practitioners, referred medical services and community health.

Total spending on referred medical services has increased at the same rate in both datasets, growing by an average of five per cent per year between 2006-07 and 2018-19. Government spending is the largest component in both datasets. It averages three-quarters of total spending in the HEA dataset, and represents two-thirds of expenditure on referred medical services in the ASNA dataset.

The ASNA dataset is higher in value than HEA estimates throughout the time series. Differences range from a total of $1.8 billion in 2006-07 (or 15.1% of ASNA estimates) to $2.7 billion (or 11.8%) in 2018-19.

There are two main reasons for the difference in the series. In-hospital MBS services (except for dental and optometry) are mainly allocated to non-referred medical services in HEA, because this is not identified as occurring in a public or private hospital. The benefit paid is attributed to the Commonwealth government, while the fee charged minus benefit paid is attributed to individuals. Spending by private health insurance funds on in-hospital medical services is allocated directly from the data supplied by APRA, and the amount is offset from individual referred medical spending. As a result, spending by the Commonwealth, individuals, and private health insurers on public and private hospital services is under-estimated in HEA. Other reasons for series divergence are national accounting adjustments and adjustments made during balancing.

The key adjustment applied to FCE is to reallocate a proportion of expenditure from residential care services for the elderly to the ASNA series for on unreferred medical services. The other adjustment is to reallocate current transfers to NPISHs from household spending to the realigned GFCE dataset. This adjustment increases the level of differences between government spending in both sets of estimates.

Differences range from a total of $2.5 billion in 2006-07 (or 19.8% of ASNA estimates) to $4.5 billion (or 18.3%) in 2018-19. The series for referred medical services are shown in Figure 12:

The ASNA series is IOPC 86010010 Residential care services for the elderly (aged care). This product corresponds to one ANZSIC class (8601), and is primary to the Health care industry. It measures expenditure on health services provided to residents of aged care facilities, even though aged care typically combines different types of services that correspond to several ANZSIC classes. These health services are provided either on-site or in specialised medical facilities such as psychiatric and non-psychiatric hospitals.

There is no corresponding series in the national health accounts for residential aged care services. AIHW explicitly measures spending on long-term residential aged care, but treats this area of expenditure as welfare rather than health.¹⁶ The scope of this series is different to the ASNA series, which focuses exclusively on the health components of residential aged care.

ASNA data comes from two sources. GFCE estimates are based predominantly on hospital services data sourced from government finance statistics. Other GFS categories mapped to this product include general practice and specialised medical services, as well as paramedical services. Prior to 2015, HFCE was estimated from the Household Expenditure Survey. The 2015-16 Economic Activity Survey has been used to determine household payments for aged care residential services after 2015.

The ABS implicitly measures expenditure on residential aged care by allocating a portion of final consumption expenditure recorded against other products to this one product. A set of weights is used to reallocate GFS data from various purpose categories to residential care services for the elderly. Section 4.1 discusses this method because a significant proportion of GFS data for hospitals is reallocated to residential care services.

The national health accounts capture the same spending under different areas of expenditure. The relevant areas (a) public and private hospitals; (b) referred medical services; (c) unreferred medical services; (d) patient transport services; and (e) dental services. Other spending is included in HEA estimates for government-financed expenditure on health, particularly activities undertaken by non-profit institutions serving households.

ASNA data comes from two sources. GFCE estimates are based predominantly on hospital services data sourced from government finance statistics. Other GFS categories mapped to this product include general practice and specialised medical services, as well as paramedical services. Prior to 2015, HFCE was estimated from the Household Expenditure Survey. The 2015-16 Economic Activity Survey has been used to determine household payments for residential aged care services after 2015.

The full value of FCE for this product has been reallocated to five areas of health-related expenditure for the purposes of this analysis, with the vast majority being reallocated to hospitals. This reallocation is based on the original set of weights used to allocate GFS data from various purpose categories to residential care services for the elderly as source data. The other adjustment is to reallocate current transfers to NPISHs from household spending to the realigned GFCE dataset. This realignment reduces the value of the ASNA series to zero for the purposes of this analysis, as shown in Figure 13:

There are two series in both accounts. The HEA series includes benefit-paid pharmaceuticals and all other medications, which are part of primary health care. The ASNA series includes IOPCs 18410010 Pharmaceutical goods for human use and 13340060 Wadding, powder puffs, pads, cotton wool, gauze and bandages. These products correspond to two ANZSIC classes (1841 & 1334), and are primary to the Manufacturing industry.

The HEA series covers two types of pharmaceutical goods. The first type includes medications listed in the schedule of the PBS and the Repatriation PBS (RPBS) for which pharmaceutical benefits have been paid or are payable. It excludes listed pharmaceutical items where the full cost is met from the patient co-payment under the PBS or RPBS. This category also includes specialised PBS medicines accessed through the Highly Specialised Drugs (HSD) Program for the treatment of chronic conditions. These medications are made available under section 100 of the National Health Act 1953.¹⁷ The second type includes medications for which no benefit is paid, such as over-the-counter medicines, private prescriptions and medical non-durables.

The ASNA series also covers two types of pharmaceutical goods. The first type includes pharmaceutical and medicinal products for human use made from both natural and synthetic sources. It also includes substances used to diagnose or monitor the state of human health. This type covers all medicines subsidised by the Australian government, as well as non-concessional medicines. The second type includes finished textile products such as woven and fabric bandages used for medical purposes.

Data for the HEA series comes from several sources. The key sources are PBS data from the Commonwealth Department of Health and data from the Department of Veterans' Affairs covering the Repatriation Pharmaceuticals Benefits Scheme. They provide records of listed pharmaceuticals for which benefits have been paid. Data for over-the-counter medicines sold at pharmacies are sourced from a private research firm, Information Resources Incorporated (IRI). Another source is the Australia Prudential Regulation Authority.

Data for the ASNA series comes from two sources. Government expenditure is estimated using GFS data for medical products, appliances and equipment. HFCE estimates are based on household spending and business sales from the Retail and Wholesale Industry Survey, which provides information regarding commodity data and trade margins paid by households.

Total spending on pharmaceutical goods has increased at similar rates in both datasets, growing by an average of 5 per cent per year between 2006-07 and 2018-19. Government spending is the largest component in the HEA dataset, averaging half of total spending on pharmaceutical goods. Non-government spending averages 55 per cent of total expenditure on pharmaceutical goods in the ASNA dataset.

The ASNA dataset is higher in value than HEA estimates throughout the time series. The two series begin to diverge from 2010-11. Differences range from a total of $2 billion in 2006-07 (or 13.4% of ASNA estimates) to $6.1 billion (or 20.7%) in 2018-19. The reasons are differences in scope and the various data sources used to estimate expenditure on pharmaceutical goods in both datasets.

Source data for recurrent health expenditure financed by government exceeded GFCE for health between 2006-07 and 2010-11. Differences between source data for both datasets reached $1.7 billion in in the latest year (or 13.5% of ASNA estimates) when GFCE exceeded HEA estimates. The reason for these differences is that the GFS dataset includes the value of the Repatriation PBS under the category of old age rather than health. HFCE is higher in value than HEA estimates for non-government recurrent expenditure throughout the time series. Differences range from $2.6 billion in 2006-07 (or 30.5% of ASNA estimates) to $4 billion (or 4.3%) in 2018-19. They imply that both datasets include a different selection of goods for this category, from non-prescription and non-concessional medications to medical sundries and related products.

There are no adjustments that can be applied to FCE, because pharmaceutical goods, like aids and appliances, are outside the scope of the Health care industry definition. Both series for pharmaceutical goods are shown in Figure 14:

Health care goods and services are produced exclusively by the Health care industry. These products are primary to the Health care industry. They include:

• Hospitals
• Unreferred medical services
• Dental services
• Other health practitioners
• Community health and other expenditure
• Referred medical services
• Patient transport services
• Residential care services for the elderly (aged care)

These products have been adjusted as part of this analysis. The attached spreadsheet provides a realignment for all eight products, but the paper has focused on a selection of these products for analysis. The series for aggregated health care goods and services are shown in Figure 15.

Total spending has increased at the same rate for categories in scope of the Health care industry definition, growing by an average of six per cent per year between 2006-07 and 2018-19.

Health-related products are also supplied by other industries. They include:

• Pharmaceutical goods
• Aids and appliances

These two products are primary to the Manufacturing industry division. They have not been adjusted as part of this analysis. Expenditure on pharmaceutical goods is significant in both ABS and AIHW datasets.

Health-related expenditure also includes:

• Public health
• Administration
• Research

These areas of expenditure do not have corresponding input-output products. They are not in scope of final consumption expenditure for health in the national accounts. Realignment is not applicable to these categories.

Administration is primary to the Public administration and safety industry, and research to the Professional, scientific and technical services industry. Public health is a heterogenous category. The various components of public health are delivered by different industries depending on the nature of individual programs. The series for aggregated health-related products and expenditure are shown in Figure 16.

Total spending has increased at different rates for health products produced by industries other than Health care, including Public administration and safety, Manufacturing, Research and Professional, scientific and technical services. Final consumption expenditure has not been adjusted in order to realign estimates in both datasets. The HEA series grew by an average of 5.1 per cent per year between 2006-07 and 2018-19, compared to 6.2 per cent for the ASNA series. 

The ABS and AIHW both produce health expenditure statistics which are used for different analytical purposes, and are underpinned by different statistical standards. By applying a series of specific adjustments, the two datasets can be more closely aligned, allowing a more consistent statistical narrative of health care provision in Australia.

This paper has addressed areas of concern raised by users around the definitions and coherence of health care statistics in Australia. This paper has explained the differences and similarities between the statistical frameworks used by both agencies, and demonstrated how differences in the data can be resolved through adjustments to enhance their coherence. The two measures provide a broadly consistent narrative of health expenditure in Australia, and a coherent set of health care statistics for use in policy making.

The ABS and AIHW welcome feedback from readers of this paper.

This section discusses the approach used by the AIHW to report health expenditure. It describes the features of the national health accounts and the System of Health Accounts used to compile estimates of health expenditure.

National health accounts

AIHW produces Australia's national health accounts annually to describe financial flows associated with the consumption of health care. It publishes the annual report, Health expenditure Australia which defines health expenditure as monetary spending on health goods and services, including investment in equipment and facilities. The NHA data published in HEA provide estimates of expenditure on health care goods and services and related investment, distinguishing between recurrent and capital expenditure.¹⁸

The NHA definitions reflect the definitions and concepts of the System of Health Accounts, but is not fully consistent with the SHA framework (OECD, Eurostat & WHO 2011). The national health accounts exclude expenditure that could be considered to have a 'health' outcome (or indirectly impact health) but is incurred outside of the health sector. Examples include expenditure incurred in educating and training health practitioners; providing long-term care in a residential aged care setting; increased production of nutritious food, road safety, or law and order.

The primary purpose of the HEA report is to measure total health expenditure in a comprehensive and consistent way, facilitating analysis of spending by governments and others across different areas of expenditure. The publication shows the sources of funding for each area of expenditure. For example, it reports the amount of net expenditure financed by government (Australia, state and territory); individual out-of-pocket payments; private health insurance funds; and other non-government sources, such as insurers providing injury cover. The HEA does not include all types of government expenditure, such as health-related expenditure by the Australian Defence Force and corrective services institutions. It also excludes some expenditure by non-government health organisations, such as the Heart Foundation and Diabetes Australia. Appendix C of HEA details the data sources used by AIHW.

The AIHW is responsible for reporting health expenditure data to the OECD and the World Health Organisation (WHO) for compilation of global SHA estimates. There are, however, some aspects of the SHA which Australia does not report on due to data deficiencies, even though expenditure does occur in these areas. For example, the AIHW has integrated expenditure on long-term aged care into residential aged care facilities, and cannot currently separate out the long-term health care component.

System of Health Accounts

The OECD's System of Health Accounts is the international standard for measuring health expenditure. SHA 2011 is the first revision to the original edition (SHA 2000). It provides a framework for analysing health care systems from an economic point of view. Annex B of the manual describes its relationship to the central framework in the 2008 System of National Accounts. The manual was developed by the OECD, WHO and the statistical office of the European Union, Eurostat. It is currently used as a basis for a joint data collection by these agencies, including expenditure data provided by AIHW for the Australian health system.

SHA is a tri-axial system, meaning that what has been provided and financed is consumed. The system therefore provides an accounting framework to detail monetary flows around the core classifications of health care functions, provision and financing. This is consistent with SNA because a key assumption of the central framework is that the same amount of a product entering an economy in an accounting period must be consumed elsewhere in the economy. The emphasis placed on each classification for each country reflects the structure of its health care system, which enables countries to focus on specific areas of health policy while continuing to support international comparisons.

The SHA manual includes a set of comprehensive, consistent and flexible accounts to meet the needs of analysts and policymakers. These accounts constitute a common framework for enhancing the comparability of monetary data over time and across countries, with links to non-monetary indicators. They therefore enable the performance of the Australian health sector to be compared to the health accounts of other nations. Nonetheless, the measurement of health expenditure differs greatly between countries, reflecting data sources and the format in which the data is structure, collected and reported. There are situations where the NHA cannot be mapped to the SHA framework, and so has not been reported.

Scope and framework of SHA

The starting point is the consumption of health care goods and services by resident households, including imports supplied by the rest of the world. This approach influences the structure of the classifications because final consumption is given priority over production. The boundary of health care expenditure is defined 'according to the functional purposes of the spending, with the boundaries being based on the categories set out in the health care classification' (SHA 2011, page 232).

The three core classifications relate to health care functions, provision and financing. The Classification of Health Care Functions (ICHA-HC) refers to groups of health care goods and services consumed by populations with a specific health purpose. The Classification of Health Care Providers (ICHA-HP) refers to the providers of health care products on the supply side, such as direct care, financing, primary care and system administration. The Classification of Health Care Financing Schemes (ICHA-HF) refers to the financing schemes of health care provision, including those administered by governments, insurance agencies, individuals and out-of-country donors.¹⁹

Key users/uses

Over 40 countries report health expenditure data to international agencies using the SHA framework. They include both OECD and non-OECD countries. Their data is used extensively in evidence-based policymaking to track policy effectiveness, and for the purposes of guiding health sector reform. The usefulness of the SHA framework is contingent on national compliance, as well as the availability, quality and accuracy of the data reported to international agencies. Australia reports data for the majority of expenditure and financing categories listed in SHA questionnaires to international agencies. Data is reported in both current and constant prices, and as a share of GDP.

The Australian national accounts are compiled from numerous data sources, and are underpinned by the supply-use framework. They reconcile three measures of GDP via this framework using the production, expenditure and income approaches. The ABS also works within the general framework of the System of National Accounts to produce internationally comparable estimates of GDP.²⁰ It uses the SNA framework to compile other measures of economic activity such as the contribution of an individual industry to GDP. An example is Australia's health care industry, with the compilation of production and expenditure statistics for health care goods and services (see Chapter 1 of the 2008 SNA).

This relationship ensures that the performance of the Australian economy (including health care) can be compared to the accounts of different nations which have been compiled according to the same statistical standards. International comparability is enhanced by use of the classification systems recommended by SNA, or other systems aligned with them. For example, to measure industry statistics, ABS uses the ANZSIC, which is consistent with the International Standard Industrial Classification of All Economic Activities (ISIC). The relevant industry for this paper is ANZSIC Division Q, which includes the provision of both health care and social assistance.

Scope of expenditure estimates

This section discusses the scope of government and household spending in ABS statistics. This discussion will help to understand the differences and similarities between the statistical and economic concepts, methods and data sources used by the ABS and AIHW.

Household final consumption expenditure

Household spending is classified by purpose or function using the Classification of Individual Consumption by Purpose (COICOP) system. Health is one of twelve COICOP categories used in the ASNA, covering a wide range of goods and services consumed by the household sector. It consists of three sub-categories:  (a) medicines, medical aids and therapeutic appliances; (b) ambulatory health care; and (c) hospital, ambulance services and nursing home care. Transactions in scope of HFCE also include the value of goods produced by households for their own consumption. The scope of household spending extends to NPISHs, which are part of the household sector in Australia's national accounts.²¹

A number of adjustments are made to the source data for scope and conceptual alignment to the SNA, as described throughout this paper. They include the following:

(a) remote and non-private dwellings which are not in scope of survey data;

(b) grants from government to NPISH units sourced from GFS;

(c) household claims associated with the health service component of workers’ compensation and motor vehicle and third party insurance sourced from APRA;

(d) household expenses associated with nursing home fees which are not in scope of the HES; and

(e) net expenditure overseas.

Government final consumption expenditure

Government spending is classified by purpose or function according to the Australian Classification of the Functions of Government (COFOG-A). Health is one of eleven COFOG-A categories in the GFS system. The government is considered to be the consumer of its own output (GFCE) because services are provided to the community free of charge, or incur charges which cover only a small proportion of costs (i.e. economically insignificant prices). An example is the delivery of some health services to households. Government output is therefore valued at its cost of production in the Australian System of National Accounts. The difference between government output and government final consumption expenditure is the value of goods and services sold by general government to other sectors.

In the core set of national accounts, social transfers in kind are treated as government final consumption expenditure and not household final consumption expenditure, even though it is the household that directly benefits. Social transfers in kind are individual goods and services provided to individual households by general government units and non-profit institutions serving households. Those products may be produced by the same units or purchased by them. Also included are reimbursements made to individual households by government units and NPISHs under a scheme that authorises purchases of approved goods and services. Examples include reimbursement for the costs of pharmaceuticals purchased under the PBS and Medicare rebates for medical services.

Transactions in scope of GFCE include net final expenditures for recurrent purposes only. A number of adjustments are made to the source data for conceptual alignment to the SNA, as described throughout this paper. They include financial intermediation services indirectly measured (FISIM) that are consumed by the general government sector, and general government consumption of fixed capital, which corresponds to the accounting identity of depreciation.²²

See Chapter 10 of Australian System of National Accounts:  Concepts, Sources and Methods for a detailed discussion of final consumption expenditure.