5310.0.55.002 - Information Paper: Implementation of new international statistical standards in ABS National and International Accounts, September 2009  
ARCHIVED ISSUE Released at 11:30 AM (CANBERRA TIME) 28/10/2009  First Issue
   Page tools: Print Print Page Print all pages in this productPrint All

CHAPTER 6 RESEARCH & DEVELOPMENT

Warning: The estimates in this publication are indicative. They are presented to give an indication of the magnitude of the impacts of proposed change to Australia's macro-economic accounts. All estimates are subject to refinement and revision in the compilation of the annual Australian System of National Accounts to be published on 8 December 2009 and in the quarterly Australian National Accounts to be published on 16 December 2009.

INTRODUCTION

A major change in the System of National Accounts 2008 (SNA08) is the recognition of expenditure on Research and Development (R&D) as capital formation. The ABS will implement this change in the Australian System of National Accounts (ASNA). This chapter explains the conceptual and practical issues involved and the effects the new treatment will have on GDP, capital formation and other national accounts aggregates.

The System of National Accounts 1993 (SNA93) acknowledged that expenditure on R&D is in the nature of capital expenditure. However it recommended that it be treated as intermediate consumption because "...expenditures on R&D do not lead to the creation of assets that can readily be identified, quantified and valued for balance sheet purposes" (SNA93, para 1.51). Since that time work has been done, in the ABS and elsewhere, to measure R&D assets. The results of ABS work indicate that it is feasible to use data on R&D expenditure in conjunction with information on the life of patented R&D discoveries to produce meaningful estimates of the stock of R&D assets. An ABS paper Capitalising Research and Development in the National Accounts (ABS, 2004), included experimental estimates and concluded that "there are no insurmountable methodological problems to the capitalisation of R&D in Australia". However, there are conceptual and practical issues as to how production, trade in R&D services, capital formation and the ownership of the resulting assets are defined and measured. The first section of this chapter examines those issues and the second section presents estimates of the stock of R&D assets, and examines the effects of the changed treatment of R&D on national accounts aggregates.

SNA08 recommends that the value of R&D should be determined in terms of the future economic benefits it is expected to provide. The definition includes the provision of public services in the case of R&D assets acquired by government. In principle, R&D expenditure that does not provide an economic benefit to its owner does not constitute a fixed asset and is treated as intermediate consumption or compensation of employees.


CONCEPTUAL ISSUES

The nature of the R&D asset

ABS R&D surveys collect data on business expenditure on research and development (BERD), higher education expenditure on research and development (HERD), government expenditure on research and development (GERD), and private non-profit expenditure on research and development (PNPERD) in respect of four types of R&D activity: pure basic research, strategic basic research, applied research and experimental development. These four surveys will be referred to collectively as the R&D survey for the remainder of this paper. Pure basic research is defined as experimental and theoretical work undertaken to acquire knowledge without looking for long term benefits other than the advancement of knowledge. Strategic basic research is research undertaken into broad areas in the expectation of making useful discoveries, applied research is work undertaken to acquire new knowledge with a specific application in view, while experimental development is work undertaken, using existing knowledge, for the purpose of creating new or improved products or processes.

SNA08 defines assets as follows: "Assets as defined in the system are entities that must be owned by some unit, or units, and from which economic benefits are derived by their owners by holding or using them over a period of time" (SNA08, para 1.46). R&D activity resulting in patented entities will meet those criteria. However, patents are not taken out in respect of all R&D and, in cases where they are not, it is less clear whether the activity will result in the production of assets. The act of protecting an R&D product, by whatever means, indicates that the owner expects it to continue to be of value into the future and wants to retain control of the product. Apart from legal protection, R&D can be protected by less formal means, for example the complexity of a product or frequent rapid product changes. The data on R&D collected by ABS in the Survey of Research and Experimental Development respondents are not asked to identify the various types of protection afforded to R&D activity.

The shaded area in table 1 shows the scope of R&D to be capitalised in the Australian national accounts. Pure basic R&D will be excluded, except in the case of R&D expenditure by businesses, while exports of R&D will not result in capital expenditure in Australia.

1 Relationship between the survey data and SNA sectors

Corporations Private non-profit institutions General Government Exports Imports

Pure basic research n.a. n.a. n.a. n.a.
Strategic basic research n.a.
Applied research n.a.
Experimental development n.a.


n.a. = not applicable.

International opinion is divided as to which types of research result in the creation of assets as defined in the SNA. The ABS has taken the view that pure basic research undertaken by government (including higher education) and private non-profit organisations is generally freely available and should therefore not be treated as capital formation. However, the ABS assumes that pure basic research carried out by private corporations would not be undertaken if it did not result in a continuing benefit to corporations.

Within the classes of assets that are identified in SNA08, R&D falls in the category of intellectual property products, along with mineral exploration and computer software. Unlike mineral exploration, R&D typically has multiple uses and spillover benefits flow to other economic units and to society generally. In this respect R&D expenditure, and knowledge information generally, exhibits public goods characteristics (Moris, 2008). The value of R&D expenditure is often embodied in other products in a similar way to other business services.

Most R&D is produced on own account as secondary output by a wide range of industries. In the case of own account production the suppliers are also the users (in R&D survey terminology, the performers are also the funders) and the market value is not observable. In these circumstances, SNA08 says that "...it may by convention, be valued at the sum of costs, including the cost of unsuccessful R&D" (SNA08, para 10.103).


Scope

Data on research and experimental development published by the ABS from the R&D surveys are defined in accordance with the OECD standard set out in the Frascati Manual 2002: Proposed standard practice for surveys of research and experimental development. R&D is defined as comprising "creative work undertaken on a systematic basis in order to increase the stock of knowledge, including knowledge of man, culture and society, and the use of his stock of knowledge to devise new applications" (Frascati, 2002). The scope of R&D in the Frascati Manual is broader than the treatment in SNA08, in that it includes capital expenditure on plant and machinery and construction works intended for use in the production of R&D services. For example the Large Hadron collider. The Frascati definition of research and [experimental] development, has been adopted in SNA08. However, the SNA definition excludes expenditure on machinery and equipment and construction works used for R&D purposes as those items are classified as expenditure on tangible assets under the relevant product groups. The estimates exclude elements of R&D expenditure reported by government and higher educational bodies, but believed to be carried out by units operating outside the general government sector. Further investigations will be carried out by the ABS to reconcile reporting of R&D expenditures in the government finance statistics and R&D surveys for general government and higher education institutions.

In estimating the amount of R&D expenditure to be capitalised, there is a potential problem of overlaps between R&D and other recognised intellectual property products, in particular computer software and databases, which are recognised as assets in the national accounts. SNA08 does not explicitly discuss the boundary issue between R&D and computer software and databases. However, it states that "Gross fixed capital formation in computer software includes both the initial development and subsequent extensions of software and acquisition of copies that are classified as assets." (10.110) which suggests that an element of R&D should be included in the software estimates. This approach is taken by Galinda-Rueda in Developing an R&D Satellite Account for the UK (2007). Galinda-Rueda gives precedence to computer software, on the grounds "...lack of auxiliary information, the complexity of the interdependencies and the wish to avoid revising the software estimates". In ABS estimates, R&D expenditure by the computer systems design and related services industry group is identifiable, and will be classified as R&D, while software R&D reported by other industries will be treated as capital expenditure on computer software.


Ownership

Paraphrasing Galinda-Rueda (2007), in the national accounts, the owner of an asset is presumed to be its beneficiary, both because of an ability to use the asset in production and because of an ability to transfer the asset to another party. However, with knowledge, the link between owner and beneficiary is less clear-cut than with tangible assets. The question of who benefits from R&D presents two conceptual problems; spillovers and joint ownership. It also presents the practical issue of deciding whether to assign R&D ownership to the performer or the funder.

The ABS R&D surveys collect data based on recommendations set out in the OECD Frascati Manual. Unlike most types of capital expenditure, for which data are collected from purchasers or users, the surveys of R&D expenditure collect data from performers, which also supply information on sources of funds. For practical purposes the ABS assumes that funders of R&D are the users and owners of the resulting R&D assets.

In the Australian context more than 90 percent of business R&D assets are produced on own account. In the case of own account production the question of ownership does not arise, as the performer is also the funder, and is clearly the owner of the asset. Where a performer purchases R&D services from an outside supplier, it will be assumed that those services are an intermediate input into own-account production of an R&D asset. Because the survey data are collected from R&D performers, who are instructed to report only expenditure on R&D performed by their own organisation, intermediate transactions are not directly observable in the survey data. The magnitude of intermediate transactions can, however, be inferred from the data on external sources of funding. It follows that R&D expenditure funded from overseas represents exports of services to an overseas purchaser. This includes the case of multinational corporations, where R&D performed in one country benefits affiliated units in other countries and those units are invoiced for R&D services. In the event that an R&D performer sells the resulting asset, the transaction will be treated as the sale of a second-hand asset.


Spillovers

From the production of intellectual property, benefits may accrue to units other than the owners of the products. The knowledge gained in the process may stimulate the production of other intellectual property products by other units. SNA08 gives examples of such spillovers as a breakthrough in the development of a new class of drug leading other enterprises to develop competing drugs of the same type, and the success or failure of mineral exploration in a particular zone informing other units with exploration rights in a neighbouring zone. These are treated in the same way as other externalities in the system. Unless there is an observable monetary transaction between parties, nothing is recorded in the accounts.

The suppliers of R&D products may also be beneficiaries of the R&D and continue to have a stake in the intellectual property. The production will make a direct contribution to GDP of the country in which they are produced, but no fixed capital formation takes place in the country of production if the owners are resident in another country. The producers of the products may retain the knowledge of the processes and may use that knowledge again. However that can be regarded as a spillover benefit in the form of development of human capital (which is not treated as fixed capital formation in the SNA). The performance of R&D requires a highly skilled/educated workforce. The maintenance and retention of that skillbase may require a substantial ongoing investment in human capital.


MEASUREMENT CHALLENGES

Valuation, prices and volumes

In principle, R&D output is valued at market prices. However, survey data indicate that over ninety per cent of research and development activity is undertaken on own account, and representative market price data for R&D products are not available. In cases where it is not possible to reliably estimate market prices, the SNA08 recommends that "...output should be valued by the total production costs incurred, including consumption of fixed capital, plus any taxes (less subsidies) on production other than taxes or subsidies on products, plus a net return on the fixed capital and natural resources used in production" (SNA08, para 6.91).

R&D is produced as a secondary activity by a broad range of industries. The data collected from R&D performers in the Survey of Research and Experimental Development Businesses, are reported on a cost basis, with costs broken down into labour costs, other current expenditure, land, buildings and other structures, and other capital expenditure. R&D will therefore be valued at cost, and input deflators will be used to calculate volume measures.

The volume of capital expenditure on R&D will be calculated by deflating the cost based expenditure values for labour and other current expenditure with input deflators. The current price estimates are deflated using deflators for labour inputs, derived from unit labour costs and other current expenditure will be deflated using a fixed weight index covering a range of items used as inputs into R&D products.

The measurement of the stock of R&D assets requires data on the flow of R&D expenditure, in volume terms, estimates of the life span of the various types of R&D assets and an estimate of the retirement distribution pattern of those assets as they become obsolete and leave the capital stock.


Trade in research and development services

The R&D survey does not explicitly collect data on international trade in R&D services. However, survey data are classified by funder, and the ABS assumes that expenditure funded from overseas sources constitutes exports of services. Because the R&D survey collects data from performers of R&D, not users, they do not explicitly cover imports of R&D products. Although expenditure on imports by R&D performers are captured by the R&D survey as part of the cost of performing R&D, they do not capture expenditure on R&D by non-performers of R&D, which fall outside the scope of the R&D survey. However, both exports and imports of research and development products are collected in the Survey of International Trade in Services (SITS). The SITS data in respect of exports are of similar magnitude to the data on overseas funded R&D services from the R&D surveys, suggesting that coverage is comparable.

International evidence suggests that a significant proportion of trade in R&D products is between resident and non-resident affiliated corporations, and around one third of importers of R&D are also exporters of R&D. The SITS covers both exports and imports of R&D services, and those data have been used for estimates of trade in the national accounts in preference to data from the R&D survey, which cover only externally funded R&D services.

The results from the trade survey of international trade in services are consistent with the results from surveys conducted in the Netherlands. The issue of trade in R&D services is discussed in "R&D Satellite Accounts in the Netherlands" (OECD, 2007) where a large proportion of private R&D activity is concentrated in a small number of foreign affiliated corporations. The Netherlands has found it difficult to determine the amounts of intragroup R&D capital service flows to and from the rest of the world, and this has complicated the estimation of domestic R&D investment.

In the Netherlands, it is assumed that R&D funded by overseas entities represents exports. However, the Netherlands R&D survey (and the ABS R&D survey) does not explicitly ask for R&D sales and purchases, or for imports and exports of R&D services. The Dutch consider that the focus on R&D performers may lead to an under-reporting of imports. In addition they consider that R&D may be transferred between multinational companies without countervailing money flows, so that even if sales information was explicitly requested it is questionable whether all exchanges of R&D services between affiliated enterprises would be reported. It is therefore unclear the extent to which knowledge capital accumulates in the domestic economy.

For practical purposes, it is assumed that imports of R&D products are by R&D performers, and that they will be used as intermediate inputs into own account R&D. It is assumed that R&D funded by non-residents represents exports of R&D services.


IMPLEMENTATION

Sourcing and adjusting data

The principal sources of R&D data are the surveys of research and experimental development conducted by the ABS. There are four surveys: the annual survey of business expenditure on research and development (which accounts for over 60 per cent of total R&D expenditure), and biannual surveys of private non-profit, government and higher education expenditure on research and development. Data on imports and exports of R&D services are collected in the ABS Quarterly Survey of International Trade in Services.

2 Relationship between the survey data and SNA sectors

Frascati Manual Source SNA sector

Business enterprise sector Research and experimental development businesses (ABS cat. no. 8104.0) Non-financial corporations
Financial corporations
Private non-profit sector Research and experimental development, government and private non-profit organisations (ABS cat. no. 8109.0) NPISH
Households
Government sector Research and experimental development, government and private non-profit organisations (ABS cat. no. 8109.0) General government
Higher education sector Research and experimental development, higher education organisations (ABS cat. no. 8111.0)General government
Non-resident sector Survey of International Trade in ServicesRest of world




Converting survey data from a Frascati to a national accounts conceptual basis.

The Frascati based ABS R&D survey data are published by sector, type of expenditure, location (state), socioeconomic objective and research field. For national accounts purposes the principal requirements will be for data classified by industry, sector and state, and special extractions from the R&D data sets will be required to provide data on that basis to the desired level of industry detail. Data on the source of funds, particularly in respect of businesses, will be required in order to identify the own account component of expenditure.

The Frascati treatment of R&D is broader than the treatment in SNA08. The Frascati Manual uses a purpose definition, which includes capital expenditure on plant and machinery and construction works intended for use in R&D, whereas SNA08 includes those items as expenditure on tangible assets under the relevant product groups. SNA08 includes R&D on computer software along with other expenditure on software as a separate class of intangible assets. For consistency with the coverage of the core accounts, those elements which are already included elsewhere will be excluded from the R&D estimates to avoid double counting.

Frascati based data will be converted to an SNA08 basis using bridging tables, which are set out in three blocks, namely output, exports and imports, and gross fixed capital formation. The following adjustments will be made to the Frascati Manual measure of gross expenditure on research and development:

Output

Output data reported in the R&D surveys will be converted to a national accounts basis by adjusting expenditure reported by performers. The following adjustments and assumptions are made:

Changes in inventories of materials to be used as input for R&D

The Frascati Manual recommends adjusting reported expenditure on R&D to allow for changes in stocks of materials used as inputs its production. However, changes in inventories for service industries will not be significant in the context of the ASNA, and as most R&D is performed as a secondary activity by a range of industries, changes in materials used for R&D purposes are subsumed in the changes in inventories of the performing industries. Changes in inventories will not be estimated for R&D products, and it is assumed that changes in respect of inventories attributable to R&D products are zero.

Acquisition of R&D to be used for R&D

In principle R&D services used in the production of R&D products will be treated as intermediate consumption, however the source data relate only to expenditure of performers and the cost structure of individual projects are not identifiable. The survey respondents report only expenditure on R&D undertaken in respect of their own activities. This minimises the possibility of double counting of expenditure and it is not possible to separate any intermediate inputs of R&D from other current expenditure. It is assumed that imports of R&D services and identifiable domestic purchases of R&D services are part of the cost of own-account research and development.

Consumption of fixed capital

Depreciation of fixed assets owned by R&D producers and used in R&D production will be estimated in a perpetual inventory model (PIM) in conjunction with capital services, which constitute part of the cost of performing research and development. Data on capital expenditure will be fed into the model, which is specified with the estimated economic lives and retirement distributions of the various classes of assets, and consumption of fixed capital will be generated as an output of the model.

Operating surplus will be contained in R&D output

The R&D survey statistics relate to expenditure and will be broken down into labour costs, other current expenditure and capital expenditure. On that basis the value of own-account R&D output, for national accounts purposes, will be calculated as the cost of labour, other current expenditure and the implicit cost of the capital services used in the production of the R&D. In the case of market units, gross operating surplus will increase by the value of own account R&D output. In the case of non-market units, the total value of output will not change with the capitalisation of R&D, but gross operating surplus will increase by the value of COFC in respect of the R&D asset.

Other taxes on production less subsidies

There are no specific taxes or subsidies on R&D products over the period from 1968-69 to 2006-07. However, a 150 per cent tax concession for R&D expenditure was introduced in 1984-85, and that concession was followed by a strong increase in R&D activity. The tax concession was cut to 125 per cent in 1995-96 and a fall in expenditure followed. Tax incentives act in a similar way to subsidies, although they are not treated as such by the SNA.

Gross fixed capital formation (GFCF) by R&D performers

Capital expenditure on machinery and equipment, buildings and other fixed assets by R&D performers will be excluded from estimates of R&D capital formation to avoid double counting. In 2006-07 total capital expenditure on R&D, as defined in SNA08, will be about $14.5 billion, or around 5.1 per cent of total gross fixed capital formation. R&D expenditure on computer software will be included with other software expenditure (apart from R&D performed by the computer systems design industry) and will be excluded from the R&D estimates. Land is classified as a non-produced, non-financial asset in the national accounts and purchases of land are not treated as capital expenditure. Land will be included with buildings and other structures in R&D statistics and the total value of that item will be excluded from the national accounts capital expenditure estimates.

Software expenditure

R&D on computer software is already included in the national accounts as software, another category of intellectual property products. The boundary between R&D and software is not be discussed in SNA08, however as the SNA definition of software includes development, it will include an element of R&D expenditure. The ABS maintains that, in general, software R&D is a direct investment in the software product, and should be treated as a software asset. However an exception is made in the case of R&D expenditure undertaken by the computer systems design industry, which has a more direct interest in maintaining R&D capability as a separate class of assets.

Exports and imports

Exports

Exports of research and development services are recorded in the balance of payments statistics. The information is collected by sample survey, in principle it includes both sales and parent/subsidiary transactions. It is not possible to separately identify related company transactions. In 2007-08 exports of research and development services was $1,068 million.

Imports

Imports of research and development services are recorded in the balance of payments statistics. The information is collected by sample survey and it is not possible to identify related company transactions. In 2007-08 recorded imports of research and development services was $626 million. It is not be possible to determine whether R&D imports are to be used in the production of R&D products or for other purposes.

Overlap with Annual Integrated Collections

The Annual Integrated Collections (AIC) include units whose primary production is research and development services. Any sales of R&D services are included with "Income from services" and expenditure on R&D services is included in "Other operating expenses". The data are not separately identifiable in the AIC, however the R&D survey includes data of expenditure by funder. If it is assumed that R&D funded by units other than performers represents income to the performers and an expense to the funders, and the transactions are therefore captured in the AIC.

The supply and use tables will indicate that sales of such products, including R&D, mostly come from the professional, scientific and technical services industry, with little secondary output from other industries. However, R&D surveys suggests that research and development are performed by a wide range of industries. This supports the view those industries engage in R&D on their own account, and that such activity would not be captured as output or sales in AIC collections. Own account production of R&D products will be valued at cost, the costs will be intermediate consumption, compensation of employees engaged in R&D activities and the imputed cost of capital services. When R&D is capitalised, the cost of own account production of R&D products is added to the value of output of industries undertaking R&D for market units, and makes a corresponding contribution to GDP. The value of intermediate consumption derived from the AIC will not change as a result of capitalising own-account production of R&D products.

The AIC collections include any purchases of R&D services with "other operating expenses", which are generally treated as intermediate consumption. In principle all current expenditure (other than labour costs) reported in R&D surveys should be included in that item, although it is not separately identified. The value of intermediate consumption derived from the AIC should not change as a result of capitalising own-account production of R&D products.

Gross fixed capital formation at current prices

The supply of R&D products is comprised of domestic production and imports of R&D services. Domestic production, in turn, is made up of own-account capital formation and products produced for sale. Products produced for sale will be deemed to be that part of R&D expenditure funded by other businesses or from overseas. All own account output, including sales of services, will be valued at cost as reported in R&D surveys. Sales of services constitute less than 10 per cent of R&D performed, and any operating surplus in respect of that component is not identifiable in survey data. The cost of production will be estimated as compensation of employees engaged in R&D plus the value of other current expenditure and the estimated cost of capital services (including consumption of fixed capital) used in the production process. On the expenditure side of the accounts, the products will be used for capital formation and intermediate consumption by private businesses, private non-profit institutions and government and exports of services.

In the case of Government units, output will be reclassified to identify R&D. The value of that output, in terms of intermediate consumption, compensation of employees and consumption of fixed capital will not change. However, the capitalisation of R&D expenditure will create an asset, and consumption of fixed capital in respect of that asset will increase the value of government output, resulting in an increase in GDP equal to the value of COFC.

Back series

Data on R&D expenditure by businesses are available annually from 1983-84 to 2006-07. Prior to that date some data are available, on an irregular basis back to 1968-69. For the period from 1968-69 to 1983-84 estimates of expenditure will be made by interpolation. In order to provide data for capital stock estimates data series will be extended back to 1948-49 on the assumption that R&D was a constant proportion of GDP prior to 1968-69.


OTHER IMPLEMENTATION ISSUES

Preliminary and quarterly estimates

All source data will be from annual surveys which means that quarterly estimates will be made by interpolating and projecting annual estimates. Labour inputs comprise approximately 50 per cent of R&D costs, and the volume estimates will be projected based on the number of persons employed in the professional, scientific and technical services industry sub-division. Those estimates will be inflated with the labour price index for that industry division to yield current price estimates.


Prices and volumes

Price deflators and volume measures will be compiled for business, private non-profit, government and higher education using information from the respective R&D surveys. Wages and salaries and numbers of employees will be used to estimate deflators for the labour cost components and a fixed weight price index will be used to deflate other current expenditure inputs. A range of product groups used in the production of R&D products will be used in the deflator for other current expenditure. The chart below shows that over the period from 1993 to 2008 the cost of all R&D production has increased at a significantly faster rate than the domestic final demand deflator.

Figure 1 - Comparison of R&D deflator and domestic final demand - 1992-93 to 2007-08
Graph: Figure 1 - Comparison of R&D deflator and domestic final demand—1992–93 to 2007–08



Consumption of fixed capital/asset lives

The value of R&D capital depreciates over time as new innovations emerge. As this occurs, earlier R&D becomes less effective in the production process and contributes less to profitability. Because of the intangible nature of the asset, the decline in value is difficult to measure and most studies use a range of assumptions based on econometric studies or the observed retirement rates for patents. The Australian Industry Commission report on Research and Development, 1995, cites work by Mansfield (1973) and Pakes and Shankerman (1978, 1984) which suggests that industrial knowledge depreciates faster than physical capital with little left after 10 years. More recent studies (Caballero and Jaffe, 1993) have suggested that the rate of technological change, and consequently the rate of obsolescence, has increased in recent years. However, data on patent expiry rates suggest considerably longer asset lives.

Data compiled by Intellectual Property Australia show that the mean life-spans of standard patents filed in Australia between 1980 and 2001 were between 10 and 13 years. The data are categorised by 'technology group', whereas R&D expenditure data are categorised by industry (to sub-division level). There is no simple correspondence between the technology group classification and the industry classification, however, there are relatively small differences between the mean patent lives for different technology groups. Given the difficulties in producing estimates for individual industries, and the fact that the estimates (based on the patent data) do not differ greatly, a single asset life distribution will be used for all R&D in the ASNA. A mean asset life of 11.0 years has been derived from a weighted average of the patent lives of the different technology groups.

Patent lives do not necessarily represent the lives of all R&D products and, in principle, an adjustment should be made to account for the fact that not all R&D is patented. Although it seems reasonable to expect that non-patented R&D would on average have shorter lives and depreciate faster than patented R&D, empirical estimates based on econometric studies vary greatly (with some of the evidence suggesting a longer life than that estimated from patents). The ABS paper Capitalising Research and Development in the National Accounts (2004) made experimental estimates of the value of R&D assets using average asset lives of 5, 10 and 20 years. In the United States in 2007, the Bureau of Economic Analysis (BEA) tested four scenarios, with the first scenario based on an 15 percent annual depreciation rate. The other scenarios were based on more rapid rates of technological change, and consequently more rapid rates of obsolescence. The assumption of shorter economic lives gives greater weight to more recent innovations in the capital stock estimates. In Developing an R&D Satellite Account for the UK: a preliminary analysis, Fernando Galindo-Rueda applies a depreciation rate of 20 percent, which is consistent with a 10 year service life. However, it is not feasible to make a reliable estimate of the life of non-patented R&D in Australia and the actual evidence from patent expiry rates has been adopted.

A mean asset life of 11.0 years is broadly consistent with international results. A recent draft OECD Handbook (OECD, 2008) states that the different approaches to estimating R&D asset lives "generally indicate that service lives lie between 10 and 20 years". However, most countries have not committed to an estimate and/or method to be used in their National Accounts (the US figures have been used in the BEA's R&D Satellite Account). None of the OECD countries use an asset life significantly shorter than 10 years. For many countries only a depreciation rate is specified, but under a standard double declining balance assumption (that is double that of a straight line depreciation) they imply similar (or sometimes longer) lives. Given the lack of evidence to the contrary, the ABS has assumed a mean asset life of 11.0 years based on patent data.

Figure 2 - Lifetimes of granted standard patents, average for all technology groups
Graph: Figure 2 - Lifetimes of granted standard patents, average for all technology groups



Incorporating R&D in the capital stock and balance sheets.

The flow of capital expenditure, the asset life and rate of depreciation of R&D assets provide the information required to estimate the stock of R&D assets. As with asset lives, the pattern of depreciation of R&D assets is difficult to measure, and will be the subject of continuing discussion. If in fact R&D assets decline in value at a more rapid rate than other types of assets, a geometric age efficiency function may be appropriate, however in the absence of authoritative information, a hyperbolic age-efficiency function has been adopted for consistency with other types of assets in the ASNA capital stock system.

The value of the stock of R&D assets will be recorded in the balance sheets as an intellectual property product, a category of produced fixed assets. The value of assets in the stock will change as a result of changes in price and the consumption of fixed capital over the life of the assets.


IMPACT ON THE NATIONAL ACCOUNTS

The impact of capitalising R&D in general government units will be equal to the value of the additional consumption of fixed capital as discussed earlier. In the case of other units, additional own account R&D output (valued by summing the costs of production) will be recorded and will appear as additional output, GFCF and GOS. Consequently the impact on GDP from capitalising R&D in non general government units will be equal to the value of the own account R&D produced.

3 The effect of capitalising R&D on GDP and Gross Fixed Capital Formation, Current Prices, 1994-95 to 2007-08
1994-95
1995-96
1996-97
1997-98
1998-99
1999-2000
2000-01
2001-02
2002-03
2003-04
2004-05
2005-06
2006-07
2007-08
$m
$m
$m
$m
$m
$m
$m
$m
$m
$m
$m
$m
$m
$m

GFCF SNA93
116 273
118 397
124 908
138 340
146 858
160 089
151 474
168 832
194 081
213 760
231 739
260 762
284 707
320 052
Private R&D
3 518
4 336
4 149
4 118
4 059
4 156
4 834
5 955
6 742
7 646
8 422
9 669
11 329
12 984
Public Corproations R&D
184
125
58
90
162
60
157
277
219
146
154
251
210
229
General Government R&D
1 731
1 791
1 842
1 900
1 967
2 108
2 246
2 344
2 432
2 543
2 691
2 921
2 917
3 191
%
%
%
%
%
%
%
%
%
%
%
%
%
%
Percentage increase to GFCF level
5
5
5
4
4
4
5
5
5
5
5
5
5
5



The initial impact of capitalising R&D will increase capital expenditure, which will reflect an increase in gross operating surplus in the production account for market units. The increased operating surplus flows through to the income accounts, where it in turn is balanced by an equivalent increase in saving. In the capital accounts the increased saving finances the increase in gross fixed capital formation of R&D assets. There is therefore no impact on net lending.


BIBLIOGRAPHY

ABS 2008a, Australian Bureau of Statistics, Research and Experimental Development Businesses, 2006-07 (cat. no. 8104.0)

ABS 2008b, Australian Bureau of Statistics, Research and Experimental Development, Government and Private Non-profit Organisations, 2004-05 (cat. no. 8109.0)

ABS 2008c, Australian Bureau of Statistics, Research and Experimental Development, Higher Education Organisations, 2006 (cat. no. 8111.0)

ABS 2004, National Accounts Research Section, Australian Bureau of Statistics Capitalising Research and Development in the National Accounts,. Fourth meeting, 17-19 march 2004, Washington D.C., USA. Canberra II Group on the measurement of non-financial assets.

Okubo S, Robbins CA, Moylan CE, Sliker BK, Shultz LI, Mataloni LS 2006, Bureau of Economic Analysis (BEA), 9/28/2006 R&D Satellite Account: preliminary estimates.

Robbins CA, Moylan CE 2007, Research and Development Satellite Account Update, Estimates for 1959, 2004.

Galindo-Rueda F 2007, Office for National Statistics, Economic & Labour Market Review, Vol 1, No 12, December 2007. Developing an R&D satellite account for the UK: a preliminary analysis.

OECD 2007, R&D Satellite Accounts in the Netherlands; a progress report, Working Party on National Accounts, OECD paper STD/CSTAT/WPNA(2007)14

Yorgason DR 2007,Treatment of International Research and Development as Investment: Issues and Estimates, Bureau of Economic Analysis/National Science Foundation, R&D Satellite Account Background Paper.

OECD 2002, Frascati Manual 2002:Proposed standard practise for surveys of research and experimental development, OECD, 2002.

OECD 2008, Draft OECD Handbook on Deriving Capital Measures of Intellectual Property Products, October 2008.

Intellectual Property Australia, website <www.ipaustralia.gov.au>