This paper analyses the rationale for, and results of, a recent addition to Research & Development (R&D) surveys conducted by the Office for National Statistics (ONS) looking at the ownership of R&D assets. The results refute the prevailing assumption that the funders of R&D are the owners of these assets, suggesting a more complex relationship than initially understood. The paper then analyses how business R&D should be allocated before concluding with recommendations of how R&D capital should be treated in the National Accounts.
Many businesses and other organisations conduct Research and Development (R&D)1 in support of their main activities, investing in innovation to give them a competitive edge. Revisions in the 2008 System of National Accounts (United Nations, 2009) acknowledge the investment nature of R&D, requiring that expenditure on R&D is recognised as formation of intangible R&D assets which contribute to production over time.
All capital assets in the National Accounts must have an economic owner (OECD, 2010:9); ‘the institutional unit [or units] entitled to claim the benefits associated with the use of the entity in question in the course of an economic activity by virtue of accepting the associated risks’ (United Nations 2009:620). Economic ownership may differ from legal ownership; a business may own a patent on an idea but could grant others access to this knowledge for use in their production process - thereby sharing the economic ownership (OECD, 2010:13).
As survey information regarding R&D ownership is not available in many countries, funding information is often used as a proxy. It is assumed that economic agents will not fund R&D without expecting ownership of the resulting R&D. The ONS decided to test the relationship between funding and ownership of R&D by including an additional question asking respondents to identify the owners of the R&D assets produced on the 2011 Gross Expenditure on Research and Development (GERD) surveys. This paper presents the new ownership question, the results, and makes comparison to information collected on R&D funding. Comparison is also made to R&D exports statistics and more detailed analysis at the industrial section level is also presented. It concludes with recommendations for the treatment of R&D assets.
In the absence of ownership information, Galindo-Rueda (2007:21) notes that it is reasonable to assume that information on the funders of R&D can provide an indication of the ownership of these assets. Typically, the funding of R&D performed in each institutional sector is aggregated into four economic agents; business enterprises, general government, non-profit institutions serving households (NPISH) and the rest of the world. Sums in each sector are divided by the total R&D expenditure by the performer to achieve ownership shares. From this, a matrix of ‘ownership’ may be created highlighting these shares.
However, Galindo-Rueda highlights two important limitations of using this data unadjusted:
Funding of R&D is constrained to the total expenditure on its production, ignoring any additional costs which the R&D purchaser might face in implementing the R&D in their business
In assuming a perfect link between funding and eventual ownership, this approach ignores various situations where this may not be the case, such as corporate donations and grants made by government where funding may be given without implying ownership rights.
The ONS question on ownership of R&D assets addresses the second of these two limitations. Respondents can indicate where they have received a grant from government or others where these bodies make no claim of ownership over this R&D.
Prior to inclusion in the GERD surveys, the phrasing of the R&D ownership question was extensively tested to ensure it was fit for purpose, and that respondents could interpret and answer it correctly. Results from the first wave of testing found that respondents had difficulty understanding the question’s relevance and were confused by National Accounts terminology, which differs from business accounting practice. Additional guidance and examples were provided to help respondents in the second round of testing, which showed that many of the issues had been addressed. Following this, the question was added to all GERD surveys. It was included on the ‘long form’ Business Enterprise R&D Survey, which covers around 400 companies that spend the most on R&D and contribute to approximately 80 per cent of total business R&D expenditures annually. Results were imputed for smaller firms. The full question, along with the extensive notes and examples provided to assist respondents, can be found in Annex 1.
However, this question is not without limitations; some respondents continued to find the question difficult to answer either due to the complex nature of their business or because they did not record the required data. In these cases, it was suggested that respondents might consider funding of their R&D as a proxy for ownership. This means our results may not completely capture the difference between funding and ownership. Secondly, the question only addresses expected sales and transfers of R&D and will exclude any unforeseen sales. It also provides no information on sales or transfers later in the asset’s life. Despite these limitations, the question still offers new insight.
The assumption that funding of R&D implies ownership of the resulting output is often used as a default position when discussing the ownership of R&D (Galindo-Rueda, 2007). Table 1 shows the implied share of the ownership of R&D assets when funding is used as a proxy for ownership.
|Funder of R&D|
|Business Enterprise||General Government||NPISH||Rest of the World|
|Performer of R&D|
|Rest of the World||74.0%||26.0%||0.0%||n/a|
Figure 1 shows that 69 per cent of the R&D performed in the UK by businesses is also funded by domestic businesses. Similarly, 86 per cent of R&D performed by General Government is funded by the Government sector and only 3 per cent of Government R&D is funded by businesses. Using funding as a proxy for ownership, this suggests that R&D is largely retained by the institutional sector performing the R&D activity.
Galindo-Rueda (2007:22) used economically intuitive assumptions to develop a measure of implied R&D ownership by each sector, applying weights to the funding assumption to achieve a new measure of ownership. The assumptions used to produce this table are available in the Annex 2, and updated results are available in Table 2.
|Owner of R&D|
|Business Enterprise||General Government||NPISH||Rest of the World|
|Performer of R&D|
|Rest of the World||67%||33%||0%||n/a|
Table 2 is broadly similar to the funding shares in Figure 1 though there are key differences. Firstly, it is assumed that Governments own less of the R&D performed by the business sector than they fund. This reflects government transfers, such as grants, which are made without any expectation of ownership of the fruits of this R&D activity to reduce underinvestment in R&D. It is also recognised that Government may procure R&D technology, so it is expected that Government will retain ownership of some of the R&D that they fund.
Secondly, Government is expected to own only 10 per cent of the R&D that it funds Non-Profit Institutions serving Households (NPISH) to perform. This reflects the grants made to the Higher Education sector and charities. Similarly, flows of R&D funded from ‘Rest of the World’ are amended to reflect that some of these may also be grants.
The residual ‘non-owned’ component of funds is then allocated to the performing institutional sector. As a result, the expected ownership of the business and NPISH sectors increases to 76 per cent and 80 per cent respectively, substantially higher than when using only the funding information.
The results from the ownership question on the BERD and GovERD surveys offer a comparison to the results achieved using funding assumptions. Table 3 shows the ownership shares implied by the ownership question on the GovERD and BERD surveys.
|Owner of R&D|
|Business Enterprise||General Government||NPISH||Rest of the World||Freely Available|
|Performer of R&D|
Table 3 introduces a ‘freely available’ category of ownership not in the funding question. This category includes R&D which has been published in journals, on the Internet or without intellectual property right. Freely available R&D amounts to 1 per cent of all business R&D. The treatment of R&D in this category is open to debate; assets should provide their owners with an economic benefit, but if R&D is made freely available the producer may expect no economic benefits (OECD 2010:46). However, Aspden (2006) highlights cases where economic benefits may be achieved whilst research is made freely available, particularly in the non-market sector. No information is available to establish if the producer still expects any economic benefit from this R&D and, as this constitutes a small share of business expenditure, it is recommended that this is included in the producer’s institutional sector. For Government all R&D is recorded as gross fixed capital formation (OECD 2012:5) regardless of whether it is freely available or not.
These results are further disaggregated in Table 4, which directly compares responses to the funding and ownership questions on the 2011 BERD survey.
|Funding Question||Ownership Question||Ownership/Funding|
|Own UK Business||65.2%||72.3%||111.0%|
|Other UK Businesses||3.5%||3.4%||97.2%|
|Other UK Organisations||0.6%||0.4%||64.1%|
|Higher Education Establishments||0.0%||0.0%||100.0%|
|Non Profit Organisations||0.0%||0.1%||22033.0%|
|Commission of the EU||0.4%||0.0%||7.7%|
|Any Other Organisations||1.8%||1.5%||81.3%|
Table 4 shows that UK Businesses fund 65.2 per cent of all domestic R&D conducted by the business sector yet own 72.3 per cent of this resulting R&D. Additionally, the UK Government is expected to own 31. 8 per cent less business R&D than they fund whilst the Commission of the EU owns only 7.7 per cent of the business R&D that it funds in the UK. These findings echo similar research from the Statistical Office of the Republic of Slovenia (Lasnibat, 2012) and are consistent with the view that Governments will provide grant support for business R&D without gaining ownership of the resulting knowledge.
Table 4 also shows a substantial difference between the funding and ownership of R&D by overseas non-profit organisations. Here, non-profit organisations own 220 times the level of R&D than they fund. This large percentage ownership of R&D flatters the more modest sum this represents in cash terms. However, such a substantial difference between the funding and ownership measures should be expected. Businesses may transfer their R&D knowledge to charities overseas for charitable reasons or because these bodies assist in the production of R&D beyond purely financing the projects. These transfers would not be captured by funding data. Analysis of the micro data behind these transfers supports this assertion.
Similarly, it might be expected that UK businesses transfer R&D to 'Other UK Organisations', a group including non profit institutions serving households (NPISH) which contains charities. However examination of the R&D ownership results finds that this sector owns less R&D than it funds. This difference may be partly explained by the existence of freely-available R&D, as charities may be more concerned with supporting particular research conducted by businesses, such as cancer or other medical research, rather than any economic gains from these. In addition, 'Other UK Organisations' are a broader group than strictly NPISH, which may obscure comparison between the two groups.
The mixed results from the ownership data may reflect the registration of patents within the UK skewing the perception of ownership towards legal rather than economic use, or may reflect companies viewing R&D as an intermediate input to further R&D. In addition, as R&D is non-rival in consumption (a business may use the R&D without reducing the ability for another firm to use the knowledge) this knowledge may be spread more widely within a business group than ownership responses suggest. In spite of these limitations, these results provide clear evidence of significant differences between funding and ownership of R&D; using the funding information alone would understate business ownership of R&D by 10 per cent and over-state government ownership of R&D by around 30 per cent.
Responses showed that 72.3 per cent of business R&D performed in the UK will be owned by their 'Own UK Business'. This category includes the not only the direct performer of the R&D activity but also any parent or sister companies that are based within the UK. To ensure that R&D capital is correctly allocated in the Supply and Use tables, it is essential to establish the industrial section1 that these businesses belong to.
To establish more accurate section allocation of R&D, it is assumed that all R&D owned by 'Own UK Business' is ultimately owned by the parent company of the R&D performer within the UK rather than any sister company. This is more likely to reflect the economic ownership of R&D than simply leaving all R&D in this category with the performer. This assumption is defensible as there is currently insufficient data available to assign ownership to any other companies within the group. Where there is no parent company, the R&D is allocated to the industry of the performer.
All businesses in receipt of surveys from the ONS are allocated a Standard Industrial Classification (SIC) number based on their principal business activity. Where businesses undertake several activities, attempts are made to identify separate reporting units of the business and allocate them SIC numbers based on their distinct activities. Where this is not possible, a general SIC number is allocated according to the activity within the business with the largest turnover or employment. This may lead to R&D activity being assigned to a unit which may not reflect the activity of the unit using the R&D. This paper assumes that all businesses are allocated SIC numbers that accurately reflect the business activity using the R&D.
Data was split into those businesses with a parent business and those without. "Parentless" businesses had ownership directly assigned to their industrial section, whereas businesses with parents had ownership assigned to their parent’s industrial section. A matrix was constructed to analyse the flows of R&D assets between the sectors producing R&D and the owners of the resulting R&D assets. It was expected that the majority of R&D activity would fit the diagonal of the matrix, indicating that R&D assets produced in an industrial sector were also owned by business within that industrial section.
Of the 18 industrial section where data are available, 11 sectors allocate in excess of 90% R&D to the same section as the business conducting the R&D, and in an additional 5 section allocating more than 70 per cent of R&D owned by the respondent’s ‘own UK business’ remains in the same section. More detailed results cannot be presented due to disclosure considerations. These results suggest that for many industries, allocating ‘own UK business’ owned R&D capital to the industry that produced it would provide a sufficient approximation of ownership.
However, there are some instances where such an allocation may not be appropriate. Using the assumption that parent businesses own the R&D may be less useful for complex business structures. For example, R&D conducted in the Mining & Quarrying section was found to be split in ownership between the ‘Mining and Quarrying’ and ‘Wholesale and Retail Trade’ sections. This is not economically intuitive, as large mining companies may operate a wholesale arm, but it is unlikely that the R&D conducted by their mining operations would be used for developing their wholesale or retail business. Here, it appears more appropriate to allocate the ownership of this asset to the Mining and Quarrying section. Analysis of the types of R&D performed supports this assertion.
In contrast, a positive case may be made for deviating from assigning all R&D capital to the industrial section producing the R&D. Table 5 summarises some of these results for R&D declared as owned by ‘own UK businesses’ which was performed by businesses in a number of sections with parents in manufacturing:
|Wholesale & Retail Sector||14.7|
|Transport and Storage||15.8|
|Professional, Scientific and Technical Activities||18.9|
|Research & Development||17|
Table 5 shows that 18.9 per cent of R&D produced in the ‘Professional, Scientific and Technical Activities’ section is owned by the ‘Manufacturing’ section for example. This is economically intuitive as manufacturers may commission R&D projects from research companies, particularly if they have experience within a particular field. In addition, separately identifiable business units from large manufacturers have already been classified to the R&D section, so flows from this section to the manufacturing section should be expected.
Information collected on the ownership of R&D provides the opportunity to compare transfers of R&D ownership overseas with other data sources. Prior to the inclusion of an ownership question, exports of R&D could be estimated using funding from overseas as a proxy. Now, ownership data may be used as a different indicator of R&D exports. These results may be compared directly, and with data from the International Trade in Services (ITIS) survey.
In 2011, £3,734m of R&D performed in the UK was funded from overseas. By contrast, the new ownership question shows that ownership of £2,958m of R&D performed by UK businesses will be transferred to organisations overseas. These compare with ITIS results of £4,444m, equivalent to 25.5 per cent of total BERD expenditure.
The results from the ownership question implied lower exports of R&D than the ITIS R&D exports data. This was surprising as the ITIS survey does not include financial corporations, and respondents often struggle to correctly identify and price flows of R&D between international branches of their company. It was therefore expected that the ownership question may make it easier for respondents to track these flows. However, ownership results only account for 66.6% of the ITIS figure.
These results do not adequately account for the international nature of R&D. Analysis of BERD employment data showed that the British workforce of large multinational corporations is typically more concentrated in R&D activities than their overseas arms, suggesting that R&D activity within these groups is centred within the UK and transferred for use by the whole company. Tanriseven et al (2007) find a similar situation in the Netherlands, implying foreign ownership of domestically produced R&D by companies should be larger.
The difference between ITIS and BERD figures may be accounted for by the difficulties in measuring intra-company flows. ITIS is a dedicated trade survey, whilst the BERD, despite its ownership question asking companies to detail foreign ownership, both internal and external to the business, is not. Incongruence between these surveys is likely as respondents may not fully consider all of these flows in the context of the BERD questionnaire. In addition, as R&D is non-rival in consumption, this knowledge may be spread more widely within a business group than ownership responses suggest; exports or use of R&D abroad may be greater than the data suggests. The Bureau of Economic Analysis, and other national statistics institutes, use ITIS as a source for R&D imports and exports instead of GERD data (Mataloni et al, 2007:2).
In addition to analysis of the business sector, data was also obtained through the GovERD survey. Table 6 shows that UK Businesses own 47.4 per cent the R&D they fund the UK Government sector to perform. Additionally, the UK Government is expected to own 10.9 per cent less R&D than they fund, in contrast to Galindo-Rueda’s assertion that Governments will own all R&D they fund. Of all Government expenditure on R&D, 16.6 per cent is made freely available to the public. Considering the primary purpose of Government, this result is intuitive. The existence of freely available R&D may also explain why ‘Other UK Organisations’ own only 10.9 per cent of the R&D they fund.
|Funding Question||Ownership Question||Ownership/Funding|
|UK Higher Education||0.8%||0.7%||79.8%|
|Other UK Organisations||3.1%||0.3%||10.8%|
|European Commission Programmes||5.1%||3.5%||68.9%|
|Higher Education Establishments||0.1%||0.0%||50.5%|
|Non Profit Organisations||0.3%||0.5%||174.6%|
|Any Other Organisations||0.3%||0.2%||51.6%|
In light of the findings of the ownership question, it is clear that using solely a funder-is-owner assumption is inadequate. This would overstate government ownership and understate business ownership with important implications for the sectoral allocation of R&D assets in the National Accounts. R&D capital should be allocated to institutional sectors on the basis of this new ownership information as this appears to give a closer representation of the ownership of R&D. Within business R&D data, R&D owned by ‘Own UK business’ should be assigned, in most cases, to the performing business’ industrial section unless there is clear economic intuition why this may not be valid, such as the cases outlined above.
Before conclusively recommending the adoption of these ownership ratios to previous periods of R&D data, it is desirable to collect data in future periods. With data covering multiple periods the stability of these relationships identified can be examined.
This paper has discussed the introduction of a question assessing the ownership of R&D assets to the business R&D survey. The results of this question have cast doubt on the commonly used assumption that funding of R&D implies ownership of the resulting R&D. Some economically intuitive results such as government funding more R&D than it eventually owns and private non-profit firms overseas owning more R&D than they fund have been found. These results are broadly consistent with the suggestions of Galindo-Rueda (2007). The new question also provides information on freely available R&D.
However, the data continues to understate trade in R&D when compared to the ITIS survey results. Further, the nature of R&D may mean that it is shared more widely than businesses reported in the survey. Further research is needed to assess the robustness of results over time. Nevertheless, these results provide important insight into the looseness of the relationship between funding and ownership, and an improved basis for the allocation of R&D to sectors in the National Accounts.
Aspden, C. (2006) ‘Freely Available R&D’ [online] Available at https://unstats.un.org/unsd/nationalaccount/aeg/papers/f4RD.pdf [accessed 12th March 2013]
de Haan, M., Kuipers, A., van Rooijen-Horsten., and Tanriseven ,M. (2007) ‘Problems related to the Measurement of International Flows of R&D’, [online] Available at: http://www.unece.org/fileadmin/DAM/stats/documents/wggna/zip.2.e.pdf [accessed 12th March 2013]
Galindo-Rueda, F. (2007) ‘Developing an R&D satellite account for the UK: a preliminary analysis’, Economic & Labour Market Review, Vol. 1, No. 12, pp18-29
Lasnibat, J (2012) Results of an additional question regarding the nature of funding in the R&D survey questionnaire for Slovenia. Paper presented at the European Knowledge Sharing Event for R&D, London, 18th December 2012
Mataloni, L., and Moylan, CE. (2007) ‘2007 R&D Satellite Account Methodologies: Current dollar GDP estimates’ [online] Available at: http://www.bea.gov/papers/pdf/rdmethods_gdp.pdf [accessed 18th March 2013]
Office for National Statistics (2009) ‘UK Standard Industrial Classification of Economic Activities 2007’ [online] Available at: http://www.ons.gov.uk/ons/guide-method/classifications/current-standard-classifications/standard-industrial-classification/index.html [accessed 14th March 2013]
Office for National Statistics (2012) Business Enterprise Research & Development 2011 [online] Available at http://www.ons.gov.uk/ons/rel/rdit1/bus-ent-res-and-dev/2011/stb-berd-2011.html [accessed 14th March 2013]
Office for National Statistics (2013) Gross Expenditure on Research & Development 2011 [online] Available at http://www.ons.gov.uk/ons/rel/rdit1/gross-domestic-expenditure-on-research-and-development/2011/index.html [accessed 14th March 2013]
Office for National Statistics (2013) International Trade in Services 2011 [online] Available at http://www.ons.gov.uk/ons/rel/itis/international-trade-in-services/2011/index.html [accessed 14th March 2013]
Organisation for Economic Co-operation and Development (2002) Frascati Manual, Paris: OECD Publications.
Organisation for Economic Co-operation and Development (2010) Handbook on Deriving Capital Measures of Intellectual Property Products, Paris: OECD Publications.
Organisation for Economic Co-operation and Development (2012) ‘Second Taskforce on the Capitalisation of Research and Development in National Accounts – Final Report’ [online] Available at http://search.oecd.org/officialdocuments/publicdisplaydocumentpdf/?cote=STD/CSTAT/WPNA(2012)29&docLanguage=En [accessed 16th March 2013]
United Nations (2009) System of National Accounts 2008. New York: United Nations.
All of your in-house R&D expenditure, including both capital and non capital expenditure, is classified as investment in intangible assets. For this question, the value of R&D assets your business produces should be allocated to owners (or future/final owners). The total expenditure allocated to the owners should equal the sum of the in-house expenditure as reported in question 11. This is a new concept and it may be difficult to identify who the owners are, so please provide your best estimates. It may be necessary to consider individual R&D projects and aggregate the results if possible.
If the R&D performed in the reporting period is (or is expected to eventually be) subject to intellectual property protection (such as patents and copyrights), the holder of that protection is the owner. For example, if the R&D your business performed in the reporting period is covered by patents owned by your UK business, the full in-house expenditure should be entered in question 14a.
If the R&D results are made freely available for others to use, this should be allocated to question 14m
If a number of organisations will own R&D assets, the in-house expenditure should be divided and allocated to the relevant owners
- If contractual arrangements exist which determine ownership in percentage terms, for example, 60 per cent owned by your UK business and 40 per cent owned by UK Government, please use these percentages to divide the in-house R&D expenditure between questions 14a and 14b.
- If no contractual arrangements exist, the value of the assets should be divided according to your best estimate of their use. For example, if your business and another organisation are both owners, but your business will make more use of the assets, ownership should be allocated accordingly. If the use of the assets cannot be estimated, the assets should be divided equally between the owners, or allocated to the lead organisation as appropriate.
- If all the R&D you perform is fully funded externally, then please allocate the owner of the in-house expenditure to the relevant question categories that reflect the funders.
The sum of answers to questions 14a to 14m (continued overleaf) should equal your answer to question 13. Please exclude any cost plus element.
14a – your UK business?
Include your parent company or other subsidiary companies operating within the UK
14b – the UK Government?
Include UK government departments, their agencies and non-departmental public bodies (for example, NHS, Ministry of Defence, Department of Health, Department for Business, Innovation and Skills, the Scottish Government)
14c – other UK private businesses?
Include non-profit organisations serving businesses
14d – any other UK organisations?
Include private non-profit organisations serving households, education establishments and trade unions
14e – business enterprises within your group outside the UK?
Include funding from your parent company or other subsidiary companies operating outside the UK
14f – other business enterprises outside the UK?
14g – other Governments outside the UK?
14h – higher education establishments outside the UK?
14i – non-profit organisations outside the UK
14j – the Commission of the European Union
14k – international organisations?
For example, World Trade Organisation, NATO
14l – any other organisations outside the UK?
14m – not owned and freely available in-house R&D
Include in-house R&D which is developed and published and made freely available for others to use. For example, publications in journals or on the Internet, and R&D with no intellectual property protection, such as patents and copyrights.
|Owner of R&D|
|Business Enterprise||General Government||NPISH||Rest of the World|
|Performer of R&D|
|Rest of the World||0.9||1||0.76||0.22|
Details of the policy governing the release of new data are available by visiting www.statisticsauthority.gov.uk/assessment/code-of-practice/index.html or from the Media Relations Office email: email@example.com