Chapter 6

IP Report 2018

Collaborative research grants

lead to better IP outcomes

The centrality of innovation to economic prosperity cannot be overemphasised. The OECD estimates that around 50 per cent of long-term economic growth in its member countries can be attributed to innovation, and this contribution is expected to grow.1 Innovative activity takes many forms, arguably the most important of which is research and development (R&D). Hence, to promote research-led innovation, the Government sponsors research conducted by a vast network of publicly funded research organisations (PFROs), which include universities, the Commonwealth Scientific and Industrial Research Organisation (CSIRO), medical research institutes (MRIs) and numerous other government and non-government bodies (Figure 15).

Figure 15: Australia's public research funding, 2016–172

The Australian Government awards funding to eligible organisations, which are mainly universities. The universities may partner with private sector businesses

In 2016–17, the Government invested an estimated $10.1 billion in research — up by around 35 per cent since 2008–09.3 Australia's PFROs are funded through a variety of mechanisms, including direct budget appropriations and competitive grants from the Australian Research Council (ARC) and the National Health and Medical Research Council (NHMRC).

The ARC administers the National Competitive Grants Program (NCGP), which is a significant component of public investment in R&D, composed of both Discovery and Linkage programs. In 2016–17, the ARC administered 4996 new and ongoing grants under the NCGP, providing over $730 million in grant payments for approved research projects.4 Grants under the Linkage program explicitly tie funding to collaborative research projects; 560 partner organisations were involved in funded projects through this program in 2016.5

To the extent that research creates new knowledge, it is an intrinsically beneficial activity. Research may not deliver economic benefits unless commercialised. Commercialisation transforms research into marketable products and processes, and patenting activity is an obvious indicator of the commercialisation potential of research output. Given that collaborative R&D activity is taken to demonstrate a greater commercialisation propensity,6 it is important to understand whether collaborative grants are associated with more patent applications than non-collaborative grants, which would indicate a higher prospect of commercialisation.

This chapter presents the findings of two related studies — one by IP Australia's OCE and the other by Swinburne University of Technology and the University of Melbourne (Swinburne) — that examined these relationships for PFROs and private businesses respectively. For the purpose of these studies, data on research grants are linked with patent applications, and then with business data from 2001 to 2014.7

The two studies employed different econometric methodologies and differed in terms of the granularity of the data used.8 Results from both studies nevertheless show that public funding of research has a positive and statistically significant impact on patent production, with the impact being stronger for collaborative grants.

Patent productivity of research organisations

The underlying hypothesis of the OCE analysis is that the amount and type of funding received by PFROs are key determinants of the number of patent applications lodged by them.9 Both Patent Cooperation Treaty (PCT) and provisional applications were considered separately. We analyse PCT applications as they are strong indicators of an intention to take a product to the international market, and so they serve as a proxy for late-stage commercialisation activities. Provisional patent applications, on the other hand, are filed when applicants still want to keep their invention secret, and plan some commercial use of the invention, so we consider it a proxy for early-stage commercialisation.10

Regression analysis11 shows that the amount of total funding has a statistically significant positive impact on the number of provisional applications (i.e. early-stage commercialisation), irrespective of past patenting behaviour.12 While the impact of funding on PCT applications is positive, the relationship is statistically significant only when the stock of past applications is disregarded, so applicants with a history of filing PCT applications do so, regardless of grants. In addition, the impact is larger for provisional applications than PCT applications — indicating stronger correlation between total funding and the choice to file provisional patent applications, an indicator of early-stage commercialisation planning.

When collaborative and non-collaborative grants are analysed we find that any additional research funding boosts all types of patent applications regardless of past activity. Nevertheless, collaborative grants have a higher impact than non-collaborative ones and a greater impact is seen for PCT applications (i.e. the later-stage commercialisation).

Finally, by calculating the predicted patent output from the estimated impact factors, we find that 2–3 additional patents may be expected from a $1 million increase in annual research funding (as shown by total ARC funding in Figure 16). While additional ARC funding is associated with more patent applications across all grant types, the results are driven by the collaborative grants where an increase in funding of $1 million leads to an increase of three PCT and six provisional applications and delivers a significantly higher patent increase than non-collaborative grants.13

Patent productivity of collaborative businesses

The Swinburne study by Jensen, Palangkaraya, Thomson and Webster14 is work-in-progress and is pioneering Australian research in the collaboration field. It uses the ARC Linkage program data coupled with the ABS business population data and focuses solely on business recipients of collaborative grants — to examine the responsiveness of business patenting activity to a proportional change in public research funding.

Figure 16: Patent filings increase as a result of $1 million increase in ARC funding

Collaborative ARC funding: 3 PCT applications, 6 Provisional patent applications. Total ARC funding: 2 PCT applications, 3 Provisional patent applications. Non-collaborative
ARC funding: 1 PCT applications, 2 Provisional patent applications.

Source: ARC (2017), Linkage Program — custom data request; IPGOD (2017) and PATSTAT (2016), custom data report by IP Australia's Patent Analytics Hub.

The Swinburne analysis accounts for multiple firm characteristics and compares the outcomes experienced by businesses who received collaborative grants with similar businesses who did not receive a collaborative grant.15

The analysis16 estimates the patenting response of collaborating businesses to an increase in annual research funding. The early results suggest that there is indeed a positive relationship between funding and patent production by collaborative businesses and that this effect occurs with a lag between the commencement of the funded research project and the later commercialisation of research outputs. Continuing and extending the analysis will further explore the relationship between research funding and the patent productivity of Australian businesses.

Conclusion

Evidence produced by these two separate but related studies shows that an increase in the magnitude of research grants, or greater availability of public funding for collaborative grants is likely to boost patenting, and the subsequent commercialisation of research outputs of both research organisations and collaborating businesses. Funding schemes that mandate collaboration with industry appear to be more productive in terms of patent applications. However, the impact of a funding increase is far outweighed by the importance of factors internal to funding recipients. In particular, the analysis suggests the grant recipient's recent experience of engaging with the patent system is an important factor in driving patent applications.

Nevertheless, the results presented here have implications for the mix of public research funding and the effectiveness of funding instruments in meeting research and innovation policy goals.

End notes

  1. Department of Industry, Innovation and Science (2016), Australian Innovation System Report 2016, Canberra, p. 23, https://industry.gov.au/Office-of-the-Chief-Economist/Publications/Pages/Australian-Innovation-System.aspx#.
  2. The ARC does not award direct funding to private sector businesses. It awards funding to eligible organisations, which are mainly universities. The universities may partner with private sector businesses.
  3. Department of Industry, Innovation and Science (2017), Science, Research and Innovation Budget Tables 2017-18, https://industry.gov.au/innovation/reportsandstudies/Pages/SRIBudget.aspx, accessed 14 February 2018.
  4. ARC (2017), Annual Report 2016-17, Australian Research Council, Canberra, p. 23.
  5. http://www.arc.gov.au/grants-dataset, accessed 14 February 2018. This figure counts organisations for each project on which they collaborate. Therefore partner organisations with more than one project can be counted multiple times, i.e. there are multiple applications per applicant and multiple applicants per application.

    Partner organisations are not funded by the ARC. Universities are funded by the ARC and they collaborate with partner organisations in the conduct of the research project.

  6. Cunningham, J & Link, A (2015), ‘Fostering university-industry R&D collaborations in European Union countries', International Entrepreneurship and Management Journal, 11(4), pp. 849–860.
  7. Grants data are sourced from the ARC and include successful grants, both ongoing and completed, from 2001 to 2014. Two additional data sources have been used for linking grants to patents — namely, the National Survey of Research Commercialisation (NSRC) and the Business Longitudinal Analysis Data Environment (BLADE). The ARC grants data then are linked with (i) the PFRO patent applications data, compiled as part of the NSRC and (ii) the BLADE, which integrates business patenting activity data from IPGOD. Grantee name has been the linking variable. This exercise has produced first-of-its-kind longitudinal datasets, which enable the tracking of grant recipients' patenting activity over time while taking into account their unobserved heterogeneity.
  8. While the OCE analysis of PFRO patenting activity used more aggregate, institution-level data, Swinburne conducted a more micro-level analysis of business behaviour using grant-level data.
  9. In econometric terminology, the amount and type of funding are the explanatory variables and the number of patent applications is the explained or dependent variable.
  10. See chapter 2 for descriptions of PCT and provisional applications.
  11. A fixed effect Poisson model is used.
  12. Since past activity or behaviour is often a good predictor of current activity, the analysis also includes a five-year moving average of past application counts as an additional determinant — to capture the inherent innovation ability of the funded organisations.
  13. Note that patent outcomes are just one tangible output from public research funding. The diffusion of knowledge from publicly funded research also takes various intangible forms which are often not measured.
  14. Jensen, P, Palangkaraya, A, Thomson, R and Webster, E (2017), ‘Does the ARC Linkage program enhance the performance of collaborating businesses?', paper presented to the 2017 Asia-Pacific Innovation Conference, Wellington, November.
  15. If more innovative firms receive grants (treatment), then their post-grant outcome (patents) can be correlated with their pre-grant characteristics rather than the grant itself. To isolate the treatment effect, therefore, it is necessary to compare the outcome of the treated group with that of a control group — matched from the pool of nonparticipants on observable characteristics such as firm size, patent status, industry and geographic location.
  16. A fixed effects panel data model is used, where log of patent application counts is the explained variable and log of funding amount is the key explanatory variable.

View the IP Report in PDF format (2MB).