R&D Review: Invention is out of scope

In December 2024, the Australian Government announced a strategic review of R&D carried out within Australia. The main reason for this review is, that relative to other OECD countries, Australia has:

The review is led by a panel that includes, Robyn Denholm (chair), Ian Chubb, Fiona Wood and Kate Cornick. It has released a discussion paper and invited submissions from anyone interested in Australia’s R&D system. Below is my submission.

Author: Paul Klink
Date: 8th April 2025

Something is not quite right about this strategic examination. Its terms of reference seem to be incongruent with the desired outcomes presented in its discussion paper. It seems to be built on the foundational premise that getting our R&D ‘right’ will lead to a similar extent of commercialisation of R&D (together with its economic benefits) as other high tech nations.

But what if that premise is either not correct or incomplete?

Last year I wrote an article “Understanding R&D Subsidies - A software perspective” which touched upon this issue but did not address it directly. Since then I have developed some further ideas as to how to address this issue at a policy level and was aiming to make a submission around these ideas. However the above dissonance was ringing in my head.

Then I read1 Richard Nelson’s seminal 1959 paper, “The Simple Economics of Basic Scientific Research” and things became a lot clearer.

There are 2 points from this paper I would like to highlight:

  1. Invention
    Richard defines invention as: “the human activity directed toward the creation of new and improved practical products and processes”2. I believe boosting invention, is the key to improving our commercialisation of R&D! Invention can be closely coupled with R&D (an input to invention) however invention is not particularly based on systematic effort. For example, I would consider the development of novel software to be more related to invention than to R&D.
  2. Social Profit of (Basic) Research
    The social profit of research is measured by the difference between its social value and social cost. The social value is defined as the benefits we receive from a given expenditure on research. However society could also redirect that expenditure elsewhere and receive some alternative benefits. These alternative benefits are the social cost. This definition allows us to answer the question of: “How much should we spend on research?”. The answer being: “We want to maximise its social profit”.

The discussion paper states that we are a high performer in research, yet we are unable to translate this into products and services at meaningful scale. This seems to indicate we are not maximising the social profit from our investment in research. Greater social value could most likely be extracted by diverting some of that investment into invention.

I would strongly urge anyone interested in R&D policy to read Richard’s paper if they have not yet done so. It is extremely insightful.

After reading Richard’s paper, I realised that my ideas were more related to invention than to research/R&D. This seems clearly out of the scope of the examination’s terms of reference. So, there seems little point in submitting these.

However, I will make one prediction:

With the current terms of reference, this examination will NOT lead to policy changes which substantially improve our economy and its commercialisation of R&D. This cannot be done unless it also brings “invention” into its scope. In (say) 7 years time, we will probably again be in the same position and need another examination/review to figure this out.

My recommendation to the panel is:

Footnotes

  1. Many thanks to Tim Anderson for his article “This Sputnik-era paper explains why basic research pays for itself” which led me to Richard’s paper. ↩

  2. Nowadays, the definition of “invention” is focused on “new” or “never existed before”. The word “innovation” is used when the focus is on “improvement”. In this submission, the word “invention” encompasses both new and improved (as per Richard’s definition) and is used to clearly delineate a process different from R&D. ↩