In 2020 a paper from economists at the University of Stanford – Are ideas getting harder to find? – argued that research productivity is falling everywhere.
In the US alone research productivity is falling by half every 13 years. As they note:
Put differently, just to sustain constant growth in GDP per person, the United States must double the amount of research effort every 13 years to offset the increased difficulty of finding new ideas.
Comparing this to Moore’s Law – the idea that the number of transistors in a microchip doubles every two years – the authors calculate that it requires 18 times more researchers to discover a new idea now than in 1970.
The implications of these findings are profound. As society pushes back the frontiers of knowledge it requires ever more people, funding, and research power, to make incremental steps forward.
Long-term R&D budgets take no account of the relative levels of productivity they are leveraging but simply the proportion of GDP spent on research. To put it another way, 2.4 per cent of GDP spent on research by 2027 delivers less productivity than today, yesterday, or five years ago, unless the UK economy goes through some exponential growth in that time.
On growth itself one of the suggestions within the Stamford paper is that “follow on” innovations have a declining level of utility compared to initial discovery.
In turn, this leads to “defensive” R&D. That is companies, who form the majority of R&D spend, become more focussed with protecting their market position than making incremental developments. The R comes to outweigh the D.
To our dual threats -innovation becoming more expensive and incentives stacking up against continual innovation – we can also add that innovation often comes from the interaction between diverse fields, but industry and academic incentives encourage subject specialisms.
The shortcomings of funding
In addition to ideas being harder to find they are not becoming any easier to fund.
Since 2015 UKRI has never funded more than 28 per cent of the total ideas put to it, and in 2020-21 had an award rate of only 21 per cent. The Grant review of UKRI points to various organisational and digital infrastructure reasons for incompatibility between some parts of the funding ecosystem while UKRI’s own Simple and Better Funding programme is aiming to reduce the time finding, applying for, and managing, research funding.
I am not drawn to arguments for less bureaucracy with no alternatives. Bureaucracy is after all the means through which we ensure the more equal treatment of people and fair dispersal of funds. I am however in favour of trying new things to address enormous challenges.
This I believe is the underappreciated aspect of the new Advanced Research and Invention Agency (ARIA).
It is my overwhelming instinct that many of its aims could be achieved within a reformed central research bureaucracy. There is nothing to stop UKRI opening more opportunities for high-risk projects with more flexible funding allocations geared toward brilliant people. However, until we have seen ARIA in action, we don’t really have any idea how well it will work or what it means to have an autonomous research funder operating parallel to a centralised research system. ARIA is a project to upend the cultures of research as much as it is a project to fund new discoveries.
Failure on its own terms
It is for this reason that it deserves a chance to succeed and crucially to succeed on its own terms. A £800m budget over the course of this parliament is not large enough to tolerate the rates of failure which will be necessary for the occasional breakthrough. This is roughly equivalent to 10 per cent of UKRI’s annual budget or less than half of the total business R&D spend on defence alone. It is a more egregious waste of public funding to set a budget which limits the scope for success by being too small than it is to prevent the waste of money in the future by having a larger budget and definitively proving whether this funding model can work.
There has also been significant scrutiny of ARIA’s independence. Again, my gut reaction is that the complexity of innovation requires investment in existing infrastructures like universities and businesses, and a project approach to bringing together disparate groups of people will only add complexity and inefficiency to existing systems.
Equally, guarding ARIA against the use of FOIs when issues of commercial sensitivity and national security are already exempt in the legislation seems pointless at best and point-proving at worst. However, again, if the Government is going to undertake a project like this it should be done on its own terms. If the point of ARIA is to be a risk taking-independent organisation it makes no sense to reverse engineer out its risk and independence.
The question then is how the success of ARIA should be judged. Firstly, it will be whether it produces any of those breakthrough findings which have so far escaped the clutches of traditional funding routes. It is hard to imagine what may come from ARIA but it should be given a fair shot of success. This might even mean more funding or expanding its funding horizon which is exceptionally short at three to five years.
The second measure of success will be what it shows the research ecosystem about the possibility for different ways of working. Despite its size, it will give insight on whether another research bureaucracy is possible. If this leads to better ways of working, less time wasted on applications, and more dynamic research cultures, then it will be £800m well spent.