Late last week saw the publication of the ARIA Framework Agreement.
The framework sets out how the Advanced Research and Invention Agency (ARIA) will work, and crucially where it will not work, with the newly formed Department for Science Innovation and Technology (DSIT).
Free to do whatever I want
During the passage of the ARIA Bill, one of the major selling points of ARIA was that it is free of the usual research bureaucracy. As far as possible, the idea is that researchers will be given money to pursue the projects that are interesting to them without the approvals, form filling, or the external scrutiny that comes with mainstream research programmes.
In short, the goal is to give talented people from a range of disciplines the maximum amount of time to focus singularly on the biggest problems that face humanity. This should mean that ARIA is not only working on new discoveries – it is also discovering new ways in which to do research.
In order to prove the test case, as I have previously argued on this site, ARIA needs the space and time to succeed or fail on its own terms. If anything, a budget of £800m might not be enough to thoroughly test a funder that can tolerate significant short-term failure in order to achieve long-term scientific breakthroughs.
There are lots of agreements of this kind that set out how semi-autonomous bodies will work with their sponsoring government departments. For example, the framework agreement between UKRI and the now departed Department for Business, Energy, and Industrial Strategy (BEIS), sets out why UKRI exists, how it is accountable to BEIS, how performance is monitored, which functions are delegated, and so on.
There are some specific clauses within the ARIA and DSIT framework that are worthy of specific mention.
Little by little
For a start ARIA is, in principle, properly independent. The idea of “maximum autonomy” is enshrined within the framework. This isn’t only autonomy over the projects it chooses but how it sets internal cultures, how it operates its finances, and how its overall project mix will work. As covered extensively elsewhere ARIA is also not subject to the Freedom of Information Act (FOIA).
The exemption from the FOIA is important not only as a principle but because it reveals a central tension on the extent to which transparency should be done away with where that transparency may slow down the actual work of an organisation. ARIA is required to publish data on its operational costs and the regional distribution of its programmes and DSIT is not exempt from FOI requests about ARIA. So it is not that there is no transparency but the transparency is tightly controlled. My own view is that even where it occasions some organisational drag more transparency is generally better than less.
Another example of the tensions around independence is that while there is a more streamlined reporting mechanism compared to some funders the Advanced Research and Invention Agency Act (2022) requires an annual report to be laid before Parliament, there is internal and external audit, and a range of broader legislation it must comply with including the Equality Act. Again, there is freedom, there is independence, but there is a limitation to that when an organisation is spending public money.
Perhaps the wider issue with the FOIA exemption is that even where there is nothing to hide there will always be the suspicion that there is. ARIA is free to control its own communications functions but there is an expectation that ministers will be involved in any announcements on significant programme funding or board appointments. However, it is made explicit that communications imperatives do not supersede the actual work of the organisations – effectively ministers can enjoy reflected glory from ARIA but they cannot tell it what to do.
This dynamic also extends to the appointment of the Chief Executive. The first Chief Executive of ARIA, Ilan Gur, was appointed by the Secretary of State. In the future the Chief Executive will be appointed by the Chair, currently Matt Clifford, in consultation with the other non-executive directors. Although not as explicitly set out within this framework under the Advance Research and Invention Agency Act (2022), the Secretary of State may remove members of the Executive under national security grounds, and non-executive members under “such other grounds as the Secretary of State considers appropriate.” Equally, while the framework agreement states the Chief Executive is responsible to Parliament he is also shielded by a Secretary of State who lays key documents in front of it.
In total, it is independent but with pretty strong failsafes. This is not to say that there is not much more latitude than would be seen with basically any other part of government. This section of the framework is instructive in this regard
It should be noted that ARIA’s objective is to pursue high-risk, long-term R&D and is therefore expected to have a significant risk appetite. Investing in novel or contentious research should therefore not constitute novel, contentious or repercussive spending for ARIA
Again, another example of the tension of independence, ARIA can appoint up to 24 roles outside of HMT senior approval thresholds while it must also have due regard for the general principles of public sector pay. These two mandates, on first reading, do not feel wholly congruent.
Aside from these pockets of tension that are inevitable with an independent body operating with public funding, perhaps the farthest reaching guidance and regulations is that ARIA will not be wholly reviewed for ten years.
It is set out in law that the Secretary of State cannot dissolve it until 2032 at the earliest. In spite of some initial scepticism, if the case was proven that a low bureaucracy, patient, high investment, funder is effective, it will gain plenty of supporters across the research landscape. The important question then is how universities may respond with their own internal bureaucracies in the face of new competition for people and funding.