Can the UK replicate the success of the ARPA model?
One of the most eye-catching parts of the UK’s plan to become a “global science superpower” is the creation of the Advanced Research and Invention Agency (ARIA). Inspired by the US Advanced Research Projects Agency (ARPA) model, the new agency was announced in the March 2020 Budget, and is expected to become operational by 2022.
The UK government has said it will invest £800 million in ARIA over its first four years. The agency will target blue-sky research areas that have potential for creating transformative technologies. It will be run by Project Managers — a feature that is core to the ARPA model. Usually hired for a short tenure of four to five years and given a uniquely high degree of autonomy, these Project Managers are highly skilled individuals equipped with a combination of cutting-edge technical knowledge, networking and coordination abilities and risk management skills, which together enable them to navigate unchartered areas of innovation and prototype their radical ideas. Efforts are being made to ensure the agency’s organisational independence, shielding it from political pressure and securing its existence for the longer term (the legislation specifies a grace period of 10 years before any dissolution can be triggered).
The first application of ARPA-related ideas in the UK’s science and innovation policy came in the aftermath of the Brexit vote in 2016, as a means of delivering new challenge-based research funding. As the UK government looked to play a proactive role to reinvigorate the economy, it looked to the once ominous idea of “industrial strategy”, and rebranded it by putting science and innovation — which has for a long time been a substitute for an active industrial policy in the UK since 1990s — at the centre. The UK’s Industrial Strategy, launched in November 2017, set out the broad directions of travel in the shape of “Grand Challenges” around which R&D activities would coalesce.
The ARPA model was first mentioned as an inspiration for a cross-disciplinary public funding stream when the Industrial Strategy Challenge Fund was launched, which sought to bridge collaborations between universities and businesses through crowding in matched funding from the private sector. But it wasn’t until Boris Johnson became Prime Minister in 2019 that the idea to replicate ARPA more comprehensively gradually rose to prominence. His once influential adviser, Dominic Cummings, has long been an advocate for the model, which aligns with his two top priorities: promoting science (and technologists), and improving how government and bureaucracy functions.
Reflecting on the history of ARPA between 1960s and 1970s — a period where the first head of its Information Processing Techniques Office, Joseph Licklider created visionary collaborations between ARPA and Xerox’s PARC laboratory — Cummings observed that ARPA’s success in pursuing high-risk research is enabled by a “low bureaucracy, high trust” structures where talented individuals are given freedom. Despite leaving the government before ARIA was launched, its current design is largely consistent with what Cummings had envisaged.
Overall then, is the establishment of ARIA a good thing? Broadly speaking, more public investment in UK’s science and innovation in any form is to be welcomed. The UK has long been a laggard in R&D investments: its domestic expenditure on R&D as a percentage of GDP is about 1.6% — significantly lower than the OECD average (2.5%), US (3%) and China (2.2%). Even so, the UK has been punching above the weight in terms of the quality and impact of its scientific research, and its field-weighted citation impact has remained world leading.
In addition, the UK recently announced a new target of spending 2.4% of its GDP on R&D. Every extra bit of spending will count towards reaching that goal, especially in the wake of losing out on European research funding following Brexit. However, some of ARIA’s initial funding allocation has come from the existing science budget, and how much of the £800 million committed by the government will be additional money remains to be seen. But as well as helping to generate additionality through its own funding, ARIA can also create opportunities for further private investments if its activities lead to major technological breakthroughs.
At this early stage however, many unanswered questions remain. Firstly, there is the question of the scale of its investments. ARIA’s initial annual budget of £200 million is about 1% of the UK’s public R&D budget, and 0.5% of UK’s total R&D expenditure. In comparison, the archetypal ARPA model it is inspired by, DARPA, spends $3.5 billion per year, which is about 2.5% of federal R&D spending and 0.6% of US’s total R&D spending. While the scale of the investments of the two agencies are more comparable in relative terms, there’s a huge difference in their absolute monetary values. When it comes to R&D, quantity matters — not least to afford the cost of the many failures commensurate with a higher risk appetite. The UK already underspends in innovation, and for ARIA to achieve what DARPA has achieved for the US, the former should seek to exceed the latter’s spend in relative terms.
Second, there is the question of ARIA’s purpose. ARIA will be sponsored by the department of Business, Energy and Industrial Strategy (BEIS), which was created in 2016 to champion the Industrial Strategy. Even though the term ‘Industrial Strategy’ has been removed from the current government’s policy agenda, BEIS will remain the key driver for science and innovation, and it makes sense for ARIA to provide continuity for a challenge-based approach to R&D. However, in the US ARPA model agencies are usually tied to specific departments (e.g. defence, energy). However, ARIA “will not be tied to a single research focus, industry, or government department customer”, and Project Managers will be responsible for defining the programmes of individual missions. This has resulted in a parliamentary report accusing ARIA of being a “brand searching for a product”. How ARIA will work with BEIS and with other government departments is far from just an academic question. In particular, the extent to which ARIA can leverage procurement power associated with different departments to drive innovation will be key. Procurement has been critical to DARPA’s success and is a key reason why the model is difficult to replicate.
Third, there is the question of how ARIA fits with the UK’s existing R&D funding structures. Over half of UK government’s R&D budgets — about £8 billion — is funnelled through UK Research and Innovation (UKRI). UKRI was created in 2017 following the recommendation of a 2014 review of the UK’s innovation funding mechanisms, with aspirations that seemingly echo that of ARPA’s. Through the partner organisations that constitute UKRI, universities obtain “core funding” based on quality-related metrics, and compete for discipline-related grants (e.g. STEM and humanities) on a competitive basis. The relationship between ARIA and UKRI has been a point of contention, and it still remains unclear. This is especially the case where there are clear overlaps in approach and objectives, such as the aforementioned Industrial Strategy Challenge Fund. What is clear is that the creation of ARIA itself and the stress on its independence of is already a statement of intent, and that the perceived bureaucracy of UKRI and centralisation of science and innovation funding may have held back “high risk, high rewards” research.
In addition, there’s a perception that while the scientific base of the UK is strong, it is bad at translating its excellent science into advanced technologies of strategic importance. The stories of graphene and monoclonal antibodies have often been cited as examples, but while both these examples may have some truth to them, this only extends to a certain point. There is some evidence that the UK’s lack of “high risk, high rewards” research and the stranglehold of bureaucracy may be exaggerated. The country’s recent success with Covid-19 vaccines is a case in point. Translating science into technology is a universal problem that many countries claim they have, and there’s always room to do better. But the UK’s perceived failures in commercialisation also have to do with more conducive and attractive environments for doing so elsewhere.
The establishment of ARIA represents an interesting development during interesting times. It may well be capable of creating sprinkles of magic, but on its own it is unlikely to be the silver bullet that transforms the UK’s science and innovation landscape.
But then again, as Zhou Enlai would have it, “it’s too early to say”.