The Dream Machine, The Imagineers of War, The Pentagon’s Brain… Perhaps not phrases you’d normally associate with a research funding agency. But these are all titles of books about the Advanced Research Projects Agency (ARPA) – an American funding body with almost mythical status in some quarters.
US President Dwight D Eisenhower set up ARPA in 1958. It was a response to the Soviet Union’s launch of the first Sputnik satellite, which had caught the US by surprise. ARPA’s mission was to make pivotal investments in breakthrough technologies for national security.
ARPA, which became DARPA when “Defense” was added to the name in 1972, is famous for helping to bring forward many of the big innovations of the last half century, from the internet to GPS and self-driving cars. It’s also well known for its enormous resources – it has an annual budget of over US$3 billion (£2.7 billion).
‘Get Brexit done, then ARPA’
Dominic Cummings, Boris Johnson’s chief special adviser, has put a UK version of ARPA high on the government’s agenda. His WhatsApp profile apparently lists his priorities as “Get Brexit done, then ARPA”. In the post-election Queen’s Speech, the government committed to “backing a new approach to funding high-risk, high-payoff research in emerging fields of research and technology”. Civil servants are now consulting research and innovation experts about how this might be done.
So what’s the appeal of ARPA, and why has Cummings made it such a high priority? His blog gives some insights – he’s written about ARPA several times, with one 2018 post accompanied by a 47-page essay.
Cummings is interested in a particular phase in ARPA’s history – the period between 1962 and 1975 – which he sees as extraordinarily productive. He focuses largely on the work of one ARPA department, the Information Processing Techniques Office, and its first head, Joseph Licklider.
Licklider, known as “Lick”, was by all accounts a visionary and highly talented manager. He’s credited with having foreseen major developments in personal computing and funded the research to make these happen. DARPA does not do its own research, but instead channels funds to researchers it sees as highly competent. Cummings explores the relationship between ARPA and Xerox’s famous PARC laboratory, which produced, among other groundbreaking innovations, the first personal computer.
He draws a few distinct lessons from ARPA’s history. The first is about the value of high-risk research. Cummings argues that even though many of ARPA’s projects failed, the “trillions of dollars of value” its successes created justify the investment. The second is about how to organise very high performing teams. Cummings sees ARPA’s success in this period as a result of its approach – low bureaucracy, high trust – and the management style of leaders like Licklider, who brought in “great people” and gave them a lot of leeway, while making sure they also connected with one another. It was its way of working, Cummings argues, that made 1960s ARPA an “extreme outlier” in performance.
Does the UK need its own ARPA?
The ARPA idea speaks to two of Cummings’ strong interests: promoting science and improving government performance. He sees ARPA as a key part of a post-Brexit strategy for the UK, making it “the best place in the world to invent the future”.
Cummings is also vocal about his frustrations with the civil service and argues that the government could perform much better by cutting bureaucracy and hiring very talented people. Another recent blog post, which received considerable media attention, described his recruitment strategy for government advisers – “hiring data scientists, project managers, policy experts, assorted weirdos … who have a one in 10,000 or higher level of skill and temperament”.
But the jury is still out on whether there really is an ARPA-shaped hole in the UK’s research and innovation system. On one hand, as experts convened by the think tank Policy Exchange have recently argued, the UK could benefit from more funding for high-risk, high-reward innovation projects, of the type that an ARPA might provide.
In addition, while the UK is already strong in basic scientific research, it could do better at translating it into commercial technologies. Again, an ARPA-style body might help fill this gap if it could build strong relationships with potential customers, particularly in the public sector, and with investors.
But there are also other important priorities for research and innovation funding, such as reducing disparities in economic performance between north and south, and helping the UK make the transition to a zero-carbon economy.
At Nesta, we’ve argued that if the ARPA idea is to create value for the UK, it should focus on a mission, like tackling climate heating, that’s clearly in the public interest. It should also boost innovation in parts of the country where the economy isn’t so strong.
An ARPA-style body in the UK could add value to the innovation system if it reaches a wider pool of innovators than existing funding mechanisms. Challenge prizes are a particularly good mechanism for this. Rather than picking one team and providing a subsidy based on the most credible proposal (the traditional R&D grants model), challenge prizes incentivise multiple teams to compete to provide the most impactful solution. The result is a greater number and variety of innovators working on a topic, and hence an ability to support more radical or unproven technologies as well as the safe bets typically funded through grant mechanisms.