The writer, a former Downing Street special adviser, is co-founder and chief executive of PUBLIC

Every government minister wants a “big bazooka”, as former European Central Bank president Mario Draghi called it, to fire money at a problem. Earlier this week, UK business secretary Kwasi Kwarteng announced he is pushing forward with Arpa — a “blue skies” funding agency modelled on the US’s Defense Advanced Research Projects Agency.

This is the government’s bazooka. Its blast, goes the hope, will address problems created by a history of low and regionally concentrated R&D, and a neglect of translational research. Its creation will also reassure innovation advocates, who feared the project would stall with the departure from Downing Street of Dominic Cummings, the former government adviser who championed it.

Should this new agency achieve what it sets out to do, it will help find solutions to some of society’s biggest challenges. It will also act as a model for other public sector innovators as they seek to build on the lessons of the current crisis.

However, for this new agency to deliver a more strategic approach to research and development, focused on long-term challenges such as achieving net-zero and building back better, its architects will need to ensure it is sufficiently well-designed and well-funded. Failure would mean it becomes just another piece of innovation theatre.

But what we know about the agency’s budget — expected to be £800m — is less than reassuring. It may simply be not enough. Put in perspective, Google’s R&D sister company, the “moonshot factory” called X, spends upwards of $3.5bn a year. It is also a comparatively small piece of the UK business department’s proposed £22bn R&D budget. Tackling “moonshot” goals is expensive, as the pandemic has laid bare.

Investment in risky projects also has the potential for expensive public failures, as shown by X’s recent closure of its Loon project, which aimed to bring the internet to the world via hot air balloons. If Arpa is to be different from the public bodies that came before, it must be willing to allocate sufficient funds and embrace a way of working more typical of the world of venture capital — that of “high-risk, high-reward” investment.

Effective prioritisation will be vital. Arpa will need to identify the key challenges it wants to address and adopt a mission-driven approach. The US’s Darpa works partly because it is so singularly focused on maintaining US military superiority. It can rightly claim to be behind the M16 rifle and lightweight mortars but also the computer mouse and the Global Positioning System. Arpa would do well to consider equally specific missions.

Key to this will be avoiding the pitfalls of siloed working that is typical of the public sector — and ensuring that the agency can cut through divisions between policy areas. While the Department for Business, Energy and Industrial Strategy will be its sponsor, Arpa will need to have a cross-government remit. Its agenda, like Darpa’s, should be shaped around a small number of grand societal challenges in which multiple departments have a stake.

Progress happens when great innovations are disseminated economy-wide. Very rarely does one firm create all the change, by itself. Shipping containers, for example, transformed the world economy and their impact and use went far beyond what their inventor Malcolm McLean could have envisaged or profited from. So it will also be important for Arpa to be embedded in a broader ecosystem of companies. Think of it as a sun around which many planets rotate, rather than a rogue star.

Then there is the question of how the agency should be run, and by whom. It will demand skills and attitudes rarely seen or incentivised among even the best civil servants. Part of Darpa’s success lies in its cadre of 220 managers and staff, who typically serve for three to five years. They are responsible for bringing new ideas and connecting with other parts of the US Department of Defense. As “Arpa-watcher” and innovation policy professor Richard Jones has argued, the British version should replicate this model.

Its leadership will need individuals well versed in high-risk investment. The government should pay well and even consider incentivising staff, venture capital-style, with some form of carry in products developed by Arpa. The government should also consider an “Arpa fund”, to sit alongside the agency, through which the private sector can match-fund. It could ensure that any intellectual property will, over time, benefit both the economy as a whole and the Treasury directly.

Ultimately, technological advance and corresponding economic growth relies on many things. But Arpa is a rare opportunity, a sandbox in which to experiment. It must be furnished with the necessary remit, skills and resources to enable it to succeed, lest it become an expensive misfire rather than a well-targeted bazooka.