If you let any conversation about innovation policy go on long enough, the story of DARPA (the Defence Advance Research Projects Agency, known as ARPA until 1972, and, weirdly, between 1993 and 1996) will likely come up, usually in a form so tortured that it is unrecognizable from the real thing.  This matters because the Business Council of Canada has just backed the idea of a Canadian DARPA as a solution to the country’s innovation woes.  This is, I think, a bad idea, but to understand why, it’s important to understand the history of DARPA.

DARPA was founded in 1958 in the midst of America’s post-Sputnik tech panic.  Its mission was essentially to perform research that would keep the United States ahead of the Soviet Union: in 1958, that meant space-projects, but the creation of NASA the following year meant DARPA got kicked off the fun stuff and on to more mundane military development stuff, which in the early 1960s meant Vietnam-related projects.  Agent Orange and some of the US government’s social science work on counter-insurgency stems were all funded by DARPA in this period.

One way to think of DARPA is as a reaction to the creation of the National Science Foundation.  Recall that during World War II, the Defense Department basically got all the civilian volunteer scientists to work directly for the Army.  None of this messing about with contracts – military people organized scientists and put them to work on highly applied problems.  When the war ended, so too did this arrangement: under the terms laid down by Vannevar Bush in his famous report to President Truman, the way forward was to keep paying academics (or rather, universities) to produce cool stuff, but without the military looking over their shoulders.  That way, researchers could have money and freedom and would come up with creative basic research which someone, somehow, by some miracle, would turn into something commercializable and hence the research would find its way into the economy.  The animating logic behind DARPA was basically: screw the universities – why don’t we just set up our own internal outfit that gives researchers money and freedom, as long as they stick to a general mandate of military application? 

This approach requires a high degree of tolerance for failure – the mechanical elephant designed to take the fight to the Viet Cong in the jungles of South East Asia comes to mind – but occasionally it created some spectacular successes.  The three most famous of these – the ones on which DARPA’s legendary reputation is based – are GPS, the internet and graphical user interface.  These did eventually make it out into the economy in various ways and had an enormous impact on the development of modern business and society – but you’ll note that the most recent of those discoveries is now nearly fifty years old and occurred before Congress clipped DARPA’s mandate because it was tired of the way it fooled around too much with non-military projects.  The only recent DARPA project which has made it big in the real world is the technology behind Apple’s Siri.

When the economist Mariana Mazzucato bangs on (incorrectly in my view) about the entrepreneurial state, she is talking about things like DARPA – government funding not just basic research but also the development of things that eventually become major parts of the economy.  And it’s not just the left that thinks DARPA is whizzy.  UK Prime minister Boris Johnson’s former Svengali, Dominic Cummings, was also deeply enamoured of the idea that a bunch of misfits working in secret with oodles of money and low levels oversight could be much more innovative than government could ever be and so spearheaded the idea of a DARPA for the UK, which is about to become operational.  Japan has a “DARPA-like” agency called the New Energy and Industrial Technology Development Organization, Germany launched a “DARPA-like” agency for cybersecurity in 2018 – and of course the Americans tried to recreate DARPA when the Obama administration created “ARPA-E” to work on energy-related projects.  Literally no one thinks any of these children-of-DARPA are nearly as successful the original, which suggests that there the original’s success might have owed more than a little something to serendipity.

Despite all this – or perhaps in ignorance of it – the Business Council of Canada (BCC) in its pre-budget recommendations abruptly called on the Government of Canada to:

…create a more robust innovation ecosystem that contributes to growth, productivity and higher living standards, the federal government should embrace a demand-side, mission-driven approach to innovation policy, setting clear national goals to translate scientific strengths into economic performance. Specifically, we recommend that Canada create an agency similar to the U.S. Defense Advanced Research Projects Agency (DARPA). The agency should focus on high-growth sectors in which Canada has a competitive advantage – agri-food, health and biosciences, clean tech and energy, advanced manufacturing and digital services – using public-private partnerships to help scale-up and commercialize research.

Little of this paragraph makes sense.  First, whatever the hell DARPA is, it is not a demand-side approach to innovation policy, which the OECD defines as things like public procurement, regulations, standards and consumer policies.  Second, demand-side, by definition, is pretty much the opposite of mission-driven.  Third, DARPA has never had scale-up and commercialization of research as a mission.  And fourth, DARPA, to the extent it ever worked, worked because it had a relatively narrow mission, not a grab-bag like “agri-food, clean-tech and digital services”. 

If you want to read this proposal charitably, just ignore the words “demand-side” and “DARPA”: what remains is the idea that the Canadian government should help scale-up and commercialize research in certain high-value areas.  It’s not entirely clear how such an idea differs from the Centres of Excellence for Commercialization and Research or the Business-led Centres of Excellence introduced by the Harper government, which the Liberals axed (sorry – “folded” into the notoriously slush-fund like Strategic Investment Fund”) three years ago.  However, it’s consistent with other BCC proposals like creating something like Germany’s Fraunhofer Society to spread excellence in applied innovation (which, again, is more or less what the Harper government tried to do with he National Research Council, but that’s another story)

But maybe the more important question is: what problem is it that these proposals are meant to solve?  The idea that an actual DARPA – even if it worked – is relevant to solving Canada’s persistent innovation and productivity problems is untenable.  It is deeply unlikely that a Canadian scientist is about to invent something as cool as GPS or the internet and even if that happened it wouldn’t be commercialized in Canada.  With the more charitable reading of “more applied research, scale-up and commercialization”, we get closer to a reasonable diagnosis: the problem isn’t knowledge-generation, it’s about all the multitude of things that go between research and a final product.  That’s not a outlandish hypothesis.  But I am not sure it is true.

I mean, this all seems pretty close to the agenda the Harper government pursued in terms of trying to increase business research and development intensity, and it wasn’t all that successful.  It wasn’t obviously worse than what the current lot is doing, but the evidence that it actually increases…well, anything – innovation, productivity, commercial activity, business expenditure on research – simply isn’t all that good. 

I would suggest that there is a deeper problem at work here.  The BCC diagnosis is not in principle that different from that pushed by the U15 or a large part of the Canadian innovation policy community.  It assumes that innovation is entirely about products and not about processes.  And it assumes that the only thing standing between our current research strengths and commercial success is a lack of government subsidies to act as a bridge between the two.  They might differ a bit in emphasis and who should get what proportion of the money, but basically the analysis is the same.

But if what you want is a more productive economy then the issue is not the number of new inventions or gadgets invented in Canada.  That’s almost irrelevant.  Only in a handful of countries can the economy be said to run on locally-developed technology.  Every country in the world is a technology-taker.  The issue is how countries use the technology, and how they apply it creatively in everyday business and industrial processes.  This is what countries like Sweden and the Netherlands and Estonia are so good at, and what we are crap at.  And at the heart of the technology adoption/absorption issues is one of skills. I suspect, for instance, that a wholesale revamping of the nation’s graduate programs to orient them more towards industry (and why not?  It’s not like academia can absorb them all) would have a much greater effect on innovation through the spread of skills into industry than any commercialization scheme could possibly have.

It’s also a question of business management, something that a “Business Council” of Canadian companies should excel at and know.  Or maybe not, given that Canadian companies are one the OECD’s weakest in terms of investing their own money in innovation.  Indeed, seeing a group ask for more government money in more government programs to solve a problem caused by their own inaction is a little bit frustrating. 

But it is very Canadian.