A few weeks ago, the Brookings Institution – America’s oldest and possibly most influential think-tank – published a paper called The Case for Growth Centers: How to Spread Tech Innovation Across America.  The paper’s problematique is the narrow distribution of tech growth in the United States (my favourite factoid here is that 90% of the growth in the country’s 13 highest-tech industries occurred in just five metro areas: Boston, San Francisco, San Jose, San Diego and Seattle) and that for the good of the country the federal government needs to encourage more spatial distribution of this growth.  The solution advanced in this paper is that the federal government should i) advance a package of federal innovation supports to establish growth/tech hubs in ten metro regions distant from existing tech hubs and ii) establish a fair and competitive process for figuring out which regions receive the support.

This is highly reminiscent of our Liberal government’s “Supercluster” concept.  Interestingly, the paper passes over the Canadian experience in silence, either because it has been so painfully slow to get going – as of last fall, three and a half years after the Government first outlined the concept, only 5% of total projected funds had been dispersed – or because describing a project as having been inspired by Canadians is the kiss of death in Washington (personally, I think it’s a bit of both).  And yet in some ways the proposal is subtly different, raising some interesting questions about the way our project has been managed to date.

Let’s start with what our government did.  To begin with, the problematique wassubtly different.  First, our government wasn’t trying to just disperse tech growth, it was trying to kick-start it, period.  Effectively, the view in 2016 was that we were starting from near-zero (Kitchener-Waterloo partially excepted) and the point was to launch any tech hubs. This was partly based on a narrative that the Harper government had ignored knowledge industries in favour of resource industries and a new Liberal government was striking out in a different direction.  The government had a strong “regional distribution” focus from a very early stage, though it tried to disguise it by inaugurating a “national competition” that everyone knew well in advance, through a series of nudges and winks, would produce five winners which miraculously would line up with the country’s five regions in a way that none would be left empty-handed.

That said, there are a lot of similarities.  Both the Brooking analysis and the one that kicked off the Supercluster program had similar understandings of what constituted successful, self-sustaining innovation clusters were. Second, both emphasized a substantial presence of direct or indirect government funding for projects that brought together businesses and higher education institutions of various size into dense networks of relationships, leading to sectoral concentrations that reached critical mass and drove self-sustaining growth.  And third, both contend that even though all these previous clusters emerged serendipitously, it should be possible for government to repeat this process deliberately.

(The first and second points there are unarguable.  The third is a leap of faith because it’s never been done before.)

Now, back to the distinctions.  The key one is that Brookings proposes supporting “metro areas”, whereas the Canadian Superclusters program effectively funded “regional consortia”.  This leads to two different focuses.  The first is greater geographic specificity.  A metro region is pretty small; some of Canada’s consortia extend over multiple provinces (Canadian regionalism undermined the spatial concentration aspect of the clusters to the point where the chance of self-sustainment drop enormously).  The second is the focus on the region as a whole, as opposed to consortia of businesses and educational institutions.  Brookings is thinking about this competition as being like the Amazon HQ2 competition, where local/regional governments themselves make the pitch, which to my mind makes a boatload more sense than we did.

There are two major advantages to making deals with regions rather than with “insta-consortia.”  First, regions have pre-existing staffing infrastructure and governance systems, while insta-consortia do not (having to build all those things from scratch is the real reason almost no Supercluster money has flowed).  Second, by dealing with actual political institutions with democratic legitimacy rather than a new organization loosely backed by a segment of local industry, you can work on more issues.  So, for instance, while the Canadian approach involves writing cheques to consortia who will in turn write cheques to other, smaller consortia for individual projects, the Brookings approach involves i) direct spending on educational institutions to expand research and graduate programs in tech areas, ii) tax and regulatory variances within a geographic area, iii) direct business financing as appropriate, iv) specific concessionary policies on use of federal land and infrastructure in a given area, as well as subsidies to improve transport infrastructure and v) comprehensive workforce upgrading, in order to ensure that lower-skilled individuals can benefit from a general economic growth that occurs when a city becomes a tech hub.  

As you can see, the Brookings conception about making a tech cluster is significantly more holistic that the one our Government has been peddling.  And as a result, it has a much higher price tag: roughly US $1 billion annually per tech hub over ten years (ours were roughly C$180 million per hub over five).  And that makes a lot more sense – both in terms of scale and scope – with respect to the kind of effort needed to create multiple new “Silicon Valleys”, as Minister Navdeep Bains ludicrously claimed his program would do at the launch event two years ago. 

The final point about the Brookings piece is the need to use clear and objective criteria to judge which bids make sense, which improves on the farcical and secretive beauty-pageant-ism that characterized the Canadian effort. The authors of the piece helpfully provide a few potential criteria and show how 35 metros across the country compare on these criteria, which on its own is far more transparent than anything our government did.

Now, I am not saying that Brooking has come up with a magic solution here.  The basic objection to superclusters – that there aren’t really any successful examples anywhere of governments intentionally creating such entities – is still there.  But I do think this approach has a better shot at creating self-generating growth than what our government did.  When it comes time for Superclusters II (and, let’s face it, this is inevitable if the Liberals are still in power come about 2023), there are some useful pointers in this proposal for how to improve our own approach to the subject.