On Tuesday, the Parliamentary Budget Office released a sharply critical paper concerning the federal government’s Superclusters project, basically saying, that a) the projects are behind schedule and b) most of the numbers used to justify the project in terms of net benefits and new jobs were utter nonsense.

It’s actually not that interesting a report.  Once you take out the executive summary and the references it’s six pages long with a lot of white space.  The broad strokes of the criticism are nothing new and are somewhat muted compared to what others – and ok, mainly I mean me – have said about the project over the last four years.  It basically focuses on two areas:

Frist, the report criticizes the project for rolling out the money late.  I think they are being generous here, by comparing the money roll-out ($30M as of March 6, 2020) to what was promised in Budget 2018 ($104M).  But if you go back to the 2015 Liberal manifesto commitment, they are *much* further behind on their timeline – like $400M behind.  That’s because the Minister of Shaking Hands with Tech Bros, Navdeep Bains, insisted on a consultation process and then a “selection process” (more on this in a bit) which ate up two years od the Liberal mandate, and then a negotiation process with successful partners that took up another eight months or so.  After which – God help us all – the clusters set up their own competition process which ate up another 6-8 months.  

Second, the report criticizes the process for overhyping the potential outcomes of clusters in terms of GDP growth and clusters.  This really isn’t hard: at the time, I noted that implicitly the feds were claiming that the annual return on investment was about 48% p.a over ten years.  Which is essentially equivalent to suggesting you’ve found a perpetual motion machine: if there were a way to make public funds return 48% p.a., then the question would not be “why are we spending $950 million on this idea” but rather “why are we not borrowing prodigiously at 1% or so in order to invest at 48%, we could wipe out the national debt in less than a decade?”. 

Where the PMO does advance a bit over previous critiques is that they seem to have the goods on exactly how vapid the economic analysis of these proposals was.  The Ministry of Innovation, Science and Economic Development (ISED) was claiming a multiplier of 25(!); the PBO points out that most economists place it between 3 and 8.  They also note – in probably the sweetest three sentences in the whole report:

In a response to PBO’s information request, ISED listed several metrics that they will monitor to measure innovation, such as increase in productivity or the development of new products and processes. To PBO’s knowledge, ISED does not have quantifiable objectives for any of these metrics, and the data reports did not contain information that would have allowed PBO to measure these metrics. As such, PBO cannot draw conclusions about whether the ISI is, or will be, succeeding in accelerating innovation.

In other words, not only did ISED utterly fabricate the projected returns, they also chose (and they had a choice) not to actually develop metrics that permits evaluation of these fabricated returns. 

(I know everyone talks about how the Harper government committed various crimes against science during its term in office, but I can’t think of any crime against social science quite as big as this one.  It’s actually deeply troubling that any public servant signed off on these figures and accountability measures).

One thing the PBO report did not castigate the program for was the tortuous and wholly nonsensical selection procedure that led to the creation of the superclusters.  ISED was basically unable to keep the politics of regional development out of the process, and so a selection process which was allegedly entirely merit-based magically led to one supercluster for each region of the country.  And moreover, those same “regional policy” instincts insured that the basic insight underpinning cluster theory – that geographically proximate firms can lead to huge knowledge spillovers – was thrown overboard in order to make sure that benefits were spread widely across each region.

And yet, and yet.

Almost everything about the roll out of superclusters has been a disaster and the Liberals deserve every bit of criticism they get for this.  But there are some ideas at the core of the concept that are worth supporting.  Of course, these programs were never going to generate multipliers of 25 – but multipliers of 3 to 8 are nothing to sneeze at.  The superclusters are designed in part to strengthen partnerships between corporations of various sizes and specialities, and this is in fact a well-established property of successful clusters.  The administration, governance and roll-out of these projects may be a hot mess, but that doesn’t mean a lot of good might still come out of these projects.  We simply do not know yet.  And, as I pointed out a few months ago, there are a lot of promising Supercluster-adjacent ideas out there which might be worth pursuing.

In short: save us from over-promisers like Navdeep Bains and his enablers in the public service, who endangered what might actually be a promising path in terms of pursuing hi-tech growth.