On the Creation of a Capstone Organization for Canadian Research
A capstone agency to manage research funding could be the bridge we need to turn Canadian technology into economic development
In a recent letter to the leadership of the Tri-Council agencies (NSERC, SSHERC, and CIHR) responsible for Canadian research funding, the Minister of Innovation, Science and Industry and the Minister of Health called for consultation around creation of a capstone agency to unite Canadian research funding efforts under one mandate. The proposed structure of his agency is lifted more or less directly from the recommendations of the 2023 Bouchard Report, which investigated the current state of Canadian research funding and made several recommendations for areas for improvement, including creation of the Canadian Knowledge and Science Foundation (CKSF) referred to in the letter as the capstone agency. In this article, I review both the Bouchard Report and the letter calling for action on its recommendations.
The mandate proposed for the capstone agency in the letter comes in three parts: Enabling multidisciplinary research that crosses the mandates of more than one tri-council agency, supporting and coordinating mission-driven research, and enabling international collaborations. While I am generally in agreement that these are all areas in which we could do better, there are some important caveats and considerations that I do not see explicitly considered in the letter.
Specifically, it is crucial that translational activity between research and new industry be a core part of the mandate and that we take this as an opportunity to rethink both how we measure mission success, and how we engage in management of intellectual property arising from industry-academic research partnerships, especially if the capstone agency is intended to focus on international collaborations.
To contribute more than extra paperwork, the capstone agency will need to define and enforce effective standards and policies for governance of research outputs and for the process of commercializing those outputs, rethinking how we define success by developing, collecting, and acting on a new set of metrics of performance that I discuss in some detail in this article.
Done well, this agency could be the bridge connecting Canada’s research outputs with early-stage domestic economic development, establishing the foundations for a new generation of small businesses that can finally capitalize on our highly productive research community. Done poorly, and it will at best add another layer of bureaucracy on top of an already-bloated system of disconnected and siloed research mandates. At worst, it will just accelerate the transfer of Canadian technologies to countries that are better at developing them economically, which at present includes effectively the entire OECD membership.
The Bouchard Report
Completed in 2023, the Bouchard Report is an excellent summary of the problems with Canadian research funding. For the purposes of this article, the high-level takeaway is that while Canada is highly productive in academic research and the tri-council agencies are to be commended for that, the approach is fragmented and unfocused, and this research productivity does not translate into economic development. The fact that we have subdivided research fields into three categories (social science, health sciences, hard sciences) each managed by one of the tri-council agencies makes interdisciplinary research challenging when more than one of these categories is involved, and there is no cohesive vision uniting efforts across the three.
The result is a high level of haphazard research productivity without any follow-up beyond publication or overall direction, and almost no downstream economic benefit from the billions spent on research.
I largely agree with the findings of the Bouchard Report and I strongly encourage readers to read at least their recommendations, particularly 16, 17, and 18 as they relate to translational activities. The primary recommendation of the report was creation of a capstone organization to address these issues, alongside several suggestions to carefully consider not just the research phase, but also what follows.
Lessons from the Bayh-Dole Act
The idea of innovation governance standardization is not particularly novel, Canada is just decades behind. One of many reasons for American dominance in technological innovation is their standardization of innovation governance, completed in 1980 through the Bayh-Dole Act, a piece of legislation that imposed a set of policies on management of IP developed using public funding by American research institutions.
The framework is relatively simple. The important provisions for the purposes of this conversation are as follows:
Title to the intellectual property right relating to output of publicly funded research are owned by the originating institution. If the institution elects not to file for protection, or elects to allow protection to lapse, they must notify the US government, which can elect to take ownership
Read: Publicly-funded research outputs cannot simply be allowed to enter the public domain without the consent of the government
If the organization elects to maintain ownership, it is obligated to patent the technology, and it grants to the government a nonexclusive, nontransferable, irrevocable, paid-up license to practice or have practiced for or on behalf of the United States the subject invention throughout the world
Read: Nothing the research institution does with the outputs of publicly-funded research can result in complete loss of access to the technology
If the institution issues an exclusive license to a technology, then the resulting products must be "manufactured substantially" in the United States
Read: Institutions cannot simply license the technology off to a multinational to be developed elsewhere. Either a license to such an entity must be non-exclusive, or licensing must favor US-based entities.
Research institutions must favor small businesses when considering potential licensees
Read: The US government recognizes that small businesses are more likely than large businesses to actually develop a product with a licensed technology.
The institution must attempt to commercialize the technology. If no attempt is made, or in a handful of other circumstances mostly related to national security, the government “march-in” rights to take control of the invention
Read: Patents cannot collect dust on a shelf as part of an IP moat. If an invention is patented, an attempt at commercialization must be made, on penalty of giving the government the option to take control of the IP.
The last point in the list is the most controversial and is the cause of headaches for lawyers all over the world, while at the same time being the key to making the whole thing work. Even though march-in rights have not once been exercised in the 44 years the act has been in force, the mere threat of the possibility has been enough to ensure a steady translation of research outputs into commercial innovation that benefits the US, and is in no small part responsible for continued American technological dominance.
I am never an advocate for blindly appropriating policy that works elsewhere. Policies work in a specific context, and they cannot be moved without consideration of that context. Nevertheless, there are key elements that contribute to the success of the Bayh-Dole Act that should inform design of the capstone agency.
The Canadian Version
A key precursor to productive execution on any project is to have a clear mission and vision. Without a cohesive plan to unite all parts of the Canadian research and innovation pipeline, our collective failure to commercialize the results of our research is not all that surprising. The lack of clear vision is contagious, and spreads from the level of government and funding oversight into Canadian research institutions. This problem is particularly evident when considering university-level IP policies. Because of the lack of guidance and direction from the government in how the outputs of research are to be managed, every single university in Canada has a completely unique, bespoke IP policy (or none at all). When they exist, these policies are administered by a tech transfer office that is typically underfunded considering the volume of research it manages, since funding for research rarely provides any resources for protection or management the resulting IP. Similarly, every university has a bespoke policy for management of conflict of interest (usually, simple refusal to engage), which adds uncertainty and complexity to the already-complex process of commercial development of technologies arising from Canadian research.
The net result is that most academic technologies never leave academia, or at best do so in a way that leads to no long-term economic benefit to Canada.
While one can debate the merits of various IP policy frameworks (and I will, in a future post), is it clear that this patchwork approach is not leading to economic development of Canadian research outputs in aggregate, and that there is a desperate need in this country to standardize how research outputs are managed.
A patent filed without intent to commercialize, or that gets stalled in tech transfer, is a waste of time and money. A paper published on a patentable idea without first filing a patent is a missed opportunity. While the Bayh-Dole approach is certainly not perfect, the results speak for themselves and are orders of magnitude better than anything achieved in Canada. With the creation of this new capstone agency, we have an opportunity to learn from the successes of the Bayh-Dole Act, to correct its failing, and to build an effective Canadian structure for governance of research outputs that bridges the gap between lab and market.
The most important contribution of the Bayh-Dole Act to translating research was to set minimum standards on how American research institutions manage IP once the research phase is finished. While there remain variations in how licensing is conducted and how the proceeds of innovation are shared between institutions and inventors, there are a minimal and robust set of requirements in place that ensure benefit to the American taxpayers who paid for the research regardless of these details, in the form of domestic economic exploitation of the technology. Defining and enforcing these guidelines for the Canadian context should be one of the core mandates of the capstone agency.
Recommendation 1: The capstone agency should impose a requirement to attempt commercialization of everything that is patented, and a requirement to provide prior justification for publication without patenting, favouring inventor-led commercialization efforts and small businesses. Where these requirements conflict with existing institutional IP policies, it should be incumbent on the institutions to change those policies.
Recommendation 2: The capstone agency should develop and periodically review a set of IP management best practices to guide institutional IP policy development. The default assumption should be that funded research will result in protectable IP, and a management plan for the resulting IP compliant with capstone agency best practices should be a required part of any grant application. Part of the funding provided for research should be explicitly allocated for the cost of IP management.
Recommendation 3: The capstone agency should retain the right to take ownership and protect publicly-funded IP that the institution or researcher elect not to protect, and to take ownership of IP that is not being commercialized. In practice this should almost never happen, but in cases where it does, the Innovation Asset Collective could be a good vehicle to manage it.
Rethinking Academic-Industry Partnership
A significant portion of the research that is conducted in Canada occurs under the umbrella of industry-academic partnerships. The model usually involves an industry partner funding a Canadian research lab, getting matching funding from the Canadian government through any number of grant programs, and the resulting IP being licensed to the industry partner for terms that are rarely made public. The research institution takes overheads as a percentage of the industry funding.
The tri-council (and most Canadian research funders) are entirely hands-off with respect to what happens to the IP resulting from funded research, leaving it to the institutions to negotiate IP licensing with industry partners. In principle, a small portion of the institutional overhead is allocated for IP management. In practice, it is not nearly enough. Combined with the patchwork of IP policies that exist in Canadian research institutions, it is not surprising that under-resourced IP management practices produce varied results.
There are three main classes of possible outcome for these partnerships:
In rare cases where a startup or small business is the controlling licensee, an attempt is likely made at commercialization. This is the best case scenario. Whether successful or not, this drives economic activity, and the subset of successes could go on to be the next generation of Canadian anchor firms, if only we could figure out how to support them effectively. On the other hand, the institution is unlikely to see much short-term licensing revenue, and so even though this is the best case scenario for long-term economic development, it registers as the worst-case scenario according to the metrics that are actually tracked.
A large business with a Canadian branch plant is the controlling licensee, and they actually try to further develop the technology. This scenario can be positive or negative, depending on the details, but is rarely economically optimal for Canada in the long run. While control of the IP is not retained in Canada, there may still be related domestic economic activity, and development of the technology does not stall. In spite of limited long-term benefit to Canada, this registers as a best-case scenario by the metrics that are tracked: licensing revenues are steady, jobs are created, and research funding continues.
A large business with a Canadian branch plant is the controlling licensee, and they shelve or fully export the technology. The patent becomes part of a law-suit deterrent or is developed outside of Canada and every dollar spent on the research is wasted. This is the worst-case scenario, but is indistinguishable from case 2 by any metrics that are actually tracked since the institution still collects license revenues, jobs are still created, and research funding continues.
While the application process for grant funding almost always involves enumerating expected benefits to Canada, it is abundantly clear from all available evidence that the metrics used to assess benefit to Canada are uncorrelated with actual benefit to Canada in practice.
To be eligible as a cost-sharing partner for these programs, companies must usually have a minimum salaried employee count and/or minimum revenues in Canada. At the same time, because the metrics considered when institutions assess tech transfer efficacy usually relate to licensing revenues at the level of the institution rather than long-term economic development at the level of the country, it is much easier to justify licensing to a larger company that can be tapped for immediate and reliable licensing revenues, unlike small businesses and inventor-led startups. Larger companies with deeper pockets can also fund larger projects for matching tri-council grants, providing larger overheads to the research institutions. In aggregate, both the eligibility requirements and the metrics used to assess the benefit of an academic-industry partnership exclude startups from being cost-sharing partners, heavily favouring large, well-established companies.
This is problematic for two reasons. First, the smaller the company, the more likely it is to actually develop IP that it licenses. Managing and licensing IP is expensive, and a startup operating on a shoestring budget is not going to license IP for which it does not have a specific purpose. Large companies, on the other hand, typically license whatever related IP they can, irrespective of intent to commercialize, since they can use the IP as part of their moat. There are exceptions, whose presence can be very valuable for the Canadian ecosystem, but they are few and far between. Second, there simply are not very many large Canadian anchor firms (by which I mean Canadian-controlled as defined by the intercorporate ownership records kept by StatsCan, not simply a CCPC) who can serve as cost-sharing partners in the first place. In combination, this means that eligible cost-sharing partners are usually the nominally Canadian arm of a multinational entity, and that most of the IP capture done by these entities is done without any intent to commercialize.
This is completely backwards, and a direct result of short-term thinking and a focus on the wrong metrics. One has only to look at the Bayh-Dole requirement to favor small businesses as licensees in the USA, or the drive to spin lab-stage technologies into startups in Israel, to see clearly that embracing the risk of failure and being content to operate in the red in the short term is a prerequisite to long-term economic benefit arising from research activity.
In other words, Canada favors precisely the wrong entities when forming industry-academic partnerships at every level of decision-making, because the metrics used to measure success focus on the short-term institutional benefits, job creation, and risk management rather actual long-term economic impact. The choice of metrics used at the institutional level stem entirely from those used by funding agencies. Until the funding agencies start to care about long term economic growth, the institutions will continue to ignore it as well.
I would like to see a study completed that examines control of IP assets, specifically considering the question of what percentage of Canadian research outputs are retained in Canadian control as a function of time, as well as the trade balance associated to the technology, broken down by originating institution. Note that this requires knowledge of the licensees who have access to or control over the IP rather than publicly listed ownership. While this data is not currently collected in most cases and is typically under NDA, this needs to change. A small scale study completed on Canadian AI assets recently suggests that the results would make abundantly clear why Canada has historically been unable to benefit in the long run from its research outputs.
In the absence of data to properly assess this today, I would like to see the capstone agency lay the groundwork for doing so in the future.
Recommendation 4: The capstone agency should require that publicly-funded research institutions publicly report the countries that ultimately have control (as defined by the StatsCan tables on intercorporate ownership) of (sub)licensees of IP arising from publicly funded research, for a period of 10 years.
Recommendation 5: The capstone agency should require research institutions that receive public funding to track and report aggregate trade balance of products and services arising from (sub)licensed IP, for a period of 10 years.
Tracking Outcomes
In my time working at the interface of academia and industry I have seen patent portfolios from all over the country that are filed and subsequently stall and lapse because the university cannot agree with the inventor on reasonable license terms; I have seen research partnerships with the Canadian arms of American multinationals use Canadian taxpayer dollars to fund development of a foreign IP moat that subsequently goes undeveloped; and I have seen ideas that could and should have been patented simply enter the public domain through publications because the professor that invented it did not want to deal with the headache. Every one of these outcomes is a near-100% loss of investment for Canada, while at the same time registering as a win from the perspective of the grant agencies and institutions: papers were published, patents were filed, overheads and licensing revenues were collected.
In other words: the metrics of success used by the tri-council to assess funding impact and to assess commercialization impact are wrong. The capstone agency should address this as a core part of its early mandate. The main issue to be overcome is that the outcomes of commercialization of research are delayed from the point at which research funding ends, often by years. The tri-council agencies can carry a technology as far as TRL 3 or 4 in most cases, but significant economic activity is rarely apparent before TRL 6 or 7, and the full impact will not be felt until well after TRL 9. If the goal of research is to seed the next generation of economic activity, then the impact of research funding and the efficacy of the programs that deliver it can only be truly assessed in terms of related economic outcomes years in the future.
Part of the letter describing the desired capstone agency’s functions refers to mission-driven research, but without long-term tracking of the eventual outcome of those missions, that will never be more than a buzz-phrase. Mission-driven research requires data-informed decision-making and course-correction on timescales relevant to the mission. Canada does not currently have the data collection practices in place to make this possible.
Recommendation 6: As part of the formation of the capstone agency, a set of metrics should be developed relating to economic activity arising from funded activities both at the national level and the level of ecosystems surrounding research institutions, to be assessed over a timeline appropriate to the TRL of the technology in consideration. These metrics should be established in consultation with a panel of stakeholders from the public sector, the VC community, private sector technology firms, and research institutions, and reviewed on a regular basis.
Recommendation 7: All research institutions that receive public funding should be mandated to build and enforce viral reporting requirements compliant with Recommendations 4, 5, and 6 into their IP (sub)licensing arrangements as contractual obligations. Funding for the associated overhead should be part of the IP management funding tied to all issued grants discussed in Recommendation 2.
In Summary
On the whole, the tri-council agencies are effective when purely academic research topics fit neatly into their program mandates. While greater efficiency is always possible, the volume and quality of Canadian research outputs speaks for itself. The potential value-add of the capstone agency will arise in the liminal spaces between tri-council mandates, and between late-stage academic research and early-stage commercial development, areas where the tri-council approach to research funding has consistently failed to deliver results. It is encouraging that both of these issues are called out in the consultation letter.
Addressing these challenges requires that we establish an overall mission and vision for our research activities on a national level, that can serve as a guide for mission-driven research generally. In addition to the issues that are explicitly called out in the consultation letter, it is also critically important that we take the opportunity afforded by creation of this agency to carefully reconsider how we manage intellectual property arising from academic-industry partnerships, and that we completely rethink our metrics of success, the timeline over which we measure them, and the quality and consistency of related data collection requirements.
The change that could be realized through this capstone agency is badly needed, but those in a position to effect it need to hear the voices of those on the ground who understand how. Given what will likely be a disruptive election that could easily derail this conversation, time is of the essence. Whether or not you agree with the points made in this article, I strongly encourage you to submit a response to the consultation call before the July 17 deadline.