Using research findings to design a better SBIR for Canada

Among the innovation-related initiatives announced in Budget 2017 is Innovative Solutions Canada, a $50 million program inspired by the US Small Business Innovation Research (SBIR) program.  Recent research has shed new light on the effectiveness of the US SBIR program, and suggests that Canada may be able to design a superior program by offering modest levels of support to young companies in emerging sectors, and letting the private sector provide follow-on financing.

Genesis of SBIR

The SBIR program was founded in 1982 to stimulate technological innovation and help meet federal R&D needs.  US federal agencies with extramural R&D budgets of more than $100 million are required to allocate a share of their R&D budget, currently 3.2%, to SBIR programs.   Currently 11 federal agencies run SBIR programs.  SBIR programs are structured into three phases.  While specifics vary by sponsoring agency, Phase I grants are approximately US$150,000K, Phase II grants approximately US$1 million, and Phase III awards involve follow-on activities that may be funded from other sources.  Programs are highly competitive—approximately 13% of SBIR applicants are selected for Phase I funding.  Since 1982, SBIR programs have awarded 112,500 grants worth a total of $26.9 billion to US companies with fewer than 500 employees.  Several other countries including Japan, the UK, the Netherlands, and Finland have developed SBIR-type programs.

Early evaluations of SBIR

Early evaluations of the SBIR program produced conflicting results.  One study, conducted by Josh Lerner at Harvard, and published in 1999, showed that SBIR had a positive effect on company revenues, employment, and ability to attract venture capital (VC) financing.  But another study, conducted by Scott Wallsten, then at Stanford, and published in 2000, showed the SBIR program had no effect on company investments in R&D or on employment.  While, from an econometric point of view, the approach taken by Lerner was less rigorous than the approach taken by Wallsten, it was more flexible and allowed the author to probe more deeply into the mechanisms that affect SBIR outcomes.  Lerner found that SBIR results are confined to regions in which there is substantial VC activity.  That is, it was not the SBIR program alone that led to growth in sales and employment for client companies, but the complementarities between the SBIR program and the availability of VC.  The results were particularly pronounced for R&D-intensive companies.

A state-of-the-art evaluation

A more recent and more rigorous evaluation of SBIR picks up on Lerner’s findings and shows that winning a Phase I SBIR award makes it more likely that a company will attract VC financing, and that this relationship is more pronounced for firms in emerging sectors.  Earlier this year, Sabrina Howell (New York University), published an evaluation of the Department of Energy’s SBIR program.  Based on data from over 5,000 applicants to the program between 1983 and 2013, Howell shows that being successful in a Phase I competition increases the average number of patents awarded to a company by at least 30%, increases the company’s chance of receiving VC financing by 9% (as well as increasing the amount of money raised and the number of deals), and doubles the likelihood of positive revenues.

Further, for those companies that have positive revenues, winning a Phase I award results in a 30% increase in revenues.  Last but not least, winning a Phase I award increases the probably of survival, and successful IPO (initial public offering), and acquisition.  These effects are more pronounced for young companies and, as mentioned, for companies in emerging sectors.

Remarkably, Howell finds no effects from Phase II awards—save for a small effect on patenting.  Howell reports that almost 40% of Phase I awardees do not apply to Phase II, and that companies that are successful in attracting VC financing are less likely to apply to Phase II than companies that do not.  This adverse selection of Phase II companies may explain the modest effects of Phase II funding.

Howell conducts further analysis to establish how the modest levels of funding associated with Phase I awards produce these effects.  First, she considers the possibility that it’s not the funding itself that is critical, but the signalling effect of having won the award.  Given the uncertainty surrounding investments in early stage companies, does winning a SBIR award signal to investors that a company is investment-worthy?  Howell dismisses this explanation, because the rank of an SBIR applicant is uncorrelated with the likelihood of raising financing, because the effects of winning a Phase I award grow over time while the effects of a signal would diminish over time, and because winning a Phase II award provides no additional signalling value.

Rather, it is the funding itself that is valuable.  A survey of 347 grantees shows that, despite grantees being unconstrained in how they can use the funds, the award is overwhelmingly used for R&D, and that companies that use the funds for technology development are more likely to raise VC financing than companies that use the funds for other purposes.

Overall, the results confirm previous findings regarding the financing constraints faced by young, technology-intensive companies.  SBIR Phase I grants are effective because they allow such companies to invest in technology development and prototyping activities that reduce uncertainties, thereby increasing the likelihood of attracting financing.

The policy implication for Canada is that the Innovative Solutions Canada program should focus on capital constrained young companies, award modest grants that alleviate the capital constraint by allowing companies to invest in technology development activities that demonstrate commercial readiness, and continue to work to increase the availability of equity financing.

Margaret Dalziel is an associate professor at the Conrad Centre for Business, Entrepreneurship and Technology, University of Waterloo and VP Research at The Evidence Network