I wrote this from the Georgia Tech Hotel and Conference Center, where I represented AUTM at the last of the Secretary of Commerce Locke’s four regional University Innovation Forums. I’ll report on this meeting next week. Earlier this week I attended the Central Region Meeting in Memphis. I was given a T-shirt emblazoned with the AUTM logo and “Central Region Rocks!!!!!!!!” and I really can’t argue with that sentiment. A great meeting and many kudos and thanks to Sue Patow and her committee. By the way, Beale Street rocks too.
So, my subject today is two recent papers in Research Policy, possibly the most respected academic journal in our space, which claim that a lot of academic technology is “going out the back door.” The papers use totally different data sources and follow totally different methodologies in coming to their conclusions.
First, I want to congratulate both teams of researchers for their important contributions to our knowledge of technology commercialization activity. These were not simple studies and required a massive effort to create their primary datasets
The first, tiled “Does policy influence the commercialization route? Evidence from National Institutes of Health funded scientists” by Taylor Aldridge, currently a postdoc at the Max Planck Institute, and David B. Audretsch, a professor at Indiana University. They looked at the top 20% of NCI grant awardees between 1998 and 2002. This gave them 1,693 scientists who received an aggregate of $5.4 billion. They then searched the USPTO database to identify which of the scientists had received one or more patents. They found that 392, or 23%, of the 1,693 scientists had received a total of 1,204 patents, an average of 3.1 patents each.
They then looked at the assignment of the patents. If a scientist had assigned at least one patent to an entity other than his institution, then they assigned all of that scientist’s work to “the backdoor.” On this basis, they found that 29.8% of scientists used the back door.
This seems to me to potentially overstate the extent of the phenomenon. It could be that only 117 patents – 1 per “back door” scientist and only 10% of the total, not 30% – were assigned to another entity. They don’t report the total number of patents assigned elsewhere.
Aldridge and Audretsch also didn’t investigate the reasons why scientists didn’t assign to their institution – e.g., whether the institution had returned the patent to the scientists or whether the scientist had made their contribution under a consulting agreement on their “one day per week.” Interestingly, they observe that:
if the scientist licenses their intellectual property they are more likely to commercialize through the technology transfer office and less likely to resort to the backdoor path for commercialization. This also suggests that those scientists not licensing their intellectual property have a greater propensity to resort to the backdoor path of commercialization.
Go TTOs!
The second study was titled “Inside or outside the IP system? Business creation in academia” and was carried out by Riccardo Fini, Nicola Laceter and Scott Shane at Case Western Reserve University and the University of Bozen, Italy.
In 2007, they sent a questionnaire to 58,646 academics who are affiliated with Carnegie I and II United States universities and employed in National Research Council-tracked disciplines – Engineering, Biological and Medical Sciences, Social and Human Science, and Mathematics, Physics and Statistics. They received 11,572 replies, a response rate of about 20%.
The questionnaire was in four parts:
- Demographics
- Academic data – publications, grant funding etc.
- Commercialization activities
- Additional questions on commercialization, only asked of professors who had commercialized.
The current study focuses on questions about whether the professor had founded one or more companies based on patents or not based on patents.
They found that only 36% of professors had ever undertaken any commercialization activities, consistent with other studies such as those by Marie and Jerry Thursby and Aldridge and Audretsch discussed above. They found that licensing accounted for almost 60% of commercialization activity with the balance being starting a company. Professors in engineering predominated among the four NRC groups in starting companies.
Their most significant finding is that twice as many professors have started companies not based on patents as have started them based on patents. There is a significant difference between the discipline distribution of the two groups – professors in Social and Human Sciences were 10 times more likely to have started a company not based on a patent as to have started one based on a patent, with corresponding reductions in the representation of the other three groups of disciplines in the non-patent group. Interestingly, while professors starting patent-based companies tend to have raised twice as much grant funding as professors starting non-patent based companies, the average grant funding raised by both groups of professors starting companies was more than double the average grant funding raised by all the respondents to the survey. This is consistent with other studies (e.g., Lowe, Cohen) showing that professors who engage in commercialization and entrepreneurial activities are generally the academic stars and superstars.
The authors make a leap to conclude that professors starting a company not based on a patent are all operating outside the TTO system, while those starting a company based on a patent are all operating within the system. This conclusion seems to reflect the same fixation on patents that critics of TTOs accuse TTOs of. TTOs in fact deal with other forms of IP – copyrights, biological materials, software, etc., and many companies are founded based on licenses to non-patent IP. However, this factor is unlikely to account for the extent of non-patent based entrepreneurial activity.
Rather, many of us are aware of professors who start companies based on their expertise and know how, utilizing their “one day per week” of consulting time. This is perfectly within their right, and the university will have no involvement with the company unless the relationship results in a financial conflict of interest with their academic research. Indeed at BU, I frequently learn of these “know how” spin-outs through COI committee meetings, often when the company applies for an SBIR or STTR grant and wants to subcontract part of the work to the professor, or wants to fund research in the professor’s lab or department directly.
I don’t believe Shane et al.s findings reflect negatively on TTO’s. Indeed I think we should welcome their findings as providing the first quantification of the extent of “know how”-based academic spin-outs of which I am aware. The findings show that there are multiple ways that technology diffuses out of universities, defusing accusations of “monolithic, obstructionistic TTO’s”. They demonstrate another important pathway of economic impact from commercialization of academic IP on the local community. These companies are undoubtedly largely dependent on the professor’s expertise for their viability and I expect they will be located close to the professor’s institution even more frequently than are patent-based companies – the Annual Licensing Activity Survey generally shows that 75% of spin-outs are located in the same state as the institution that is the source of the IP.
What should be the institution’s role in supporting such faculty? My guess is that most faculty probably won’t want to involve their institution if they don’t have to – “let sleeping dogs lie” as it were – so they probably won’t be looking for much help from the institution. The institution probably isn’t likely to support these ventures financially if it doesn’t have an ownership stake in them, but my guess is that TTOs and others will be willing to help them with contacts and networking, recommendations of professional advisors and potential sources of funding. Entrepreneurship centers and business schools will surely welcome the opportunity to help them develop their business plans. They’d be candidates to be a tenant in a university run incubator.
Finally, I do feel that one of the Shane et al. conclusions misrepresents their data, and unfortunately, it is a conclusion that I think will get traction and may be cited against technology transfer offices and entrepreneurial activity. The conclusion concerns professors’ time distribution. Shane et al. report that professors starting companies based on patents spend less time on teaching and research than their counterparts who start companies not based on patents. This is headline grabbing stuff, hence my concern. But when you look at the data in detail, it certainly shows that the patent-based professors spend 10% less of their time on teaching than the non-patent based, but it also shows that they spend 8.5% more time supervising Ph.D. students. Supervising Ph.D. students isn’t part of teaching? And while the data shows that they spend 6.5% less time on research than the non-patent based professors, it also shows that they spend 6% more time writing grants. Grant writing isn’t part of research? I think these time distribution results merely confirm Shane’s observation elsewhere in the paper that the patent-based professors are commercializing the results of laboratory-based research, and these professors spend their time differently than professors in the social sciences and humanities.
These are my thoughts -- your comments? Please start a thread below – it’s easy.