1	Investigators Research choices IP Acquisition Use of IP
2	Motivations for Patenting
3	Academics’ Commercial Activities Substantial commercial activity Industry funding: 19% have some industry funding Patenting (in last 2 years): 22% Business activity (e.g., startup, negotiations, licensing, commercialization of discovery) : 35% More for those doing “drug discovery”
4	Reasons for Choosing Projects, Academic Respondents
5	Reasons for Not Pursuing Projects
6	Patents and Project Choice Project choice driven primarily by scientific objectives, interest and access to funding Scientific competition can also redirect projects Prospect of patent on research output or commercial potential have little impact (though higher (~20%) for drug discovery) Redirection does not seem to be a problem for basic research Potential SDR bias
7	Extent of Involvement in Patenting by AAMC Medical School Faculty, By Department
8	Patenting as a motivation for research? Azoulay and Sampat (2005): Key factor affecting the probability that an AAMC faculty member patents in a given year is volume of lagged NIH grants NIH grants can be interpreted as an “input” into research and/or as a measure of scientific prominence Azoulay, Ding, and Stuart (2005): Patenting positively related to publication activity by academic scientists Publications by patenters similar in “quality” to those by others But, publications by patenters tend to be in journals with a higher share of industry-authored publications Generally, these (and other) results suggest that traditional motivations for research prevail Patents are by-products of “normal” research activities The key unanswered issue is whether/how patents affect dissemination
9	Means of Disseminating Patented IP – All (n=117)
10	Means of Disseminating Patented IP – Academia
11	Means of Disseminating Patented IP – Industry
12	Investigators Access to IP Citations to Patented Inventions
13	Cause of Difficulties in Obtaining IP
14	Effects of Difficulties in Obtaining IP
15	Length of Time to Negotiate IP
16	Roads not Taken Main reasons: Over 40% of respondents rate funding, time constraints, infeasibility, and scientific importance as more than moderately important Patent-related reasons Too many patents on tools and other inputs: 3% No different for those involved in drug discovery Terms associated with rights to inputs, including materials: 10% 21% for those involved in drug discovery “Patent thicket” on inputs has little impact on decisions to choose, or not pursue projects anti-commons does not seem to be a problem
17	“Pure IP” and Academic Research Does the existence of disembodied “pure IP” impede academic research? Awareness of patents on knowledge inputs? Effects? Delay Modify research approach Abandon project
18	Awareness of Patents on Research Inputs 8%, or 32 of 381 respondents, believed they needed knowledge or information covered by patents Given burst in research tool patents, why so few? Only 5% check regularly for patents on knowledge or material inputs (little change since Madey) 22% received instruction from institution (v. 15% 5 years ago) AAAS study: 14% of universities give instructions BUT, instruction does not change behavior (6% v 4%)
19	Impact of “Pure IP”
20	Case Studies: At Risk Fields Prior results provide base-rate data But, even rare result might have major social welfare impacts if it affects important technology To probe this, we collected data from researchers in three fields that have high scientific importance, and varying levels of patenting and commercial activity EGF, NF-kB, CTLA-4 Lots of research activity (foundational paper had over 1500 cites for first two, around 900 for CTLA-4) Many patents (760, 90, 60, respectively) Drugs in market or clinical trials
21	Case Studies: At Risk Fields Pure IP: Adverse affects rare, although slightly more common than base rate More likely to know about patents 3% had abandoned a project (v. 0% for random sample) Access to materials even more problematic 26-32% did not receive last request (v. 19% for overall) NF-kB and EGF well above norm in terms of projects abandoned or delayed due to not receiving requested inputs (CTLA-4 near norm) Thus, even in high risk areas, the impact of pure IP is small, while the impact of withholding tangible property is even greater than the base rate
22	Summary Commercial activity by academics is substantial, but little growth Little evidence project choice affected by commercial incentives or anti-commons But beware of SDR bias
23	Summary: Impact of “Pure IP” Few are aware Even if aware, has little impact on research Even in high risk fields (EGF, NF-kB, CTLA-4) minimal impact (3% abandoned a project in last 2 yrs) Few academic institutions have policy of notifying faculty And, even if notified, does not seem to change behavior Earlier qualitative study: similar result “Law on the books” is not the same as “law in action”
24	Do Academic Genomic Patents Curtail Subsequent Research? Key findings: For genomic “techniques,” a patent grant has no statistically significant effect on citations to corresponding articles For genomic “sequences,” a patent grant is associated with a 12 percent decline in citations to corresponding articles, all else equal (p<.05) Murray and Stern (2005), using a different empirical approach and focusing on papers from Nature Biotechnology find that patents cause a 9 to 17 percent decline in citations to corresponding articles
25	Do Academic Genomic Patents Curtail Subsequent Research? Caveats: Even if patents do “hinder” research, welfare implications unclear This empirical exercise relies on strong identification assumptions which are difficult to test directly
26	Do Academic Genomic Patents Curtail Subsequent Research? Reconciling results with Walsh et al. Perhaps citations are affected more than subsequent research? But this still requires “awareness” of patents Perhaps results are driven by non-academic citers? Currently testing for this
27	Institutions Licensing IP
28	Licensing IP
29	Licensing IP II
30	Licensing IP III
31	Licensing IP IV
32	Licensing IP V
33	Exclusivity by Company Size
34	Exclusivity and Outside Interest
35	Nonfinancial Diligence
36	Nonfinancial Diligence II
37	Diligence
38	Diligence II
39	Diligence/Termination
40	Diligence/Termination II
41	Market Activity
42	Summary
43	A few Observations Based on the Policy and Interview Portion of Survey The NIH guidelines about the sharing of research tools are widely known and are often adopted by technology transfer offices. There are even more usages of the term of “exclusive” than known initially, including, “Exclusive for use with company’s own patented technology”, is still called “exclusive”. Those schools interviewed reported always ensuring a research exemption for themselves, and starting about recently-about 5 years ago, putting in a general research exemption. Sample: Nothing in this Agreement will limit the right of University to publish any and all technical data resulting from the research performed by the University relating to the Invention and to make use or practice the Invention, Licensed Product, Licensed Service, Licensed Method an associated technology and allow other educational and non-profit institutions to do so for educational and research purposes.
44	A few Observations and Anecdotes Based on the Policy and Interview Portion of MIT Survey Those schools interviewed note that smaller companies want broader and more fields in their “exclusive” licenses than do larger companies. Discernable consensus on need to patent and license with some degree of exclusivity therapeutics. Less on consensus on diagnostics and targets. Interesting definition of Research Tool: “ If you know it’s specific utility, it’s not a research tool”.
45	MIT Exclusive and Nonexclusive Patent Licenses 1980-6/30/98
46	MIT Exclusive and Nonexclusive Patent Licenses 1980-6/30/98 II
47	Count of Programs by Income and Program Age
48	More information on MIT Survey
49	Investigators and Institutions Material Transfers
50	Sharing Material Research Inputs Where others’ tangible inputs necessary for research activity itself, may have different impact from pure IP Examples Cloned gene, organism, cell line, protein, drug, unpublished information, etc. About 75% of our academic respondents requested materials in the prior two years Average # of requests (last 2 years) 7 to other academics and 2 to industry
51	Difficulties in Accessing Tangible Research Inputs 19% did not receive last requested research input Change over time? For academic to academic exchanges in genomics, percent of requests not received: 2003-04 (Walsh, et al): 18% (+/-3.7%) 1997-99 (Campbell, et al): 10% So, appears to be some increase in recent years Delay research (>1 month): 8% of requests (v. 1% for pure IP)
52	Impact of Not Receiving Research Inputs (Academic to Academic)
53	MTA Terms, Negotiations About 40% of transfers require MTA More common if request drugs (64%) Fees 93% from academic, no charge, < 2% over $1000 85% from industry, no charge, 7% over $1000 Terms (requested) Reach through-38% Royalties-17% Manuscript review-30% Drugs to Academics: 70% of final agreements
54	MTA Terms, Negotiations Except for royalties, academic respondents doing drug discovery tend to be more subject to restrictive terms than those doing basic research Industry suppliers tend to impose more restrictive conditions than academic suppliers 26% of MTAs (11% of requests) take more than one month to negotiate
55	Why Do Scientists not Provide Materials? (multivariate regression) Concern about SDR bias in self-report data: use multivariate regression From point of view of academic supplier: Commercial orientation (business activity) Industry funding Scientific competition (# competing labs) Burden (requests/lab dollar) Total budget Publications Drug discovery Male
56	Negative Binomial Regression for Number of Times Respondent Does Not Fulfill Research Input Requests
57	Why Do Scientists not Provide Materials? Main predictors Scientific competition (# competing labs) Prior business activity Burden (requests/lab dollar) # Publications (Eminence or opportunity cost?) Insignificant Industry funding (modest pos. effect) Drug discovery
58	Summary: Tangible Inputs Access to tangible research inputs more problematic than access to pure IP About 10-20% of Ac-Ac requests not fulfilled Refusals increasing Adverse affects due to scientific competition (and cost/effort of compliance), as well as commercial incentives But, social welfare impacts of denials and MTA terms ambiguous
59	Summary So, if there is a problem, it’s one of access to tangible—not intellectual—property, and the constraints on access turn more on cost/effort, scientific competition and commercial activity than on IP per se Culprit may not be IP on materials, but Bayh-Dole and IP-related legislation that fosters commercial activity among academics more generally But need to weight benefits of such legislation against any costs Solutions should be tied to problems Changing patentability rules may not address problem Facilitating tangible input sharing (to reduce cost/effort) may be key