A few years ago, the University of Washington claimed to have started a bunch of companies. I went through their lists of startups for two years and found they were making it all up. They claimed to have started 35 companies, but at best it looked like they had been involved in about 13, and that was being generous. They were counting pretty much anything they had done business with, and were indifferent to when the company had started and why. So they counted non-startups as startups. They counted companies that had started for other reasons as startups because UW provided a license at some point. They counted companies as startups that hadn’t been named, that were nonprofits, that were little more than web sites covering for typical university research or distribution activities. Virtually no operations, no products, little investment, no economic development–all a matter of bullshit metrics, as I called it in an article that responded to the university president’s swipe at my efforts to get the university to come clean.
Eventually, just as the money from the big hit Hall patents ran out, the university president, who had made this pattern of behavior his signature program, left, and the vice provost in charge of all the hype and apparently (according to the hype) at the peak of success, suddenly was out of a job. They had blown through over $100m in Hall money and state money. They had burned through their reserve funds, too. If it was a green thing, they had eaten it, and when things began to look brown, they left. And nobody caught them.
As I was restoring links in my articles after the port to researchenterprise.org, I had the opportunity to chase down more information involving Ennaid Therapeutics, one of the startups that UW claimed for FY14. I added the discussion there, but thought it was worth putting the discussion here as well, where it was more visible. The technology that apparently went to Ennaid illustrates the fragmentation of ownership in research technology that has resulted from the Bayh-Dole Act’s interventions as well as how when multiple institutions each claim a share of tech transfer activity, AUTM metrics easily and conveniently get bloatified. The result is Fragmented Ownership Institutionally Licensed technology–FOIL technology.
Ennaid reported that it had acquired technology from Tulane, Rockefeller, and Florida Gulf Coast University. The University of Washington also claims to have provided technology–even started Ennaid.
How do all these universities get involved? Let’s start with a publication in PLoS from 2012, co-authored by sixteen scientists from Tulane, Florida Gulf Coast, National University of Singapore, Duke-NUS, and Purdue, with one author affiliated with the University of Texas Medical Branch at Galveston. The authors indicate that Tulane has filed patent applications on the invention. A USPTO patent search brings up patent 7,416,7333 “Flavavirus fusion inhibitors” with a co-inventor (Garry) who is also a co-author on the PLoS paper. The patent is assigned to both Tulane (technically, to “The Administrators of the Tulane Educational Fund”) and Rockefeller universities. The Rockefeller assignment appears to come from the affiliation of another co-inventor (McKeating) with that institution. Garry and McKeating also collaborated on a separate invention involving a treatment for diseases caused by hepatitis C. The patent application was filed in 2003 and issued in 2008. A divisional from this application was filed in 2008 and issued in 2010. A further divisional was filed in 2010 and issued in 2012–three patents from one invention report, each patent claimable by two universities. AUTM would report this as six patents! And that’s just the US activity. Holy statistical bloat, Batman!
Meanwhile, Florida Gulf Coast was also filing patent applications. One is patent 8,541,377, filed in 2008 and issued in 2013, with six co-inventors, including Garry (Tulane) and Samudrala (Washington) and Jenwitheesuk (Thailand), but assigned (according to the patent, only to FGCU). A second patent application was filed on the same day as the first and issued in 2014 as patent number 8,637,472, with five co-inventors, and assigned jointly to Florida Gulf Coast and to the University of Washington. Three of the co-inventors are FGCU (Costin is also a co-author on the 2012 PLoS paper), one is from Thailand, and one (Samudrala) from the University of Washington. Samudrala, according to his bio at Wikipedia (always a sketchy romp), was recruited to UW via a state economic development funding program that was to function “as a bridge between cutting-edge research and education, and new economic activity.” He left UW in 2014 for SUNY Buffalo. Costin and Samudrala were co-authors (with nine others, some of whom also end up as co-inventors on FGCU patent applications) on a PLoS paper from 2010 on the design of peptides to inhibit dengue virus activity. One of the co-authors listed on the Tulane and FGCU patents (Jenwitheesuk, listed in the patent as a resident of Thailand) shows up on the 2010 PLoS paper giving a past affiliation with University of Washington (according to faculty listings, he was a “senior fellow” in the Department of Microbiology in 2007) and his present (as of the publication date) affiliation is with the National Science and Technology Development Agency of Thailand.
So here are two patents with applications filed on the same date, with six inventors (five in common), suggesting a common inventive step split up as a patenting strategy into two filings, with ownership of one of the patents split between UW and FGCU and the other solely FGCU. The odd thing is that the sole FGCU patent lists Garry, from Tulane, as a co-inventor–why would not Tulane also have an interest? We have then twenty authors listing affiliations with eleven institutions resulting in patents listing nine co-inventors (three of whom are not in either of the co-author lists for the papers above), with four institutions claiming a share of five US patents–Tulane, FGCU, Rockefeller, and Washington–and other organizations–Purdue, Texas, CDC, FortéBio, Duke, NUS, and the Thai NSTDA apparently have no interest in the patents. Of course, inventorship is decided by inspection of the claims drafted for an invention, and the owners of a right to patent can determine what goes into any given set of claims. Thus, even if these other co-authors had also invented something related, the universities controlling the patent applications can choose to omit those related somethings, and thus leave those otherwise co-inventors (and their institutions) out of the patenting flurry.
In all, if four institutions are claiming patents on the work of twenty scientists spread across eleven affiliations, there is plenty of room for any of these to claim that their people were involved in good science resulting in patentable inventions. The core of the teams recognized as inventors were at Tulane (Garry, especially) and FGCU, with the UW contribution via Samudrala involving aspects of computational techniques involved in the development of peptides to work against dengue fever. If each of these four institutions reports to AUTM that they have patents, then one can see how quickly AUTM patent counts can exceed reality. But the question at issue here is whether UW’s contribution to the science (and presence as co-assignee on one of these patents) is sufficient to support the claim that UW started Ennaid Therapeutics, and did so in FY14. It’s hardly a credible claim. UW contributed science to an area of technology, given it a claim to ownership on one part of the overall technology (its part–using protein prediction software, apparently). Because the rights to the overall technology are fragmented across the four institutions, Ennaid, to gain an exclusive position, has to get a license from each, for each patent in which an institution has even a co-ownership position. Thus, there could be as many as nine licenses if each institution licenses its interest in each patent as a separate transaction.
One technology development results in 5 patents and up to 9 licenses, with up to 4 startups reported. This is the administrative muck any company has to wade through to try to get a clear license to a university technology. This case is not so very atypical. Most inventions at universities are co-invented, and author lists in the sciences often do run off the fingers of both hands, and collaborations are rampant (and apparently encouraged).
The University of Washington reports its grant income–over a billion dollars per year–but generally does not report that its actual expenditures are significantly lower (but still wowable) because UW serves as a general contractor for a bunch of institutions. UW has the reputation to win government grants, so everyone rallies behind UW to win the grant, and then UW spread the money out via subcontracts by which the money flows, often, out of the state of Washington and to other schools, which report the subcontract amounts as their grant income, after Washington has counted it as its grant income. Expenditures are a much more meaningful–and less deceptive–measure of economic activity. Any model of regional economic impact that takes grant income rather than expenditures in the region as an input, with a multiplier used to estimate impact, is fundamentally flawed, perhaps even willfully so. Deduct the subcontracts. Deduct the major instrument purchases from out-of-state manufacturers, deduct the money spent in field work and talks out of state, and the level of economic activity triggered by research funds goes down. So, not as dramatic, where “dramatic” is a nice cover word for “deceptive.”
As long as the federal government pushes for such research collaborations, and spreads funding in a given area of technology around, inventions will often have fragmented rights. That means that invention metrics will be inflated as institutions count the same events multiple times, each claiming their role, and worse anyone hoping to use such fragmented-ownership institutionally licensed (FOIL) technology, will have to go through multiple license negotiations. On the back side, the institutions involved will often burn even more energy trying to work out patent administration agreements–which university will file the applications, who will control patent prosecution, who pays for the patent work, who leads “marketing” efforts, who negotiates any licenses and who has a right of review, not to mention controlling law (ever seen a spat between lawyers at state universities in different states refusing to back down on which state’s laws must control–because no state will be subject to the laws of another state–gollwhuz those discussions hurt).
The creation of FOIL technology is another result of university technology bulimia–like a dog that cannot control its appetite, when university administrators see invention rights that they can just *take* (because faculty often do not have the money or access to smarts or the energy to resist), the administrators cannot help themselves. They take everything. They write policies that make it appear that it is their right to take everything. That taking is a natural consequence of employment, especially when employment includes non-employment. Taking, for the bulimic administrator, is natural. So, apparently, is burping the technologies back out, but mostly when they have been ruined by sitting in the dens of university licensing offices.
University licensing is not a matter of invention -> patent -> license -> public benefit. That story can be imposed on most anything. The effect of university dealing is primarily one of collaborative technology -> fragmentation -> exclusion -> delay. That is, the game sparked by Bayh-Dole (though it’s not in Bayh-Dole–it’s just that if you walk around the playground giving out cigarettes to kids, they will light up) is one of trying to convert non-market/networked technology into market/individual technology. The result is FOIL technology–fragmented, hard to acquire for use, best used for speculation in the form of startups, with speculators aiming to get rich by attracting in even more speculators until established companies buy them all out, or when the startups fail, for patent trolling. If the patent can’t encourage development and use (because development and use ignored the patent altogether), then at least the patent can get used to shake down industry for cash. Note that Bayh-Dole has nothing to say about the naked use of patent rights–the objectives of the Act focus on the practical application of subject inventions. That’s the only use of the patent system authorized by Bayh-Dole. If the patent system is not necessary for the practical application of subject inventions, then Bayh-Dole’s interest is exhausted.
Combine that with the requirement that universities act as trustees for the beneficiaries of federally funded projects with regard to any patents that the universities acquire with federal funds (including, then, as a compulsory condition of working on a federal grant), and there is a strong argument that if a university can’t license patent rights to a subject invention prior to someone practically applying that invention, then the effect of federal regulations is to prohibit trolling. Once there’s practical application without reliance on the patent system, then Bayh-Dole’s standard patent rights clauses void any further assertion of rights to block that use. An amendment to Bayh-Dole would make this implication clear and do a great deal to take universities and their research foundations out of the business of patent trolling.
In any case, it’s not credible for UW to claim that Ennaid is a UW startup. At best, Ennaid got a license from UW to sew up an exclusive position on one of a number of patents it presumably had to license to support its venture financing. At the time of the UW license, it appears that neither of the co-inventors of its bit of the stuff being licensed was even still at UW. Jenwitheesuk left in 2007, and Samudrala left in 2014. By the time UW is claiming to have started Ennaid in FY14, its inventors to a bit of the inventions were already gone or nearly so.
The more egregious implication of C4C’s claim, however, is that the startups it is claiming will contribute to the Washington state economy. That’s what Samudrala was recruited to UW for by state legislative money in the first place. But the license to Ennaid does not appear to have anything whatsoever to do with biotech industry in the state, manufacturing in the state, further research in the state, new venture capital coming into the state, or jobs in the state. That’s all simply untrue. But C4C puts it out that Ennaid is such as startup. That’s the whole context for the claim that starting companies will result in a boost to the economy of the state. Instead, at best, UW gets royalties, if Ennaid ever develops a commercial product (or sells the rights off to another company and has to pay a share of the sale of rights–more speculation–to UW). But that income is not what is put out there in a report of a record number of startups. The public is to believe that research at UW is sparking an economic revolution. That’s what the C4C 2010 strategic plan argued. Further, that plan argued that UW would make its money from the cash out of equity in startups, generating so much money that UW could change its financial model. All of it was pure bunk and snake oil, but university administrators washed down their technology habit with large tankards of the stuff, gulping and gullible, if not complicit.
It’s a plain deception, in the context of state economic development through research technology, to claim Ennaid as a UW startup. Ennaid may be a Tulane and FGCU startup–that makes some sense. But UW’s role appears to be one of parasite–taking a financial piece of the action, contributing next to nothing administratively.