Here is a short description of the Bayh-Dole Act at a US university tech transfer office web site. There are many things wrong with the four paragraphs here. Consider:
In 1980, the Bayh-Dole Act (PL 96-517, Patent and Trademark Act Amendments of 1980) created a uniform patent policy among the many federal agencies funding research.
Generally correct. A good start. This would be a good place to end, too. But no, we go on.
As a result of this law, universities retain ownership to inventions made under federally funded research.
Only in a strange reading of “as a result of this law” to mean “taking advantage of ignorance about the law” can this statement be somewhat true. Certainly “retain ownership” here is meant to mean “take ownership” and not “may elect to retain ownership if the university has obtained that ownership through assignment from the inventors.” The language in the statement is the usual conflation of “elect to retain title” with “elect title” meaning “take title.” After Stanford v Roche, it is clearly untrue. That it persists in university web publications, where it could be readily revised, is something akin to contempt of court.
In return, universities are expected to file for patent protection and to ensure commercialization upon licensing.
There is no such bargain in Bayh-Dole. Bayh-Dole does not hand invention rights from inventors to institutions. The bargain is *if* a contractor obtains ownership of a federally supported invention, it may keep that ownership subject to a number of restrictions so that agencies do not have to waste time fussing about it on a case-by-case basis. The law certainly does not stipulate that universities must “ensure commercialization upon licensing.” The law defines instead “practical application” and everywhere one might find a reference to sale one also finds reference to use. The fundamental condition is that the benefits of a subject invention are available to the public at a reasonable price. Making benefits available does not require “commercialization” at all. Commercialization requires the selling of new products. That is a form of practical application, but it is certainly not the one mandated by Bayh-Dole.
The royalties from such ventures are shared with the inventors; a portion is provided to the University and department/college; and the remainder is used to support the technology transfer process.
This is a funny reversal of the requirements in the standard patent rights clause authorized by Bayh-Dole. See 37 CFR 401.14(a)(k)(3). There, expenses of licensing are recovered first. Payments to inventors are included as costs. Thus, the entire income stream after other costs could go to the inventors! Second, under the standard patent rights clause, any remaining funds are to go to “scientific research or education.” Here, however, the list is reversed. First the inventors, then some bit for research and education (conveniently dropping the “scientific”), and the remainder is to “support the technology transfer process.” It is worth looking at the language in (k)(3):
(3) The balance of any royalties or income earned by the contractor with respect to subject inventions, after payment of expenses (including payments to inventors) incidential [sic] to the administration of subject inventions, will be utilized for the support of scientific research or education;
The allowable payments for expenses are for “administration of subject inventions” not “support the technology transfer process.” What’s the big deal? The patent rights clause makes it clear that royalties from federally supported inventions may support only federally supported invention expenses, not “technology transfer” operations generally. In (k)(3), everything is in the plural, so it sure looks like a contractor is authorized to spend income from one subject invention on the expenses of another subject invention. But that’s the limit. After expenses, everything is to go to scientific research and education, not simply as slush money dropped on departments. I wonder what a federal audit of royalties on subject inventions would turn up. Probably gross indifference to tracking the use of funds, on the claim that anything the money is spent on at a university is “education”–which I would argue is not true in the slightest.
That gets us through the first paragraph. Now for some revisionist history:
From a historical perspective, there was a need for reliable technology transfer mechanisms and for a uniform set of federal rules to make the process work.
The problem that was raised at the time was how agencies approached inventions made under research contracts. The issue was not so much “reliability” as it was consistency, and the problem of consistency was a matter of administrative convenience for invention management agents such as WARF and Purdue Research Foundation, not the agencies. A succession of administrations from Kennedy on had endorsed the need for flexibility in agency management of inventions, but some agencies held onto legal opinions that they had to take ownership of all inventions made with government support. Notably, the Department of Defense and the National Science Foundation did not. Health, Education, and Welfare did, and it was promoted as a great achievement that HEW agreed to “Institutional Patent Agreements” as a way of allowing university invention management agents to hold onto inventions made with federal support. The process that was a problem, therefore, was how some agencies released ownership claims on inventions as contract deliverables.
This issue is not one of “technology transfer” as it involves only rights going against the direction of transfer–back to the sponsor rather than out to the field. That is, the Bayh-Dole Act has nothing to say about making any “process” of “technology transfer” “work.” It deals with how federal agencies may claim inventions as deliverables under contract. It is not clear, even now, that a “uniform” approach across all agencies, for all contracts, makes much sense, especially for the transfer of technology. It does make sense for resetting competing administrative bureaucracies at the agencies, and it does make sense for the administrative convenience of invention management agents wanting to get their hands on more and more inventions. But not for actual movement of new technology from discovery to practical application.
It was tough for the federal government to transfer technologies for which it had assumed ownership. In 1980, the federal government had approximately 30,000 patents of which only 5% led to new or improved products.
This repeats and inflates the rhetorical arguments that were floated when Bayh-Dole was being debated. There, it was 28,000 patents and less than 5% and the figures were claimed as of 1976, not 1980. But as Rebecca Eisenberg and David Mowery have pointed out, many of the patents that the government held were in weapons systems, and were held by the Department of Defense, which had a policy that allowed contractors to own inventions made with federal support. It’s just that the contractors did not want to own! And it might make good policy sense for the government not to leave such inventions with contractors or to encourage “commercialization” of such inventions. The 5% had to do not with “new or improved products” but to “commercially licensed”–which means, I would think, licensed to commercial firms. Neither licensing to commercial firms or the production of “new or improved products” has anything to do with the actual use of the inventions. It is just rhetoric. I have yet to find anything that documents how the 28,000 figure was arrived at or how the 5% figure was determined. It may just have been estimates. It may have been simply a plausible rhetorical gesture.
All that aside, estimates of the commercial activity around any patent are in line with a 5% figure. Other than the rhetorical context, claiming 5% of patents are seeing commercial activity would appear to be nominal, if not quite good. The advocates for Bayh-Dole played apples and oranges, though. They compared the accumulation of federal inventions and a sub 5% licensing rate with the handful–a few hundreds–that university affiliated patent agents had begged back from the federal agencies–where the licensing activity (again, not reporting any new products) was on the order of “25% to 30%.” Well, of course–if one is highly selective, aiming to manage only what one thinks one can license for commercial use, then maybe one will get higher licensing rates. Even then, does the licensing result in practical application, so that benefits are available to the public at a reasonable price?
If one compares 25% with 5% it sure does look bigger. But if one considers that the university agents, even when they were sure they had something worth licensing, were only getting it done 1/4 of the time, one has to wonder why the university rate wasn’t 50% or better. If an invention is all that good, why isn’t it possible to put it into use? That is, it would appear that the university agents weren’t doing all that great a job with federally funded inventions. And we might add, as Rebecca Eisenberg points out, WARF had got itself sued multiple times by the government for antitrust behaviors involving the patents it had under management. Folks had reasons to worry about private management of research inventions where making money from exclusive licenses could become more important than ensuring access to a lot of stuff, for a public benefit. The argument for exclusive licenses turns out to be, essentially, that such licenses limit the risk exposure of speculators, for which they are willing to pay to gain an advantage over industry practitioners.
Many patents were not being used as the government did not have the resources to develop and market the inventions.
This is too funny! The government did not have resources to develop inventions? What was that industrial-military complex all about, building bombs and bombers and radar and computers and moon rockets and the internet? No, the government had plenty of resources to develop inventions, and had no need to “market” them. This bit of history is pure bombast–but funny bombast! So of course, the patents were not being used–if by that meaning to keep others from exploiting the inventions. But the inventions that mattered were being used, apparently, patented or not. If the point intended is that there were not civilian uses, or uses in retail products, or uses that allowed for a speculative market in patents, then fine–but that’s a different thing, and perhaps quite removed from either the government running low on money or a university feeling the need to get in bed with speculators as a matter of public good.
Thus, Bayh-Dole gave universities control of their inventions.
Completing the revisionist history, we get the conclusion, which restates the false premise that universities under Bayh-Dole get control (or ownership, or the right of first refusal, or a mandate to own, or any other goofy claim) of “their” inventions. “Because the government was claimed to be doing a rotten job of attracting speculators to make money from faculty inventions, Bayh-Dole stripped faculty of their inventions so that universities could go over to the dark side and try their hand at attracting speculators and making money.” A lot rides on that “their”. Since when did inventions made by faculty collaborating with the government become the “university’s” inventions? That is, “their” should refer to *the inventors*, but here, clearly, it doesn’t. Bayh-Dole restricted agency claims to inventions as contract deliverables, and thereby allowed inventors to make arrangements with approved agents for the development of inventions made with federal funds–for the development of their inventions–here properly referring to the inventors. One such approved agent was of course the university that hosted the research. But other agents are also approved–including any organization with a primary function of managing inventions.
The reason that the Bayh-Dole act is so instrumental to university technology transfer is that it speeds up the commercialization process of federally funded university research and helps new industries to develop quicker.
This is just bizarre. There is little to indicate that Bayh-Dole has sped up anything. Every unlicensed patent is a barrier. Industries develop because there are economies of scale with shared platforms. See Sylvia Kraemer’s account of the problems of patent fragmentation in the early airplane industry as an example of the problem. There is a good argument that university practice under Bayh-Dole destroyed the emerging nanotech industry, and has done a lot to delay the emergence of 3d printing. Imagine if the internet and web had been subjected to university patenting. Nothing in the statement of Bayh-Dole objectives has to do with speeding up “commercialization” or helping new industries develop “quicker” (quicker what? or, er, even “more quickly”).
Examples range from Stanford’s Cohen-Boyer patent on the basic gene splicing tools – to the Axel patents, from Columbia University which provided a completely new process for inserting genes into mammalian cells to make protein.
Cohen-Boyer and Axel neither were Bayh-Dole subject inventions. As “examples” they may show how a licensing program generates revenue, but neither is even a good example of what the university-affiliated patent agents argued. They argued for *exclusive* licensing, which the government was loath to do. Cohen-Boyer and Axel were both licensed non-exclusively, and in the end, Stanford handled Cohen-Boyer’s expiration gracefully, while Columbia did not. In neither case, furthermore, is it apparent that the patents were necessary for development of the technology. In both cases, the inventions were research tools that could facilitate discovery and development of products. The benefit of the patents, then, arguably was to keep ownership of such a tool from the exclusive control of any single company. That was something that an industry found worth paying for, for the right tools, in the right circumstances. But that is not the “commercialization” process–but rather a process of “common platform deployment”–namely what the federal government was doing anyway. The main difference was that the universities were charging for access. Somehow that does not equate immediately with speeding things up.
Bayh-Dole has also enabled laboratory advances to become a significant factor in U.S. and Canadian industrial growth.
Huh? What does Canada have to do with it? What are “laboratory advances”? Handwaving.
The Bayh-Dole act is also vital to the university as a whole. University gross licensing revenues exceeded $200M in 1991 and by 1992 that number had risen to $250M. In FY 2000, U.S. and Canadian institution and universities Gross Licensing Income is reported in the AUTM survey at $1.26 Billion.
Notice that only gross revenues are reported, and not specific for federally supported inventions, and from a long time ago. What is not reported is the net revenues over time as a function of total costs. The amount that benefits the university “as a whole” is a tiny share of the total gross revenues, after deduction of expenses for the tech transfer programs. For most universities, the amount made from patent licenses tiny compared with the income from donations and research contracts, and even less when considered as a fraction of the university budget. Even if one takes the $1.26b figure from 2000 as accurate in some way, that works out to, on average, about $4m/university. Pull off the average expenses and one is left with maybe $2m/university. That’s not a lot for all the delay, expense, pain, oppression, bureaucracy, and bitterness heaped on everyone by the compulsory, monopolistic, institutionally self-centered version of Bayh-Dole, or, er, the faux Bayh-Dole.