Given that university patent administrators don’t write about the reasons why a comprehensive, compulsory invention assignment policy is such a good thing, especially when coupled with a monopoly-seeking exclusive licensing program, I will identify some possible arguments in favor of such a happy combination and sketch the problems with them. Of course, I can be accused of constructing straw man arguments, not the actual arguments that have been used. But those actual arguments, apparently, are secret, if not fantasies, so until those get trotted out to be considered, these will have to do.
1) Compulsory assignment is required by federal law for inventions made with federal support.
This claim is made, for instance, in the patent policies at the University of Southern California and the University of Iowa. The US Supreme Court expressly rejected this claim in Stanford v Roche. There is no basis for the assertion.
2) Compulsory assignment is required to comply with federal law for inventions made with federal support.
This claim is made by patent attorneys and university administrators. The University of California made this claim in requiring all employees to sign a revision of the patent policy that included a present assignment. Both NASA and the NSF make this argument in their policy documents. There is nothing in Bayh-Dole or in the standard patent rights clause authorized by Bayh-Dole that requires university ownership of inventions made with federal support. There is a requirement, however, that universities require their research personnel to make a written agreement to protect the government’s interest, including assigning or licensing their inventions to the government–but universities do not comply with that requirement. Instead, they substitute a requirement that research personnel assign to the university. That is, the reason that the requirement to assign to the university appears necessary is that the university has refused to comply with the terms of its federal contract. “Because we choose not to comply, you must assign so that we can comply anyway.”
3) Universities make more money when they own the inventions made by their personnel.
I hear this argument in informal chats. That if university inventors are left on their own, or work with external invention management organizations, the university receives less in royalty income than when the university hires its own invention management staff. I don’t see any data that supports the argument. The argument does involve an impossible comparison–we can’t run experiments in which an inventor uses both university and external invention management agents at the same time. Given that external invention management agents typically take costs + 40% and a university policy typically takes costs plus 67%, there is a superficial truth to the argument. But the actual money received by a university is a matter of both the fee structure and the effectiveness at licensing. If an external agent is better at placing inventions, or places them in better situations, or is better at collecting royalties, then the total income to the university after the inventor’s share, even after the agent has taken 40% , may still be much greater than the university’s own efforts.
When Bayh-Dole was being debated, university folk claimed that they had a 30% licensing success rate, and the government’s rate was more like 5%. Most of that licensing success rate was the result of Research Corporation’s work, since most universities did not have their own licensing operations in 1980. Recently, the University of California reported that it had a success rate of 0.5% in licensing inventions that led to a commercial product on the market. Now, rate of licensing is not the money getting made, but without licensing, there isn’t much money. If the licensing rate for external agents was 30%, and the licensing rate for the federal government was 5%, and the current rate at a top tier university licensing program is now 0.5%, then the burden of the argument is to show that a university is a better owner of inventions than is either the federal government or an external invention management organization.
Finally, we can ask the simple question in response–why is the university’s making of money from patents so compelling? Is making money the main social purpose a university should be dedicated to when it takes ownership of inventions for the purpose of patenting them? Perhaps that argument can be made. I haven’t seen it. Few university policies go that far. Wake Forest does, for instance. But even there, there’s no explanation for how making the “maximal possible” money from the disposition of patents results in public benefit. Oh, yes, the “invisible hand.” Magic. There has to be something with more substance.
4. University ownership of inventions is standard practice (or best practice).
No doubt university ownership of inventions is now standard practice. That’s clear from a review of university patent policies. But that’s not the issue. The issue is whether university ownership is a good practice, a beneficial practice, a justifiable practice. “Everyone is doing it” is the lemming argument. It’s the argument that got made right before the mortgage meltdown. I once was in a discussion with another university licensing official regarding licensing practice. He argued that I might be right, but that he could not take the risk of doing something that ran counter to what everyone else was doing. If he did what everyone else was doing, he didn’t have to explain his decision, and if he did what was right, what was indicated, what was otherwise possible, he would be on his own. He couldn’t bear the idea that he might be accountable for a decision and would have to base the decision on reasoning from the circumstances and the available tools. Much better to follow the herd and deflect any criticism with “this is a best practice.”
The problem with “best practice” as an assertion is that it means, generally, that a practice is regarded as “best.” But to substantiate the claim, there’s got to be something substantive that shows how a given practice is “best.” What makes it “best”? That it makes more money? That it is compliant with law? That it respects the social norms and purposes of a university faculty? That it advances research enterprise? There’s a great deal packed into “best” that hasn’t come out in any discussion.
It might, for instance, be helpful to see the email threads between inventors and university administrators to evaluate the practice. One might want to see the status of each invention–has it been patented? has it been licensed? has the license resulted in commercial sales or use? are the results of those sales or use available to the public on reasonable terms? has research (at other institutions, in companies) been advanced or adversely affected by the license? We can’t answer these questions now because universities do not make public such metrics and Bayh-Dole exempts this information from federal public disclosure law when federal agencies request it.
5. University ownership reduces the liability from breach of research and licensing contracts.
Universities routinely offer private sponsors of research a non-exclusive license to inventions made in the sponsored project and a first right to negotiate an exclusive license (or even ownership of inventions). If an inventor were to own an invention personally, then the university couldn’t offer a license at all–the inventor would have to do it. At this point, university administrators drop their QEDs. But this can’t be the shape of an actual argument, since the liability arises from the university offer of rights in the first place. If the university anticipated that it would have no rights, then a sponsor would either require the university to get the rights or would do business directly with those conducting the research. Bayh-Dole’s standard patent rights clause doesn’t even bother deciding. It insists that universities authorize their research personnel to deal directly with the federal government with regard to their inventions made with federal support, and if they assign to their universities, then the government will deal with the universities. The same strategy works for private sponsors of research.
One might put it this way, “Our inept and willful licensing behavior creates such liability for the university that the only way to mitigate it is by demanding ownership of all inventions made by university personnel.” Preserving crappy licensing practice is not a good argument for a insisting on a non-voluntary approach.
There are additional sources of liability in licensing–dealing with background rights and improvements, tracking licensing commitments across multiple research projects, double licensing. But these liabilities are increased by university ownership, not decreased. If a university does not have a robust patent portfolio, then it does not have a vast array of claimed but mostly undocumented property to worry about. A comprehensive, compulsory approach to inventions–especially if the scope of invention includes a random list of anything that might have value that’s not “traditional” scholarship–creates a huge liability for university licensing. If a university licenses an “invention” or “technology” rather than a specific patent (or licenses both the invention and a patent), then the scope of the license reaches beyond the claims of the patent to anything that university policy might construe as owned by the university. Technology in other labs might be within the definition of the licensed invention and still be well outside the claims of the licensed patent. A reduced university ownership position greatly reduces the uncertainties about what is licensed, and what is owned, by the university.
The upshot is that a comprehensive, compulsory approach to university invention ownership creates great liability, introduces unnecessarily complexity and uncertainty, and operates, generally on a tenuous streak of luck and corporate good will. If universities are going to own patents, it were good that they own only a few, and only patents, not inventions of whatever nature, even if not patentable.
6. Multiple approaches to inventions are confusing to inventors.
I’ve heard this argument many times. Show me a confused inventor. We operated with multiple approaches at the University of Washington for over a decade and the only people who claimed to be confused (or that there was confusion, or there could be confusion) were the patent licensing folks. They never produced a confused inventor. Whatever uncertainty in having some inventions claimed and others not arises from poor policy drafting and fits thrown by patent administrators. For instance, Stanford for a time had a patent policy that inventors owned their inventions “whenever possible.” Only when Stanford had a legal obligation to own (by law, by contract) did the policy require assignment. That policy creates two approaches–one when Stanford has legal obligations and another when it doesn’t. If there’s confusion, it arises when Stanford contracting officials don’t make clear to all what obligations the university has.
But most university patent policies even now retain at least two approaches to inventions. The first is (typically) that the university claims to own everything. The second is that the university can either choose not to assert its claim or having asserted its claim can regret it and re-assign inventions to the inventors. Once there is a waiver or re-assignment, the inventors are back to a voluntary approach–but often after delays, expenses, fuss, threats, and bombast. But here’s the rub: most universities I know of refuse to release inventions or make the process of release so difficult, with so many conditions, that inventors mostly give up.
Other university patent policies allow the university to manage an invention directly or to assign the invention to an external agent–an internal and an external approach. In the past, a university might direct inventors to assign to the university or to an invention management agent, and might even allow the inventor to choose the agent. Such policies involved a compulsory use of an invention management agent, but did not require university ownership. Again, there appears to have been little confusion in multiple methods.
Of course, to be able to ascertain the degree of inventor confusion (and the source of that confusion–including bad drafting, ditty patent managers, misinformation) we would have to have research universities with multiple approaches in action. But we don’t. The reality is that university patent managers apparently cannot tolerate competition. They cannot imagine their operations would continue to exist if inventors had a choice. If that’s the case, then it argues that university patent management operations are so bad they should be shuttered.
7. University licensing offices cannot recover their costs if they don’t get all inventions.
There are variations. If inventors have a choice, then all the really lucrative inventions will “go out the back door.” If inventors have a choice, then the university’s office won’t have enough volume to justify hiring specialists. If inventors have a choice, then they might just give their inventions away and the university wouldn’t have a chance to make money on them first.
Here’s a related statement to consider: University licensing offices cannot recover their costs in their present approach. One license deal every decade or two might make millions. Stanford had three lucrative deals in 36 years and has been the envy of most everyone. Then the millions might make up for the millions spent in patenting, salary for patent management staff, legal review and litigation, operating costs, insurance costs, costs in delays in research and contracting, opportunity costs as a consequence of the approach. But most university licensing operations don’t break even, even with comprehensive, compulsory ownership. One reason: they have claimed so much IP and non-IP that they are swamped with disclosures (which they report as success), swamped with patenting logisitics (which they report as a sign of future potential), and they are swamped with trying to document, inventory, and market all of what they have. They are so busy on the loading dock that they hardly have time to get anything out into service.
The compulsory approach creates such costs in volume and complexity of work that it makes it next to impossible to break even. If licensing is mostly a matter of focus, luck, and industry good will, then a university would do much better being highly selective or having inventors deal with any reputable invention management agent of their choice and asking only for a share of licensing income if an invention has been made or developed in a manner that supports a finding that the university has an equitable interest in the invention. Help inventors find good agents, provide substantial resources upon request, and request a share of the income from those inventions in recognition of that help. Few costs, no ownership, nice returns–readily greater than expenses.
8. The university must develop new sources of revenue.
This argument was basic sixty or seventy years ago and echoes along on its own even though circumstances have changed. Prior to the formation of the National Science Foundation and the conversion of a number of national foundations from prevention and public information to research, most universities did not have many sources for extramural research. Some industry work. Donations. Consulting. The federal government preferred to deal with two handfuls of universities and not much more. Cottrell, in setting up Research Corporation, envisioned that patents licensed under the direction of industry to industry could generate a research fund that could fund research nationally–at universities and elsewhere. The university-affiliated foundations that were created on the Research Corporation model dispensed with some of Cottrell’s design–they were provincial not national, and aimed to take inventions only from the affiliated university and provide a share of royalties only back to that university; they were not directed by industry but by alumni or faculty or wealthy boosters and potential investors; they operated locally, with offices near their university rather than near where someone might have an interest in what the university’s inventors had made. But they retained the idea that royalties could support research where there weren’t many other sources of funds.
MIT and then the University of California made the case that university ownership of inventions and university patent licensing could create research funds better than an external agent model. Stanford created such an office, too. Research Corporation encouraged the universities it worked with to create “technology transfer” offices–staffed by people who would inform inventors of Research Corporation’s services and help them document their inventions and send them to Research Corporation for review. “Technology transfer” meant “transfer from the lab to Research Corporation.” When Bayh-Dole hit, and university officials were given to understand that federal law required university ownership, these technology transfer offices were given licensing duties as well. But the same claim, that research funding from royalties was critical, didn’t hold up. By the time of Bayh-Dole, at most universities, federal money had swamped out all other sources of funding–often 60% to 75% of extramural research funding came from the federal government. Industry funding often was less than 10%. Foundations and state sources made up the rest. And Bayh-Dole perpetuated the argument for research from royalties by stipulating that universities (nonprofits–but the target of the legislation was universities and their research foundations) use any money after the costs of managing federally funded inventions for “scientific research or education.”
So the argument for patenting to make money for research is built into Bayh-Dole. It’s just that universities are floating in research funding. Universities are constantly expanding their research facilities to compete for still more money. At one point about a decade ago, some faculty at Stanford asked–publicly–whether it might be a good idea for Stanford to cap its extramural research budget and focus on really challenging and important projects, to go for quality with the infrastructure in place rather than quantity that required still more construction, more administration, more debt, more diversion of available funds and the future diversion of funds to pay for the new additions. But that idea, after some squabble, went nowhere.
Most patent royalties received by universities go into slush funds. If there’s a big hit deal, perhaps a building gets constructed, or an investment fund gets formed–use money to make more money and put that money in administrative slush funds. The University of Washington made hundreds of millions from the Washington Research Foundation’s licensing of a hepatitis B vaccine and related inventions, but put the money each year into administrative slush funds, some of which supported research that couldn’t get federal or industry support. I doubt anyone can point to anything of consequence that came of those hundreds of millions. It was like grinding up a truck load of gold bars every year and scattering the dust over the university.
The thing is–patent royalties mostly support the patent licensing office staff and the patent attorneys who do the patent work. While the argument that patent royalties can support research may have been true three generations ago, it’s not true now. Patent royalties support the careers of university patent administrators, and not much else, most of the time. The “royalties for needed research” argument is just a cover for “royalties to seek more royalties.”
9. Inventor personal ownership of inventions creates personal conflicts of interest.
This is the argument university patent administrators take to the university conflict of interest folks. University resources must not be used for personal gain; therefore, university resources cannot be used to invent or develop an invention that would be owned by the inventor. That would be a misuse of resources and a betrayal of the public trust. So, the argument goes, a university must demand ownership of inventions to save the inventors from the shame and discipline that must come if they are caught out claiming to own for their own, personal, selfish financial benefit. The university’s compulsory assignment policy is then merely an administrative remedy to avoid ethics problems created by private ownership of inventions. “We have no choice but to be totalitarian about it–anything else would condone unethical behavior.”
There is nothing inherently unethical in owning something that is used or improved in university activity. This happens all the time with scholarly manuscripts, with artistic works, with scripts for screenplays derived from novels written by faculty teaching creative writing or literature or dental hygiene. This happens all the time with collections of rocks or plants or shells or photographs or recordings–made by individuals and brought into the university to assist research, teaching, or just to stimulate discussion. These are not unethical acts–they are part of the very reason a university exists. The point of faculty appointments is to provide faculty with the resources to do what faculty do–teach and study–without having to take side jobs to get by. The universities resources provided to faculty are not the sacred bones of the public, to be touched only by priests, but rather are a form of subvention, a gift-in-aid to members of the faculty. That faculty have personal ownership of what they think, write, design, make, and invent is not unethical–it is a consequence of what a university is.
Even if personal ownership of inventions did create a conflict of interest (or an appearance of a conflict of interest), that conflict is manageable. Decisions regarding a given invention within the university can be referred to someone who doesn’t have an ownership interest or financial interest in the invention. One might own an invention and make it freely available to all. Where is the personal conflict? Oh, well, at UCSD patent administrators argued that it was a conflict of interest to release software open source because then, maybe, someone might pay someone who had worked on the software to help them with it. Thus, releasing software open source was–went the argument–the equivalent of cutting the university out of its licensing money in order to create opportunities for personal consulting.
University of Washington patent administrators a few years ago went even further than UCSD. They argued that it was a conflict of interest for faculty to use university resources to improve their personal know-how, because they might then use that know-how in private consulting–even if that private consulting was as a volunteer. Thus, the university must also own every faculty member’s know-how. The patent administrators didn’t rightly know how they owned know-how, but they were insistent that they did own it, and it was an ethics violation otherwise. (They ignored the exception in state ethics law for faculty involved in research–it was all about asserting their administrative power, never actually about ethics).
One might see in the UCSD argument the makings of an institutional conflict of interest, or a professional conflict of interest for the patent administrators. Their jobs depend on the money they bring in from licensing. Thus, when they construe the personal ownership activities of faculty as unethical, it is their own conflicted opinion that they assert, not some truth about the university and its resources. In an employer-employee relationship, the employer’s goods might be used only for the benefit of the employer. That’s an argument for fiduciary duty. But in a subvention, the provider’s goods are made available as aid to what those benefiting are trying to achieve anyway. The deal is: if you agree to join the faculty, we will make these resources available to you to do what you do. The deal is not: you now become an employee and everything you do will belong to us or you are an unethical cheat.
The challenge in deciding how a university’s resources are to be used, and what obligations a faculty member has to the university, is a matter of a university’s agreements with the faculty member–what has the university committed to, not just what the faculty member has committed to.
The claim that personal ownership of inventions creates a personal conflict of interest is bluster. Ownership is not the issue. Actions based on ownership may represent a conflict of interest or misuse of university resources, or may not. But it is not a conflict of interest to own. It is not a conflict of interest to dedicate inventions to the public domain (and so deny the university patent officials from the opportunity to make money from those inventions). And if there were to be a conflict of interest arising from actions, those conflicts are often manageable. The greater conflict of interest arises when patent management officials are permitted to make allegations in the name of the university about conflict of interest where their own jobs and careers depend (as they see it) on the university’s assertion of ownership for everything they can.
10. It is natural for the employer to own the work of the employee.
This is a most pernicious argument. It sounds so true, and yet it’s not true for patentable inventions, and it is not true for faculty, who are not, generally, employees when it comes to their scholarship and research.
Patents in federal law have the attributes of personal property. The Constitution gives the federal government to secure rights for inventions to their inventions for limited times. Courts have made it clear that even when an employee invents using the employer’s resources–equipment and time, say–the employer has no natural right to own the employee’s invention. The employer gains a “shop right” in the invention–an implied license to practice the invention in the employer’s business. If the employer desires to gain ownership of an invention, the employer must bargain for it. One bargaining chip, of course, is continued employment. Another is an offer of payment. A third is an offer to help in some way, serving as a test site for the invention. All this gets reduced to the advice that an employer must require a “patent agreement” with each employee that requires the employee to assign all inventions to the employer.
An employment patent agreement conditions employment and continuing employment on a promise to assign inventions within the scope of the agreement to the employer. Key elements are that there’s employment, there’s a clear statement of scope, that scope is within the tolerance of public policy, and there’s consideration for the agreement. For a business, the scope should have to do with an invention that pertains to the employer’s business or foreseeable business, or was made or developed with the employer’s resources.
Some states have laws that set such outer limits of an employer’s claims on employee inventions. If a business were to state that whatever its business might be, its business also includes trying to make money on any invention an employee might make, then the business would easily appear to circumvent these laws–though one might expect that there would be strong legal pushback. Yet this is just what universities have done. They claim not only inventions made within the “course” of employment, but also any invention made within the “field” of an inventor’s work, or the “discipline” in which the inventor works.
And of course, universities make this claim whether or not an inventor is employed–they assert their ownership via policy statements that they claim form agreements with inventors even when those inventors are students, visitors, volunteers, or even collaborators on research. In such situations, there isn’t an employment agreement, and there isn’t a patent agreement. There’s just a university assertion of a claim.
For university faculty, there’s another problem. While faculty are employees of the university, they are not employees for everything that they do while at a university. They do not accept university assignment or direction or approval or control with regard to their choices of professional development or research. Even if a university might penalize them for not doing research in expected areas, that’s different from a university assigning their work. In a similar way, a university might take into account in promotion decisions the awards a faculty member has won, but there’s no obligation for a faculty member to win such awards. Faculty, for much of their professional development, activities, and research, are not employed by the university unless they choose to be. Without an express agreement to be employed, faculty aren’t employees. They are appointees, members of the faculty. They are employees of the university for those activities that the university may assign them work–teaching assignments, for instance. The scope of a faculty member’s employment is remarkably narrow and unlikely to create patentable inventions.
There’s a second problem for patent agreements based on employment. When a faculty member is granted tenure, a university has acknowledged that it does not have a right to terminate the faculty member at will. There has to be more to it. So a patent agreement based on employment has problems when it comes to faculty.
Finally, putting the patent agreement in a policy statement has a huge limitation. If the university insists that it has the right to change policy at will, then the policy is fundamentally an agreement to agree–something that’s unenforceable as a contract. The university merely asserts a right to require whatever it wants to require. As the court in the Shaw case made clear, a university may change policy, but it can’t change a contractual agreement it has made with an employee without the employee’s consent. When a university makes a change in its patent policy and asserts that it can impose the change on all faculty, it is not offering the change for continued employment. It is threatening to terminate employment because a faculty member rejects the change.
faculty appointments and faculty members rather than faculty employees. Some university policies even now do not use “employee” to refer to faculty members in the faculty handbook portion of the policy. They are members or appointees, not employees. Employees are staff and administrators. The distinction is utterly lost in most university patent policies, as is the meaning of employee–since volunteers and visitors and students are also often defined as employees “for the purposes of this policy” (which is as much to say, “we ignore other policy in writing this policy” or “this policy matters more than any other policy”).
11. University ownership of inventions is more efficient.
Let’s see. If a faculty inventor wants to start a company to develop her (or his) invention, in the present university approach, the inventor must disclose the invention to the university, assign the invention to the university, hire someone to negotiate with the university, have that person obtain a license to the invention, and the company must agree to pay the university’s patenting costs but have no control over the patent prosecution nor the choice of patent law firm nor the billing practices. The university will impose limits on sublicensing, reporting requirements, often commercial milestone dates and payments, and sometimes a play for equity in the company. University exclusive licenses typically require the company to indemnify the university for any problems and carry insurance to cover the cost of any liability related to such indemnification. Potential investors in the startup have to take into account this license, along with all the different ways that the university can cancel the license or interfere in the management of the invention.
Now, if the inventor has a choice in the matter and assign the invention directly to the startup. The university might have a financial claim on a share of the money the inventor makes, but otherwise, there’s no assignment and license pathway through the university back to the company. The company controls patent prosecution, uses a law firm of its own choosing, which may then advise the company on other matters regarding the commercial deployment of the patent right, and the company has no obligations to a licensing university to report development progress or pay fees prior to selling product, just to keep the license from being terminated. How is this situation “less efficient”?
Similar things happen with licenses to companies. If the license is from the university, then university attorneys require all sorts of clauses to “protect” the university. These clauses not only cause delays in licensing and drive up the costs of licensing for companies, but also make their offered licenses less attractive.
A university’s licensing operation may be more efficient than really badly run licensing operation, but even an exceptionally well run university licensing operation is not necessarily more efficient than a well run professional licensing agent that specializes in the industry and technology represented by a given invention. And for startups, running an invention through a bureaucracy, no matter what bureaucracy, is almost always less efficient than dealing direct.