The IPA Public Covenant

[revised to add a discussion of Kennedy’s patent policy statement and distinguish it from COGR’s account of it; added an account of the various federal agency approaches to ownership of inventions]

Advocates for Bayh-Dole practice odd forms of historical revisionism. They claim that until Bayh-Dole, universities weren’t involved in patent commercialization–which is only deceptively true. Before Bayh-Dole, nearly all universities weren’t involved–as institutions–because administrators in those days were smarter than they were greedy and pushed patent work to external invention management agents. And when universities did take ownership of patents directly, often they did so to ensure broad access to everyone–so “commercialization” was not the objective, especially if “commercialization” means “profit-seeking monopoly licensing to the exclusion of research uses.”

And “commercialization” may–and does–take place in ways other than setting up a monopoly in which one company or one group of investors sets out to dominate a market–and has to come up with the money to do all the development in isolation. The idea that a monopoly is necessary as a remedy for “market failure” is just a goofy policy meme that gets retold around administrator campfires. It’s just that most administrators believe what they hear around those campfires, or at least believe that they will suffer ridicule, lose their jobs, and grow warts if they don’t at least pretend to believe the meme. The warts bit, no doubt, is true. 

The revisionistas also claim that before Bayh-Dole, the federal government was “presumed” to own inventions made with federal support. Here, “presume” is a weasel word, since it fails to identify who did the presuming and the basis on which they presumed. (Answer for the perplexed: it is the administrators who do the presuming, on the basis of self-interest.) The historical reality is different.

Different federal agencies adopted different regulatory practices. The military departments allowed company contractors to own the inventions they made under contract. The Department of Agriculture and Department of the Interior claimed government ownership of inventions that were part of their effort to develop proto-products that could then be released generally for commercial production–a new fertilizer, say, that the federal agency had demonstrated to be more effective than conventional fertilizers. The Atomic Energy Commission and NASA were governed by statutes that required federal ownership of inventions in atomic and nuclear energy and in space technologies, respectively, and so could release the government’s interest only after making a determination that doing so was in the public interest. The Public Health Service, as a matter of regulatory policy insisted on ownership of inventions made in its funding directed at matters of public health, and then expected to release those inventions for all to use and develop. Think commons or standards or open innovation rather than moving directly to products based on a private monopoly position.

In any of the cases in which a federal agency as a condition of providing funding claimed ownership of any inventions made under that funding, a contractor that obtained assignment of such an invention could appeal and the federal agency might release its claim. In the case of the IPA program, the NIH (and later the NSF) by-passed the need for a determination procedure after an invention had been made, and even by-passed the requirement that a federal agency head could determine in advance for any given research contract that the contractor that obtained ownership of an invention made under that contract could retain that ownership–under the IPA program nonprofits were required to obtain ownership of inventions when they decided to file patent applications on those inventions. The IPA stood alone as a master research contract without any research involved, consisting instead of only the patent rights clause for all subsequent research contracts between the federal agency and the participating nonprofit.

In short, then, there was no uniform presumption that the federal government owned all inventions made in federally supported research. The revisionists are wrong. If there had been such a presumption, then federal policy on the matter would have been uniform–and the revisionistas assert that it was Bayh-Dole that made federal invention policy uniform. They are wrong about Bayh-Dole, too, but that’s a different wrongness.

In short, too, the NIH–where Bayh-Dole eventually emerged as the rough beast that it is–really did operate within a federal agency policy framework that asserted that the federal government owned all inventions made in federally supported research directed at matters of public health. But this “presumption” was not government wide. Furthermore, the NIH itself in the IPA program reversed this presumption whenever a university decided to file a patent application on an invention that the NIH otherwise would be constrained by policy (and later, federal procurement regulation) to claim. So even for the NIH, it simply wasn’t true that the federal government presumed to own all inventions–the NIH itself circumvented that presumption. The only thing not uniform in the NIH’s own invention management program was that apparently not all nonprofits receiving funding from the NIH were in the NIH’s IPA program–though a great many were. The IPA program was terminated when the NIH sought to make the program available government wide. The program was reviewed, found to be both ineffective and operating against the public interest, and shuttered. Led by Norman Latker, legal counsel at the NIH, Bayh-Dole immediately thereafter began to slouch toward legislation.

Take a look at the Institutional Patent Agreement, helpfully archived at Bayh-Dole Central. The federal government did not presume that it owned inventions made with federal support. In fact, it presumed that inventors did–a good thing, given common law ownership of inventions. The IPAs required universities to obtain assignment from the inventors:


The IPA requires a promise to assign, conditioned upon the university deciding that it wants to file a patent application on. That is, if the university decides to file a patent application, then it will obtain assignment of the invention. There is no obligation to obtain assignment of every invention made with the support of the federal agency on the other end of the IPA. Only inventions that the university decides to attempt to patent are within the requirement to obtain an assignment. No need for a present assignment of all future inventions. No need to bottom trawl the university–kill most everything to try to catch one invention a decade. The IPA in this regard was smart.  There was honor among thieves in those days.

Here’s how the IPA assignment requirement plays out. First, the university has the right (in the federal contract–with regard to the federal agency) to decide to file a patent application on an invention made with federal support–on a “subject” invention. But for a university to file a patent application, administrators still have to get the cooperation of the inventors. That is, there’s nothing in the IPA that cuts inventors out of the situation–whatever “right” a university gets by contract is constrained to the mutual agreements in that contract.

Compare this to the nonsense that university administrators spout about having to use a present assignment of all future inventions in order to comply with Bayh-Dole. The IPA retains an expectation of selectivity: the only business a university has is with inventions made with agency support that the university decides to seek a patent for. Anything else is between the federal government and the inventors. Here’s the assignment clause:


Read carefully–the university (“Grantee”) is to require assignment when it decides to file an application. If the university decides not to file, then the university has no right in the contract to take ownership anyway–not even to pass that ownership on to the federal government. If the university is not going to file a patent application, then according to the IPA, it’s none of the university’s business what happens otherwise. That is, in the IPA, the university agrees with the federal government on what it will do with all subject inventions, regardless of whether the university decides to administer some of themThe university by contracting with the federal government gives up whatever freedom it would otherwise have to do something different with subject inventions.

In this case, the entire right of the university is bound up in the federal contract. The contract does not read–“university may do anything it wants with subject inventions, except if it decides to file a patent application, then it must get ownership, too.” Rather, it reads–“university may only get ownership if it decides to file a patent application, and then it must get that ownership.” The IPA is clear–if the university does not choose to file a patent application, then whatever happens with the subject invention is “subject to the disposition” of the funding agency “in accordance with its Regulations then in effect” (IPA, Article III). Slopping reading, of course, creates much more room for institutional self-interest–as Vannevar Bush put it, the “hardening of administrative consciences.”

The IPA dealt with the right of the government, in providing funding, to make delivery of subject inventions to the government a condition of funding. That “right” of the government is not a matter of patent law, but of money and federal agency regulations–what an agency requires of itself in entering into research contracts. If a university (or a faculty investigator) wants federal money, then the university (and faculty investigator) must accept the terms on which the money is offered. If the government is going to relax its opportunity to claim inventions as deliverables, then it may dictate what it wants instead of ownership.

The IPA gives the university the right (under the IPA) to elect to file patent applications and to administer the invention–but subject to the provisions of the IPA. This arrangement has entirely to do with standing in the contract, not some general right endowed by law. There are two parts to consider: the election and the administrations. Let’s look at both.


“Elect” is an odd verb here. Why not just “The Grantee may file patent application[s] and administer subject inventions pursuant to the provisions of this Agreement”? What does “elect” do? One clue is in the IPA recitals:


The Grantee represents that it wants an agreement under which it has “a first option” “to retain principal rights” in subject inventions (subject to the provisions of the IPA). So “elect” is tied to “electing an option” to retaining “principal rights.” That is, one “elects” the option offered by the government; one does not “elect” ownership of a subject invention.

What are “principal” rights, then? Why not just “Grantee desires to have the choice whether to own subject inventions rather than allow inventors to assign those inventions to the government or manage such inventions personally.” It would appear–the document is not a brilliant use of English–that “principal” is used here because the federal agency asserts a common interest in any patent that the Grantee university obtains on a subject invention. The Grantee-university gets “principal” rights–main rights, first rights, lead rights–but the rights are shared through the contract between the university and the federal government. The university gets the right to lead the dance, but it’s a partners situation, and the government is the university’s dancing partner. The government gets not only a non-exclusive license to any patent that issues, but also places a public covenant on the university’s “administration” of the subject invention and has the right to take over the administration on thirty days’ notice–if the university doesn’t perform or a federal agency feels the whim. The university, then, does not rightly “own” a subject invention even though it is required to obtain assignment of “all right, title and interest in and to each subject invention” for which it “elects the option” offered by the government. That’s an option to administrate, not an option to own outright, a subject invention. There’s no “quiet possession” in an IPA.

That is, the university in obtaining assignment owns the subject invention, except that the university in electing the option in the IPA also agrees to share the rights of ownership with the federal government. The university administrates principal rights within the requirements established by the federal government. It’s not rightly “ownership” if one lacks significant control over the property in one’s possession.

Let’s look then at the public covenant established by the IPA. There are multiple components. First, the government must review and approve a university’s patent policy and practices. The HEW regulations permitted the agency to allow a Grantee to manage inventions:


But the Grantee must have an “approved established patent policy” and that policy may have to be modified to meet the public covenant required by the federal government. The recitals then state the conditions for the review of university policy:


It is not just the university’s policy but also the university’s practices that must be reviewed. Further, these policies and practices are “subject to the provisions” of the IPA. When the government was asked to consider a university’s request for an IPA relationship, it asked for a range of information and statements, including ten years of licensing history.

This is not the freedom of private ownership to exploit a patent right. The IPA moves a review that’s case-by-case forward to prior to any invention, on the condition that all future cases are constrained by the same review and requirements. In such a case, the pressure of the IPA is to establish the most conservative course, because an up-front approval of policy and practices must cover all future cases of subject invention included within the scope of the IPA. Our defaults will be conservative, with the burden on the university to justify any exception.

The IPA anchors the requirements with two statements:

  1. “in the public interest”
  2. “the stated objectives” of the Kennedy patent policy statement


“In the public interest” is broad and open to all sorts of interpretations. University administrators have a habit of asserting that most anything they do must be in the public interest because they act for a university which by (their) definition must be acting in the public interest. The IPA does not mean that, however. If it did, there would be no need to review and approve a university’s patent policies and procedures, and further require modifications to those procedures, and further subject those approved procedures to a President Kennedy’s “Statement of Government Patent Policy” and further apply the express requirements of the IPA. So “in the public interest” means what all these stipulations require, not whatever floats in an administrative brain.

Let’s look at Kennedy’s policy statement. Some of its provisions move directly into the IPA, and later, in sausage-like fragments, into Bayh-Dole. The statement opens with “basic considerations”; especially, that significant federal funding has resulted in many inventions, which constitute a “valuable national resource.” Federal policy therefore should ensure that


Note here how the various parties that might have an interest are handled:

  • inventors–stimulated
  • government–needs met
  • contractors–recognize equities
  • public–serve their interest

Within this framework, the policy should encourage “expeditious development and civilian use of inventions.” But there are competing public interests:


The “incentives to draw forth private initiatives” (patent monopolies) must be “weighed” against “the need to promote healthy competition in industry.” And this is the crux of government patent policy. The Kennedy patent statement indicates that the government’s purpose in acquiring patent rights is to promote competition, and contractors may acquire patent rights when the purpose is to “draw forth private initiatives.” The Kennedy patent statement makes clear that the goal of government ownership of patents is to make new inventions available “through dedication or licensing.” Doing so creates a federal patent commons. The purpose of patents owned by the government is to prevent anyone from holding a private monopoly on inventions that ought to be available to all. This reasoning, of course, is beyond the comprehension of patent attorneys such as Howard Bremer, who argued that such practice ran against the very purpose of the patent system:

When the Government holds the patent under the aegis that the inventions of the patent should be freely available to all, much the same as if the disclosure of the invention had been merely published, the patent system cannot operate in the manner in which it was intended. The incentives inherent in the right to exclude conferred upon the private owner of a patent, and which are the inducement to development efforts, are simply not available.

Bremer does not indicate who actually intends how the federal patent system should operate. There is no prohibition against licensing patents non-exclusively–even WARF used that strategy. There is also no requirement that an inventor or patent owner find an incentive in charging a fee for the use of an invention. Licenses to patents are traded, for instance–cross-licensing–or pooled–as in the creation of a standard. In such cases, opportunity is created not by exclusion but by making inventions “available to all.” Unlike dedication to the public domain, the existence of patent provides a basis for engaging competitors in common practices that might otherwise not come about. In the early history of the aerospace industry, for instance, the federal government intervened to broker a patent commons because the many companies fighting over patent rights prevented any one of them from building an airplane competitively with European industry counterparts.

It is clear that the Bayh-Dole advocate’s claim that the government failed to commercialize inventions then is disingenuous. The point of government ownership was to stimulate competition, not to serve a particular private initiative. Each invention did not have to be “commercialized” by exclusionary dealing. That direction led to gridlock. The patents the government allowed a contractor to exploit were the ones for which commercialization followed from the contractor’s established commercial position and did not give the contractor a dominant position.

This division of labor between the federal patent commons and private initiative is utterly lost in most accounts of the role of Bayh-Dole. Bayh-Dole did not enable commercialization of otherwise rotting federal inventions–it destroyed a federal commons intended to promote healthy competition. If federal policy was not promoting healthy competition, then the change in policy was not to give up to monopoly interests but rather to exercise greater care in what inventions were sorted for promotion of competition and which were to good for stimulating private initiative.

The question, then, raised by Kennedy’s patent statement is: when is it appropriate for a contractor’s private initiative to take precedence over healthy competition?

The Kennedy patent statement recites the outer bound of antitrust law as a constraint on the exploitation by a contractor of patent rights in inventions made with federal support. But there are other concerns as well, both in the statement and in how the IPA program justifies its existence within the framework of the statement.

While it is not clear how inventors are stimulated by the policy set forth in Kennedy’s patent statement, it is clear that inventors are included as a key element in the policy. They are not alienated from their inventions, for instance, as faux Bayh-Dole would have it. And under the IPA, they do receive a share of royalty income, if a university’s patent policy permits it (Bayh-Dole requires royalty sharing with inventors–the heck with university patent policies).

The governments needs are met with non-exclusive licenses or securing “principal rights” in inventions made with federal support. The Kennedy patent statement sets out four circumstances in which the government “will acquire or reserve the right to acquire the principal or exclusive rights” in inventions made with federal support. The government should obtain “principal rights” in inventions made with federal support when the research is to create a product for public use, either commercial or required by federal regulation:


In such a case, the government is doing the development. There is no need to stimulate a private initiative or to permit exclusive exploitation of a patent right. To allow a monopoly here would run against public interest–it would simply allow a private party to exploit for profit a government project. It makes sense for the government to hold the patent right and have folks compete on some basis other than patent.

The government should also obtain title to patents when the work is directed to public health or welfare:


If the public health or welfare is at stake, why should the government allow any contractor to have a monopoly? For instance, no patents may be enforced on surgery techniques as a matter of public policy. It makes some sense–why should any doctor be prevented by a patent from aiding a patient? All the more so for patents on inventions funded by the federal government.

The expectation set out here in (2), however, is the central objection by the patent middlemen that led first to the IPAs and then to Bayh-Dole. To get to the IPA, however, people had to construct a protection for the use of patents in promoting the public health and welfare that followed from the government’s interest in public access while permitting some forms of monopoly. We will see how this comes about.

There are two other conditions in which the government should expect to obtain principal rights to inventions. The first of these is when the government has been the primary developer of the field or there is little expertise in the field outside government and a contractor holding patents would have an inappropriate position, such as in competing for future federal contracts or in preventing the development of expertise elsewhere in industry:


If the government’s interest is in building a new capacity in industry–or in creating a new industry–then allowing any one contractor doing work for the government to own patent rights potentially frustrates the government’s efforts. Certainly allowing such a contractor to enjoy any rights of patent–to suppress all private use, to use a dominant position to force others to give up their technology, to ruin emerging competitors with infringement lawsuits–makes no sense. To allow a contractor to hold rights in such inventions requires something more than an invisible hand of self-interest. There has to be some sort of public covenant to achieve on behalf of government–and with better success than government–what the government seeks to achieve. One has to serve that interest, not self-interest, and it is not likely (or possible) that an unconstrained self-interest in a patent monopoly will magically work itself out as a new industry full of competition and diversity.

The fourth area of government ownership is straightforward–where the government contracts for someone to operate a government lab or coordinates work for the government. Here, the Kennedy patent statement recognizes the prospect for an exception:

kennedyexcept1a kennedyexcept1b

The government should acquire rights to inventions “made in the course of or under” the federal contract. But the government can allow broader rights “at the time of contracting” if doing so “will best serve the public interest” or later, after there’s an invention, if the invention is “not a primary object of the contract” but then only when necessary to “call forth private risk capital and expense to bring the invention to the point of practical application.”

This is the ground zero of the middlemen’s efforts to get at the results of faculty research. Universities take federal money to coordinate the work of faculty, who have proposed research to the government. The government can allow private ownership after an invention has been created when it was not an objective of the contracted work and does require private investment to develop.

The policy here does not say that private interests can buy their way into a profitable monopoly but that a private monopoly is necessary to the development of an invention for practical use. If the only purpose of the monopoly is to make others pay for use or suppress the development by others, then there is no point in the monopoly. The federal government may hold the rights and allow use, since the private investment happens without a monopoly–and perhaps happens only because there is no monopoly.

The IPA must navigate this condition in order to establish that an invention can be managed by the contractor rather than the government. But there’s more work to do to move the decision whether the university can manage such inventions to the “time of contracting” rather than after an invention has been identified. For that, the IPA must deal with the “exceptional circumstances” that permit a federal agency to give up a claim on principal rights upfront, before any work has been done. And the IPA has to take it even one step further, and allow for a federal agency to give up its claim on principal rights even before any contract has been offered–that is, for all future contracts, not contract-by-contract and not invention-by-invention.

The Kennedy patent statement then turns to conditions upon which a federal agency should allow a contractor to acquire principal rights:


If the contractor already has expertise, patents, and a “commercial position,” and the contract asks the contractor to provide work product within that expertise to the government, then the government should allow the contractor to acquire principal rights, subject to a non-exclusive license to the government. In other words, in these situations, the government should contract only for the delivery of rights necessary for government use, not to the invention and its patent rights. Put another way, the risk capital and incentive is already present in the contractor, and to demand exclusive government rights would be a disincentive for private development. This appears to be the part of the Kennedy patent statement that recognizes the “equities” of the contractor.

The Bayh-Dole revisionista history in which the government claimed all rights to inventions doesn’t hold up, at least not for the Kennedy memo. Add to that the IPA exploit adopted by 77 universities–and which Bayh-Dole dismantled–and it’s clear that there were plenty of opportunities for private contractors to use the patent system to develop inventions made with federal support. What there weren’t much of were opportunities for patent speculators or other middlemen to get a piece of the action. Bayh-Dole would become a law for the benefit of middlemen.

If a contractor does not have an established commercial position or expertise, then agencies are told to manage inventions on a case-by-case basis.

Kennedy’s patent statement then proceeds to set up the outline of a public covenant when principal rights “remain” with a contractor:

  • the contractor will provide written reports on commercial use
  • the contractor has three years from patent issue to take “effective steps” to (i) bring an invention “to the point of practical application” or (ii) make the invention available for licensing royalty free or on reasonable terms (i.e., the practice of the federal patent commons) or (iii) or “show cause” why the contractor should “retain” the principal rights. Otherwise, the government can require non-exclusive, royalty-free licensing.

If we break this second requirement down, we see it framed by the federal patent commons. Either the contractor within three years has developed the invention to practical application or can explain why exclusive rights are still necessary to such development, or the patent should enter the federal commons, either by contractor action or government action.

Kennedy’s patent statement defines “point of practical application”:


This definition requires that the invention is (i) used so that one can (ii) establish that the invention is “being worked” and (iii) its benefits are (iv) reasonably accessible to the public. This is the heart of the public covenant for privately held patents on federally supported inventions. It is not sufficient that an invention is being used or that a patent is being exploited–the use has to be “establishable,” has “benefits,” and these benefits are “reasonably accessible” to the “public.” This definition is incorporated by reference in the IPA and shows up slightly modified in Bayh-Dole.

The property rights offered by a patent held by a contractor run three years to establish reasonably accessible public benefits arising from the use an invention made with federal support. Otherwise, the contractor has participate in the federal commons or show a reason why not. And these requirements extend to a licensee or assignee. The covenant runs with the patent.

It is clear that the Kennedy administration, at least, that the federal patent commons is a core repository for federally supported research. Exceptions to the federal commons were allowed “in special situations.” Even when private capital is necessary to develop an invention or a contractor has an established commercial position or the federal agency determines that it is in the public’s interest that a contractor hold principal rights in an invention, the contractor still has only three years from a patent issuing–perhaps seven years overall (one year to file, three years for patent prosecution, three years after issue)–to establish practical application. These are the requirements that the IPA must meet.

Now let’s look at how Kennedy’s patent statement was interpreted by the universities [document since removed from Bayh-Dole Central] by considering the summary of points in a memo created by the Council on Government Relations. Doing so helps to understand from the university side the IPA’s public covenant. Let’s review the university synopsis.

It’s in the public interest for the federal government to fund research at universities. According to COGR, universities accept this funding to improve their ability to serve the public interest.


There’s nothing in the Kennedy patent statement to support this contention that federal funding is provided to strengthen university “capacity to serve the public interest.” It surely was not a matter of federal policy. As Vannevar Bush has it, funding to universities supports the activities of faculty–in training and research–and it is their activities that serve the public interest. The university is the steward or trustee of the activity. The university serves the public by enabling its faculty to serve the public. This contention that places the responsibility with universities, however, frames the argument for university ownership of inventions.

The question that the COGR account then takes up is the proper role for the university. For this, the COGR memorandum sets out a list of six items that “should be considered.” Think of them not only as talking points but also the building blocks of university patent policy and the rationale for the IPA and later Bayh-Dole.


Since federal agencies don’t have the responsibility to introduce inventions into public use, someone else will have to do it. The Kennedy patent statement, by contrast, anticipates both that the federal government may contract to develop products for commercial use and that many inventions will not require a monopoly position to attract public use. Thus, the COGR memo sets up the appearance of an absence in federal policy when there is none.


Many universities have policies and procedures that foster public use of inventions made by inventors. Well now–if that’s the case, then these universities, whatever their approaches, have resources the government ought to support. It’s worth noticing that the emphasis is on “public use”–not commercialization, not exploitation of patent rights to raise institutional money for research, not to amass a pile of patents in the hopes that someday a wealthy company will stumble across a claim and have to pay up or face an infringement suit–or stop using the invention. The subtly in this second point is that there’s no indication that the policies universities have developed have been responsible for the development of successful products. Cottrell at California and Steenbock at Wisconsin developed their ideas about patents without the benefit of university policies. Many university policies subsequently directed inventors to use such external agents. It is not at all well established that the university policies have done anything other than provide a basis for university administrators to claim a share of royalty income. That is, just because a university has a patent policy and there is later a successful product does not mean that the policy is responsible for the product. The product may have just as well been developed despite the university’s policy, or more slowly, or at greater expense than otherwise. There is work to do to show that a policy is responsible.


The university policies are varied, not uniform. This is an attribute worth calling out. Varied policies mean “flexibility of approach” and “variety of interest.” These attributes are necessary to meet “the broad spectrum” of inventions coming from research. Nothing here about forcing a uniform policy on everyone. Nothing here about how a pharmaceutical industry love of monopoly requires every university to make that their goal and impose monopoly licensing on all other aspects of university research results. But alas.


This part is worth a whole article on its own, for another time. If universities have figured out how to introduce research inventions to public use, and they have done so by adopting a varied bunch of policies and approaches and interests, and so can handle a spectrum of inventions, then it makes sense that one university might refer an invention to another university better suited to handling that bit of the research spectrum of results. Of course, when this idea was raised a few years ago (perhaps I had something to do with it), AUTM rushed to mock it and got university administrators to say it would never work.

AUTM strongly opposes “free agency,” a concept which would allow university faculty to shop discoveries to any third party for licensing—regardless of where the research was conducted.

AUTM, the most anti-innovation of innovation-facing organizations. The Cortez or Sheriff Brown of opportunity (“…what a killer, “…kill them before they grow“). As Research Enterprise has pointed out, free agency was how university tech transfer started, with Research Corporation and inventors choosing where to take their inventions, once they had decided that patenting should be part of the effort to encourage public use. Free agency is what creates markets, is what encourages brokers to compete to provide the best services. Free agency is what allows an inventor to choose a broker whose specialty and capacity match the inventor’s strategy and the invention’s requirements. What is it about free agency that so scares university bureaucrats? Funny, university inventors do not appear afraid. Nor does the public cry out to bureaucrats–“please, please protect us from university inventors who might choose the best organization to help them achieve a vision for the public use of what they have discovered!” No, that cry does not exist. Just the pontificating of university bureaucrats.


And so here we have a fundamental statement of inventor free agency. Decisions about whether to patent an invention and encourage public use “are best decentralized and made by those concerned with the discovery or invention in an immediate and practical sense.” The memo here does not designate university bureaucrats as those “concerned” with an invention in “an immediate and practical sense.” But the COGR memo also cannot bring itself to allow inventors to make decisions about what to patent, or how that patent might be used, and to what purpose.

The arrangements are left to be decided within an approved policy framework that is “dedicated to the public interest.” This is the core of the rationalization to permit universities to step in and perform services involving patents that otherwise the federal government would administer. University patent policies were indeed varied–many were based on inventor decisions to seek patents and voluntary choice of what broker, if any, to use. That is, of those universities with a patent policy, many–if not most–practiced some form of inventor free agency. It was not that inventors were free to do anything with an invention–but rather they were free to choose who might help them do things that were “in the public interest.”

The spin that the IPA puts on all this is that the particular requirements of the IPA are within the objectives of the COGR memo:


This item is the strangest of the six. The patent system’s purpose is to promote the progress of the useful arts. There’s not a premise of financial incentives. The COGR memo recites Bremer’s assertion–that a patent system cannot be used for anything other than creating private monopolies that attract investment. The COGR memo constructs a rationale for the patent system that is not there. Inventors may patent and never license; may demand unreasonable terms; may never work their inventions; may make inventions freely available to gain benefits from the sale of related goods and services; may trade patent rights to gain access to others’ technology; may dedicate patents to a standard. The COGR memo suppresses all such uses of the patent system, for better or worse.

The thrust of this item 6 is not to focus on commercialization for profit, but to make clear that as far as federal policy goes, it is acceptable that inventors and universities and companies acting in the public interest also may “be allowed” “reasonable incentives” to stimulate their activity to introduce “into public use” inventions made with federal support. In short, item 6 argues that subject patent licensing does not have to be royalty free–the public is not failed if there are “reasonable” incentives. What constitutes a “reasonable” incentive is a question left unaddressed by the memo. This is the crucial argument that attends a claim by the university to manage inventions made with federal support. Where an academic inventor may have no such interest in “reasonable incentives”–they are, after all, academics who have chosen public service–a university administration may require such incentives. This, too, is strange. If universities are dedicated to the public interest–perhaps as a matter of policy–then what financial incentives arising from monopoly positions do they in addition require?

From these considerations, however, we can see that there is sense that the federal government is prepared to impose a uniform patent policy on universities. Rather, the COGR memorandum acknowledges the value of diversity and flexibility, anticipates that inventors might seek out university programs best suited to their opportunities, and that in all the flutter there is an argument that inventors, their universities, and the companies they worked with should expect a reasonable profit for their efforts. Not unlimited profit, not the incentive to do anything with their patents–but profit and incentives appropriately in the public interest.

In the Kennedy patent statement, the government policy is to “recognize the equities of the contractor.” For a commercial concern, these equities have to do with its commercial position. A university, however, has no such commercial position. What “equities” might a university possibly have? The government reimburses the university for the salaries of its researchers, pays their direct costs, and compensates the university for its administration of each award. A university incurs no costs and makes no investment in the federal work (unless, of course, it is inefficient in its administration or chooses to add its own money to augment the federal work). In short, a university has no “equities” in a federal contract. It has no right to own inventions, no commercial position from which to develop inventions, no special position from which to claim a preference over the federal government–other than that of asserting a public covenant that constrains its patent practices and arguing that these constrained patent practices will lead to public benefit.

The IPA follows the considerations of the Kennedy patent statement in its development of a public covenant to run with any subject invention to be administrated by a university. The COGR memorandum runs along side and sets the stage for a competing rationale. The Kennedy patent statement will be transformed into Bayh-Dole, while the COGR memoradum lays the foundation for Faux Bayh-Dole.

Here is the basic condition of the IPA:


The procedural stuff is straightforward. Administer subject inventions in the public interest. File patent applications. Obtain assignments from inventors. Require assignments from any subcontractors. Report subject inventions to the government promptly or within 90 days of a bar date if the university does not choose to file a patent application. Grant the federal government a non-exclusive license. Give the government power of attorney to inspect patent prosecution documents. Include a federal funding statement in issued patents.

It is worth noting that there is a difference between “administer inventions” and “administer inventions in the public interest.” If any use of the patent law were “in the public interest” there would be no need for the qualification here. Similarly, there is no need for a qualification if a university can do nothing other than administer inventions in the public interest. “In the public interest” means, then, what the Kennedy patent statement sets out as the public interest–and what the IPA then develops.

Here’s the rest of the sentence started in (c):


There’s your default public covenant from the Kennedy memoraundum. Nonexclusive licensing, royalty-free or reasonable royalty. Not the “best royalty rate you can get” but a “reasonable” royalty. Nothing here about suing for infringement or threatening industry to prevent use or encourage use in work-arounds and alternatives to subject inventions. It’s funny, too, that the IPA imagines “applicants” for licensing, something that the Kennedy patent statement also describes. Most universities never see more than one “applicant” for a license, and most university licensing operations are set up to offer monopoly positions as their vastly preferred default.

The IPA then provides a procedure for departing from a non-exclusive licensing program, set forth in paragraph (d). We work through it part by part:


To move to an exclusive license, a university must make a determination that nonexclusive licensing won’t work for a “commercial market.” That determination is not mere whim, nor policy pronouncement. A university has to do at least one of three things:

  • make “reasonable efforts” to license non-exclusively, or
  • determine an exclusive license is “necessary as an incentive” to develop an invention, or
  • show that “market conditions dictate” exclusive licensing.

The first two conditions are express in the Kennedy patent statement. The third is not. There’s nothing in the Kennedy policy regarding “market conditions.” A contractor allowed to acquire primary rights has an expectation of bringing an invention to the point of practical application–regardless of how it does so.

But we are not done with the requirements on exclusive licenses:


An exclusive license has its own defaults in the public interest. The term of an exclusive license will be no more than 3 years from first commercial sale or 8 years from the date of the license–whichever comes first. Plus, the license must require diligent efforts to get to a commercial sale sooner rather than later.

But just as exclusive licensing is an extension of the public covenant default of nonexclusive licensing, so too there is the possibility of extending the limited term of an exclusive license:


But the federal government has to approve any extension of the 3/8 year limits on an exclusive license. After the expiration of exclusivity, the university has a positive obligation to offer non-exclusive licenses, still with a reasonable royalty, and not more than the royalty rate of the exclusive license. (There’s no overt requirement in the IPA that an exclusive license has to carry a “reasonable” royalty rate–but item 6 of the Kennedy patent statement discusses “reasonable incentives”–presumably it’s a “reasonable incentive” for a company not to have to pay an unreasonable royalty.

There is a further, general stipulation on licenses, regarding royalties:


This is an oddly stated requirement. The exclusion of the federal government is expectable, though why the federal government should be exempt from the public covenant imposed on universities is worth a discussion some time. Invention management organizations are dealt with separately–any contract with an invention management organization has to be with a nonprofit and be approved by the federal government–and the patent management organization is subject to the IPA requirements as well. So whatever is being referenced here in (e), it’s not either the university or an invention management organization. One would think, then, that the licensees are those aiming to develop a subject invention and have the right to sublicense or exploit the licensed patent.

The central requirements set forth here stipulate:

  1. safeguards against unreasonable royalty
  2. safeguards against repressive practices
  3. all sales to the federal government are royalty free

In general, a “royalty” is any consideration for a patent license. The university, under the IPA, is already obligated to ask no more than a reasonable royalty. So this paragraph (e) must not apply to the university’s practices, but to the practices of others that the university licenses to. A company, for instance, acting in its own interest rather than the public interest might require an unreasonable royalty in a sublicense, even if its has cleverly negotiated a merely reasonable royalty with the university.

A repressive practice must be something other than demanding an unreasonable royalty or it would not be treated separately. One possible repressive practice would be to refuse to license the invention–or to take money to prevent the use of the money, or to license to a company with the knowledge that the company intends to keep the invention from the market. This would be, perhaps, the “nonuse” referred to in Bayh-Dole’s statement of policy and objectives. But repressive practices might also include the tying of a license to the purchase of other products of the licensor, collusion on pricing of product, and dividing up markets to prevent competition.

What the actual safeguards in a license might be to prevent unreasonable royalties (pricing?) and repressive practices (of all sorts) is not set out. One might expect diligence to develop, easy downgrade to non-exclusive or termination of the license, regular reporting–that is, the things one would expect with “march-in” by the university-licensor to enforce the public covenant, rather than march-in by the government. Again, the difference with Bayh-Dole is stark–under the IPA, the university undertakes march-in responsibilities to defend the public covenant. Under Bayh-Dole, march-in is shifted to the federal agencies and made next to impossible–leaving universities free to exploit subject inventions any which way they care to, from nonuse to trolling to monopolies to patent pools to unreasonable royalties to indifference. Universities must require protectionist clauses–exclusive US licenses must require substantial US manufacture; any license must give preference to small businesses–but what happens under the license is whatever. Under Bayh-Dole, the government has to step in. Under the IPA, universities were expected to step in. That’s a huge change, if you think about it.

The third requirement of paragraph (e) is the most difficult. The government already has a royalty-free license to the subject invention from the university. Paragraph (e) is concerned with sales to the government by the university’s licensees. But to comply with (e), a university then can require only a royalty on sales–any other form of royalty, such as upfront payments or payment in-kind (such as discounts on company products) or payment with company equity (as with startups) would make it impossible for the company to know how much to deduct in sales to the federal government. If the university requires a 1% royalty on sales, then a company selling a patented device for $10,000 expects to pay $100 to the university for each sale. The government, so paragraph (e) stipulates, would pay $9,900 and the company would pay no royalty to the university for the government’s purchase. That much appears easy.

But if the company had paid $100,000 up front to the university in addition to a 1% royalty on sales, how does the company calculate how much of that $100,000 is represented in each sale to the government? It’s even worse if the entire royalty owed the university is reflected in warrants to purchase stock and the value of the stock could far exceed the total purchase price paid by the government. If the government is the only purchaser (as might be the case with weapons or space gear or currency printing equipment), then how does the government’s price get established? It would appear paragraph (e) on this point is intractable. But everything makes sense if a university is restricted to asking for no more than a reasonable royalty on sales. Nothing more. Indeed, making a university wait for commercial sale before getting anything ensures that the university is ready and willing to cooperate with the company to develop a subject invention–and also ready to terminate or downgrade the license if the company fails on its end.

There is one further stipulation in the public covenant, having to do with the university’s disposition of royalties. Inventors may be paid a share of the royalties, but only within limits stipulated by the IPA–no more than 15% beyond $13,000 (1968 dollars–about $90,000 now). A university may keep the rest, but with restrictions:


There are two restrictions, in fact. First, royalty income can be used to recover the cost of the administration of subject inventions–but subject inventions only. There is no provision for a university to use royalty income from subject inventions to cover any other licensing costs. There must be a separate accounting for the “expenses incident to the administration of all subject inventions assigned” to the university. That is, expenses can be recovered only when a subject invention has been assigned, and subject inventions can only be assigned when the university has made a commitment to file a patent application. Any part of the licensing program that involves management prior to a decision to patent and obtaining an assignment is not recoverable from IPA-based royalty income. Second, any remaining royalty income must be used “for the support of educational and research pursuits.” Not to be invested in real estate or stocks to make money; not to pay administrators; not to pay for the rest of the tech transfer program–but for education and research. Pretty plain stuff until it hits the hardened consciences of university administrators. But it’s absolutely clear that no royalty income is to be used for any patent or licensing work involving any inventions that are not the subject inventions the university has claimed under the IPA. Income from royalties on subject inventions cannot pay for the rest of the university’s patent licensing program.

One might see in this how different the definition of subject invention is in the IPA from that of Bayh-Dole. In Bayh-Dole, an invention is not a subject invention until it is owned by a contractor. The definition of subject invention in Bayh-Dole stipulates an “invention of the contractor [made] in the performance of work under a funding agreement” rather than an “invention [made] in the performance of work” for which a contractor might later obtain assignment. To make the Bayh-Dole definition work, the drafters of the implementing regulations had to require universities to add research personnel as parties to the funding agreement (via the (f)(2) agreement), and therefore within the definition of contractor, so that when they invented, they would be the owners of the invention, and as contractors they would produce a subject invention. Not that anyone in AUTM will ever understand this.

The public covenant, then, entails a university policy and practices that have been approved by the federal government and comply with government policy; claims only on subject inventions that the university chooses to file patent applications on; for these inventions, non-exclusive licensing at no more than a reasonable royalty, exclusive licensing only with a determination that it is necessary and only then on a limited 3-8 year window, with safeguards for unreasonable royalties and repressive practices by licensees–essentially university march-in requirements as the second line of defense.

There is nothing here that presumes university ownership or federal ownership of subject inventions. Nothing that requires a uniform federal policy or uniform university policy for either compliance or optimal performance–indeed variety and flexibility are the themes. Nothing presumes public use requires commercialization or that commercialization requires monopolies. In all, a workable approach that over seventy universities adopted prior to Bayh-Dole.

There is one further step in the public covenant: government enforcement. Here is where it all gets most interesting.

IPA additional licenses

This provision tracks the requirements of the Kennedy patent statement. A university has three years after a patent issues to take “effective steps” to get a subject invention that has been “left for administration” to “the point of practical application.” That is, about seven to eight years from the time an invention is reported. If not, then either the university releases the invention non-exclusively on reasonable terms or “shows cause” why the university should continue to hold rights or the government can march in and require any of three actions:

IPA additional 2

Again, these requirements are taken from the Kennedy patent statement. A university administrating subject invention has to perform or the government can take title, cancel exclusive deals, or compel non-exclusive licenses on whatever terms the government determines are reasonable. The protocol for government march-in is remarkably simple, if not blunt. It happens upon notice once the conditions have been met. The government can also march in at any time if it determines that the subject invention is “required for public use by governmental regulations”–makes sense–or, more generally, the public is better off with non-exclusive licensing:


Again, the determination is at the discretion of the federal government–a university is hardly an owner of a patent right as “personal property” under the IPA. The university must have a conforming patent policy and practices; must secure ownership of subject inventions it decides to patent; must report the status of each invention; has a limited time to achieve practical application; must default to non-exclusive licensing and justify anything else; and must use any royalty income after expenses for educational and research purposes–and if there’s nothing within three years, then the government can march in.


Even if there is practical application within three years, the government can still–as it wishes, for the most part–march in if it determines that the public interest is better served by non-exclusive licenses. All the university and licensees gets is thirty days’ notice to request a hearing. Jaw-jaw but no assurance anything will change.

Other than holding title, there is very little about university administration of subject inventions under an IPA that has to do with a property right. The university holds “principal rights” to administrate an invention, not to behave like any random owner of a patent right. Whatever patent law might permit an inventor (or patent owner) to do–prevent all use, troll industry that stumbles over claims, create monopolies, charge exorbitant fees–the IPA creates a public covenant that “runs with the patent” that significantly limits the property rights in a patent–achieve practical application, introduce the invention into use, expect only reasonable incentives for doing so, and use any institutional royalty income after assigned subject invention expenses for education and research only.

Anyone obtaining an exclusive license to a patent on a subject invention has to get to commercial product within three years of a patent issuing to expect to retain exclusive rights, and even then, exclusivity runs no more than three years from first commercial sale or eight years overall–whichever is shorter, unless the government approves otherwise. While the patent system is used under an IPA, it is a short-term patent with a ton of restrictions on licensing and use of royalty income. The “principal rights” of the IPA are not “title” but merely to step in and do something that the government has no mandate to do itself. If a university can’t do it either, then things go non-exclusive, if not free, at the government’s determination.

Even then, what’s sad is that the outcomes were not nearly what one might expect from reading the hype put out by advocates of Bayh-Dole. Here’s the DHHS statistics as of 1981 (these appear to be cumulative over the history of the DHHS IPA program):


Let’s see–universities claimed 89% of what they saw. They work with 96 inventions, with 76 issued patents and 22 more pending. Result: 10 total licensees and 4 products. That’s, um, a 4% practical application rate–about the rate the federal government was accused of having–but the government got to 4% without much trying and including all the inventions that its contractors had declined to administer but a federal agency had sought patents on anyway. The IPA system was no better than the government’s approach. One might consider that federally funded inventions managed by universities–even when selected for desirability–were 8x less productive than were privately funded inventions chosen by inventors for patenting and placed with invention management brokers of their own choosing.

Now, after 35 years of Bayh-Dole, it appears that the licensing rate is about 0.5% with 1 invention in 1,000 resulting in a really successful commercial product. Not only did Bayh-Dole lead to the destruction of a private network of invention agents claiming a 25% or better success rate but also to the dismantling of an IPA infrastructure serving over seventy universities that implemented a public covenant and functioned 8x better than universities are doing under Bayh-Dole. Bayh-Dole has led to invention management practice nearly 100x less productive than the voluntary, selective, diverse, flexible private network of invention support that it destroyed.

Despite the bombast from Bayh-Dole advocates–and it is bombast–Bayh-Dole has offered no improvement over the IPA system in terms of productivity, and surely has created a massive new bureaucracy, one that believes it is entitled to midwife every new idea, separate it from its presumably dysfunctional parents as quickly and incisively as possible, and offer it for profit to speculators, all in the name of public interest. What a disaster, what a killer.

This entry was posted in Bayh-Dole, History, Policy and tagged , , , , . Bookmark the permalink.

1 Response to The IPA Public Covenant

  1. Pingback: Senator Bayh’s inventor-loathing faux Bayh-Dole Act | Research Enterprise

Comments are closed.