From time to time, I revisit territory. I wrote about this issue almost two years ago, now. I provide here a different angle that gets at the same point.
Here’s Sean O’Connor proposing that a flawed assumption in the U.S. Attorney General’s 1947 report on government patent practices led to the Stanford v Roche dispute and Supreme Court ruling (footnotes removed):
The Biddle Report not only framed the debates over government patent policy for both intramural and extramural R&D for decades to come, but it also introduced a crucial mistaken assumption that all government contractors were routinely securing patent assignments from their employees.
Almost all government work at the time involved for-profit contractors, primarily in the defense industry. It appears only a couple of handfuls of universities were involved in government contracting. The report’s assumption is reasonably accurate that all the government contractors that mattered in 1947 were securing patent assignments.
For contracts with universities, the Navy’s 1946 contract with the University of Southern California for research pertaining to the physiological effects of acceleration might illustrate the strategy (quoted in the Mine Safety Appliances v The United States decision):
(b) Contractor agrees to and does hereby grant to the Government an irrevocable, non-exclusive, non-transferable and royalty-free license to practice, and cause to be practiced for the Government, throughout the world, each Subject Invention in the manufacture, use and disposition according to law of any article or material, and the use of any method * * *.
The Mine Safety Appliances court reasoned:
The Government desired and paid for a royalty-free non-exclusive license resulting from that type of research (among others), and neither U.S.C. nor the inventors could avoid that federal right by unilateral action.
That is, the court doesn’t care whether USC owns the patents or the inventors do–USC agreed to grant the license in a federal contract and USC controlled its inventors. The court asserts that the license exists without the need to determine who, as between the inventors, USC, and the company that the inventors formed, actually grants the license:
Even in the absence of a patent-license clause in the contract, the same general policy has been applied by this court to give the United States an implied license, as against the contractor-patentee, to use inventions resulting from a federal experimental and development contract.
This finding was not without a dissenting opinion with regard to whether the scope of the government’s implied license was broad enough to include a helmet developed by the inventors and which USC claimed was outside the scope of the Navy contract. But the primary point is that in government thinking about university inventions, it was sufficient that the government have a royalty-free license, and it simply did not matter what sort of patent assignment agreement a university had with its inventors.
Thus, it appears that at least the Navy’s assumption was that the federal government obtained whatever rights it needed by a required license, and it was up to the contractor to manage its affairs so ensure that this license was granted, regardless of the ownership. Such a license could be granted in a number of ways. The university could acquire ownership and grant the license. The university could permit the inventor to own the invention on the condition that the inventor granted to the university sufficient rights to grant the required license to the federal government. The university could require the inventor to grant the required license to the federal government. The university could make the inventor a party to the contract, so the inventor had the same obligation to grant a license to the federal government as the university did. Doesn’t much matter. My point: the assumption throughout much federal government contracting is that if the government has a license to the invention, that’s enough. Ownership of inventions is then a separate matter.
Back to O’Connor, giving a reason for the assumption that contractors routinely obtained assignments for inventions made in federal contracts.
This arose from a faulty conclusion in the Report’s summary findings that glossed over complicated evidence raised in seventeen monographs provided as appendices to the Report. While the detailed monographs showed that private firms and most nonprofit research organizations were securing assignments, they also indicated that only some universities were doing so.
The general policy at many universities was not to claim ownership of inventions made by faculty inventors. Archie Palmer’s 1948 survey of university patent policies makes this clear. Universities did claim ownership when they hired people to invent, and they might claim ownership when the use of resources or other “equity of the circumstances” indicated ownership. But even where a university claimed ownership of an invention, it almost always pushed actual ownership to a patent management firm–most often Research Corporation–and most often then only when the faculty inventors choose to pursue patenting. The vast bulk of federal research contracting circa 1947 was not with universities. There would be no reason for the Biddle report to worry about university patent assignment practices. University research was not the problem for government contracting for patent rights. Put another way, the thing going on in the Biddle Report was not a faulty conclusion about university invention ownership practices but rather a disregard for university research altogether in shaping federal patent contracting practices.
Nonetheless, the Report concluded that contractors were securing title to employee inventions and thus federal policy need focus only on the allocation of rights as between government agencies and their contractors.
That’s perhaps a way of reading the Biddle Report, but here’s what the Report finds and concludes:
Where patentable inventions are made in the course of performing a Government-financed contract for research and development, the public interest requires that all rights to such inventions be assigned to the Government and not left to the private ownership of the contractor.
It’s this “not left to the private ownership of the contractor” that appears to bother O’Connor. But the Report’s focus is on that the government should obtain rights–so what might get left to the contractor or to the contractor’s inventors isn’t relevant. It’s not a faulty assumption regarding university invention assignment practices that is in play, but rather a finding that it is not in the public interest that contractors, including universities, have any patent interest in inventions made in the performance of federal contracts. The Report goes on to list the arguments supporting this position:
Public control will assure free and equal availability of the inventions to American industry and science; will eliminate any competitive advantage to the contractor chosen to perform the research work; will avoid undue concentration of economic power in the hands of a few large corporations; will tend to increase and diversity available research facilities within the United States to the advantage of the Government and of the national economy; and will thus strengthen our American system of free, competitive enterprise.
Well, there’s plenty there to bicker over. But it would appear that no one cares about these things any more. By the time of Bayh-Dole they are largely meaningless window dressing. The Biddle Report then recommends (VI.B.1):
As a basic policy, all contracts for research and development work financed with Federal funds should contain a stipulation providing that the Government shall be entitled to all rights to inventions produced in the performance of the contract.
Again, the Report doesn’t care where the ownership is–it cares where the ownership ends up. If all contracts stipulate federal ownership, then it is up to each contractor to secure for the federal government that ownership. It simply does not matter what pathway the contractor takes to comply with this stipulation, just as it doesn’t for licenses to the government. The contractor can take assignment of inventions and assign these inventions on to the federal government, or the contractor can require its inventors to assign their inventions to the federal government, or the contractor can join the inventors as parties to the contract, so they have a direct obligation to assign their inventions, without further involvement of the institutional contractor.
O’Connor continues, now without factual support:
This recommendation subsequently became a bedrock principle upon which all government patent policy was based up through the passage of the Bayh-Dole Act of 1980.
Sadly, this bit is simply wrong, especially with regard to universities. And wrong in multiple ways. First, the Public Health Service instituted the Institutional Patent Agreement program in the early 1950s. Under the IPA program, participating nonprofit organizations could own inventions subject to a license to the federal government. Under the 1963 Kennedy executive branch patent policy, contractors were to be permitted to own their inventions where they had the requisite capability and position. The Kennedy patent policy distinguishes between government work directed at commercial products, public health, compliance with regulations, where the government has the primary technology or is the primary user (space, atomic energy), and contracts to direct the work of others (where the government should expect to obtain rights), but:
In other situations, where the purpose of the contract is to build upon existing knowledge or technology to develop information, products, processes, or methods for use by the government, and the work called for by the contract is in a field of technology in which the contractor has acquired technical competence (demonstrated by factors such as know-how, experience, and patent position) directly related to an area in which the contractor has an established non-governmental commercial position, the contractor shall normally acquire the principal or exclusive rights throughout the world in and to any resulting inventions, subject to the government acquiring at least an irrevocable non-exclusive royalty free license throughout the world for governmental purposes.
This basic policy was reaffirmed by Nixon in 1971 and codified in federal procurement regulations in 1976. Here are the key provisions in the FPR long form patent rights clause:
(b) Allocation of principal rights. (1) Assignment to the Government. The Contractor agrees to assign to the Government the entire right, title, and interest throughout the world in and to each Subject Invention, except to the extent that rights are retained by the Contractor under paragraphs (b)(2) and (d) of this clause.
The Contractor must be prepared to assign rights in inventions to the federal government. How does the standard patent rights ensure that the Contractor can do so? Here, at (e)(3):
The Contractor shall obtain patent agreements to effectuate the provisions of this clause from all persons in his employ who perform any part of the work under this contract except nontechnical personnel, such as clerical employees and manual laborers.
It is pretty clear, then, that there is no misplaced assumptions about contractor patent agreements, at least by 1976. A contractor had to get a patent agreement that ensured that the contractor could assign inventions to the federal government. It also then makes sense of how a contractor could “retain” rights under the FPR patent clause–the contractor had to obtain ownership of inventions in preparation for “effectuating” the provisions of the patent rights clause, and if the federal government agreed that the contractor did not have to assign a given invention to the government, then, why, the contractor would “retain” that invention.
Bayh-Dole, five years later, and drafted by an attorney who also worked on the FPRs, by contrast, bungles all of this with its new definition of “Subject Invention” that includes contractor ownership (“of the contractor”), no requirement that contractors have patent agreements with their employees, a provision for employee-inventors to retain their ownership, and a limited requirement that contractors assign inventions that they own only when they choose not retain title or when they fail in various administrative aspects of the patent rights clause–disclosure, filing and prosecuting patent applications, maintaining and defending issued patents. Bayh-Dole certainly is a botch job, but the botch does not stem from the Biddle report and cannot possibly be the result of a mistaken assumption by Norman Latker, among others, that somehow contractors, especially university “contractors” receiving subvention grants and generally having policies that allowed faculty ownership of inventions, just always got assignment of inventions.
From 1963 to 1981, then, the bedrock principle on which government patent policy was based was that the federal government had the right to acquire title to inventions made with federal support unless for policy reasons (and later by regulation) it was in the public interest that a contractor “retained” ownership of a given invention made in work receiving federal support. Thus, federal policy indicated it was in the public interest that companies with market positions involved in federal procurement contracts should be allowed to own the inventions they acquired–unless special policy requirements also applied, such as that the federal agency sought to create a product ready for commercial release (a new fertilizer, say) or that the research involved matters of public health or safety or access to the results were needed to comply with federal regulations.
The problem as laid out in the Kennedy and then Nixon patent policies was not the bit about commercial contractors with market positions “shall normally acquire the principal or exclusive rights” but rather whether pharmaceutical companies met the criteria for ownership of inventions under the policies’ exclusion for research addressing matters of public health. The Public Health Service argued that they did not. Federal research in medicinal chemistry was clearly within the requirements for government ownership of inventions. The pharma companies boycotted NIH inventions as a result. That was where the battle was, where the Harbridge House identified the federal determination as the cause of the industry boycott. The NIH then used the Harbridge House report to revive the IPA program, this time with an express clause requiring nonprofits to take assignment of each “subject invention” that the nonprofit decided to manage under the IPA.
The requirement that nonprofit contractors assign their inventions was express in the IPA master agreement (IV(a)):
The Grantee shal obtain patent agreements from all persons who perform any part of the work under a grant or award from the Department of Health, Education, and Welfare, exclusive of clerical and manual labor personnel, requiring that such persons promptly report and assign all subject inventions to Grantee or its approved patent management organization.
The Grantee shall require assignment to it of all right, title and interest in and to each subject invention in which it elects to file any patent application for administration by it in accordance with and subject to the terms and conditions herein set forth. Assignments from the inventor to the Grantee under U.S. patent applications shall be promptly obtained and recorded by the Grantee in the United States Patent Office, and copies of the recorded assignment shall be furnished to the Grantor.
That is, the nonprofit was not required to obtain the assignment of every subject invention, but rather was required to obtain the promise of inventors that they will assign if the nonprofit or its patent management organization elects to file a patent application. (Here, subject invention means any patentable invention made with HEW support. Unlike Bayh-Dole, which defines subject invention to be “of the contractor” and so is very different. “Elect” in the IPA program means “choose to file a patent application” not “choose to retain title”–first a nonprofit has to decide to file a patent application; then it has an obligation to obtain assignment of the invention under that patent application; it is required to have required inventors to promise to make such assignment, under those conditions. There is nothing here about taking ownership of every invention as it is disclosed). Obviously, there is absolutely nothing in the IPA program that makes any sort of faulty assumption that nonprofits always get assignments of inventions.
The revived IPA program ran until 1978, when it was shut down during the attempt by the NIH to make it government-wide, despite executive branch policy providing otherwise. Thus, for nonprofits in the revived IPA program–over 70–the bedrock principle was that if they elected to file a patent application on a subject invention, then they had to get assignment, and for that purpose, they had to get a promise to assign if the nonprofit filed a patent application from every qualified inventor as a condition of their participation in the work.
The Bayh-Dole Act did not reverse the Biddle Report recommendations with regard to nonprofits in the IPA program, nor did it reverse the Biddle Report recommendations for contractors involved in commercial enterprise. And given that the same people who drafted the revived IPA program template also drafted Bayh-Dole, there is no way that they relied on some “faulty assumption” of the Biddle Report from 1947 to develop their new statutory language. The purpose of Bayh-Dole was to revive the IPA program a third time in order to reestablish a patent monopoly pipeline from the NIH to the pharmaceutical industry. This time the IPA program had to be statutory, beyond the reach of the executive branch, difficult for Congress to change, with such advantages that any threat to it would be met with widespread objection. Thus, the IPA program was expanded to include small companies (for no good practical reason), and was expanded to cover all federal agencies (to obscure the patent monopoly pipeline for the pharma industry).
It’s just that in drafting Bayh-Dole, Norman Latker botched the text. He pulled the definition of “subject invention” from the 1976 patent regulation model clauses for government contracting rather than from the IPA agreement–so a subject invention became a patentable invention “of the Contractor”–that is, an invention already owned by a contractor, rather thana patentable invention made with federal support under a funding agreement. The burden of Bayh-Dole is to determine when a contractor, having obtained ownership of an invention made with federal support, may keep that ownership if faced with a federal agency request for title. There is simply no faulty assumption here based on the 1947 Biddle Report. Latker knew nonprofit practices. Latker drafted the IPA template to require conditional assignment. For some reason he failed to do that when escalating the IPA program to become part of the federal patent law.
I am told there are boxes of files from Latker’s time at the NIH. Perhaps there is a record there of his drafting. I am also told that he took a great deal of effort to prevent government knowledge of his role in drafting Bayh-Dole. So perhaps not. But consider the dilemma Latker faced in moving the IPA program from regulations to statute. As regulations, the federal government could stipulate the form of its contracting provisions for acquisition of rights in inventions made under its contracts. But as patent statute, the government could not so easily announce that contractors owned inventions made with federal support. It would have to require contractors to require inventors to assign inventions. But even that disrupts the common law premise that inventors own their inventions and the federal government’s authority is to grant those inventors for limited times exclusive rights to their inventions. There’s no easy, unobjectionable way for the patent statute to be twisted around to vest ownership of inventions in nonprofit contractors, no even to force nonprofit contractors to take that ownership. Even the IPA program did not do that–the contractor was to acquire ownership after it had committed to file a patent application, and the assignment was directed at that application, at the claimed invention.
Latker, then, was left with just what O’Connor gets right–that Bayh-Dole deals with ownership of inventions as between federal agencies and nonprofit and small business contractors. But the Biddle Report has nothing whatsoever to do with it. The NIH IPA program subverted executive branch patent policy to enable the pharmaceutical industry to secure patent monopolies developed with public money. This was a most wonderful outcome for the pharma industry. When the venture capital industry discovered it could get ahead of pharma and do deals with universities and then sell whole companies to pharma later–quite apart from whether the originally licensed invention from some university was viable–the venture capital folks also came to love Bayh-Dole, since it created a market for speculative investment rather than simply an opportunity to commit to the development of an invention supported by the promise of monopoly pricing. One could trade on the prospect of future monopoly pricing without having to get more than a small percentage of the way there. The upshot was that much of the future value of a patent on a medicinal compound came early in the screening process, so speculation could fund early screening, just as public money could fund discovery.
Pharma could not be more pleased to be the apex predator in such an environment. It is no wonder that BIO sings the praises of Bayh-Dole, the gift of public funds that keeps on giving, that precludes competing approaches, that preserves patent monopolies for a couple of decades worth of exploitation.
But Latker could not find a place in the federal patent law to stick a requirement that inventors must give up ownership of their inventions to mere government contractors. Yes, such provisions could be stipulated as model clauses in government contracts, and Bayh-Dole does include just such required clauses. But the IPA program was rogue from executive branch patent policy, arguably non-compliant with federal regulations. Arguably not in the public interest. Certainly not so per the Biddle Report argument, nor the Kennedy patent policy. I don’t see how Latker could have included an assignment clause as part of the statutory language of Bayh-Dole. Thus, it appears that his strategy was to leave the assignment clause for the standard patent rights clause drafting. So Latker moved to a position in which he could draft that the patent rights clause. I’m told there was a huge fight among the federal agencies over this clause and various procedures. I expect Latker did not get what he wanted.
Instead, there’s an (f)(2) written agreement requirement that clearly leaves out any requirement to assign. Inventors are instead delegated to report inventions and execute papers to permit patent applications to be filed and to establish the government’s interest–not the contractor’s interest. And of course, it would be stupid to require inventors to assign inventions defined as subject inventions, since subject inventions must have already been assigned. Instead, the (f)(2) agreement makes inventors parties to the funding agreement and therefore also contractors, and therefore when inventors own their inventions, those inventions become subject inventions. It’s just that they are not “inventions of the institutional contractor.” Institutional ownership has to be established entirely outside of the Bayh-Dole apparatus, unlike the IPA program approach, which was simply rogue government contracting.
Thus, the folks around Bayh-Dole created the faux version of the law, in which the law mysteriously upended federal common law, and used “retain” to mean “take and keep” rather than “keep what was previously got.” All that got repeated so often that university folks were led to believe it was true, even while they dutifully still went out and obtained assignments the old fashioned way–but often telling inventors they had no choice, that federal law required them to assign (and all such stuff was untrue). Had Bayh-Dole been properly represented to the lower courts, there would have been no issue for the Supreme Court in Stanford v Roche. The courts would have seen that there was no obligation at Stanford for its roving post doc to assign inventions that Stanford had released by policy, contract, and by its own actions.