Here is a question: What should the federal government do with patents it issues to itself?
In the 1940s and 1950s, as the United States government contracted for research services associated with the development of weapons systems and the procurement of goods to meet its needs, a debate arose about whether the government should include delivery of patentable inventions as a contractual requirement or whether the government should instead require a non-exclusive license to patentable inventions for its own use, whatever that may be. It’s a limited dichotomy, of course, but as with most political things, things got polarized and formalized into this form. In science policy terminology, these were the “title” and “license” alternatives for government contracting. The government could take title to inventions, or take only a license.
Beyond the dichotomy were other alternatives, which also were used. The government for instance, could limit the scope of what could be privately patented–which it did for a time with atomic power inventions and space technologies. The government could not even require a license to inventions made with its support–which it did and still does for fellowship funding. Fund a project, take a license. Fund a person, take nothing. And the government could take title to an invention but grant a license back to the contractor, so the contractor could practice the invention outside work for the government.
For the most part, early federal funding for research took the form of procurement from commercial vendors and from contract research organizations. A commercial vendor was a company that sold product, and the government’s contract asked the company to build a version of the product for the government. A contract research organization’s product was the research it was contracted to do. In the case of a contract research organization, it was easy to manage invention deliverables, since contract research organizations make it a practice of delivering invention rights along with everything else in the research. If the federal government restricted its research contracting only to CROs, the executive branch patent policy would be easily uniform–require delivery of rights to patentable inventions along with everything else.
For commercial vendors, the discussion gets more complicated. An invention may be in the direction of a vendor’s business, or it may be entirely skew from that direction. A company making weapons electronics might invent a circuit that would be most useful in kitchen timers–and have no interest in manufacturing kitchen timers. Similarly, an invention may be specific to a federal agency’s goals or may be skew from those goals. A vendor’s invention might enable a product feature that the federal agency wants included in all similar products available from all vendors in the industry (for public safety purposes, say). Should a vendor keep title to patentable inventions that it has no interest in manufacturing and selling? Should a vendor keep title to patentable inventions that the government wants other vendors also to use, either in supplying work to the federal government or selling into a government-regulated private marketplace?
Put it another way. When the government chooses a vendor for research (as distinct from simply purchasing available product), the government is also creating an opportunity for that vendor to establish a patent monopoly position in the area of work specified by the government contract. Choosing a vendor becomes, then, choosing who might get to dominate an area of technology development–and because the government specified and paid for the work. It is one thing to select a vendor based on technical capability and price. It’s another thing to include in that selection that the vendor has government help to establish monopoly positions and so prevent other vendors from even competing for federal contracts in the same area of work or from supplying what the government specified in the private marketplace for goods and services or from practicing what the government, if it made inventions a deliverable, could allow anyone to use. Letting such things happen violates a basic goal of procurement contracting–never let your organization become dependent on a sole supplier as a result of your contracting. This holds for the federal government as well. The monopoly supplier has the potential to become expensive, lazy, risk-averse, entitled, controlling–or could fold up shop and leave you with no alternatives.
The Kennedy patent policy in 1963 navigated this problem by distinguishing vendors with a commercial position and capability to work an invention from other vendors. A vendor with position and capability could retain ownership of an otherwise deliverable invention; other vendors should give up any claim to ownership and treat patentable inventions as contract deliverables. In special cases, a vendor might make a case to a federal agency to keep rights, but the burden was on the vendor to make the case. Federal agencies were authorized to do what was in the public’s interest–whatever that might be.
Within this geography of contracting, universities show up first as contract research organizations. They have no commercial position or capability; they have no business operations for which patents would be meaningful. Their only business–should they have one–would be to trade on patent rights. So it is easy to see that university contract research is like any other contract research, except done without an apparent profit motive. Whatever university administrators might want in their hearts, and however uselessly idealistic they might think the university’s charter, it is entirely fair that within the formalisms of federal contracting, universities have no profit motive, no need to make more money from federal research contracts than their costs and the benefits arising from hosting talented researchers. If a university wants to deal in a patent on an invention made in a federal research contract, then it ought to have to make a case for that dealing, relative to what a federal agency might determine to be in the public’s interest. The university has to argue–“Allowing the university to deal in a patent on this invention is, for you dearest federal agency, the best way for you to meet your obligation to best meet the public’s interest in the invention.” If the federal agency is not persuaded, then the invention remains a research deliverable.
You might see in this arrangement the roots of the argument that sometimes it may well be in the federal agency’s interest in serving the public interest to allow a contract research organization to deal in patents on federally sponsored inventions. Sometimes, if a federal agency remains unpersuaded by a contract research organization’s arguments, it may make the wrong decision, or at least a less good one. Other times, of course, a federal agency might make the right decision, even if through indifference or incompetence or accident or political shenanigans. Bayh-Dole cuts through all this with the claim that it is always in the public interest for contractors to own inventions made with federal support, regardless of the contractors’ purposes, capabilities, policies, and practices; and regardless of the federal agencies’ purposes, capabilities, and policies.
Subvention and Independence
There is a further part of the cascade that matters–and that has to do with research that is not based on procurement but rather on subvention. That is the special form of research that Vannevar Bush proposed in Science the Endless Frontier, that it was an appropriate role for government to fund research for which the government had no programmatic purpose. Such research funding was to be a grant-in-aid, like a fellowship, and not a procurement, not justified by an established government need. The government need, if there was one, was that there might be more independent investigation into nature, to add to the available tools that others might draw on to devise new things, new products, new industries.
Bush’s subvention thinking is that federal funding might reach to university faculty outside any contract research activities, so that the nation gains the benefit of faculty cats exploring their neighborhoods with independence. At best, argued Bush, the government should get a license to use whatever the faculty cats might find–but that whatever is found lies with the faculty inventors, not with any institutional contractor. Here, Bayh-Dole destroys the idea of free inquiry and establishes that university subvention research be another form of procurement, as if it were, for patent purposes, procurement even if for programmatic purposes the government has nothing to procure and the universities have nothing to vend. Strangeness.
Inventions as Government Contract Deliverables
With this context, let’s look at the problem for the inventions that the federal government requires as contract deliverables. What should the federal government do with these inventions? Again, keep in mind that the inventions in this class of deliverables may represent very different things in terms both of vendor capabilities and of federal agency purposes. The only common feature is that these inventions should be contract deliverables and not stay with vendors. There’s a similar thing that happens when the federal government requires assignment of inventions from federal and government laboratory inventors–the government then has rights in these inventions.
What should the federal government do when it has rights to patentable inventions? The government could
- return the rights to the inventors
- dedicate the inventions to the public (that is, publish and choose not to patent)
- suppress publication and hold as a government secret
- patent and license non-exclusively
- patent and license exclusively
- patent and assign the invention to a non-federal entity
- patent and sit on the patent, doing nothing
The first option–returning rights to inventors–undoes the policy of requiring inventions as a deliverable. The next three options suggest disposition of an invention without a profit motive. If an invention is published, then the purpose of the patent system has already been met without any need for a government offer of exclusivity for a limited time as an inducement to publish through the patent literature. If the government chooses not to publish the invention–for defense classification reasons, for instance–then the whole publication in exchange for exclusive rights has no interest. The government has decided to restrict practice of the invention altogether.
A patent may be licensed non-exclusively and royalty-free for public purposes, such as to control quality, safety, and truth in advertising. Such licensing may run in parallel with federal regulations requiring use of technology that includes the patented elements. The federal government does not need a financial interest in such dealings–and the federal patent system was created with an eye for preventing the government dealing in patent monopolies as a source of revenue. A patent position also provides the government with a means to control imports of the invention, thus allowing it to protect domestic industry from what may be unfair trade practices by foreign companies and governments–dumping products in the U.S. market, for instance, to damage domestic industry and steal American jobs.
Finally, holding a patent gives the government a seat at the table in dealing with what may appear to be abuses of the patent system by companies and speculators. In this, we might find that the patent system is not perfect and is open to exploitation that might serve private interests that do not end up, Adam Smith style, in providing the best of all possible economic outcomes. Self-interest can sum to uselessness, gridlock, and suppression of new technology–even if using the patent system to do so. Thus, the government might use a patent position to organize an industry standard, or to exclude companies from practice who assert patent positions to block further research or to block cumulative development of an area of technology or who attempt to block competitors from bidding for federal contracts. This is the world of cross-licensing to mitigate the effects of solitary patent positions for technologies that often require scores if not hundreds of patent positions to be readily available for use. Such technologies simply cannot develop if hundreds of patent owners all hold out against the community in an attempt to maximize their own financial interest.
Thus, for a number of reasons royalty-free non-exclusive federal government licensing might make good sense. We might note, too, that this analysis puts a huge hole in the university commonplace that a “non-exclusive license is just a tax.” That’s nonsense. Non-exclusive licenses do not have to involve a fee.
We are left with three alternatives to consider. The last is to do nothing with a patent. Just accumulate patents as signs of technological advance and use their presence in the patent literature to raise the bar for further inventive work in the same area. The federal patent, even though the underlying invention is not worked by the patent owner and not formally licensed to anyone, has an effect–the new technology is published in a common location, and the standard for inventiveness is raised. And of course, if the federal government does nothing, then it also does not enforce the patent and anyone can make, use, and sell the invention without formalities of a license. In a sense, sitting on a patent is a defacto form of non-exclusive licensing without the bother of formal contracts, without any added controls on the making, using, or selling that might involve the patented invention.
We are then down to exclusive licensing of an invention and assigning an invention. These are related practices. An exclusive license can also be an assignment, if all of the substantive rights in the invention are licensed exclusively, so that the licensee has the right to enforce the patent. For an exclusive license not to be an assignment, at least one licensed substantive right–make, or use, or sell–must not be licensed exclusively. Thus, an exclusive license might involve making and selling, but use is non-exclusive. Or making and using, but selling is non-exclusive, or selling is exclusive, but not making or using, or making is exclusive, but using or selling. You get the idea. In an exclusive license, the patent owner breaks up the substantial rights of the invention, and thus, also of the patent monopoly. The patent owner must deal with any infringement–no exclusive licensee may do so. The patent owner must coordinate the various players who have licenses–there may be non-exclusive licenses to make but only a single license to sell; there may be many non-exclusive licenses to make and use, but only one license to sell; there may be only one license to make, but many licenses to use (as for methods), and many licenses to sell product based on the invention. This sort of complication involves a keen awareness of the shape of a given area of public activity, industry activity, and commercial markets for the sale of product.
Government Dealing in Monopolies
When the federal government issues a patent to itself, it then gives itself the right to deal in monopolies, despite the patent system’s effort to exclude the federal government from such dealings. The government may deal in patents to make money for itself, to play favorites with people and companies, and to impose more requirements on patents than those that the patent system imposes (when combined with antitrust law).
It is here that we face the tough questions, as any of these three conditions appear to allow the federal government to circumvent the patent system itself. The federal government can just spread its research funds around, attracting many of the inventive people from the arenas in which they would invent–out in industry where they would have particular problems to solve, in garages where they would tinker and gadgeteer–and put them to work on whatever problems they think sound the best to government committees. Then the government can acquire all those inventions, and trade in patents on those inventions. The patents do not issue to inventors. The government has an interest in monopolies, has an interest in making money from patent positions, and has an interest in re-issuing these patents to favorites. The government also may impose by federal contract whatever additional requirements it wishes, regardless of the requirements established by federal patent law. Thus, the government may impose a working requirement–that a licensee or assignee must work the invention to retain the license–even though the patent system has no working requirement. Or that a license or assignee must pay the government a share of the upside created by holding a monopoly position in market–even though the patent system provides for no such payments to the government.
In using a research contracting premise to acquire rights to patentable inventions, the federal government in essence forestalls the market in patentable inventions. When the government holds patents for assignment to another, it begins to trade in monopolies that otherwise the patent system would preclude. The government issues patents to itself, not to inventors based on the inventive merit of what they have done and on their decision to pursue a patent. The government makes inventors pursue a patent, and the government obtains the patents on these inventions, based on its contract position not on anything of technical merit that the government has done. If the government then assigns an invention to a non-federal entity, it is in effect circumventing the patent system while exploiting that system. The patent re-issues to whomever–not to the inventors, not to the federal government–from a federal agency operating its own ad hoc patent office, adding working requirements, financial interest, and open to playing favorites with monopolies.
It is just a further variation on the theme that a federal agency allows a contractor without any commercial position or capability to trade in patents on inventions. In this case, rather than having a patent issue to itself, the federal government has a patent issue to a contractor, not because the contractor has a private deal with inventors that the government allows to stand despite a contract that makes inventions contract deliverables but because the government allows the contractor to stand in place of the federal government. The patent re-issues to the contractor on behalf of the government. The government cedes the money interest and favoritism and working requirements to the contractor, to act on the government’s behalf. The same result would come about if the government had a patent issue to itself and then assigned the invention to the contractor for management.