In 1983 Senator Bob Dole wrote a letter to Senator Charles Mathias, Jr. regarding a bill Sen. Dole intended to introduce to extend the “Dole/Bayh bill” (as Sen. Dole called it) to large businesses. I rather like the construction Dole/Bayh. Perhaps we should use Dole/Bayh when we mean the statutory version of the law and Bayh-Dole Act when we mean the faux version of the law, the version concocted by university administrators, along with Bremer, Latker, Allen, and Bayh.
It’s the faux version that gets promoted by AUTM, COGR, and others, and it’s the faux version that has killed the research commons, stifled collaboration, and stripped university inventors of access to their inventions (which they would have, at least, if the federal government obtained title and dedicated the inventions to the public or licensed non-exclusively and royalty free; and which they would have if university patent administrators did not deliberately overclaim which inventions came within the scope of, um, Dole/Bayh). The statutory version of the law–Dole/Bayh–does not mandate university ownership, does not require commercialization, does prohibit assignments masked as exclusive licenses and does forbid university administrators siphoning off licensing income to pay for non-subject invention management. But Bayh-Dole, in its faux glory, rejects these things and makes up its own version of the law. That’s the version the Eagle Forum has just finished presenting to legislative staffers on Capitol Hill. No one asks for briefings on Dole/Bayh these days. Post-truth, fake news and all.
Anyway, here’s Sen. Dole discussing the licensing problems before Dole/Bayh:
With the government now funding approximately seventy percent of the basic research done in the United States, we can no longer tolerate the abysmally low rate of commercialization that accompanies federal ownership of new inventions.
What Sen. Dole does not point out (and perhaps doesn’t consider) is that basic research is not intended to produce commercializable properties. Vannevar Bush was clear on this point, that basic research is research conducted to advance scientific knowledge, not to be done under pressure to produce commercial products. Some definitions of “basic” or “fundamental” research depend on the idea that there won’t be patentable inventions. Here’s one:
Fundamental research; it often produces a wide range of applications, but the output of basic research itself usually is not of direct commercial value. The output is knowledge, rather than a product; it typically cannot be patented.
So the whole idea that the results of federally supported basic research should produce commercial products is just strange. The outcome of basic research should be the advancement of scientific knowledge. The big problem is what sort of advances we should value. It’s easy enough to study Windows 95 for decades for what we can learn, and when we surface, our imaginations will not be a whole lot better off. Sure, the universe is mirrored in every butterfly wing, but it’s not so easy to get around from that starting point.
Vannever Bush, in imagining science as an endless territory to be explored and mapped in the z direction, argued that basic research needed to be conducted by “free play of free intellects” and not encumbered with institutional demands for immediate application–not by government, not by industry, not by established professional interests (military, medicine), not by university administrators. Of course, now, the trendy thing in science policy is to argue that Bush was a liar or just wrong and the best people to direct science are generals, CEOs, and activists–especially if they listen to science policy wonks. Imagine the policy mice struggling to get at that block of cheese.
A low rate of “commercialization” implies “commodification” and so is even more removed from practical application of inventions made in basic research. Before one gets to commodification, there’s still use of inventions in research, use by technologists and professionals who have the capability and resources to use, including use by industry. None of these uses require anything to be sold, nor any commodity version of an invention to be designed and produced.
Sen. Dole’s concern in introducing a new bill is to extend Dole/Bayh to large businesses as well. This was the goal all along, apparently. But for many federal contractors–especially those doing contract research for the federal government–it didn’t matter, in part because they routinely did not want the rights, and in part because a number of federal agencies permitted them to have the rights they did want. What was left were the inventions that the big contractors didn’t want, generally. For those orphaned inventions, the companies involved were too smart or too focused on federal research to want to spend time commodifying inventions into products for markets they didn’t know or didn’t care about or did know and knew the inventions made for the government had little value there. So low commercialization rates for orphan inventions–how will those rates change just because law makes arbitrary the right of large companies to own stuff they don’t care about? And if the large companies obtain patents on that stuff anyway, given the overhead of Dole/Bayh, does anyone really imagine the commercialization rate will change at all? No.
And consider the IPA program, at least the version from 1968 to 1978. Under an Institutional Patent Agreement, a university got all the baggage of Dole/Bayh but the commercialization rate appears to have been, um, about 4%–pretty much the federal rate. The issue was not one of whether the federal government owned the patents arising in its research or whether a university middleman did–it was an issue having to do with the nature of the research the federal government supported at universities; or having to do with the role of patents in having anything to do with encouraging use; or having to do with the nature of the use that happens, perhaps not requiring so much commodification, not of interest to a consumer market.
Consider even further. If I am a university-based scientist and I have an idea for basic research and I know that the government and university bureaucrats are scheming to take my invention rights and stuff them into a machine to try to commodify inventions for profit, and I don’t want that to happen, I don’t include in my research proposal my inventive thoughts. I save those out for consulting, or for after I leave the university for a startup. Only the foolish, the desperate, or the indifferent would push their best inventive ideas into federally supported research to be appropriated by patent brokers. And to be indifferent prior to Dole/Bayh meant that if the federal government obtained a patent, the rights would be dedicated in some way to the public 95% of the time or better, including made available to the inventor and all other scientists at the university and elsewhere. Before Dole/Bayh, indifference was okay. After Dole/Bayh, especially in the age of Bayh-Dole, indifference means that the inventors think the inventions are going to be pretty useless things.
Think about it–Dole/Bayh sends you inventors foolish, desperate, indifferent, or worse who naively think that it is your job and commitment to make them rich merely because they produce some patentable inkling. Are these the inventors you are looking for? Hardly. Is this the profile for pushing up the commercialization rate on their bits of inventing? Hardly. Dole/Bayh, sadly, becomes another perverse incentive to aggregate in institutions the control of a bunch of patent dreck under the premise that bureaucratic control is just the thing to induce commercial investment in commodification of the dreck, the growth of markets for the dreck, and folks getting wealthy and economies growing fat on the dreck.
One more thing, just to show the extent to which politics flattens all things into mental roadkill. In basic research, there are potentially four outcomes that might involve invention. Let’s work through them.
There’s discovery–the thing that advances scientific knowledge. This is the pleasure of finding things out, the thing that scientists don’t need bureaucratic incentives to seek after, the happiness of priority and recognition, the thing that secures a solid place in the shiftiness of academic life. The problem is, discovery is often serendipitous, not the thing one is looking for or trying to do. Discovery may involve accident, mistake, sloppiness, idle curiosity, misdirection, inattention, strange mental juxtaposition. In fact, one might think of basic research as a stage on which to make huge mistakes. The question is less what one claims to attempt as it is the nature of the stage on which one might fail in making the attempt. So discovery may well be outside the scope of what is being studied. One might seek to understand sand flea sexual identity and come away with a blunderingly new idea about thin films on silicon dioxide.
There’s application. One might be studying those sand fleas and realize that a sand flea in heat (I’m making things up here) is a great biomarker for chemicals released by decaying kelp, which might have potential in sensors to detect stale sushi. Those detectors are far outside the scope of research, and may be as well far outside the expertise and capability of the “inventor” of the application to produce–and outside the capabilities of many other people as well–how to figure out how sand fleas are attracted by rotting kelp, how to biomimic that sensing mechanism, how to incorporate in a sensor, how to integrate the sensor into a detection system that works in a sushi kitchen, how to persuade anyone in that kitchen that they actually need this sensor. You can see all the potential for patenting–but it’s patenting a mostly fictional sense of the future. Neat. What sort of commercialization rate would you expect for this sort of stuff? It’s the age of feuilleton, as Hesse imagined it in Magister Ludi, adapted to basic research. Patentable? Perhaps. Commercialization rate? Going to round to 0% Number of state-based “seed funds” willing to fund this sort of stuff anyway? Oh, close to 90%.
There are tools. Now we get to the most likely outcome of basic research–that there will be new lab equipment, instrumentation, methods, algorithms, recipes, and data sets (themselves a kind of viewing device) by which natural phenomena are isolated, observed, predicted, modeled, and explained. Whatever else, we can expect tools. When Lee Hood started thinking about how to automate the tasks of gene sequencing, he was thinking tools. Tools have their most immediate application in the lab that develops them. They are tools when they get used, and before that, they are imagined as possible tools. Might get made, and might not, might get made and not used. Research tools are never “early stage” or “high risk”–if they aren’t being used, they aren’t tools, and if they are being used, then they are already workable.
Tools may have a number of outcomes. They can be used to evaluate claimed findings (research “of”); they can be used to learn more about the discovery (research “on”); they can be used in research not directly related to the research in which they were developed (research “with”); and they can be used to study and extend the tools themselves, which is not so much research as “development” of a sort. All of these uses are research uses. This is how science advances, that not only is there a finding about moons orbiting Jupiter, but there was a telescope adapted for the purpose of observing these moons. The object of the research is to discover new worlds and thinking about the implication of something not orbiting the earth, as it were. But the outcome is also an improvement in the use of the telescope–a new objective lens, perhaps.
All sorts of research-developed tools might also be inventive, and so patentable. But what is the role of a patent in such cases? To prevent other scientists from using the tool without payment? To limit the ability of rogue scientists from improving the tool ahead of the originating lab’s own efforts (better funded folk, or with too much time on their hands, or interested only in the tool, not in the discoveries it enables, or with way more tool-making capability)? Or at least forcing those rogues and their patent broker buddies to dedicate their improvements back to the cause or face the full wrath of patent law? Certainly the uses for patents on research tools are not those for ordinary patents. Perhaps to enforce a commons; perhaps to encourage a commodification of the tool that would run in competition to local implementation.
Tool creation may be an invention deliverable in basic research if the statement of work anticipates the need to create the tool to achieve the goals of the proposed study. And even if not expressly identified, tool creation might be done “in the course of” getting to the results anticipated. In such cases, in federally funded work, an inventive tool would be a subject invention. But a tool might be developed on the side, or imported and adapted for use in the research, and so not within the scope of the funding agreement and not a subject invention. Either way, the commercialization rate for such tools is an entirely secondary concern. The primary concern is whether the tool has value in further research. Scientists are weavers of the imagination and their tools are the looms that weave theory and evidence into fabric. If a tool can improve the loom, then the measure is the nature of the fabric produced. Whether a commodity product ever gets produced is beside the point.
For Vannevar Bush, there was no linear model of basic research to applied research to development to commodification. There was exploration and mapping and producing an infrastructure of communication so new territory could be known and navigated. That territory then might be used in some initiative made possible or advantaged by access to that territory. My blog depends on WordPress depends on the web depends on SGML depends on the internet depends on TCP/IP depends on open standards depends on DARPA messing around with folks at UC depends on all sorts of other things depends on digital computers depends on Fine Hall at Princeton depends on John von Neuman depends on folks like Vannevar Bush giving Johnny room and resources to play with. We can work the account backwards to various sources, somewhere usually ending with the free play of free intellects. But we can’t work it forwards as institutions demanding basic to applied and the like.
If there’s a great fabric spun from new looms, then the Vannevar Conjector (I’m coining that expression here) is that we will discover new things to do that lie outside the definition of progress set out by the elites of any established order, be it the military, medicine, government, or university administrators, or scientific consensus. Trying to “commercialize” basic research inventions, other than to produce more efficient research tools, is a fool’s game only bureaucrats could embrace. Patents on basic research discoveries, applications, and tools get in the way of science and technological adoption. Trying to by-pass or exclude science and technological implementations in favor of monopolies on commodification efforts is the work for gougers, exploiters, speculators, and folks who don’t actually care about any further advance in science, don’t care how the loom works or the fabric is woven. They would steal the golden goose as a gosling, and to do that, they would kidnap all goslings, to make sure any hoped-for future golden one never gets away.
And there’s realization. Beyond discovery, applications, and tools, a researcher can come to a realization–that the line of inquiry is unproductive, or that a similar line of inquiry could be really interesting focused on a different topic or objective, that a tool developed here isn’t the one to develop elsewhere, but now there’s a good idea of what a different tool might look like. Realizations are not “conceptions” of invention–they are a developed sense of what to do next that might be worth one’s mortal existence or career or might relieve the dull boredom of being a scientist dumping out the same thing again and again with minor variations. Realizations might start new areas of inquiry–digging mine shafts where no one has dug before, the root of the concept of “innovation” back to the ancient Greeks. Realizations might start new companies, might challenge existing scientific claims and consensus. Realizations produce actions that feel new, whether to stop doing something that’s unrewarding or start doing something different that one hopes will be rewarding. Realization also carries with it capability–an awareness of how tools work, how observation works, how discovery comes about (by getting balled up in difficult situations that produce really great accidents and still maintaining one’s prepared mind).
After spending a million dollars on federally supported basic research, what do you do next? Perhaps trying to get another million in federal grants is the least interesting thing, a sign that no realization took place in the previous million.
That’s the long way around to the idea that the commercialization of inventions made in federally supported research is one of the least interesting things to do. Perhaps there’s a realization in that for you.
Back to Sen. Dole’s letter. Having complained about low commercialization rates, Sen. Dole gives us the oft-repeated commercialization rate figures:
For example: compared to a licensing rate of 33% for university developed inventions, the government has licensed less than 4% of inventions owned by it to the private sector for commercial use.
I have yet to find the data that back the claims for these rates. Everyone cites them, and cites others who have cited them, but no one trots out the data. But let’s work with them anyway. Sen. Dole shifts from commercialization rates to licensing rates. Perhaps he means the same thing. Likely he does, but the data are the problem. But here’s the thing. Under the IPA program from 1968 to 1978, about 4% of the subject inventions universities claimed became commercial products. That is, the IPA rate was pretty much the same rate as the federal rate that is so roundly criticized. The 33% rate (or 25% rate by other accounts) is the rate for inventions that were a) not federally supported and b) largely voluntarily placed with university-affiliated patent brokers. The 33% rate depends on faculty inventors choosing to pursue patenting, deciding on the broker to work with, and setting out the conditions for their collaboration with the broker. The broker, in turn, is free to decline the work, to negotiate the terms of agency that will make the relationship work, and to return any invention that proves ill suited to the effort.
But Bayh-Dole (the faux thing) made university ownership mandatory, forced dreck in which golden goslings, and let university administrators decide on the goals (exclusive licensing, making money “for research” <and themselves and their programs>, holding all things, letting go of nothing). That 33% comes about because voluntary decisions by inventors lead to high-quality opportunities for brokers lead to industry interest in adopting. Dole/Bayh forces things where it should have enabled, cozies with patent brokers when it should have been supporting inventors and science and research and entrepreneurs, fussing about how to franchise off loom parts in two-decade monopolies for cash from speculators rather than considering how scientists might build new or more powerful looms.
But Sen. Dole attributes the problem to government bureaucracy:
This is primarily because of chaotic end inefficient agency patent procurement policies that strangle innovation with red tape.
There was indeed red tape. The Harbridge House report in 1968 made that much clear–but the red tape was only in some federal agencies, for a handful of inventions. In others, such as HEW, the issue was not inefficiency or red tape–it was a bona fide disagreement between the director of HEW and the pharmaceutical industry over whether the federal government was funding medicinal chemistry research as a subsidy to corporate interests in creating monopolies that support high drug prices and huge profits. HEW argued for public availability and limitations on any exclusive license (as the IPA program at HEW stipulated, following the Kennedy patent policy).
The pharmas boycotted the compounds discovered with federal funding. A boycott is not red tape. It’s an industry refusal to accept a commitment to develop new drugs at less than a monopoly profit. Whatever you think about the importance of monopolies for the special case of seeking profit in alleviating human suffering, perhaps you can still accept that Sen. Dole didn’t get much right about the cause of the low commercialization rate. Even for the IPA program, which in Dole/Bayh style allowed (heck, required) universities to take ownership of any invention (but only after the university decided to file a patent application), the commercialization rate was no better than the federal government’s rate, and the universities claimed they had a profit motive, where the federal government had no such profit motive.
In 1964, a university organization drafted a commentary on the Kennedy patent policy, asserting that the basis for patenting is to seek money, even though the federal basis for acquiring patents is clearly not monetary:
The patent system of the United States reflects the basic premise that financial incentives are necessary and desirable to secure the development and commercial production of discoveries and inventions, so that where the educational institution believes this necessary, reasonable incentives should be allowed to the inventors, the institution and the producer to encourage the initiation of patent actions, defray expenses and stimulate the successful development and introduction into public use of discoveries and inventions made with the assistance of federal funds.[1]
The university patent brokers, of course, wrote themselves into the position of leadership, in place of the faculty inventors. “Where the educational institution believes”–as if abstractions have minds. No, “When patent brokers authorized by the university to do its believing for it believe…,” then patent brokers should have “reasonable incentives” to undertake the patenting work, and then of course all sorts of good things will happen. Bayh-Dole (I use it in this form), the dream of university patent brokers.
We could go off on “title certainty” as well–but there, too, the problem was a failure of universities to state a clear case for the public interest in obtaining title to inventions made with funding from federal agencies that otherwise expected to own and dedicate such inventions to the public. The universities (and here we mean, the patent brokers) simply failed to make compelling cases. They might argue that they could make money when the federal government wouldn’t, but making money is not a reason the federal government takes title to the patents that it issues to itself, just as making money for itself is not a basis for the federal government issuing patents to any inventor (renouncing a long history in which patent royalties were paid to the monarch granting the patent, hence the term). Call the failure of the patent brokers “red tape” and “inefficiency”–sure. But the problem appears to start and worsen on the side of the patent brokers, who could not show that their management of patent monopolies advanced the public interest any better than did the government. The primary difference in allowing university patent brokers to take private federally supported inventions was that they believed they could make enough money to keep themselves in the business of brokering patents. Even in testimony running up to Dole/Bayh, advocates for the law admitted that their track record was not very good–but over time, they claimed, at least there would be a whole lot more of them! And in that, they were visionaries.
What we don’t have, as a result, is a continuing discussion on how patents might advance scientific knowledge, how patents might enable more effective and useful research, and when and how institutions might assist in the management of patents arising from basic research. Instead of responsive federal policy, we have arbitrary policy. Instead of distinguishing procurement and subvention, we have conflation. Instead of valuing the free play of free intellects in the context of institutional support, we let the institutions play the role of predators. Instead of imagining ways in which technology changes, we stipulate that the proper, only, first, primary way for that change to come about must involve speculation on the future value of patents. Narrow, inefficient, inflexible, foolish statutory dreck. And some folks want the new administration to “protect” Bayh-Dole.
Sen. Dole didn’t get his bill, but there were strange amendments to Dole/Bayh passed the next year. And besides, in 1983 President Reagan had already used an executive order to extend Dole/Bayh to large company federal contractors.