In my last post, I ranted on about the problems surrounding contributions by academics to molecular modeling, and cited as a particular issue the Bayh-Dole Act (BD), which enabled and encouraged academic institutions to patent the intellectual property arising from government funded research. Lo and behold, the December 9 issue of Nature features a couple of articles on BD, which of course prompts a further rant on the subject. The first article, by Bhaven Sampat of Columbia University (“Lessons from Bayh-Dole”, pg755-756), was wildly positive on the consequences of the act (no surprise; Columbia has benefited disproportionally from BD). The second, by Jeffry L. Furman, Fiona Murray and Scott Stern (“More for the research dollar”, pg757-758), is more circumspect, at least citing some of the adverse consequences of restricting the sharing of information.
Let me be less circumspect: as far as science is concerned, there has NEVER been a worse Act of Congress in the history of the United States.
Let’s look at the raw numbers. The total income from patents owned by universities is now claimed to be in the vicinity of $2 billion a year. A kingly sum, you might think, until you consider that the operating budget of Stanford this year is about $3.8 billion, that of Harvard a little more than three billion. So Bayh-Dole, it would seem, effectively enables the funding of half the operating cost of a major university in the U.S. Except that isn’t really accurate. Typically one third of that two billion goes to the inventors—the scientists who are supposed to be in this for the love of the pursuit of knowledge. (To be fair, some of that money gets fed back into their research.) In addition, all universities active in applying BD now have Offices of Science and Technology; they even have their own society (the Association of University Technology Managers [www.autm.net]). Those offices, even if only minimally staffed, cost money. Then there is the cost of actually getting a patent. This can be relatively inexpensive, but if there is any contention regarding the invention, the cost can skyrocket. For example, for Emory University’s patenting of Emtriva came at the cost of a multi-million dollar lawsuit that, at its peak, was the most heavily litigated (i.e., costly) case in the nation.
Do I know how exactly much these direct costs subtract from the $2 billion in patent income? I don’t, but let’s make some educated guesses. First, subtract out a fraction of the one third of $2 billion that goes to inventors. Let’s be generous and assume 50% goes back into research. That leaves about $1.7 billion. Next, consider the tech transfer offices of which there were already more than to 250 over ten years ago. Let’s assume the budget for the average office, including a staff of three professionals, a support staff, and the usual university overhead, is around $800,000, i.e. for a total of $200M. Second, patent costs: In 2004 there were 11,089 academic patents filed, a number that has increased year upon year since BD and is now closer to 20,000 a year. Each probably cost a few thousand each to file, say $10k on average, leading to a further cost of $200M. Last year about three thousand academic patents were granted. It costs money to see these things to fruition so let’s make the cost of this $100k per granted patent, $300M in total. Then there is defending, litigating and licensing- harder to guess at these costs but another $100 million doesn’t seem unreasonable (litigating a single patent can be > $1million, then consider the legal costs for contentious licensing).
So the direct cost, in all, to get that $2 billion a year in patent income is: $200M + $200M + $300M + $100M = $800 million, possibly a conservative estimate. Subtracted from $1.7B we are still $900 million in the black—although some estimates have put things closer to breakeven. Keep in mind, as well, that these numbers don’t even take into account the patents that would have been filed anyway, even if BD had never been foisted on the academic world. For example, much is made of the Cohen-Boyer transformation patent from Stanford and the University of California, a basic patent for the biotech industry; this patent was filed in 1980—it was going ahead with or without Bayh-Dole. Similarly, the Axel patent at Columbia a couple of years later was not influenced by BD. Columbia, Stanford, the University of California already had organizations capable and active in patenting. Still, let’s be generous and assume that patenting rates are much higher now, say five times higher than had BD not been passed. Revenue would still have been at least $400 million (probably more since only high quality patents would have been considered) and this sum would have come without a lot of the associated cost of current practices (patent everything, make deals on everything you can, etc.), so I’m going to assume the net revenue without BD might have been closer to $300 million.
So my back of the envelope calculation is that the net effect, monetarily, of BD is about $600 million dollars. Not an inconsiderable sum: A bit more than the annual operating budget of the University of Idaho ($442M). About 1.5% of the combined budgets of the NIH, NSF and DOE. The question is: what has this actually bought us? I would claim it has (1) disgraced and dishonored the principles and traditions of scientific endeavor, (2) held back progress by preventing the free transfer of information between academics and (3) held back commercial realization of ideas because of raised barriers between academia and industry.
Disgrace and dishonor! Academia was not founded on the concept of personal enrichment, but for the pursuit of knowledge and understanding of the natural world. Universities have, throughout the ages, been regarded as sanctuaries from the worlds of commerce and politics, and those who work there of having a higher calling not unlike that of religious orders. These are the standards passed down to us from the Greeks, from the great Arab institutions of the later part of the first millennium, and from the great European universities of the Middle Ages. You don’t find many examples of academics in any of those times getting rich; on the contrary, there was then a tradition of the rich supporting science and knowledge. Consider the state of affairs today, where more than $200 million is shared by three academics at Emory, more than $140 million goes to Robert Holton at Florida State, not to mention the fortunes of Cohen, Boyer, Axel and others. Just as the professions of law, medicine and banking have become besmirched and belittled because of obvious conflicts of interest between the society they serve and outrageous personal compensation, how long can we expect academia and, more importantly, science to remain untouched? Even if none of these scientists pursued their research to get rich (and I happen to know this is not the case for at least one of my examples), this is corruption. Corruption of an ideal that has served mankind well since the emergence of Greek culture 2,500 years ago. Knowing you can make money from your research cannot but affect, even subliminally, the choice of research. And we all know the excuses we make: “it’s money I can put back into research;” “but mine are ideas that ought to be commercialized for the betterment of mankind,” ad nauseum. Yeah, maybe. Surely by now we as a community have figured out that most spectacular advances do not come from expected directions, from directed research? They come from the discovering the unexpected, from the blue sky, from just doing something because it’s cool.
My second point is one that at least is debated: to what extent is the need for non-disclosure hurting the progress of science? It’s clear that the answer is “somewhat,” but we really don’t know how much. (This is the question the Furman, Murray and Stern article tries to address.) But let me posit that if we can point to the traditions of the ancient Greeks as a rebuke to today’s practices, then we can go even further back to critique the consequences of reducing information flow. How did modern culture evolve at all? Why are we the dominant species on earth and not just a pack of wandering savages? We are anatomically identical to our distant ancestors, the Homo Sapiens; it is only our culture that makes us any different, and that culture arose very sporadically over the last hundred thousand years. Why and how did that culture arise? One interesting theory is that the persistence of culture only came about when the population density was sufficient to support it. (See J. Henrick, American Antiquities, 2004, or Powell, Shennan and Thomas last year in Science, June 6.) The central idea here is that critical density is necessary for the persistence and evolution of concepts. As Powell et al. demonstrate, the benefits of density have to be scaled by the effectiveness of that density; a lot of people not talking to each other is equivalent to not many people talking freely. If you think this is sophistry or idle speculation, consider this: When Galileo was at his prime at the end of the fifteenth century, the center of science in Europe was the free state of Venice. When he made the mistake of thinking himself safe from the Roman Catholic Church and returned to Italy, his imprisonment and recantation under threat of torture of his work on the Copernican view of the solar system so chilled scientific debate in that part of the world that for two centuries scientific advancement in the south of Europe essentially ended. People stopped talking. The center of science moved to northern Europe, where it would see its greatest flowering. Is there any doubt that this is why, even to this day, the industries of northern Europe are stronger than those of southern Europe? Silence in science costs money in the marketplace.
Finally, it seems almost heretical to suggest that BD may actually reduce industrial-academic interactions. Surely there are now so many more license agreements, so many more commercial interactions? Well, yes and no. There are more interactions of a commercial nature, for sure, but this has not typically been how industry and academia have communicated. As I have mentioned in my last blog, I think molecular modeling is a field poorly served by academia precisely because of a lack of communication. When I hear from colleagues in industry that it is easier to talk to other companies—and in an industry as paranoid as the pharmaceutical industry, no less—than to establish relations with academics because of the licensing issues over any emergent intellectual property, something is very, very wrong. Some industries (for example, the chip design and manufacture industry) do not patent or license much at all from the academic world—and yet there is much more interaction (and success) between the two worlds.
In conclusion, I am tired of the laziness of authors like Columbia’s Sampat, who trumpet the success of a misconceived Act of Congress without any attempt to rationalize either the direct costs or the indirect costs. I think even a cursory evaluation casts plenty of doubt on the value of the Bayh-Dole Act and I fear a careful analysis would suggest that we, as a community of scientists, have been poorly served indeed.