Our patent system is operating in a very beneficial manner--for lawyers. For others, including inventors, engineers, scientists, and legitimate companies generating and/or using new technology, it is a costly, time-consuming quagmire. In some areas it is significantly slowing technological progress. What are the problems? Can they be solved via fine tuning, or should we consider a very different approach?
Patents were first seriously implemented in the eighteenth century to replace a harsh, dictatorial system whereby royal personages in Europe and England imposed monopolies on emerging technology via such means as privileges granted arbitrarily to favorites. To the extent possible, secrecy was employed to maintain monopolies, and there were even efforts to prevent the emigration of those knowledgeable about new technology--particularly in the textile industry.
The fundamental purpose of the patent system, as stated in the US Constitution, is "to promote the progress of science". Article I, section 8 authorizes (does not require) accomplishing this by "securing for limited times to...inventors the exclusive right to their... discoveries". The idea was to encourage invention and the dissemination of technical knowledge. Inventors who made their ideas publicly available were to be rewarded by allowing them to profit directly from those ideas via monopoly privileges granted for roughly fifteen years.
For a long time, there were strong restrictions on what could be patented. Basic scientific and mathematical concepts were excluded. Patents on developments obvious to practitioners in a field were also rejected. During the past few decades there has been a considerable expansion of the area in which patents could be obtained, and the degree of novelty required has been significantly reduced. There is now great controversy over what is or should be patentable. There are different answers in different countries, and these answers appear to be changing over time.
Creative people are not driven by the prospect of material rewards. Claude Shannon, David Huffman, and Amos Joel (who obtained over 70 patents) enjoyed the process of developing brilliant solutions to challenging problems. They reaped no particular rewards as a consequence of the patent system. Nor did the organizations that paid their salaries. (This situation is changing to the extent that many universities are now deriving income from patents obtained via faculty research.) In the absence of the patent system, universities and companies such as AT&T (via Bell Labs) would have had the same need to employ such people to develop needed technology. The patent system is a time-wasting nuisance to engineers, generating tedious paper work and often putting restrictions on their activities.
With respect to most employers (including those just mentioned) the patent system constitutes a substantial overhead cost and burden to management, with little in the way of compensating benefits. Some, perhaps most, large companies, such as IBM (and the old AT&T) use(d) patents in a defensive manner. They are most concerned about being forced to pay large fees to use technology essential to their businesses. They therefore make great efforts to patent inventions necessary for their operations, and also use their patents to trade with other companies for patents that they need. Altho they do sometimes sell or lease patents, this is not their principal motivation for developing new technology.
There are independent inventors who set out to devise new devices to accomplish useful ends, with a view, not to manufacture the devices, but rather to sell or lease patents to companies that will do this. I doubt that they constitute a major factor in technology today. (Which is not to say that their interests should be ignored in any proposed reforms.)
There are also companies that use patents to hamstring actual or potential competitors by denying them licenses. Since it is common for useful devices or processes to be developed by elaborating on prior versions, denial of access to a basic invention may prevent, or greatly delay, improvements.
Still other companies, called "patent trolls", exist solely to extract fees for patents that they purchased (often from companies going bankrupt) solely for this purpose. The trolls seek out victim companies producing products that incorporate technology allegedly infringing the patents in question. The patents held by the trolls are often very weak, but, due to the high cost of patent litigation, a company attacked by a troll may find it expedient to pay hundreds of thousands of dollars to avoid a law suit.
Is the patent system essential? One way to find out would be to abolish the patent system and then see what happens. But this would be a rather risky procedure, which, fortunately is not necessary. As is often done to answer questions of this type, we can gain valuable information by looking at special situations that occurred in the past and also by looking at other places where different rules apply.
There is currently a great deal of patent activity in the field of computer programming. This is in contrast to the situation two decades ago when software could not be patented. Given that the patent incentive did not exist in this field until, say, 1989, was there a paucity of invention in this field prior to that date? Not exactly. Most basic programming ideas antedate 1989. Consider assemblers, compilers, parsers, structured programming, simulators, emulators, hashing, random number generators, sorting routines, data base systems, just for starters. While copyright protection was tentatively introduced before that date, it was not a major factor. There was a significant amount of cooperation among computer users, one important manifestation being SHARE, the IBM user's group. Many important, widely used, items such as the emacs word processor, compilers for C and C++, and the linux operating system were developed cooperatively by the free software movement.
I don't believe anyone would argue that a surge of creativity in the software field started with the introduction of software patents. A stronger case could be made that software patents have impeded development in that field. Patents have been issued for so many simple programming techniques, that it is apparently difficult to produce a nontrivial piece of software today without infringing somebody's patent. As Richard Stallman put it:
Developing software is [now] like crossing a minefield. With each design decision, you might step on a patent that will blow up your project.
It seems clear that patents are a negative factor in the software area, adding to costs and discouraging, rather than stimulating, innovation. On the other hand, the open software approach seems to be flourishing.
Patents play a major role in the pharmaceutical industry. Once a drug has been approved for sale, and production is underway, the marginal cost of production is a small fraction of the prices set by the companies. They are able to keep the prices so high solely because competition is precluded by patent rights.
This is not the place for an extensive discussion of the
pharmaceutical industry, but I will mention a few salient points.
The industry claims that high prices are justified by high R&D costs,
and the expensive testing process necessary to get new drugs approved.
Critics argue that:
In addition to helping inflate the prices of drugs, for the benefit of pharmaceutical corporations, the patent system also often impedes drug development efforts, as illustrated by the following (from the NY Times):
"You start work in a certain therapeutic area doing some screening and then a patent will issue,'' said Ronald A. Pepin, a former Bristol-Myers executive. ''You're stuck. You either shut down your program or you try to get a license.''
Peter Ringrose, chief scientific officer at Bristol-Myers, has said there are more than 50 proteins possibly involved in cancer that the company was not working on because the patent holders either would not allow it or were demanding unreasonable royalties.
Altho the principal objective of the patent system was to promote the free dissemination of scientific and technical information, it doesn't seem to be working that way today. Companies such as Microsoft and those manufacturing e-voting systems refuse to make available their source code, which is, of course, essential for understanding and improving the programs.
About five years ago, the secrecy side of the patent system was brought home to me rather directly in connection with a graduate course in logic circuits. I wanted to include a short discussion of an interesting pipelining technique devised by one of our students as part his doctoral research. Before doing so, I checked with his advisor, who asked me not to teach this material until the following year, since earlier exposure might interfere with a pending patent application. This was not a serious problem, and reflects no discredit on my colleague, but it nicely illustrates one of many ways that the patent system impedes rather than promotes progress.
The fact that an idea has been patented does not automatically make it easily accessible to others. Published, refereed articles are all too often very difficult to read, but the worst of them are models of clarity when compared to most patent descriptions.
The patent route is optional, not mandatory, so many companies choose to shield their technology thru old fashioned trade secrecy. Under certain conditions, this can be done so as to acquire legal status. So secrecy comfortably co-exists with the patent system.
Overall, I think we would be much better off if the patent system did not exist at all. But because it is now so well entrenched, with so many people and organizations relying on it in various ways, it would not be feasible to just legislate the patent system out of existence overnight. There would have to be a long phasing out process with grandfathering of existing patents, etc. But what should replace it?
Recall that the objective of the patent system is to encourage useful innovation. If we want more R&D, the most straightforward approach is to pay people to do R&D (as opposed to paying lawyers to harass such people). More government research laboratories might be set up and more government funding might be provided to university researchers. Generating scientific knowledge and new engineering techniques seems like an appropriate function for publicly funded organizations. An added benefit of having more R&D conducted by non-commercial entities would be that more experts would be available free of conflicts of interest when advice is needed about such matters as product safety or environmental effects. A system of awards for excellent technological innovations (akin to Nobel Prizes) might be set up. Money for these purposes could, for the most part, come from what is saved by the downsizing and eventual elimination of the Patent Office.
The ideas generated by the above-mentioned government funded operations would be made freely available to those wishing to exploit them. Since the ultimate funding source would be the American taxpayer, means should be devised to give an edge to US-based companies. For example representatives of such companies might be invited to special symposiums to get early access to R&D results.
Of course nothing should prevent companies from operating R&D facilities focussing on areas important to them, or for consortiums of companies to operate such facilities jointly. Elimination of corporate patent departments and reduced expenses for patent litigation would help offset the costs. There already exist precedents for laboratories financed jointly by more than one country.
There are additional problems with the patent system. It is hard to imagine a contemporary invention that does not rely heavily on a complex web of knowledge developed, often over centuries, by numerous other people. It is very common for the same invention to be conceived of independently by several people over a relatively short span of time. Indeed, much patent litigation entails disputes over priority. A further complication is that, when groups of people are involved in developing the idea to be patented, it is often not clear which of them should be designated as co-inventors. These factors call into question the justice of heaping substantial rewards on one, or a few, individuals found to have the best claim for taking the final step in formulating the device or process in question.
Note that I haven't used the term, "intellectual property", which refers to copyrights, and trademarks as well as to patents. This is because I agree with Richard Stallman's argument that these are very different concepts in principle and in law and, therefore, should not be lumped together.
Michele Boldrin and David K. Levine, Against
Intellectual Monopoly, January 22, 2007. (Chapter 9, on the
pharmaceutical industry, is particularly interesting.)
Pierre Desrochers, "The Case Against
the Patent System", September 2000
Richard Stallman, "The
Danger of Software Patents",
Speech at Cambridge University, 25 March 2002
Comments can be sent to me at unger(at)cs(dot)columbia(dot)edu