Continuing to liveblog from PatCon 3. I'll continue updating this post throughout the debate.
A Debate About the Patent System
The Honorable Judge Richard Posner of the United States Court of Appeals for the Seventh Circuit
Professor Richard Epstein, Laurence A. Tisch Professor of Law, New York University School of Law, and James Parker Hall Distinguished Service Professor Emeritus of Law and Senior Lecturer, University of Chicago Law School
Moderated by Olivia Luk of Niro, Haller & Niro
For another perspective on this debate, check out Lisa Larrimore Ouellette's post over at Written Description.
Judge Posner: wrote "Why there are too many patents in America."
Longstanding academic interest in intellectual property; since becoming a judge, he's volunteered to handle litigation at the district court, and for the last ten years he has done almost exclusively patent cases. He's conducted bench trials, Daubert herings, claim construction, and so on. He's prepared several cases for jury trial, but they all fizzled out before trial.
In his view, we all probably agree that patents should not be granted unless there's a significant invention that, unless patented, would not be developed at all or would be developed much later. The reason is that a patent is a monopoly and a monopoly causes economic distortions. Beyond that, the administration of a patent system is very costly. The reason for this is that a patent right is very unlike a traditional property right. An invention is an abstraction. One of his cases involved a patent that had never actually been used to make any product, while the prior art that was being asserted had also never been used to make a product. So there was an extreme abstraction of the invention involved.
The clearest cast for patents is for pharmaceutical drugs, because pharmaceutical drugs are very costly to develop. Companies may spend hundreds of millions of dollars to develop a product for sale. But once the product is sold, the cost of duplication is relatively small. So here is a situation where, if there is no patent, the drug creator would never be able to recover the cost of its drug. This is the clearest case. But it is also relatively rare.
If you look across the economy more broadly, an inventor really doesn't need the patent as an incentive to invent. Why?
- In many areas, the incremental cost of invention is relatively small. Engineeres will tweak products without the incentive of patent.
- In other industries, the cost of duplication are actually quite high. One example is the learning curve phenonmenon. Once a product is invented, it's relatively costly at first to produce the product. But over time, the inventor figures out how to produce the product efficiently. When a copyist comes in, however, the copyist is going to have to figure out the production learning curve on its own. And thus will be initially competing at a disadvantage as compared to the inventor.
- Customers many also adapt to the original product and get accustomed to the original product before the competitor enters the market. There is an inertia here.
- Many products have a very short shelf life. By the time that a competitor realizes that this is a profitable product, and decides to duplicate the product, it may be too late.
- Trade secrecy may also make it difficult for a duplicator to figure out the way to duplicate. Keep in mind that trade secrecy is a two-edged sword: it's a cheaper way to protect the invention but it makes it more difficult for the public to learn how the technolgy works.
What he's learned from handling patent cases:
- The concept of "inventiving around" or partial duplication is not a congenial concept for jurors. Jurors think that inventing around is a sneaky thing; it's copying. This is a reason why patentees like juries. The other reason is that juries like inventors.
- Because of the attraction of the patent as a source of monopoly power, there has been a huge proliferation of patents. And these patents are being asserted and wheeled out as a basis to challenge their competitors. One of the reasons this has become so intense in the software industry is because of rapid change and high volatility. A company that is dominant today may be gone tomorrow. Once you believe that your survival depends on agressive assertion of your patents, that will cause your competitors to react in the same way.
- Patent trolls – had several patent troll cases. He likes the patent trolls: they're bucaneers, privateers. Can be very skillful. In some industries, it can be very difficult to do a patent search.
He has become very conscious of the difficulty of patent litigation for judges. He's found patent lawyers that he's dealt with to be very cooperative; it's a very elite, capable bar. Still, there's a tremendous challenge inherent in litigating patent cases. How do you explain a mathematical algorithm or software or how components of a handheld device work to a jury?
- Give the juries an anchor. He's decided that in every jury case he will appoint his own experts, who will be neutral. That expert will explain the technology and may also have a view on issues such as prior art and validity.
- Instructions. Have the instructions conference before trial and rule on the evidence before trial. 100 page instructions with jargon aren't going to work for a jury. Rewrite into 10 page instructions without jargon.
- Patent Pilot Program: 10 year experiment to see if district court judges who volunteer to take on patent cases can make a different in the patent litigation process.
There are some improvements that can be made in the process, but more fundamentally we need to think about the patent law more broadly. There are some industries in which patents should have shorter terms or no patents at all.
Richard Epstein (apologies for this not being nearly as Epsteinian as Richard Epstein actually delivered it)
The question of perspective matters a great deal in how you view these things. Regards litigation as the tip of the iceberg. His view is focused on how we can establish rules so that as few cases get to litigation as possible.
So how good is the current system? Hard to say because we're in the midst of a transition. In his view, the 1952 patent act, for all its limitations, was a great system.
How do you design the system and what might ought to be done?
No one thinks that a patent system is as easy to operate as a system for land – although here, you must be careful, because land systems can actually be quite complicated to administer.
Similarly, in patents, applying the patent law to different areas of technology can be difficult. Even in pharmaceutical patents, there can be a great deal of variation.
How do you work within the patent system? Two basic moves for a patent system to make:
Basic move is to talk about patent eligibility. Important because it's designed to weed out broad areas from the patent system so you don't have to do a case by case study.
Then you need to think about obviousness: those things that could be easily duplicated by many individuals ought not to receive patent protection because the gain isn't worth it.
Judge Posner argued that there are many areas in which patent protection is not necessary. This is an argument for keeping the patent system, not for getting rid of it. This is because the operation of the patent system itself should weed out these areas in which it's not worth getting a patent.
Applying this to patent improvements: Gleevik. Indian court basically concluded that there wasn't enough of an improvement over the existing technology to justify a patent. This is the wrong way to think about it. If you have something that is patented, and that patent has expired, and you now decide to make an improvement, then your patenting of that improvement will not be successful unless the improvement is sufficiently great for it to be commercially successful over the now freely available product. In this situation, he doesn't see a strong enough case for changing the way these particular patents operate.
Other kinds of patents: software patents. here, the argument is that the patent thicket creates the kind of blockade due to the defensive use of existing patents. In his view, the patent thicket argument is severely overated. This is because each patent represents an incremental gain, and is a substitute for the things that are already outthere. It's like a superhighway that goes over the street. A new patent thins out the thicket by creating an alternative. In his view, the novelty outweigs the blockade situation.
The immense number of licenses and cross-licenses also show that the system is working. Made even more efficient due to the valuation of pools across the pools themselves. This leads to more accurate pricing.
He thinks about patent reform from a perspective of how the patent system has become too lax in its protection of property rights. Two ways in particular:
1) Injunctions. The decline in the use of injunctive relief in the context of ordinary patent disputes. It's very difficult to measure actual damages to a patent holder; measuing consequential damages is always hard. The benefit of an injunction is that it avoids the need to calculate damages.
One issue here: When you're granting injunctions, they shouldn't be thought of only in absolute terms. Injunctions can be varied in two ways: the can be tailored in absolute terms and they can be married with damages.
If you don't have these kinds of injunctions, you'll have people flee the licensing system and dare someone to sue them.
2) Limitations on licensing. The law of licensing in patents isn't just reliant on consensual notions of licensing. There are all sorts of restrictions on the ability to license. These, to him, seem to simply increase the costs of licensing.
On the AIA: Not clear that we want to experiment with rules when the outcome is not clearly positive. That's what the AIA did. His basic take is that we had a lot of reform that basically put everything in reverse.
Post-presentation dialogue: Posner, responding to Epstein's position that a patent holder should be entitled to exclusive right. A patent is both a property right and a weapon to protect something that you've invented.
Epstein takes the position that small incremental improvements are either not valuable, in which case inventors will not file for a patent or the improvement will not matter in the marketplace because a non-infringing alternative already exists.
For Richard Epstein: what are your views on Judge Posner's view on different term lengths depending on technology. Response: Epstein is not particularly opposed to different patent terms. The value of a patent in different areas is going to diminish and thus he doesn't have a strong opposition to this.
Ted Sichelman: Would you talk about the roles patents play in incentivizing pre vs. post-invention activity.
Posner: To the extent that patents play a role in post-invention activity, this strengthens the case for patent protection.
Epstein: The stuff that you make commercialization contingent on receipt of a patent: the fact that you're dealing with a patented product vs a non-patented product shouldn't make a difference.
Mark Lemley: Recent data by Colleen Chien suggests that large numbers of patents are being filed by PAEs and that the suits overwhelmingly get resolved for relatively small amounts of money, much of which goes to pay lawyer's fees. How do you feel about fee shifting in these types cases?
Epstein: I think this is a serious problem. I love the German system: it requires the plaintiff to name a number. You can't go above that. Then, depending on how much the plaintiff wins of that amount, lets you calibrate the amount of fees.
Posner: The problem with fee shifting in general is that the incentive effects can be ambiguious. If you think you have a good chance of winning, and there's fee shifting involved, that means that your actual monetary victory is higher than it would actually be, and thus you're going to ask for more than you might otherwise.
Audience question: Would you talk more about remedies?
Posner: There are serious problems with damages remedies. Many patents are patents on components, not devices. Trying to value a component can be extremely difficult. You can do consumer surveys, maybe. The problem with the injunction is that it can be a disaster. If you've been found to infringe a component, and now you've been enjoined, that means that your entire inventory is blocked. You have to throw away a huge inventory or redo your production process.
Epstein: There is no quesiton that this is a very important difficulty. There is an advantage to having huge number of components in that you can look at royalty licenses for these components and see the values. This will help rein in the amount of the damages. Another thing you can do with injunctions is ask whether this is a marginal component or an integrated component. In the first case, you would be tough on the injunction while in the second case would not be so tough on the injunction.
My broader takeaway from Richard Epstein's views on this subject: underlying Epstein's approach to patents is that they are not just exclusive rights, but something like mutually exclusive rights. In other words, each time an invention is patented, that takes nothing away from what the public already has. Rather, it instead adds something new that would not have existed but for its creation by the patentee. For this reason, he is not terribly concerned about the patenting of marginal improvements because if they are truly marginal they will have a disruptive effect on the market because of the pre-existence of a non-infringing alternative.