As I mentioned yesterday, one of the events at the PatCon 4 conference is a debate on the questions of whether hostility to patent trolls is "well
justified theoretically or empirically" and whether said hostility
"will likely result in bad law." Arguing that the hostility is not well justified are Professors John Duffy and David Schwartz. Arguing that it is are Professors Mark Lemley and Michael Meurer. (Note, however, that Professor Lemley's views in particular are somewhat more nuanced, as reflected in his article with A. Douglas Melamed, Missing the Forest for the Trolls.) Although I wasn't planning on live-blogging today's proceeding, I thought this debate in particular would be of interest to readers, so here goes.
Professor David McGowan introduced the panel. Professor Duffy then went first. Alienability of patent rights should not be changed; alienability is a feature of any property system. It's not just a matter of property rights generally, but applies to patents in particular. Inventors may not be good at business. If you're not hostile to ATT Bell Labs, but you're hostile to patent trolls, you're making a point about how corporations are integrated. Should they be integrated into a large corporation, or should we have atomistic competition with middlemen? The old model of having everything integrated into one corporation isn't necessarily good, but at any rate we should be agnostic about the matter. Second, litigation costs are high. Patent trolls are more efficient at dealing with those costs of litigation. They are capable of extracting value for these rights. Moreover, if you believe in relatively narrow patent rights, you need a secondary market to aggregate those rights. Patent trolls are developing a secondary market in patent rights and have an incentive to build better tools to evaluate patent value. That is a huge public benefit.
Professor Meurer went next. A hypothetical: suppose we were evaluating global warming. My opponents would argue that we should wait and see, etc. I would express my surprise at their skepticism and say we should apply the precauutionary principle. I concedee that teh case against patent trolls isn't that extreme, but it is strong. First, in 2000 there were few troll lawsuits, not they are significant. Second, small businesses hae motivated state legislatures, Congress, and the executive to act. Third, Peter Detkin of Innovation Ventures thinks that troll research shows that trolls are gaming the system (admission against interest).
Trolls impose a tax on innovation. Before the 1990s, the patent system was already performing badly, as documenting in Meurer and Bessen's Patent Failure. Research [I missed name of researcher] shows that being sued affects spending on R&D. Moreover, patent trolls typically lose their lawsuits. Bessen and Meurer estimated $29 billion loss according to one methodology, $80 according to another. The harm to innovation captured in Katherine Tucker's paper re Acacia. Litigation halted research relating to imaging software. Colleen Chien: 55% of unique defendants are small firms. Robin Feldman similar. Further research by Tucker.
Professor Schwartz: the particular form of the entity that holds the patent is not the problem. Are the suits frivolous? There's a lot of anecdotal information, but not a lot of good data. The data are not overwhelming. Most of it asks the wrong question. PAE litigation grew dramatically, but are the cases bad? Some guideposts for empirical study: first, what is a PAE? Meurer suggests it's anyone who doesn't practice a patent, but that includes universities and individual inventors and failed startups. And it includes aggregators and speculators. About a quarter of the NPE suits are individual inventor suits. Second, you need a baseline. Are the problems endemic to the litigation system? If R&D spending goes down when there is litigation, it's not necessarily attributable to PAE litigation. Third, we need to change the way we approach this research. The data should not be proprietary but publicly available.
Lemley: refers to article with Melamed, noted above. So why am I on this side of the debate? While trolls are not themselves the problem, they are a symptom of real systematic problems with our system. Takes too long to get a patent, can't tell what it means (particularly software patents), incentive to draft broadly, and the combination of these things means I can make a plausible claim that I'm entitled to 5% of your revenues; and it takes years to litigate. Bottom feeder model: taking advantage of enormous costs and uncertainties to extort nuisance value settlements. People sell to trolls not because of efficiencies but because trolls exploit these aspects of the system. Lots of discovery, little of it used; troll goal is to raise cost of discovery.
Is hostility to trolls creating bad law? Consider eBay; greater sophistication in damages; eliminated willfulness game; expertise in PTO, PGRs; reducing forum shopping. What might we do? Giving courts more discretion to punish frivolous suits; more detail about what your patent covers; transparency; reducing cost of discovery; make you sue the manufacturer rather than mom & pop shops as a way of increasing royalty base. So the hostility has led to good law, not bad law.
Meurer: my job now is to rebut Duffy and Schwartz. Like John, I worked at Bell Labs and I think decentralized research is a good thing. If we look at performance of small patent owners over that time, they were able to enforce their patents without intermediaries (startups like Apple, Google, Microsoft). Intermediaries not needed in biotech or pharma. John emphasized intermediaries in the patent licensing market, not necessarily a benefit to innovators who want access to technology. With regard to Dave's comments about whether PAE litigation is frivolous, Allison Lemley & Walker show they lose 90% of cases that go to court. Should we be more protective of universities? Yes, but they are infrequently represented in litigation. With regard to transfer of rights, encouraging alienation of rights is great, but we need more transparency to make that work; don't need the inappropriate strategic litigation benefit.
Duffy: I want to rebut the precautionary principle. We study innovation, change; shouldn't be scared of it. Maybe with global warming, we should be worried about change; not with patent law. Trolls are an innovation in business and law. In other fields with robust property rights, they are secondary markets. They play an important role. If my company goes bankrupt, I should be able to sell my IP. Other aggregators like ASCAP are beneficial. It's an oddity that we haven't had it in patent law. As for the asymmetries of litigation, defendants also take advantage of this. Meurer wrote in a paper about why large firms don't file DJ actions, waiting for evidence that patent owner has resources to litigate. Finally, the evidence Meurer cites includes event studies. Suit filed; see if stock price drops. On page 138 of his book, one of the reasons the price drops is the risk that patent will be invalidated. That's a gain for the public. Moreover, there's a drop in value on both plaintiff and defendant sides.
Lemley: John says we should welcome innovation, and I do. But the kind of innovation John wants to encourage is innovative ways to use the legal system to extract revenue from the people who do the technical innovation. Distressing fact: if you ask the software industries, startups, VCs, they tell you overwhelmingly that the patent system is a cost, not a benefit. That doesn't mean all software patents are bad or that all patents are bad. But it suggests that perhaps we have too much innovation in the legal models, and it's imposing a tax on the technical areas. Moreover, what do we want the secondary market to accomplish? Is the secondary market putting patents to productive use? These are patents that would have lain fallow. But is there any reason to believe that people innovate in the belief that even if my company goes over I'll be able to sell my patent to a troll. Secondary market argument is implausible. Traditionally, 98% of patents not enforced. What happens when these secondary markets mean we're enforcing 50-60-70% of all patents? Not a recipe for tech transfer, but rather exclusivity transfer.
Schwartz: Lemley talks about bottom feeder cases. Duffy and I are against frivolous suits. But the data is weakest here, and this is the linchpin of the argument. Allison/Lemley/Walker paper is the main one, but that study has its limitations. Many of the patents most litigated are outliers. Not the right way to make inferences. Kesan, Cotropia, and I doing further research. In addition, just losing at summary judgment doesn't mean the case was frivolous. Selection effects: only about 10% of cases reach the merits phase. Considering the 10% to make statement about the 90% may not be right. Possibility of an injunction affects settlement negotiations, and NPEs now can't get injunctions. NPEs have only one way to settle a case, by money.
McGowan: First (directed to Lemley and Meurer), isn't Duffy right that you don't want patent law to distort minimum patent size; different outcomes based on whether you practice or not? Second (directed to Duffy and Schwartz), isn't there an aggregation effect at issue here--when you have a portfolio, you can sue serially, which is different from, say, a one-on-one Apple v. Samsung case.
Lemley: should we single you out because you're an NPE? No. But that leaves room to recognize that motivations among parties. Moreover, we need proxies in the absence of perfect information. eBay does it pretty well.
Meurer: we all agree not to tailor patent law based on status. But some tailoring based on a proxy might be appropriate. Something like eBay allows a court to find the right context and craft the right remedy. Probably for many university plaintiffs, injunctions might be fine. But what we need to push the courts to do is to be more conscious of strategic behavior.
Duffy: By and large, no one is saying we should treat these inventors differently as a theoretical matter, and that is my basic point.
Schwartz: on the portfolio issue (company can serially sue) is not unique to PAEs; big firms can do that too. In response to Lemley and Melamed, serial enforcement might be a problem; royalty stacking. There may be solutions but potential for mischief.
Duffy: if the problem is that there are many junk patents, that's the problem to address, as in KSR. Moreover, both parties have an incentive to avoid litigation. Filing suit on a large number of frivolous claims can be a bad business model.
Meurer: John cited my book but not the underlying article. One thing he should have said is that one reason pharma companies lose value is that investors think they have rock-solid patents, and filing suit leads investors to question that. Second, assertion by NPEs as more efficient: the borderline patent that will be asserted will be of borderline quality. Shifting assertion to NPEs automatically means there will be more borderline patents asserted. And we still have problems after KSR. PAEs are asserting patents that in the 1990s would have been asserted, if at all, by PEs. PAEs aggravate the harm. I don't like NTP v. RIM; that was not frivolous, but it's disastrous for the patent system. I'm not as concerned about whether they are frivolous; it's the size of the patent tax that grows with the effectiveness of PAEs.
McGowan: Do you concede or agree there's a positive correlation between R&D spending and being sued? Why the spike in NPE suits in 2001? Mike and Mark: how confident are you that the data we have now is sound?
Duffy: the bigger you are, the more likely to be sued. More you spend on R&D, more likely to be sued. Firms on the cutting edge are the ones that get sued for infringement. As for the spike, there is innovation going on, and IBM (a PE) made money from licensing. They show us new ways of doing things.
Schwartz: Re the increase in litigation: growing view that patents are valuable assets, and contingency fee suits. More litigation doesn't necessarily mean the system is bad.
Lemley: Worth saying that we know less than we would like to about suits that settle. About 10-15% of cases do go to judgment, usually summary judgment. They may be unrepresentative in some respect. But academic literature suggests they should be about 50-50. The fact that trolls lose when they go to judgment is still striking. And when we say 90% of those cases fail, that's a significant chunk of all the cases in the system. Colleen Chien survey on settlement and payout: how much do you pay? 90% have payout of $10 million or less including attorneys' fees. Suggests that most of the cases are settling for something like the cost of litigation. Doesn't necessarily mean frivolous, but perhaps small stakes. There are limitations to the data, and I agree. Maybe we should require disclosure about patent settlements. Finally, it's worrisome if the more I spend on R&D the more likely I am to be sued: whether the problem is trolls or not.
Meurer: What should the presumptions be? The political economy of this problem is crystal clear. We need to reform. If we push too far, the concentrated interests will push back. It's a rare opportunity that Congress and the executive care about patents. Sensible thing to do is to reform. There will be a continuous force in the opposite direction.
Question from Oskar Liivak: Imagine two inventors who come up with process patent they won't manufacture. Can try to license, or wait until someone else invents and then sue. The inventor who actually pushes the technology out is more socially beneficial than the one who puts it in the drawer and waits. The patent system puts them both on the same footing: they get a reasonable royalty.
Duffy: First, I agree that ex ante licensing is great. The inventor may not be the best person to do it. I might want a PAE who's ready to license you. Second, patents don't get put in drawers; they get put on the Internet.
Lemley: We want technology transfer. We used to get it through the patent office because the diffusion of information was slow and the PTO was fast. No longer. If you're asserting a software patent, your technology is probably obsolete. Or you're trying to recast what you invented as something on the market now. That's not tech transfer, that's patent rights transfer.
Question from Nancy Linck: In AIA, Congress told GAO to make a study about effect of PAEs on litigation. Study found no effect?
Lemley: Study concluded that; compared data concluding in 2011 before AIA joinder change. Asked only how many lawsuits were filed, not how many people were sued. Most lawsuits filed in 2000s were by PEs, but most people sued were sued by NPEs.
Meurer: The GAO study told us there's a serious problem with software patents.
Schwartz: The study did look at number of people sued as well.
Professor David McGowan introduced the panel. Professor Duffy then went first. Alienability of patent rights should not be changed; alienability is a feature of any property system. It's not just a matter of property rights generally, but applies to patents in particular. Inventors may not be good at business. If you're not hostile to ATT Bell Labs, but you're hostile to patent trolls, you're making a point about how corporations are integrated. Should they be integrated into a large corporation, or should we have atomistic competition with middlemen? The old model of having everything integrated into one corporation isn't necessarily good, but at any rate we should be agnostic about the matter. Second, litigation costs are high. Patent trolls are more efficient at dealing with those costs of litigation. They are capable of extracting value for these rights. Moreover, if you believe in relatively narrow patent rights, you need a secondary market to aggregate those rights. Patent trolls are developing a secondary market in patent rights and have an incentive to build better tools to evaluate patent value. That is a huge public benefit.
Professor Meurer went next. A hypothetical: suppose we were evaluating global warming. My opponents would argue that we should wait and see, etc. I would express my surprise at their skepticism and say we should apply the precauutionary principle. I concedee that teh case against patent trolls isn't that extreme, but it is strong. First, in 2000 there were few troll lawsuits, not they are significant. Second, small businesses hae motivated state legislatures, Congress, and the executive to act. Third, Peter Detkin of Innovation Ventures thinks that troll research shows that trolls are gaming the system (admission against interest).
Trolls impose a tax on innovation. Before the 1990s, the patent system was already performing badly, as documenting in Meurer and Bessen's Patent Failure. Research [I missed name of researcher] shows that being sued affects spending on R&D. Moreover, patent trolls typically lose their lawsuits. Bessen and Meurer estimated $29 billion loss according to one methodology, $80 according to another. The harm to innovation captured in Katherine Tucker's paper re Acacia. Litigation halted research relating to imaging software. Colleen Chien: 55% of unique defendants are small firms. Robin Feldman similar. Further research by Tucker.
Professor Schwartz: the particular form of the entity that holds the patent is not the problem. Are the suits frivolous? There's a lot of anecdotal information, but not a lot of good data. The data are not overwhelming. Most of it asks the wrong question. PAE litigation grew dramatically, but are the cases bad? Some guideposts for empirical study: first, what is a PAE? Meurer suggests it's anyone who doesn't practice a patent, but that includes universities and individual inventors and failed startups. And it includes aggregators and speculators. About a quarter of the NPE suits are individual inventor suits. Second, you need a baseline. Are the problems endemic to the litigation system? If R&D spending goes down when there is litigation, it's not necessarily attributable to PAE litigation. Third, we need to change the way we approach this research. The data should not be proprietary but publicly available.
Lemley: refers to article with Melamed, noted above. So why am I on this side of the debate? While trolls are not themselves the problem, they are a symptom of real systematic problems with our system. Takes too long to get a patent, can't tell what it means (particularly software patents), incentive to draft broadly, and the combination of these things means I can make a plausible claim that I'm entitled to 5% of your revenues; and it takes years to litigate. Bottom feeder model: taking advantage of enormous costs and uncertainties to extort nuisance value settlements. People sell to trolls not because of efficiencies but because trolls exploit these aspects of the system. Lots of discovery, little of it used; troll goal is to raise cost of discovery.
Is hostility to trolls creating bad law? Consider eBay; greater sophistication in damages; eliminated willfulness game; expertise in PTO, PGRs; reducing forum shopping. What might we do? Giving courts more discretion to punish frivolous suits; more detail about what your patent covers; transparency; reducing cost of discovery; make you sue the manufacturer rather than mom & pop shops as a way of increasing royalty base. So the hostility has led to good law, not bad law.
Meurer: my job now is to rebut Duffy and Schwartz. Like John, I worked at Bell Labs and I think decentralized research is a good thing. If we look at performance of small patent owners over that time, they were able to enforce their patents without intermediaries (startups like Apple, Google, Microsoft). Intermediaries not needed in biotech or pharma. John emphasized intermediaries in the patent licensing market, not necessarily a benefit to innovators who want access to technology. With regard to Dave's comments about whether PAE litigation is frivolous, Allison Lemley & Walker show they lose 90% of cases that go to court. Should we be more protective of universities? Yes, but they are infrequently represented in litigation. With regard to transfer of rights, encouraging alienation of rights is great, but we need more transparency to make that work; don't need the inappropriate strategic litigation benefit.
Duffy: I want to rebut the precautionary principle. We study innovation, change; shouldn't be scared of it. Maybe with global warming, we should be worried about change; not with patent law. Trolls are an innovation in business and law. In other fields with robust property rights, they are secondary markets. They play an important role. If my company goes bankrupt, I should be able to sell my IP. Other aggregators like ASCAP are beneficial. It's an oddity that we haven't had it in patent law. As for the asymmetries of litigation, defendants also take advantage of this. Meurer wrote in a paper about why large firms don't file DJ actions, waiting for evidence that patent owner has resources to litigate. Finally, the evidence Meurer cites includes event studies. Suit filed; see if stock price drops. On page 138 of his book, one of the reasons the price drops is the risk that patent will be invalidated. That's a gain for the public. Moreover, there's a drop in value on both plaintiff and defendant sides.
Lemley: John says we should welcome innovation, and I do. But the kind of innovation John wants to encourage is innovative ways to use the legal system to extract revenue from the people who do the technical innovation. Distressing fact: if you ask the software industries, startups, VCs, they tell you overwhelmingly that the patent system is a cost, not a benefit. That doesn't mean all software patents are bad or that all patents are bad. But it suggests that perhaps we have too much innovation in the legal models, and it's imposing a tax on the technical areas. Moreover, what do we want the secondary market to accomplish? Is the secondary market putting patents to productive use? These are patents that would have lain fallow. But is there any reason to believe that people innovate in the belief that even if my company goes over I'll be able to sell my patent to a troll. Secondary market argument is implausible. Traditionally, 98% of patents not enforced. What happens when these secondary markets mean we're enforcing 50-60-70% of all patents? Not a recipe for tech transfer, but rather exclusivity transfer.
Schwartz: Lemley talks about bottom feeder cases. Duffy and I are against frivolous suits. But the data is weakest here, and this is the linchpin of the argument. Allison/Lemley/Walker paper is the main one, but that study has its limitations. Many of the patents most litigated are outliers. Not the right way to make inferences. Kesan, Cotropia, and I doing further research. In addition, just losing at summary judgment doesn't mean the case was frivolous. Selection effects: only about 10% of cases reach the merits phase. Considering the 10% to make statement about the 90% may not be right. Possibility of an injunction affects settlement negotiations, and NPEs now can't get injunctions. NPEs have only one way to settle a case, by money.
McGowan: First (directed to Lemley and Meurer), isn't Duffy right that you don't want patent law to distort minimum patent size; different outcomes based on whether you practice or not? Second (directed to Duffy and Schwartz), isn't there an aggregation effect at issue here--when you have a portfolio, you can sue serially, which is different from, say, a one-on-one Apple v. Samsung case.
Lemley: should we single you out because you're an NPE? No. But that leaves room to recognize that motivations among parties. Moreover, we need proxies in the absence of perfect information. eBay does it pretty well.
Meurer: we all agree not to tailor patent law based on status. But some tailoring based on a proxy might be appropriate. Something like eBay allows a court to find the right context and craft the right remedy. Probably for many university plaintiffs, injunctions might be fine. But what we need to push the courts to do is to be more conscious of strategic behavior.
Duffy: By and large, no one is saying we should treat these inventors differently as a theoretical matter, and that is my basic point.
Schwartz: on the portfolio issue (company can serially sue) is not unique to PAEs; big firms can do that too. In response to Lemley and Melamed, serial enforcement might be a problem; royalty stacking. There may be solutions but potential for mischief.
Duffy: if the problem is that there are many junk patents, that's the problem to address, as in KSR. Moreover, both parties have an incentive to avoid litigation. Filing suit on a large number of frivolous claims can be a bad business model.
Meurer: John cited my book but not the underlying article. One thing he should have said is that one reason pharma companies lose value is that investors think they have rock-solid patents, and filing suit leads investors to question that. Second, assertion by NPEs as more efficient: the borderline patent that will be asserted will be of borderline quality. Shifting assertion to NPEs automatically means there will be more borderline patents asserted. And we still have problems after KSR. PAEs are asserting patents that in the 1990s would have been asserted, if at all, by PEs. PAEs aggravate the harm. I don't like NTP v. RIM; that was not frivolous, but it's disastrous for the patent system. I'm not as concerned about whether they are frivolous; it's the size of the patent tax that grows with the effectiveness of PAEs.
McGowan: Do you concede or agree there's a positive correlation between R&D spending and being sued? Why the spike in NPE suits in 2001? Mike and Mark: how confident are you that the data we have now is sound?
Duffy: the bigger you are, the more likely to be sued. More you spend on R&D, more likely to be sued. Firms on the cutting edge are the ones that get sued for infringement. As for the spike, there is innovation going on, and IBM (a PE) made money from licensing. They show us new ways of doing things.
Schwartz: Re the increase in litigation: growing view that patents are valuable assets, and contingency fee suits. More litigation doesn't necessarily mean the system is bad.
Lemley: Worth saying that we know less than we would like to about suits that settle. About 10-15% of cases do go to judgment, usually summary judgment. They may be unrepresentative in some respect. But academic literature suggests they should be about 50-50. The fact that trolls lose when they go to judgment is still striking. And when we say 90% of those cases fail, that's a significant chunk of all the cases in the system. Colleen Chien survey on settlement and payout: how much do you pay? 90% have payout of $10 million or less including attorneys' fees. Suggests that most of the cases are settling for something like the cost of litigation. Doesn't necessarily mean frivolous, but perhaps small stakes. There are limitations to the data, and I agree. Maybe we should require disclosure about patent settlements. Finally, it's worrisome if the more I spend on R&D the more likely I am to be sued: whether the problem is trolls or not.
Meurer: What should the presumptions be? The political economy of this problem is crystal clear. We need to reform. If we push too far, the concentrated interests will push back. It's a rare opportunity that Congress and the executive care about patents. Sensible thing to do is to reform. There will be a continuous force in the opposite direction.
Question from Oskar Liivak: Imagine two inventors who come up with process patent they won't manufacture. Can try to license, or wait until someone else invents and then sue. The inventor who actually pushes the technology out is more socially beneficial than the one who puts it in the drawer and waits. The patent system puts them both on the same footing: they get a reasonable royalty.
Duffy: First, I agree that ex ante licensing is great. The inventor may not be the best person to do it. I might want a PAE who's ready to license you. Second, patents don't get put in drawers; they get put on the Internet.
Lemley: We want technology transfer. We used to get it through the patent office because the diffusion of information was slow and the PTO was fast. No longer. If you're asserting a software patent, your technology is probably obsolete. Or you're trying to recast what you invented as something on the market now. That's not tech transfer, that's patent rights transfer.
Question from Nancy Linck: In AIA, Congress told GAO to make a study about effect of PAEs on litigation. Study found no effect?
Lemley: Study concluded that; compared data concluding in 2011 before AIA joinder change. Asked only how many lawsuits were filed, not how many people were sued. Most lawsuits filed in 2000s were by PEs, but most people sued were sued by NPEs.
Meurer: The GAO study told us there's a serious problem with software patents.
Schwartz: The study did look at number of people sued as well.
No comments:
Post a Comment