By Mike Dorf
The latest episode of the NPR radio show This American Life addresses the question of patent trolls by following up an episode from two years ago, focusing on the efforts by Nathan Myhrvold's company Intellectual Ventures (IV) to enforce a 1998 patent that claimed invention of an internet storage system--and more. It's highly entertaining radio but, in my view, mostly goes after the wrong target.
Okay, here's the basic setup (spoiler alert!). Two years ago, the TAL/Planet Money team went out to talk with folks at IV, trying to determine whether IV was simply a gigantic patent troll--as its critics charged. The people at IV said no, it's a company that helps little-guy inventors get paid for, and bring to market, their inventions. When the reporters asked for examples of such little-guy inventors, they were given the name Chris Crawford. They track down Crawford but he won't talk to them and as of two years ago it looked like IV had sold Crawford's patent to what was mostly a front company (Oasis somethingorother) to bring lawsuits against large tech companies.
As the story picks up with the current episode, we discover that most of the defendants settle for undisclosed amounts, even as they regard the IV/Oasis/Crawford patent as bogus; still, they conclude that it's cheaper to settle than to roll the dice and go to trial. However, two companies refuse to settle and they do go to trial. They win, but just barely--and not because it's crazy to think that in 1998 anyone could have reduced to practice a patent on a functioning internet backup system. Instead, the defendants win because Crawford's invention was actually the result of work he did with three other people, so the patent falsely claimed the invention in Crawford's name alone. There is a jaw-dropping segment of audio from Crawford's deposition in which he claims not to know the meaning of "apostrophe s", thus disclaiming a document that referred to one of the other inventors. It's better than "it depends on what the meaning of 'is' is."
The show has a two-part takeaway. One--with which I agree--is that the PTO gives out too many patents, with coverage that is too broad, thereby allowing people to extract millions from companies that fear patent liability.
Second, there is a strong suggestion throughout the show that patent trolls like IV are bad actors. To my mind, this misses the point. IV and other firms that bring lawsuits opportunistically probably do stifle innovation but the question is why. TAL, various commentators and a fair number of scholars--including my colleague Oskar Liivak and Eduardo Penalver in a forthcoming article in the Cornell Law Review--lament patent trolling, by which they mean the accumulation of patents that are acquired simply for the purpose of suing alleged infringers, not to make products. I think that's a mistaken view. If we could fix the granting of patents by the PTO and their enforcement by the courts, we would necessarily fix the problem of trolls.
Here's why. Suppose IV or some other patent-accumulating company holds a patent for widgets and let's assume that the PTO properly granted the widget patent. Now IV goes to Acme and claims that Acme's gizmo infringes IV's widget patent. Should Acme pay IV? How much? The answers depend on whether the gizmo really does infringe the widget patent. If it does, then Acme should pay. If it doesn't, then it shouldn't--except maybe for the nuisance value of the lawsuit. That's a social loss, but the same is true of nuisance suits in every field.
Should anything in the foregoing paragraph depend on whether IV is IV rather than a company that actually makes widgets? I don't see why. IV wants to monetize its patent. It can do so by making widgets or by selling licenses. And if the litigation system is functioning well, then valid claims will settle--or ideally, they won't even be brought, because companies like Acme will purchase licenses in advance rather than waiting to be sued.
Part of what fixing the patent system would entail is making it more readily searchable. The defendant companies discussed in the TAL episode didn't discover the IV/Oasis/Crawford patent before they made their own products because it didn't occur to them to search for internet backup patents as far back as 1998, but even if it had, they might not have found it. That's because the proliferation of patents and the breadth of claims makes it very difficult for companies to determine in advance whether they need a license. And a lawsuit has holdup value, so that a company that builds an infringing product may end up having to pay more to settle an infringement suit than it would have had to pay ex ante for a license.
But solving the problems of over-patenting, granting overly broad patents, and opaque patents solves the troll problem, because fixing those problems means that there's no difference between litigating against infringers and granting licenses--and in such a world, licenses will be preferred because they've got lower transaction costs than litigation.
In short, patent trolls are a symptom. Fix the underlying substantive problems with the patent system and there will be no distinct problem of patent trolls.
Postscript: After I posted the foregoing, I became aware of an article by Mark A. Lemley and A. Douglas Melamed that is forthcoming in the Columbia Law Review and contains (far more detailed) analysis along similar lines. It has the wonderful title "Missing the Forest for the Trolls."