pageok
pageok
pageok
[Adam Mossoff, guest-blogging, May 7, 2009 at 10:49pm] Trackbacks
Patent Thickets, Patent Pools and Antitrust:

This is my final blog posting, and I'd like to briefly raise an issue with the readers of this series about a complex policy and legal issue: antitrust and its impact on patent pools.

Among all the differences between the nineteenth century and the modern age that have been identified and discussed by VC commentators and others over the past week, this is perhaps one of the most consequential. Interestingly, it's a difference in law, not in science or technology. Yet this may have the biggest impact on patent thickets, because it makes its private-ordering solution that much harder to form.

As I discussed in my posting about the Sewing Machine Combination, this patent pool and commercial trust was attacked in the popular press as a "grinding, pitiless monopoly." In the Sewing Machine Combination's patent infringement lawsuits, defendants repeated such arguments, such as calling its litigation war chest an act of "oppressive conduct." Such arguments, however, fell on deaf ears in the courts. In one 1862 case, for instance, a federal court rejected this claim, asking rhetorically, "why they might not make a common fund for the purpose of protecting their common rights by prosecuting those they thought had infringed them, I am at a loss to conceive."

After the enactment of the Sherman Antitrust Act in 1890, this was no longer a rhetorical question for a court to ask, as the federal government embarked on a campaign of "trust busting" (to use Teddy Roosevelt's famous term).

I am not an antitrust expert, and so I am very much hoping for some substantial feedback from the VC readers. It's my understanding that, for much of its history, antitrust law prohibited patent pools as anticompetitive -- they were deemed to be a form of collusion for the purpose of squelching competition. In more recent years, this hard line against patent pools has softened a bit, as the federal government now applies a "rule of reason" analysis to determine if patent pools are pro- or anti-competitive. This requires the courts to engage in a market-specific, context-sensitive, patent-specific analysis of the nature of the pool and its function, asking such questions as whether the patents in the pool are complementary or substitutes for each other, whether prices will rise or fall, whether the pool will facilitate a product coming to market or not, etc.

Now, I think there is a legitimate philosophical question to ask about whether such concepts are even valid, but that is beyond the scope of both my paper and this blog posting. Here, my inquiry is pretty limited in scope.

Antitrust scholars acknowledge that rule-of-reason analysis is "indeterminate," which means that it is next to impossible for the firms to predict ex ante whether their actions will run afoul of the antitrust laws or not. So, here's my question: How does this indeterminacy impact the use of patent pools to resolve paten thickets? If so, does this make patent thickets more or less problematic today? In other words, is one of the reasons for the problem of patent thickets today exogenous to the patent system -- is antitrust mucking up the workings of the patent system here?

I don't know the answers to these questions yet, but it seems commonsensical that antitrust would be relevant to the patent thicket problem. The question would seem to be how much and in what way. I'm really interested in hearing from people who know more about antitrust than I do -- I'm currently researching it as part of my writing project on patent thickets -- and so any ideas, suggestions, pointers, cases, etc. would be greatly appreciated.

With that said, I'm now signing off from my guest-blogging stint here at The Volokh Conspiracy. I wanted to thank Eugene again for inviting me to do this. I had a lot of fun, and I hope you (the readers) did, too. I know that my final paper will be far improved as a result of this experience. In fact, I really enjoyed my engagements with the commentators, who are to be commended for their thoughtful and substantive remarks. The comments stayed on track and did not get sidetracked, and for that, thank you.

einhverfr (mail) (www):
IANAL, but having followed some of the issues in Microsoft's sabre rattling regarding Linux and patents, I think that the indeterminate nature of the problem is one thing that helps reduce the scope of the problem.

Basically in this case, Microsoft is trapped between three bad options:

1) They can sue Linux developers, service providers, and users, and thus earn a great deal of negative PR. In such a suit, the fact that they have market power will be used against them and they might end up losing enforceability of patents in question. Defeats here would be bad for the company for a number of reasons and could weaken the company in other antitrust lawsuits due to issues of collateral estoppel. Even if they win a few, this has all of the problems associated with #3 below.

2) They can do nothing and watch Linux take over additional markets that they covet. They can thus watch their monopoly erode while doing nothing.

3) They could disclose the list of patents they think Linux infringes on as a way to scare people away from the system. This is bad too: 1) Software engineers will quickly work around these patents, 2) some of the patents might get challenged in re-examination proceedings, and so forth, etc.

The indeterminate nature of the antitrust elements of the patent question is one aspect which helps keep the problem in check. Microsoft may issue press releases but they don't really want to sue because this carries with it substantive legal risks.

I would expect similar considerations would go through IBM's legal department too. And this may be one reason why major players don't seem to be terribly interested in initiating patent lawsuits.
5.8.2009 12:01am
Soronel Haetir (mail):
I would think another reason large companies don't assert patent rights would be that even if you can force a particular person to stop infringing it would be extremely difficult to prevent determined distributed infringement that is not constrained by the need to physically transport anything across national boundaries.

It is far easier to intercept infringing widgets at the border when customs is inspecting the shipment than it is to monitor every ftp server in existance.

So even if you had some critical patent that could not be worked around and without which no other system could compete in the software market against your licensed product it would be next to impossible to keep infringing products from being available.

While not patented technology I think the DVD experience is illustrative for the above proposition.

As for the Sewing Machine Wars, thank you for an interesting series on a mostly forgotten slice of history.
5.8.2009 1:01am
Mike C. (mail):
Thank you for a most interesting series of posts. I found myself reading every one of them, and I'm sorry to see it come to an end. Very interesting, and well-written enough to allow us non-lawyers to follow along.
5.8.2009 7:36am
Ken Arromdee:
So even if you had some critical patent that could not be worked around and without which no other system could compete in the software market against your licensed product it would be next to impossible to keep infringing products from being available.

What's to prevent you from just suing the competitor (they have to be located somewhere)?

You mention DVDs, but those are pirated 1) by individuals, who aren't competing on the market, and 2) by companies that are headquartered in places which don't enforce copyrights.
5.8.2009 9:56am
Soronel Haetir (mail):
Ken Arromdee,

I am thinking in terms of commercial product vs free product, where the commercial product maker owns Critical Patent. Given the distributed multinational development nature of free software I believe it would be next to impossible for the owner of Critical Patent to assert their rights in a way that would keep infringing products from being easily available. About the best you could likely accomplish would be keeping infringing products out of distributions (at least physically packaged distributions, network distributions would likely be hard to stop as well).

I understand that Critical Patent is in fact extremely unlikely to exist, the point being that even if it did exist it likely wouldn't do much good against the forces of free software.
5.8.2009 10:41am
Ken Arromdee:
Having something available but not in a commercial product, or even in a free product that has an easily recognizeable company as a source (like most Linux distributions), is a really really big restriction on it. The only people who will use it will be those willing to download it themselves (or know someone who they trust with their computer, who can do that).
5.8.2009 11:11am
einhverfr (mail) (www):
Ken Arromdee:

What's to prevent you from just suing the competitor (they have to be located somewhere)?


Well, let's explore the big differences between something like open source software and something like automobile manufacturing.

In automobile manufacturing, you need a factory, inventory, expensive equipment, hired help, etc. This is expensive. If you have an injunction, you are out of a LOT of money beyond your immediate living expenses. And there is centralized competition.

In open source software, you need equipment you already have around the house, software that doesn't cost anything, and you can do it yourself without hiring anyone. You can do it from home. And the same software might be collaboratively built by a large number of small commercial entities coordinating their effort. These networks may span across national borders, and some of those countries may not protect software patent claims.

In this case, the "competition" is not "located" in a specific spot, or even in a specific company but is spread around. Furthermore by the time you file your suit, the other companies will probably just work around your patent anyway, maybe by changing encryption algorithms, moving to other compression systems, etc. So you end up suing someone with no money who doensn't really control your competition, etc. The distributed nature of the product development makes asserting patents very cost-ineffective. Furthermore.... the community includes a great number of skilled engineers who may coordinate on re-examination proceedings. So suing open source software competitors over patent infringement is a losing strategy: you gain virtually nothing and you have a great deal to lose including the patent.

A few larger open source businesses (Red Hat in particular) have begun to build defensive patent pools of their own, but the fact is that although driving Red Hat out of business would hurt open source software, it wouldn't end ANY of Red Hat's products....

Open source software is largely unaffected by patent problems because:
1) There are few if any viable targets for lawsuits
2) The community can rapidly respond to infringement claims and work around them.
3) The proprietary software industry is dominated by a few players, each of which has market power in their own domains. Inevitable antitrust defences and/or counterclaims, if successful, would impact more than just patent claims but also set up such a business for more legal problems later.

I suppose on further reflection (and thinking through some of Massoff's posts), the danger to open source software from software patents are overblown and ignoring claims is still a viable strategy. While the scope of the problem may be new, offsetting elements (low barrier to entry, global economy with variable scopes of protections to patents) are new also and these undermine the patent protections fatally when used against open source software.
5.8.2009 12:56pm
einhverfr (mail) (www):
Ken Arromdee also wrote:

Having something available but not in a commercial product, or even in a free product that has an easily recognizeable company as a source (like most Linux distributions), is a really really big restriction on it. The only people who will use it will be those willing to download it themselves (or know someone who they trust with their computer, who can do that).


A good counter-example would be PostgreSQL, which does not have an identifiable company behind it but rather a number of independent companies (Red Hat, Command Prompt, Green Plum, EnterpriseDB, etc). Since Sun has been a big sponsor, it will be interesting to see whether in a year we will be adding Oracle to that list.......

PostgreSQL really is THE open source RDBMS for enterprise database management. Because there is no single company behind it, you don't see a lot of flashy press releases, but you do see solid use where it matters.
5.8.2009 1:04pm
Ken Arromdee:
Those companies producing PostgreSQL are big enough that they can be targeted for lawsuits, and are *not* too hard to sue out of the market. It would take more than one lawsuit, since there's more than one source, but it's not like suing a million individual DVD pirates.

Incidentally, there was recently an article on groklaw.net about a new patch to Linux which prevents it from writing FAT32. Red Hat is trying to find prior art, but if nobody can, the feature will have to be taken out of the kernel. The fact that it's impossible to sue every individual isn't going to help; if no major Linux distribution can use it, that's still a huge benefit to Microsoft in helping kill Linux.
5.8.2009 3:38pm
zippypinhead:
Antitrust scholars acknowledge that rule-of-reason analysis is “indeterminate,” which means that it is next to impossible for the firms to predict ex ante whether their actions will run afoul of the antitrust laws or not. So, here’s my question: How does this indeterminacy impact the use of patent pools to resolve patent thickets?
As a practical matter, not much at all. DOJ's Antitrust Division has had a "business review" procedure in place for several years that permits JV's, formal patent pools and others to submit proposed combinations to the Antitrust Divison in advance. DOJ reviews the proposal and issues a formal opinion (usually including some fairly detailed legal analysis) as to whether it might take enforcement action. DOJ has a fair amount of information, including the text of previous business reviews, on its web site. Experienced counsel who practice in that area have a pretty good idea what does and does not fly.
5.8.2009 5:32pm
einhverfr (mail) (www):
Ken Aromdee:

Those companies producing PostgreSQL are big enough that they can be targeted for lawsuits, and are *not* too hard to sue out of the market. It would take more than one lawsuit, since there's more than one source, but it's not like suing a million individual DVD pirates.


However, suing the companies out of the market is quite a different thing than suing the project out of the market. Remember that many of the individual developers have moved from one company to another as some companies come and others go so. And the community is big enough that a new generation of businesses would just arise to take their place, probably after they have versions that don't infringe on the patents. I could probably name half a dozen central PostgreSQL companies which have entered and exited the market in the last 10 years.

If you sue EnterpriseDB out of the market, does that mean that Bruce Momjian will stop contributing? If you sue Red Hat out of the market does that mean Tom Lane will stop contributing? Probably not..... So really, one would end up replacing the larger companies temporarily with small private individuals, and from them, the next generation of companies would arise.

So once again, patent suits give some short term gains, but don't really offer long-range security.

Now, with PostgreSQL, the general fear though is that the few companies which offer proprietary versions of the software (EnterpriseDB, Green Plum) could be especially vulnerable to patent suits which is why that project is careful about patents. However, this really only addresses certain segments of the market and is not the same thing as saying the project as a whole is endangered by the patents.
5.8.2009 5:43pm
Adam Mossoff (mail):
zippypinhead: Thanks for the feedback, but I'm not entirely convinced that the antitrust issue is as easy as you make it out to be.

First, just because the DOJ thinks a particular business practice, such as a patent pool, is pro-competitive, this does not mean that the firms are in the clear. The FTC may disagree, and it may decide to initiate enforcement actions itself. This has happened, for instance, in the context of the reverse-settlements for generic drug manufacturers. The DOJ's position has been that this is not an antitrust problem, and the FTC has been continually bringing cases and filing cert petitions to the Supreme Court. The amount of money spent on lawyers in all of these cases and filings by the FTC has to be huge.

Second, just because the DOJ or the FTC signs off on a business practice, private parties are always free to bring their own antitrust lawsuits, right? It's no surprise that competing firms bring these lawsuits as a way to get rid of their competition.

It's my general sense, based on the early research I've done, that these two issues add up to a whole boatload of uncertainty about the costs of a future patent pool. So, potential patent pool participants have to take into account the costs of possible governmental action, costs of possible private action, the costs of defending the lawsuit, and the probabilities of a court ultimately finding that the practice is unlawful.

This is a whole lot of predictive cost-benefit analysis that goes into forming a patent pool -- with lots of potential huge costs -- that didn't exist before antitrust, and this has to have an impact on whether companies ultimately conclude that the transaction costs may swamp the benefits to them in forming the pool. I understand that this is somewhat generalized, but, as I said in my posting, it seems commonsensical.
5.8.2009 10:45pm
Ken Arromdee:
However, suing the companies out of the market is quite a different thing than suing the project out of the market.

Once they win a suit against at least one, any other one, including future ones, will become very reluctant to use the patented technique. That FAT patent, for instance, isn't really aimed at Red Hat, it's aimed at makers of devices like MP3 players and cameras which use mass storage. But if Microsoft can succeed in suing them, Red Hat will no longer be able to have it in Linux.
5.8.2009 11:34pm
Montana:

Once they win a suit against at least one, any other one, including future ones, will become very reluctant to use the patented technique. That FAT patent, for instance, isn't really aimed at Red Hat, it's aimed at makers of devices like MP3 players and cameras which use mass storage. But if Microsoft can succeed in suing them, Red Hat will no longer be able to have it in Linux.

The FAT patent is not a very representative example, and it's actually somewhat of a problem for Microsoft: The only reason anyone still uses FAT is that everyone else is using it. It isn't as though it's even a decent file system, it's just that it's so old that everything is compatible with it. If Microsoft puts restrictions on who can use it, they destroy the universal compatibility which is the only reason it still has any value in the first instance. Worse yet, if the device makers adopt an open source file system in response, Windows will end up requiring fickle third party file system drivers to make the devices work.

In addition to that, because use of the patented file system is required to achieve compatibility with Windows, and that is the only reason anyone is actually using it, antitrust prosecutors may have a strong argument that enforcing the patent against that behavior is anti-competitive.

A patent with no Windows compatibility implications might be easier to enforce, but then Microsoft has to worry about OIN (see www.openinventionnetwork.com) and worry about the possibility that they own a patent Microsoft is infringing.
5.9.2009 5:02am
einhverfr (mail) (www):
Ken Arromdee:

Once they win a suit against at least one, any other one, including future ones, will become very reluctant to use the patented technique. That FAT patent, for instance, isn't really aimed at Red Hat, it's aimed at makers of devices like MP3 players and cameras which use mass storage. But if Microsoft can succeed in suing them, Red Hat will no longer be able to have it in Linux.


First, I don't think that Red Hat gets much use out of the FAT filesystem driver in Linux since they are not interested in the desktop market.

Secondly, this highlights the antitrust problems that Microsoft would have in the event of using patent litigation to forestall interoperability with Windows.

Any legal discussion of patent rights and abuse in this context must start from the recognition that Microsoft has a monopoly on critical portions of the market, and has a past pattern of illegally maintaining that monopoly. This cannot be legally disputed in this sort of case because it was decided as a necessary part of previous legal action. Therefore, there is a hightened presumption of anticompetitive action on the part of Microsoft than there would be from, say, The SCO Group. So Microsoft starts off on a bad foot here. In short they risk being unable to enforce their patent on antitrust grounds anyway, and a loss would make it far harder to assert ANY OTHER patent claims against interoperating products.

Now, Microsoft could start by suing companies like Olympus if they won't license the patents, and once the vast majority of companies using FAT are licensing it, then come around and ask for similar deals from Linux vendors (who would then have to stop distributing FAT drivers under the GPL).

However.... There are a couple of problems with this. Windows does not preclude other filesystem drivers, and in fact provides a driver model for these sorts of things. For this reason, it seems reasonable to suggest that the industry might move to better filesystems (like ext3, ext2, jfs) which are supported on most platforms with readily available drivers. Furthermore a Linux user like myself could STILL download the original source from a country where software patents are not yet recognized and include that functionality if I need it myself (I do this with mp3 player software, for example).

So for Linux desktop users (we aren't really dependant on Red Hat or any other vendor), suing Red Hat over the FAT patent would do very little to accomplish anything but would clearly hurt Microsoft. For server users, FAT is irrelevant anyway.

Is there some risk? Sure. However the risk falls disproportionately on Microsoft's shoulders.
5.9.2009 11:12am
zippypinhead:
There's a lot more going on "under the hood" of antitrust enforcement than meets the eye, that makes the intersection of IP and antitrust less opaque from the inside than it otherwise might appear. A few relevant datapoints, according to some of my hard-core antitrust/IP geek friends who follow such things for a living:

1. There's never or almost never been an FTC enforcement action against a venture following DOJ's issuance of a favorable business review letter, provided the venture is operating in the manner described in the review application (and if you materially diverge from the business review's operating assumptions without first seeking further guidance from DOJ, the FTC generally defers to DOJ per their standard interagency clearance protocol). The reason is fairly self-evident: existence of an explicit green light by the other Federal enforcement agency is powerful evidence that "reasonable minds can differ" as to the competitive effect of the described practice, and that's pretty devastating to any rule of reason case. It's unlikely that the plaintiff (whether the FTC or a private litigant) could get past summary judgment on those facts.

2. I am told by my antitrust geek drinking buddies that the professional antitrust bar is actually quite small and surprisingly specialized. And most of the big guns in the private antitrust bar have done tours inside one of the Federal agencies, and maintain pretty close relationships with the enforcers (in fact the antitrust bar is not unlike the proverbial small-town bar where half the patrons are cops and the other half are drug dealers they went to high school with, and they all constantly buy each other drinks). In addition to the business review process, full-time antitrust/IP practitioners are fully familiar with other guidelines that are often overlooked by general pratitioners, including:

-- FTC/DOJ Joint Venture Guidelines and Guidelines for Collaborations Among Competitors.

-- FTC/DOJ Intellectual Property Guidelines. These have been somewhat controversial recently and the antitrust geeks tell me that insiders can detect a few marginal differences between the agencies when they carefully study the 2007 update that followed a series of public hearings, but as long as you're not pushing the edges of the envelope, they still give decent general guidance as to where the safe harbors are.

-- Other helpful statutes like the National Cooperative Research and Production Act, 15 U.S.C. §4301, et seq., which not only operates similarly to business reviews for standards-setting, joint R&D and even joint production ventures, but also eliminates the possibility of treble damages in private litigation (thus all but eliminating such actions, since the difficulty of proving a RoR case without the potential treble damages payoff just isn't worth it).

-- Public speeches, reports, studies, and testimony by the FTC Commissioners and DOJ/ATR officials that are intended to announce substantive policy guidance (yes, speaking from a pure law enforcement perspective, using things like speeches in this way may sound a bit odd, but that's the way the antitrust regulators have long operated and the insiders are used to it).

-- Enforcement actions that (whether successful or not) announce what conduct the agencies think is clearly out-of-bounds. And counsel advising a venture is generally well-advised to pay attention to such cases. Thus, the lesson conservative counsel draws from the FTC's Rambus mess is don't play hide-the-ball with undisclosed IP interests while participating in standards-setting organizations, and the lesson to be drawn from the generic drug "pay-for-delay" cases is that the private contractual extension of market power, if coupled with blatant exclusionary intent, may come back to haunt you. The lesson of U.S. v. Microsoft is don't be a thug and use your IP to artificially maintain your market power like a mobster uses a gun. And the lesson of cases like FTC v. BMS, U.S. v. BMS, and U.S. v. Bodnar is that if you lie or cover up how you're really using IP to hose competition, it's going to cost you, and you might just go to jail...

Bottom line, according to my antitrust geek friends: While there are gray areas at the margins, skilled antitrust/IP practitioners are more than capable of advising their clients how to structure ventures to take advantage of established safe harbors and generally stay out of trouble. Or more cynically, the bottom line really is "don't hire a patent lawyer to give you antitrust advice..."

Hope all the links work - being a mere pinhead who doesn't get this InterTubes Thingie, I'm never sure. But all the stuff is right on the public DOJ or FTC web sites, so it's easy to find.
5.9.2009 11:15am
einhverfr (mail) (www):
Here is a counter-example though (a patent that was discovered and taken seriously, and why).

A couple of years ago, the PostgreSQL team became aware that the memory caching algorithm they were using, called ARC, was patented by IBM. The developers (at places like Red Hat and EnterpriseDB) quietly replaced the caching algorithm with one called 2Q (very similar to ARC but preexisting it and missing one distinctive feature: adaptive lengths to the cache queues). Then they told everyone to upgrade.

IBM issued a very unhappy press release because they felt they had been cut out of the loop of the process. "We wish you had talked to us first and maybe we would have given you the permission you needed" was the substance of the press release. However, the Pgsql team made the decision they did because:

1) PostgreSQL is licensed under a license that allows closed-source spin-offs.
2) They felt it was unlikely that IBM would grant an acceptable patent license covering those closed source spin-offs.

(Personally I think it would have been better to open discussions with IBM AFTER releasing the upgrade because if ARC WAS better, then maybe something could have been worked out without the threat of legal action hanging over everyone's head. Sort of a "negotiate with everyone, but from a position of strength" perspective.)

However, the basic issue was that the patent was addressed not because there was concern over IBM (which makes a competing product, UDB) suing the project out of existence, but because it might cause some additional liability for a specific segment of the community (companies like Green Plum and EnterpriseDB, not companies like Red Hat or Command Prompt).
5.9.2009 11:41am
Adam Mossoff (mail):
zippypinhead: Thank you for the additional thoughtful and substantive responses to my counterpoints. I really appreciate the explanation, and the links, of course, go far beyond the call of duty here. You've given me some great new material to chew on in my research, and so thank you again -- and please tell your "antitrust geek drinking buddies" that I owe them a beer if they're ever in the D.C. area. :)
5.10.2009 1:08am
einhverfr (mail) (www):
Speaking of stupid patents that IBM has.

Any of you who schedule 40 minute meetings are hereby forewarned: You are violating IBM's business process patents!
5.10.2009 2:14am

Post as: [Register] [Log In]

Account:
Password:
Remember info?

If you have a comment about spelling, typos, or format errors, please e-mail the poster directly rather than posting a comment.

Comment Policy: We reserve the right to edit or delete comments, and in extreme cases to ban commenters, at our discretion. Comments must be relevant and civil (and, especially, free of name-calling). We think of comment threads like dinner parties at our homes. If you make the party unpleasant for us or for others, we'd rather you went elsewhere. We're happy to see a wide range of viewpoints, but we want all of them to be expressed as politely as possible.

We realize that such a comment policy can never be evenly enforced, because we can't possibly monitor every comment equally well. Hundreds of comments are posted every day here, and we don't read them all. Those we read, we read with different degrees of attention, and in different moods. We try to be fair, but we make no promises.

And remember, it's a big Internet. If you think we were mistaken in removing your post (or, in extreme cases, in removing you) -- or if you prefer a more free-for-all approach -- there are surely plenty of ways you can still get your views out.