We have just heard of a proposed settlement of an EC antitrust action against a private company named Rambus. Rambus has reportedly tabled an undertaking to reduce its royalties for DRAM patents, which would lead, technically, to a decision according to Art. 9 of EC Regulation 1/2003. The Commission has published it for market testing, inviting comments by the middle of July. Although the proposed licenses and royalties still raise questions, it is in the line of successes of EU antitrust activities where USA initiatives have most remarkably failed. But more importantly, it is the first signal to the world of standards that the recent free rides by some players in that field are going to be under vigilant scrutiny of the antitrust authorities.
Protecting the standards
Ghosts haunt the standardization process. They go by several names and come in different forms: “standards abuse”, “standards hijacking”, “patent ambush”, “royalty ambush”, “patent trolling“. The standardization world has never been so much under fire. Some companies try to bend the standardization process to fit their own selfish interest, without any regard for the common weal. Some others just sit and wait until some of their patent claims are “necessarily infringed” by a standard, the industry is locked in, and then pass the hat to collect the high toll that standard-abiding companies are forced to pay, in spite of the licensing rules of the standard setting bodies (SSB) that would require Reasonable And Non Discriminatory conditions (RAND) as a prerequisite for inclusion of any patented contribution into the standard. Others do the same, but in addition they actively seek to seed the standards with their own patented technology. Some of them discriminate in their licensing, to give their own downstream business and advantage, and to restrict competition in standard-compliant products.
One of the most colourful disputes in the panorama of standardization abuses concerns Rambus. The company has allegedly remained silent in the standardization process for SDRAM, failing to disclose the relevant patents it held over the technology under discussion. Was it an intentional act or an ex post decision? Many evidence suggests the former is the case. It appears that when disclosure became inevitable in the standardization process, Rambus left JDEC (the relevant SSB), arguably in an attempt to avoid SSB's IP policy rules. But allegedly it kept some insiders in the works of the SSB, which allowed it to shape its patent applications, also arguably to be in a position to charge on the implementers. In fact, when the SDRAM standard was issued, Rambus demanded royalties on the patents it held. Litigation esued. Among other initiatives, some competitors pressed antitrust charges. They argued that Rambus acquired market power by keeping silent when it had a duty to speak, and abused that market power by charging royalties that it could not have charged had it not breached its duty to speak. Rambus argued in response that their technology was so good, it would have been chosen anyway. If Rambus really thought that, why did it not disclose its patents? Was it because they knew that the SSB would select a different, royalty-free technology?
Standards are a sport for gentlemen
Why do I care? Standards are one of the most beneficial and honorable institutions in the world, and they allow a great deal of simplification in everyday's life. From railroads and aviation to electricity, from paper size to the Internet, innumerable parts of our life are made workable by standards, of varying complexity.
Standards are a win-win game, if all players abide by the rules. The rules must be made to ensure that anybody is playing fair, and those rules must exist within the standardization process as well as outside the standardization process.
For a standard to work, everybody in the market should volutarily adopt it, because working on different standards is a guarantee to non interoperability and high inefficiency. Sometimes different standards merge in order to resolve interoperability problems. But the game of standardization is easily disrupted, because in order to work, a standard must be neutral, not be favouring one entity over another, it shall proceed by consensus on a good faith cooperative effort. Which is difficult to achieve, because many competing entities must agree upon a common line. Many are the standards, a comparatively small fraction of them are successful. Some are utterly successful, the largest majority remains on the paper or little more.
But some companies try to get a free ride. While big players have often more to lose than to gain from playing dirty, some “technology-only companies” feel less constraints in deciding what to do. Sometimes big players become fringe players, and might decide to play less nicely, or in some cases their portfolio is just put on the market in a bankruptcy proceeding and sold to a litigation company. And in some cases, they base a business model on controlling and milking standards. So the standard game could be turned into a big win (for the merry few) and a big loss (for the rest of us).
I have already mentioned the need to achieve consensus in order for a standard to work. Cooperation and “democracy” flow from that pricinciple. Also vendor independence is an advantageous and willed consequence, as nobody would agree upon a standard that would favour only a competitor. There are exceptions. In some cases, competitors strive to impose their own competing specifications on the market, to make them de facto standards, and we have a classical example of two standards–bad, as I wrote in an earlier article of mine (“There Can Be Only One”, see there for more background). In other cases, one competitor tries to bend the standardization procedure through hijacking the approval process. We all know about a quite recent example of this.
The case of Rambus is different.
The US Federal Trade Commission (FTC) was of this opinion. was of the opinion that Rambus engaged in an antitrust violation. But after a decision of the Court of Appeal, which was contrary to some findings of the FTC and arguably wrong, when the US Supreme Court refused to look at the case upon a certioari motion, the FTC decided to drop the charges, leaving the antitrust issues to private litigation. The European Commission could have followed suit, but it has not, fortunately.
There are some people who argue that these standards ambush should not be addressed by antitrust rules, but by standards organizations themselves or by Government regulation. These arguments may be well-intended or may be designed to “bury” the problem in a mountain of regulatory paperwork. But I worry that this is totally unrealistic. As I said, SSBs work by consensus. This means that the companies that engage in these kind of strategies are perfectly capable of blocking new rules by the SSBs to prevent these abuses. So it seems that antitrust is the best and probably the only way to solve this: establish a few strong and clear precedents that give a clear signal to the market, and then perhaps regulate. Antitrust rulez!
Kudos to Commissioner Kroes, once again!
While the European Commission and Rambus do not agree on the allegations, a final decision has not been made, and it can be argued that the propsed settlement would allow Rambus to benefit from its breach of the rules (charging less, but still charging when it could not have charged anything had it complied), it is nonetheless important that this case has been brought close to an end with restrictions on Rambus. This suggests that the application of antitrust to standard abuse is not beyond the realm of possible at all. The case could indeed be a convincing precedent for followup initiatives by the Commission.
Patent technical vs. legal exploitation, more failure, more remedies
We are not safe home, though. There are other businesses who just wait and see when their more or less innovative technology is incorporated (or mandated) in industry standards and then start demanding royalties. Do these deserve the title of “patent trolls”? “Patent trolling” refers quite pejoratively to the questionable practice of collecting patents for no other purpose than for litigating against those who by chance infringe a part of them and have become dependent on them. Others call it “hold-up” (“your money or your business!”). Technology management companies — I am sure this naming convention would be more acceptable to those companies — most of times avoid seeking active licensing before a product embodying their patent hits the market. They keep quiet until their victim is well and truly locked in. The game theory again explains what happens quite easily.
Ex ante, or beforehand, negotiation is a guarantee of fair dealings, because it occurs when the SSB (in case of standardization) has choice, and in general before huge investments are made and the choices are irrevocable. Because alternative strategies exist,the bargain power of the technology provider depends roughly on the appreciation of the inherent value of the technology, on how it fits the needs of the prospective client and the price that is demanded. The patent owner can never ask for more than the benefit that the licensee can derive from using the patent instead of the next best alternative. Ex post negotiation is just collecting ransoms. It is like negotiating with a handgun clearly sported on the table: injunctions, seizure of goods, treble damage awards, attorney fees, etc. The paradigm is not what you gain from the technology instead of the next best alternative, but how difficult is to get rid of the product, and how deep the cul de sac where the defendant has put itself in, most of the time innocently. In a holdup, the victim has a stark choice: hand over all the profits (perhaps keep a tiny bit) or get out of business.
I said “innocently”, yes. There are obscure outfits which just collect patents, whose description and claims are often very obscure and difficult to parse, without making any use of them beyond “licensing”, most frequently in a contentious setting and ex post.because they are kept invisible in the form of secret patent applications (using what is called “patent continuation” procedure to delay publication) until the entire industry is locked in, and only then are they published. As in WW2 they just await for a vessel to pass by and attack it without an early warning. Whether this practice is legal or not — I have some doubts at least for the most extreme cases — it is clear to me that it should not be allowed at all. If patents have some utility is because they teach something, they contribute to the advancement of technology in an open system where knowledge is freely achievable and its implementations knowingly made after having negotiated with the relevant rightsholder a fair price. This is why they are registered and published: in order to bring innovation to the market. The manufacturer has the option to implement a given innovative technical idea in its products or to find a way to invent around if licensing from the patent holder is inconvenient on whatever account. Clearly defined patents, limited and clear-cut protection, small number and high quality of granted patents are the prerequisite for this ideal scenario, which makes the technology advance. Patents are not certainly conceived to increase the level of litigation by exploiting them in court rather than on the market. Sometimes I have a feeling that the innovativeness of patents are more in the cleverness of the attorneys finding a way to describe a problem and obfuscate the solution rather than in the protected idea itself, of course when the patent is not outright ludicrous.
Any day more cases are reported of companies attacking existing implementations of standards. The quarrel around the royalties demands of Qualcomm concerning some patents for the 3G standards is very close the case of Rambus. Qualcomm did not hide its patents, but is accused of misleading the SSB about how much it would charge, and making the companies pay through the nose after lock-in (and monopolizing the downstream market as well). Or look at the litigation activities by a patent portfolio management company named IPCom GmbH (not to be confused with IP.com Inc.) are reported. Other cases concern MP3 patents and affect even big and overcautious companies (when it comes to due diligence their technology) such as Microsoft — which has been slapped with a billionaire award for infringing some Alcatel MP3 patents while the company was reportedly in good faith believing it had already licensed them from Thomson Licensing. Those cases and others sprouting here and there suggest that the standard and the patenting system need a fix.
I am still undecided on whether a system granting patents for inventions has any sense in the current era and on whether the current regime is at pace with the times. What I feel reasonably safe assuming is that cases like those quickly examined here are an evidence that the system is not par with its charter. The earlier the standard setting bodies adopt a serious, binding and irrevocable patent discolusure and licensing policy (like for example the one from W3C, not perfect, but already something), the better for the standardization's sake. Unfortunately, the most relevant SSB are far from entering such a sound policy, for a number of reasons most of times independent from the will of their officers.
And yet I have not mentioned the software patent issue at all, which deserves an entire separate discussion.
A very good insider's introduction to the interference between patents and standard can be found at http://www.consortiuminfo.org/ipr/ , authored by Andrew Updegrove of Gesmer Updegrove. The article shows how over-simplistic my recount of patent problems with standards is.
Commentary: Richard Wolfram, “Analyze This!” Deconstructing Rambus Following the Supreme Court’s Denial of Certiorari – The Mechanics of How the D.C. Circuit’s Decision ‘Jumped the Tracks’.
“The Gridlock Economy: How Too Much Ownership Wrecks Markets, Stops Innovation, and Costs Lives”, by Michael Heller.