Tuesday, October 16, 2012

Google opposes fixed license fees for SEPs, insists on percentage of end product price

Last week I mentioned that Microsoft and wholly-owned Google subsidiary Motorola Mobility filed motions in limine (motions to exclude testimony and arguments) in their FRAND obligations lawsuit in the Western District of Washington. I wanted to await the parties' opposition briefs (just like in a parallel Apple v. Motorola FRAND case in Wisconsin) and, for the time being, commented only on an interesting part of Microsoft's proposed findings of fact, which related to a pirated Katy Perry video.

Late on Monday by local time, Microsoft and Motorola filed their responses to each other's motions in limine. In my opinion, the two most interesting items are

  • Google's (Motorola's) fundamental objection to per-unit fixed-rate license fees for standard-essential patents (SEPs) and

  • Google's (Motorola's) attempt to exclude from next month's trial any references to a 2003 valuation of its IEEE 802.11 patents that was performed by a consultancy firm, InteCap (later acquired by Charles River Associates), on Motorola's behalf.

There's less that I can say about the second item than the first because most of the information is redacted. But in the publicy-accessible part, Motorola at least concedes that "the 2003 InteCap study was intended to explore possible licensing opportunities during the infancy of 802.11 deployment in the market". Microsoft argues that Motorola's fight against the use of that patent valuation study conducted on its own behalf is just another example of its attempt to capture hold-up value of patents. Judge Posner and others say that the correct approach to SEP valuation is to determine the pre-standardization value of those patents, so as to eliminate post-standardization hold-up value.

This week I became aware of a very interesting empirical study conducted by three European researchers (Rudi Bekkers, RenĂ© Bongard and Alessandro Nuvolari) on "the determinants of essential patent claims in compatibility standards". Their key finding is that "the involvement in the standardization process is a stronger determinant than the technical value ('merit') of the patent". I've watched many SEP assertions and I've been consistently underwhelmed by the technological merits of the SEPs-in-suit. This study confirms what I felt, and this is exactly why a 2003 valuation of Motorola's IEEE 802.11 patents is, in general, more likely to reveal the FRAND value of those patents than any hypothetical negotiation approach based on today's relevance of the IEEE 802.11 standard.

The first item listed above has to do with the Entire Market Value Rule (EMVR), which says that unless your patent drives demand for the product as a whole, your demands can only be based on the smallest saleable unit your patent reads on. It's a patent damages theory, and Motorola has been saying for a long time that damages theories shouldn't apply to FRAND rates, but the EMVR provides a framework for solving the royalty base problem.

Not only will the court have to set a FRAND rate (both a range and a point) for a Motorola-Microsoft SEP license deal but in a subsequent step, which based on Judge Robart's current plans would involve a jury, it will also have to determine whether Motorola's original demand (2.25% of the price of the relevant end product, such as an entire computer) was so outrageous that it cannot be considered a good-faith deviation from FRAND. In order for this comparison to be made, be it by a judge or a jury, some normalization will have to occur: one will either have to convert Motorola's percentage to a dollar amount and then compare it to a reasonable dollar amount, or convert a reasonable dollar royalty to a percentage of the price of an end product and compare it to the percentage Motorola demanded two years ago (and ever since, though it has already backtracked on the royalty base in a PR stunt styled as a settlement proposal back in June). Normalization is the only mathematical step that makes sense to enable the comparison that has to happen. But Motorola doesn't want this discussion to be too mathematical. Mathematics is an exact science, while the law is not. In mathematics, 1+1 equals 2; in the legal world, you can at least try to find an expert who will say that 1+1 equals 3, or argue that it cannot be ruled out that it might equal 3 under some unreal circumstances. If you have to defend a totally outrageous royalty demand, you want to argue verbally, not numerically.

In an effort to avoid the EMVR and a comparison of normalized rates, Google (Motorola) even claims that "[r]equiring licensing parties to strictly adhere to the EMVR would effectively preclude licensees and licensors from using experience and their sound judgment in efficiently negotiating a license". This is an exaggeration. If two parties can agree on a license deal without needing help from the courts in determining one for them, there's no problem with any royalty base they use, as long as both believe it's a fair deal. But once there's a situation in which the courts have to set a rate or analyze a demand that was made, the proper royalty base must be applied lest the derivation be inherently incorrect.

In a recent submission to an ITU roundtable, Nokia argued that "as 1% of a base of 100 leads to the same result as 10% of a base of 10, which royalty structure is actually used has no economic relevance, as long as the result reflects the appropriate value for a license". This is pretty much also Motorola's argument. But Nokia's proposal makes sense only if two parties can agree on a license deal without a legal dispute. If they can agree, let them agree on anything. But if there's no agreement, and if we assume that Nokia's example results in a rate that is FRAND in that fictitious scenario, arriving at 10% from a base of 10 would be a correct result by design, while arriving at 1% from a base of 100 would be a correct result by mere coincidence. In a legal dispute over FRAND rates, however, the proper rate must be calculated by design. That's why companies making royalty demands in situations that could lead to a dispute have to apply the proper royalty base or be held responsible for using an incorrect approach that has more to do with hold-up value than with pre-standardization technological merits.

Apple once argued in a letter that all car drivers pay the same highway toll, whether they own a jalopy or a new sports car. That's because a highway toll is meant to be compensation for use of a road, not a tax. If the idea is taxation, then one can certainly argue that those who can pay more should make a greater contribution to infrastructure shared by society at large. However, industry standards are not meant to authorize SEP holders to impose private taxes on the rest of the industry. The whole idea of FRAND rates for SEP licenses is to ensure that right holders contributing to standard-setting will have an economic incentive to do so and will not be disadvantaged as compared to a scenario in which their patented inventions are not included in a standard. As a result of inclusion of their technologies in a standard, those right holders are provided with a market potential that is usually many times greater than if they marketed those technologies on their own outside of a standard. That should be enough of an incentive. But if they additionally had the right to charge more on a per-unit basis than otherwise, the whole system would be completely out of balance, and innovation would be stifled.

If we rejected the notion of a private tax on implementations of standards and stay focused on the idea of FRAND licensing, fixed per-unit license fees make a whole lot of sense. If Apple buys computer memory, it also doesn't pay more on a per-unit basis than a maker of low-end devices (unless the latter gets a volume discount, but even in terms of volume it's hard to beat Apple these days). Chip makers can't tax computer manufacturers -- they can simply sell. Intellectual property used in a smartphone, tablet computer, gaming console or operating system is also simply a component that you can buy and incorporate into your product.

The most illogical paragraph in Motorola's opposition to Microsoft's motions in limine is the following one, which argues that fixed rates are unrelated to the EMVR, while they actually are perfectly consistent with it:

"What is more, Microsoft's reliance on patent pools exposes even more negotiated licenses that are not limited by the EMVR. The MPEG-LA and VIA pools relied on by Microsoft -- while irrelevant to [F]RAND for other reasons -- nevertheless include a fixed per unit rate regardless of the product's price or whether the 802.11 or H.264 functionality is a primary or minor feature of the licensed product or is the basis for customer demand. Thus, the fixed rates of pools -- on which Microsoft relies -- are further real-world evidence that negotiated licenses regularly use end product price as the royalty base, having nothing to do with the EMVR."

There's no contradiction between the EMVR and fixed rates. For the reason I explained further above, the requirements of the EMVR cannot be satisfied by merely lowering a percentage that uses an incorrect royalty base. That approach must be rejected just like a spreadsheet refuses to mutiply an ERROR field with a percentage. But a reasonable percentage of a reasonable royalty base complies with the EMVR, and there's nothing wrong -- nothing whatsoever -- with simply multiplying a correct percentage with a correct base and using the result. And that's what the fixed-rate approach of MPEG LA's and Via Licensing's pools effectively does.

It's unbelievably wrong to say (as Motorola's brief does) that "fixed rates [...] are further real-world evidence that negotiated licenses regularly use end product price as the royalty base". Fixed rates are like the per-unit chipset prices that Apple pays for the memory it incorporates into its computers and other devices. They are, by definition, related to the smallest saleable unit. If they were in any way related to the "end product price", the rates wouldn't be fixed. Obviously, even a fixed rate could be outrageous. But that would be a problem of the percentage, not of the royalty base. And market acceptance of the pools in question indicates that those fixed rates work for makers of low-cost and expensive devices alike.

I doubt that Google (Motorola) will impress the judge by stating in one sentence that those pool rates apply "regardless of the product's price" and then claiming in the very next sentence that those rates "use end product price as the royalty base".

I have no idea how Judge Robart is going to adjudicate the parties' Daubert motions and other motions in limine. Since the November 13 trial will be a bench trial, he doesn't have to worry about juror confusion. If there's something he doesn't buy, he can throw it out now or he can let a party present such an argument anyway even if it's unlikely to succeed. The EMVR issue is a key one that could be addressed ahead of the trial, and it appears ripe for a decision.

If you'd like to be updated on the smartphone patent disputes and other intellectual property matters I cover, please subscribe to my RSS feed (in the right-hand column) and/or follow me on Twitter @FOSSpatents and Google+.

Share with other professionals via LinkedIn: