Chipping Away at the Presumption of Patent Validity

The Supreme Court granted certiorari today in Microsoft Corp. v. i4i Ltd., a case in which I signed on to an amicus brief supporting grant. While there are a lot of things the Court can do, the basic question is a relatively simple one: how much weight should the US Patent and Trademark Office (PTO) determination of patent validity count when the patent is later challenged in litigation.

This is an important question in an age where many think that patent quality is at a recent, if not all time, low. The current law presumes all issued patents to be valid, and consequently requires clear and convincing evidence of invalidity. This is theoretically a high standard, but more on that later.

At the other extreme, the law would give no deference to PTO determinations, and require the patentee to prove validity as part of the elements of a cause of action. This is not unheard of; trademark owners must often prove that their marks are distinctive, for example. Further, given that patent applications are presumed valid at the PTO unless an examiner can prove that the patent is not novel or obvious means that any granted patent always has a taint of not being fully vetted when it is granted. High patent grant rates - well over 60% or even  70% in some subject areas - support this position.

However, examiners often find prior art and reject patents or require amendments to narrow claims, and it makes little sense to have an examination system if patent grants are given no deference. Indeed, the U.S. had a system where all patents were granted in the early 1800's and it was rejected in favor of examination in 1836. 

The question, then, is just how much deference should be granted. Like many professors I know, I favor a system where patents are presumed valid, but that they can be invalidated with new evidence that was not before the PTO. The defendant would have to prove such invalidity by a preponderance of the evidence. If, however, the PTO considered a piece of evidence, then the patent would could only be invalidated by clear and convincing evidence.

This sounds fine on the surface, but is a bit more complicated than it might seem. If only one piece of prior art is at issue, the system works fine - either that reference invalidates the patent or it does not. However, many patents are challenged as obvious - where one prior reference does not teach the invention. In such cases, multiple references are combined or "general skill" is used to build on a single reference to fill in any gaps between the prior art and the patent.

It is not at all obvious (pun intended) how the burden of proof should be allocated in such situations. It is unclear why judge or jury should be in the business of second guessing a trained patent examiner determination about what is or is not obvious. Further, what if a new reference is combined with an old one? This might imply a lower burden, but if the new reference is similar to an old one, then courts will have to determine exactly what the PTO considered just to assign a burden of proof. 

Furthermore, sometimes examiners mentally consider and reject obviousness arguments without recording them in any detail. It will be very difficult to determine whether the PTO actually considered an issue if the record is unclear. Of course, patentees can pepper the record with confirmatory statements that the PTO rejected arguments, but such CYA submissions both increase the cost of the system and are unhelpfully self-serving.

I suspect that the Court will issue a ruling similar to its recent minimalist and flexible patent rulings that says something like, "Courts should give as much or as little deference as justified based on what happened before the PTO," and leave it to lower courts to figure out the difficult questions I pose here.

In the end, I wonder how much the burden of proof really matters. When it is a single prior art reference, I think not much. If something invalidates, it is usually clear to all involved once the court has determined what the patented invention is. There are cases on the margin where it is unclear if a prior art reference invalidates; such cases usually occur because some piece of information is either vague or unknown. If so, however, the only thing that would change results is to shift the burden over to the patentee, and that's never been the law (nor do I expect that result in this case).

Instead, a primary difference between the standards is about what evidence may be credited and considered. The clear and convincing standard often requires corroborating evidence. Thus if the prior art is uncorroborated (e.g. oral testimony about what a document said or how a product operated), then it might be sufficient to invalidate under a preponderance standard but not a clear and convincing standard.

This was the very issue in the Microsoft v. i4i dispute. The parties each presented testimony about how a prior product worked. Neither had corroborating evidence. The Federal Circuit held that Microsoft needed corroborating evidence to invalidate, and that i4i did not need such evidence to sustain its patent. Thus the jury's finding that the claim was valid (crediting the i4i witness) was easily affirmed.

Even so, it's not clear that the difference matters. Even if the burden were preponderance of evidence and even if i4i held this burden to prove validity, the court would have affirmed the jury's finding on competing oral testimony. That's what juries are for - resolving competing testimony that might conflict.  Even if the Microsoft witness were the only one, the jury could have disregarded the testimony if it lacked corroboration. Thus, it's not clear at all that changing the standard will help Microsoft in this case.

In contract, the standard should theoretically have a much stronger impact on obviousness determinations.  Lowering the standard would presumably make it easier for defendants to argue that patents would have been obvious. The "gaps" in the prior art can be more easily bridged under a preponderance standard.

Here to, though, a change in standard will not clearly make a difference, for at least three reasons:

1. Obviousness determinations are notoriously difficult to make, and so long a juries are allowed to make factual findings that underly the inquiry (the Court denied cert. on this issue a couple of years ago), then their findings will likely be upheld one way or the other, as they were in Microsoft v. i4i. Of course, the verdict form in that case was pretty generic, and the Federal Circuit presumed all sorts of facts that jury likely did not consider. Even so, such presumptions are unrelated to patent presumptions and are based on longstanding appellate review principles.

2. To that end, the jury instructions hardly imply an insurmountable burden. For example, the Northern District of California Patent Jury Instructions (now used by several districts) say this: "To prove invalidityof any patent claim, [alleged infringer] must persuade you that it is highly probable that the claimis invalid." I objected at the time that this instruction does not have nearly the same meaning as "clear and convincing," and I still believe that today. (I do note, though, that in Microsoft v. i4i, the court did say "clear and convincing" in the jury instructions.)

3. Since the Supreme Court's ruling in the KSR case, lower courts have had much more leeway to find patents obvious. Indeed, in the Court's obviousness jurisprudence through the years (like 150 years), it has found patents obvious with little more than hand waving, despite a clear and convincing standard. As a result, lowering the standard may not have much of an effect on court rationale at all. Of course, this may just imply that the standard never was clear and convincing.

In sum, it makes sense in theory to lower the standard of proof where the PTO has not considered an issue. In practice, however, I think that doing so will make very little difference except in a few special cases where the evidence is right at the margin and it is the presumption and nothing else that is protecting the patent.