By Jason Rantanen
The number of patents involved in litigation
The number of patent suits filed each year is well known. But what about the number of patents involved in those cases? Mark Lemley and others have estimated that about 1-2% of all issued patents have ever been litigated. To test this estimate, I searched the LEXIS utility patents database using the LIT-REEX field to identify all patents that were flagged as the subject of litigation filed in the year 2009.* I ended up with 3404 patents, or approximately 2% of the average number of utility patents that the PTO issued annually over the previous decade.
This result provides further confirmation of the conventional belief that the vast majority of patents are never litigated – especially since I suspect it overstates the percentage of patents involved in litigation, as it assumes that the litigation flag is limited to just infringement suits (as opposed to inventorship disputes, for example) and it does not take into consideration the fact that some patents are the subject of multiple suits filed in different years.
How old are these patents?
The below chart shows the age of the patents involved in suits filed in 2009. More than half (55%) of these patents were at least five years old as of the litigation filing date; 24% were over ten years old.
What about reexamination requests?
The above chart also shows the number of patents involved in suits filed in 2009 that have ever been the subject of a reexamination request. Of the 3404 total patents, LEXIS reports that reexams were requested at some point for 578 (17%). Many of these requests were filed in 2009 or later (likely as part of a litigation strategy): 398, or 11% of the total number of patents involved in suits filed in 2009 were the subject of a reexam request in 2009, 2010, or 2011.
*I used 2009 because the 2010 and 2011 results were affected by the numerous false marking suits filed in those years.
There you go again with the “syncophant” rhetorical nonsense again, MM. Frankly, when you resort to personal insults and invectives like that, that tells me you’re not interested in a rational debate.
You also don’t seem to know when no one is interested in playing your “game” anymore. As I said before “go pound some salt,” it might make you feel better. And please do take your bat and ball, and go home!
MM never tries to answer questions put to him.
That’s an easily disproven lie, sockie. In fact, I answer far more questions here than you ever did or will. Shall I disprove your lie, sockie, with direct evidence? Let me know. That’s a question for you.
That’s why it’s ironic that he is imploring for answers to his question.
No, sockie. The “irony” is that the biggest blowhards in the patent blogosphere (Gene Quinn and Kevin Noonan) refuse to answer a very simply and straightforward question about patentability, eligibility and enforceability of claims reciting a single, old patent-eligible step and a single step of thinking a new, non-obvious thought. That’s the “irony,” sockie.
Of course, the “irony” disappears once you take a few seconds to understand that they refuse to answer the question because it raises conflict issues which they also do not want to discuss. Isn’t that right, EG? Oops, I forgot that the syncophants never betray their heroes.
MM’s reply does not make any sense. The point is that MM never tries to answer questions put to him. That’s why it’s ironic that he is imploring for answers to his question.
Couldn’t have said it better myself, anon.
But everyone who reads the comments knows, EG, that I’ve always tried to answer any questions you’ve asked me. Certainly that’s been the case with respect to this 101 issue which you (and your heroes, Gene and Kevin) keep pretending doesn’t exist.
Couldn’t have said it better myself, anon.
“Wasn’t one of the points made”
Additional credit Paul F. Morgan, as this point (limited timing) was made by you prior to Ned’s posted reply to you.
“If your position is that there is no insight to be gained from any calculation that includes errors, then fine. But you should probably stay away from most technical disciplines, then.”
Nice strawman Leopold. No, my position is not that there is “no insight”, but rather, that “insight” from a poorly constructed and known error induced protocol, shaped for unstated purposes, with an unsupportable basis of “simplification” as being better merely because it is simple, is dubious at best.
Your implication of staying away from technical disciplines, then, is baseless, for I know of no technical discipline that willingly engages in such avoidable error. Yes, simplification occurs and certain known minimal error may be countenanced, but most definitely not the type and degree of error you seek to introduce.
So, once again, I call on you to understand the posts below and recognize that your math is off. Your induced subjectivity of future litigation is unsupported at best, and fallacious at least.
If your position is that there is no insight to be gained from any calculation that includes errors, then fine. But you should probably stay away from most technical disciplines, then.
“average annual number of patents that I used on that figure for the purposes of this post”
What is the purpose of this post?
What do you mean to say with “average annual number of patents”?
Some posts indicate that your number of patents is understated, why choose only a decade? Why are you annualizing the decade’s number?
I guess it comes down to the thoroughness of the data mining beyond the AO 120 forms LEXIS gets from the PTO. If they are diligent, then the LIT-REEX database is probably accurate. It’s hard on the outside, though, to know how thorough they are.
I may be missing something, but the number of utility patents that the PTO issued between 1998 and 2009 remained relatively constant in the 150-160,000 range, see link to patentlyo.com, so I think it’s a fair assumption to base the average annual number of patents that I used on that figure for the purposes of this post. Your point that some of the litigated patents come from a period when the PTO was issuing only around 100,000 patents a year (1988-1997) is a valid one, however, and I could do a more refined analysis that broke out the number of patents issued each year that were litigated in 2009. That said, I would not expect the result to change much, given that those pre-1998 patents made up less than 24% of the total.
Ultimately, though, as with any empirical study, this one comes has its limitations. At best it presents a snapshot based on what occurred in 2009. Perhaps, if we were able to look back to 2000, we might find a higher percentage of issued patents being litigated. At the present time, however, I don’t have that data available.
Leopold,
Your math is off – see the posts below.
Your attempt at including “or will be is fallacious and introduces an impossible “guess” factor into otherwise solid objective numbers (it is what it is). Simple calculations aren’t better because they are simple if they are wrong.
“Better examination would make patents a more definitive and less debated property. People only venture into litigation when they believe they have a chance of winning that litigation”
+1
Do it right the first time.
+1
Prof Rantanen, if you are using an annual number for litigation data, and are attempting to reach a statement for percentages of more than those patents litigated in that particular year, your logic and your math will fail you. In any given year, all live patents are subject to be asserted through litigation.
The annual numbers will change, but taking a straight line approximation will readily induce errors given the “horrific” (for some) increases year over year in issued patents. The net gain in number of active patents is not steady state.
Actually, I believe that even more than live patents may need to be counted, given the six years of past infringement coverage – for naturally expired patents – but I am shooting from the hip on that.
“I just want to make sure everyone remembers the exact point in the discussion”
EGAD – If we were to attempt to do the same with you MM on all of the unanswered questions put to you, the blog posts would triple (it would be like having the whole world – but where would you put it?).
“recognize the impossibilty of co-mixing compact prosection with BRI as currently practiced“.
Fixed.
I think that BRI (or at least the wanton and reckless application of it) is ripe for fixing. I don’t think the Office needs to wait for legislation to effect such changes either.
“When a patent goes into the assertion mode from the decorate the wall mode, the patent owner can then spend the big bucks on a search and can now polish the patent with a supplemental examination if necessary”
Wasn’t one of the points made about supplemental examination the fact that the period available for supplemental examination does not match up with the timeframe (age) of an asserted patent?
In other words, the time for polishing is long past when the time for assertion arises (at least, I think that’s the point being made).
Can’t seem to let go of that “old” question can you, eh MM?
I could “let go” of it if I wanted to, EG. I just want to make sure everyone remembers the exact point in the discussion where you and your heroes Kevin and Gene (and Denise, too) suddenly clam up and turn into frightened children, afraid to confront and address very simple facts about what is and is not eligible/patentable, and why.
Can’t seem to let go of that “old” question can you, eh MM? Did Santa put coal in your stocking this year? Or maybe the Grinch stole your Christmas, possbily? I suggest “pounding some salt” might help you clear your mind of it so you can celebrate the New Year.
Paul, we have had some threads here on the topic of IDSs, where it is noted that examiners rarely cite any art in an IDS, even if, in the opinion of the applicant, that art is better than the art the examiner uses. This seems to indicate that spending big bucks on searches is a waste without a more effective way of focusing the examiner’s attention on truly relevant art. I factor into this notion the fact that BRI at times focuses the examiner’s attention in the wrong direction leading him to ingore the art submitted by the applicant.
When a patent goes into the assertion mode from the decorate the wall mode, the patent owner can then spend the big bucks on a search and can now polish the patent with a supplemental examination if necessary. If the patent reaches the assertion mode and the best art has not been considered by the examiner, it is too late. The claims may have to be reexamined after assertion, thereby invoking intervening rights if the claims are amended to narrow them.
I think supplemental examination may help the patent owner to fix errors not only by the applicant, but by the examiner. But I would also focus on fixing the original ex parte examination by adjusting examination practice to recognize the impossibilty of co-mixing compact prosection with BRI.
The explanation here, like the math, is found wanting.
EG do you have a valid and accurate database to work from (i.e., the facts) to make rational conclusions from (i.e., the interpretation)?
Are you ready to discuss “facts” and “rational conclusions”, EG, or are you still deeply invested in shilling for the eligibility/enforceability/validity of [oldstep]+[newthought] claims, regardless of “facts” and “rational conclusions”?
I see that I misattributed this effort to Dennis – thanks for the post, Jason.
According to LEXIS, the notice of litigation entry in the LIT-REEX field is populated both through notices from the district court that are filed with the PTO and its own mining of docket notices and pleadings. Because the latter is based on electronic filings, it’s going to be more comprehensive the later in time one goes. I used 2009 because by that point many district courts were using ECF, and thus it provided the most comprehensive set of data before getting into the false marking case years. That said, one weakness of using this source is that it may still be undercomperhensive.
One other thing to note: I was also told that LEXIS is constantly updating this field as additional electronic data becomes available. So it is possible that the 2009 number may vary if additional cases are added.
I used the average annual number of patents issued because the litigation data was for just a single year. Comparing the number of patents litigated in a single year to the entire population of issued patents (or even just those currently in force) would only tell us the percentage of those patents litigated in that particular year.
The purpose of this analysis was to look at the percentage of issued patents that are subject to litigation. You raise a good and separate point, however, that the percentage of patents currently in force that are litigated is necessarily higher than those that issued, and I’ll think about whether there’s a way to get at that question.
“compared to the great economic burdens and increased public product costs for those sued for patent infringement. Hence, the AIA concentrated on new reexaminations and litigation issue simplification for those few patents of real importance‘
Exactly the wrong message to take away – thanks for drinking the kool-aid (the major assumption made is that only “real important” patents are ever litigated.
Totally baseless assumption.
Better examination would make patents a more definitive and less debated property. People only venture into litigation when they believe they have a chance of winning that litigation (and thus, one can infer that murky, less well examined patents) drive the litigation “nightmare.” Focusing on after effects (blown out of proportion to the typical patent), rather than on something that would better all patents is simply the wrong direction to go (and have gone).
These statistics may not be perfect, but they are about the best that can be obtained from the available data, and a lot better and more useful than mere speculations.
Yes, improving examinations of all patent applications would be desirable, but it does not make economic sense to raise patent application fees further to give much better examinations to the roughly 98% of patents that are never litigated compared to the great economic burdens and increased public product costs for those sued for patent infringement. Hence, the AIA concentrated on new reexaminations and litigation issue simplification for those few patents of real importance. Run the numbers. For example, if you required a fairly decent $20,000. prior art search for each of the roughly 500,000. patent applications per year, that would cost a rediculous amount more per year. And the amount of prior art needing searching is more than doubling every ten years.
I’m not sure I follow you, Feed. If 98% of patents are never litigated, doesn’t that suggest that improving the error rate during prosecution is less important, rather than more important, compared to a scenario where, say, 10% are litigated?
At the very least, one can say that 98% of patents do not face the many issues that seem to raise alarms with the general public (patent trolls, Eastern Texas forum shopping, excessive litigation awards, inequitable conduct charges, and on and on).
Perhaps if this fact were more in front of those minding the shop, “patent reform” would actually look at issues of more importance, like how to get more effective and efficient (i.e. decreased Type I and Type II errors) patent application examination.
You correctly point out the problem of such statistical analysis, anon: do you have a valid and accurate database to work from (i.e., the facts) to make rational conclusions from (i.e., the interpretation)?
…soooo, since patents do not expire annually, the real number is more like 0.2 % (generously).
Uh, in fact, somewhere well north of 100,000 patents do expire each year, including in 2009, the year captured by Dennis’s calculation. That number was likely significantly less than the 10-year average annual issuance figure, but Dennis’s rough estimate is much closer than yours, and his approach is just fine, even though it assumes a steady state for litigation and issuance.
To redo Dennis’s calculation slightly differently:
20 years worth of patents is about 20 * 150,000 per year, equals a rough estimate of 3,000,000 in force in . That’s on the high side of course, since the 90’s weren’t as “productive” as the 2000’s, and many patents have undoubtedly lapsed early, but a high estimate makes the final number conservative, since the 3M is going in the denominator.
20 years worth of patent litigation = 20 * 3404 = 68080 patents in suit. This is on the high side if you’re only looking backwards, since there’s been a pretty good ramp up of litigation in the last decade. However, if we’re looking at all patents in force in 2009, then you need to project forward as well. So unless you think that patent litigation trends are going to reverse over the next decade, then this is a pretty decent estimate of the number of patents active in 2009 that have been or will be in suit.
68080/3000000 = 2.27%. So, roughly 2% of all patents-in-force in 2009 are or will be litigated.
Dennis’s calculation was much simpler: 3404/150000 = 2.27%
“”approximately 2% of the average number of utility patents that the PTO *issued annually*over the previous decade.””
The mathematical problem with this is that it is not 2% of all existing and unexpired patents. That number has yet to be calculated. If one wants to manufacture a number, the one proffered here is an inacccurate and misleading comparison of one year of patents versus a larger body of patents that may be asserted.
I enjoy the blog but not this type of weak mathematical content.
I think the comment has to do with the hysteria that “trolls” are running amuck due to the “giveaway of patents…
(when more than 99% of patents do no such thing)
Paul, my research indicates that the AO 120 forms fail to catch an additional 38.7% patents litigated as compared to those reported. Actually, I made that up. That’s more or less my guess, though.
The problem of non-reporting and under-reporting appears to be most acute in judicial districts that see relatively few patent cases. I suspect that the clerks in these districts are often unfamiliar with the requirements of 35 USC 290. More than a few times, a clerk has forwarded a patent appeal to a regional circuit, even though the appellant was very clear in the notice of appeal about how the appeal was to the Federal Circuit.
PTO Statistics: In FY 2009 [your study year] 258 inter parte requests were filed of which 68% were “known” to be in litigation.
That would mean that most of them were already in litigation before the reexamination was requested, and relatively few were filed by anyone before they were sued.
And there would have been more reexaminations if patents from applications filed before 11/29/99 had not been excluded from current inter partes reexaminations. [Which will end later next year with the new AIA inter partes reexaminations, although the PTO is apparently going to cap the number it will accept.]
?
Thanks for this study.
Of particular interest for the coming AIA Post Grant Reviews, which are only available for the first 9 months after a patent issues, and seem only likely to be used [if at all] by someone being sued, how may patents did you find being sued on within only 9 months of their issue dates? I did a quick LEXPAT sample review several years ago, and it looked like a very small number. What did your study find?
I agree with Andrew Dhuey above that there has been a problem with some District Court clerks failing to send in patent litigation notices to the PTO, but with with 3404 patents identified as sued on in 2009 they must be catching most of them?
But with so many of the patents being sued on being older patents, that must be more than 2% of the un-abandoned patents being maintained in force by payment of maintenance fees?
“approximately 2% of the average number of utility patents that the PTO issued annually over the previous decade.”
…soooo, since patents do not expire annually, the real number is more like 0.2 % (generously).
I wonder about the completeness of the LIT-REEX database. I don’t know the source of it, but it might be that the LEXIS people go to the PTO and scan the AO 120 forms (Action Regarding a Patent or Trademark) district courts send in when a patent case is filed. What I discovered in my research in FLFMC v. Wham-O, however, is that very often the court fails to send in that form or sends it in with missing information.
I don’t question the main point about how relatively few patents ever get litigated, but the numbers might be significantly higher than the LIT-REEX database indicates.
Comments are closed.