“The sheer number of patents in the U.S. is fueling frivolous litigation and drastic action is needed to make patents more difficult to obtain and easier to invalidate.” according to Judge Richard Posner of 7th Cicuit Court of Appeals, in comments made at a recent forum with retired Federal Circuit Judges Arthur Gajarsa and Paul Michel. Judge Posner was particularly critical of the glut of patents on relatively minor software innovations, which “creates confusion and uncertainty about what one can do without running into patent minefields.”

Judge Posner’s view is that most patents are unnecessary because they grant a monopoly to exclude others from practicing what are often incremental improvements that take little time and investment to develop. Why, in other words, grant a monopoly on an invention unless the invention wouldn’t have been made without giving the inventor a monopoly?

While the other two judges agreed that the number of low quality patents that are issued indicates that the PTO has problems that need to be addressed, they were unwilling to go as far as Judge Posner in calling for major reforms to the patent system, such as eliminating the presumption of patent validity and shortening the 20 patent term, at least in technologies like software that have a very short halflife.

Judge Posner’s comments will resonate among many patent and industry observers. However, it may be unfair to blame the “patent minefield” problem on the PTO alone, especially given the rise of patent-troll litigation in recent years. This note takes a look at current PTO statistics and trends to see whether Judge Posner’s critique of the PTO is justified.

Is there evidence that the PTO is granting more lower-quality patents?

One measure of patent quality is the number of patents issued by the PTO, calculated as a percentage of total applications filed. Data for this analysis is available from U.S. Patent Statistic Chart for Calendar Years 1963-2012, accessible at http://www.uspto.gov/web/offices/ac/ido/oeip/taf/us_stat.htm.

Looking at the past 20-year period only, there has been a steady year-by-year increase in number of utility applications filed and the number of patents granted. In 1993, 174,743 utility applications were filed and 98,342 were granted; by 2003, those numbers had increased to 334,441 and 169,023, respectively; and by 2012, the numbers were at 542,815 and 253,155. Overall, the number of utility applications filed increased nearly 70% and the number of utility patents granted increased by over 60% in the past 20 years.

To examine changes in the “quality” of patents being issued over the 1993-2012 time period, the percentage of granted patents relative to the total number of applications filed was calculated for each year. For this calculation, the average period of pendency for an application was assumed to be 2 years, so the percentage of applications that mature into patents in any year was calculated based on number of applications filed two years earlier. Thus, for example, the percentage of patents issuing in 2012 was calculated as the ratio of patents issuing in 2012 divided by the number of applications filed in 2010. As a measure of patent quality, the numbers are imprecise for a variety of reasons in addition to prosecution-lag time: are patents being granted with narrower or broader claims and is the overall quality of inventions changing over time? More importantly, to Judge Posner’s point, is there a significant variation in quality among the different art units?

Click here to see graph

Putting these uncertainties aside, one can see from the plot above that standards for patentability have been on a roller coaster for the past 20 years. In 1998, the grant rate peaked at over 75%, then started a 10-year descent to a low of around 37% in 2008 and 2009 before trending upwards again to the current rate of about 50%. The trough in the 2008-2009 period would corresponds to KSR and its aftermath; the cause of the peak rates in 1998 and 1999 is less certain, but could be related to patent exuberance associated with the dot.com bubble.

But to return Judge Posner’s criticism, it would appear that the quality of patents issued since 2005 has been better on average than in the years leading up to 2005. In fact, however one feels about Judge Posner’s proposal to do away with the presumption of validity, a patent challenger might very well argue that the presumption should be given less weight for a patent granted in the late 1990’s than in the period since 2004.

Has high turnover in the USPTO Examining Corps led to lower quality patents?

A report in 2010 on examiner turnover in the PTO found that the percentage of examiners with less than 3 years’ experience has grown from less than 50% in 1996 to about 80% in 2009, while the percentage of examiner’s with 10+ years has decreased from 20% to less than 10% in the same period. Those numbers are not reassuring, since 3 years is a reasonable estimate of the time required to become a proficient and competent examiner. (see, for example, http://www.patentlyo.com/patent/2010/02/patent-examiner-experience-levels.html).

So how does the quality of patents granted by the USPTO stack up against patents granted by the EPO and JPO, both of which have better paid and more experienced examiners? An Encore Report on Patent Office Governance and Patent System Quality, for example, presents statistics from 2008 showing that USPTO examiners are expected to search and examine more than twice the number of claims as their EPO or JPO counterparts, while receiving a total compensation package that is only about 58% of EPO examiners and less than half that of JPO examiners. (the Encore report is accessible at http://www.ecore.be/DPs/dp_1306940512.pdf).

Using the percentage of patents granted/ number of applications filed in any year as a measure of patent quality, the USPTO appears to be holding its own.

For the years 2010 and 2011, the percentage of patents granted by the USPTO is slightly greater than 44%. In the EPO, the percentage of granted patents in the same two years is 38% and 43%--lower than the U.S. numbers but not strikingly so, considering that the percentages in the U.S. were below 38% in the two years following the KSR decision. Moreover, the percentages don’t reflect specific factors independent of quality that could explain the lower grant rates in the EPO, such as the EPO’s stringent “written description” requirements.

Surprisingly, the JPO has a more generous grant rate than either the USPTO or EPO. For the years 2010 and 2011, the percentage of patents granted by the JPO were 64% and 76% respectively, comparable to the US only in the anomalous 1998-2001 time frame. Does this mean that Japan is granting lower quality patents, or (as seems more likely) that the JPO is granting more patents, but with narrower claim scope?

In short, the USPTO seems to be doing a respectable job in examining applications for patentability, despite the funding challenges that have limited the Office’s ability to retain experienced examiners.

What do affirmance rates on appeal say about the quality of patents issued by the USPTO?

Perhaps the ultimate test of overall patent quality is the affirmance rates for patent validity on appeal before the CAFC or when challenged in a post-grant proceeding. Data on invalidation rates by district courts and the Federal Circuit can be found, for example, at https://www.morganlewis.com/pubs/Smyth_USPatentInvalidity_Sept12.pdf.

Data from 2002 to 2012 shows that most patents being litigated survived an invalidity challenge. Invalidation rates by district courts varied between 12% (2006) and 26% (2007 and 2008), with an average over the 11-year period of slightly more than 20%.

Over the same period, the percentage of patents that were held invalid by the Federal Circuit varied between about 8% in 2002 and 26% in 2012, consistent with the increased scrutiny that court is giving to the several requirements for patentability.

Data on the rates of clam cancellation and claim changes for patents tested by ex parte and inter partes reexamination are found, for example, at www.aipla.org/committees/committee_pages/IP...in.../1%20Hill.ppt. Summarized in the table below are the percentages of reexamined patents whose claims have been confirmed, cancelled, or changed. The data show a surprising vulnerability of granted claims when reexamined in an inter partes proceeding. Either the PTO is harder on its own patents than the courts are or a more vulnerable group of patents are subject to inter partes reexamination than to a court challenge.

Click here to see chart.

In any case, the patent “failure” rate on reexamination is not too different from that observed for Opposition proceedings in the EPO. Statistics for the years 2000-2006 show that in nearly 60% of opposed patents, granted claims are either amended or rejected, and in 25% of the cases, all of the claims are rejected and the patents revoked.

Based on these various indicators, the quality of U.S. patents doesn’t seem seriously out of balance with the quality of patents granted by the EPO or JPO.