Bi-Level Technologies

From the SelectedWorks of Ron D. Katznelson

November 17, 2014

A Century of Patent Litigation in PerspectiveRon D Katznelson, Bi-Level Technologies

Available at: https://works.bepress.com/rkatznelson/77/

November 17, 2014

1

A Century of Patent Litigation in Perspective

By Ron D. Katznelson1

ABSTRACT

When comparing patent litigation rates or “rarity” across decades, one must take into

account the proportion to the actual scale of commercial activities that give rise to

patent disputes. Such normalizing scales are preferably national metrics of

commercial activity such as (a) the number of patents issued in the year, (b) the total

number of patents in force over which disputes may arise, (c) the total number of

Federal civil suits, or (d) the economic scale of the Gross National Product (GDP) in

real dollars. This paper marshals for the first time information on all patent litigation

in Federal district courts spanning almost a century. The patent lawsuit filing

information is newly obtained from the Judicial Conference Annual Reports going

back to 1937 and further collected from the weekly Official Gazette of the Patent

Office going back to 1923. In addition, an estimate for the number of US patents in

force in each of the years covered is derived. Using non-parametric statistical tests, it

is shown that, with the exception of the AIA-caused litigation anomaly of 2011-2013

explained in the paper, for all four normalizing metrics, patent litigation intensities

during this century had not exceeded those experienced during the 20th century. High

patent litigation intensities in the 1920s-1930s and the 1960s have been comparable

to, if not higher than, those in the 2000s. These litigation activities are thought to be

consistent with major shifts in technological developments such as the development of

radio and electronics and chemical advances in the 1920s-1930s, the development of

semiconductor transistor electronics in the 1960s and the wireless communications

and internet-based technologies at the turn of the 20th century.

Keywords: patent litigation, GDP, patents in force, intellectual property lawsuits,

judicial statistics.

JEL Classifications: K41, N12, O34.

1 The author wishes to thank Susan Hager for her tireless support in entering the patent lawsuit

records from the Official Gazette of the Patent Office into the lawsuit database. Thanks are also due to Jim Hirabayashi of the USPTO for his guidance on the Office’s legacy TAF database used in this study to estimate patent maintenance rates.

November 17, 2014

2

1 Introduction

Recent years brought a proliferation of articles that describe and quantify a purported

“explosion” in patent litigation.2 The causes for this purported explosive rise are often

attributed to technological change, the Federal Circuit “patent-friendly” court, the

growth of patenting, poor “patent quality,” and to litigation tactics of non-practicing

entities (NPE), or the illdefined “Patent Assertion Entities” (PAEs), pejoratively called

“patent trolls.”3 A recent White House report on PAEs states without supporting

sources that “[t]he increased prevalence of PAE suits, and patent suits in general, in

recent years stands in contrast to the 20th century, when suits for patent infringement

were relatively rare.”4 However, without the context of the general scale of commercial

activity, this statement is meaningless because the number of all federal civil suits “in

recent years stands in contrast to the 20th century,” when civil suits “were relatively

rare:” about 283,000 Federal civil suits were filed in 2010 compared to only about

54,600 filed in 1950;5 in Federal district courts, about 3,700 trademark suits were filed

in 2010 compared to only about 200 filed in 1950;6 and about 2,000 copyright suits

were filed in 2010 compared to only 165 filed in 1950.7

Therefore, when comparing patent litigation “rates” across decades, one should cover

periods of major technological change, and one must take into proportional account the

actual scale of the economy and the commercial activities that give rise to patent

disputes. Normalization measures for the number of patent suits are preferably the

growing user base of new technologies as manifested by commercial activity indicia

such as (a) the number of issued patents in the year, (b) the total number of patents in

force over which disputes may arise, (c) the total number of Federal civil suits, and (d)

the Gross National Product (GDP) in real dollars. All four of these normalizing

measures are used in this study.

Previous studies on historical numbers of patent lawsuits relied on the limited data

from the Federal Judicial Center, Federal Court Cases Integrated Database8 which

unfortunately does not cover the years before 1970. As a result, these studies do not

cover some major technological development eras of the 20th century.9 Moreover,

2 James Bessen and Michael J. Meurer, “The Patent Litigation Explosion,” 45 Loyola U. Chi. L. J. 401

(2014); Adam. B. Jaffe and Josh Lerner, Innovation and its discontents: How our broken patent system is endangering innovation and progress, and what to do about it. Princeton University Press, (2004) p13.

3 Colleen Chien, “Patent Trolls by the Numbers,” Patently-O Blog (March 14, 2013), available at www.patentlyo.com/patent/2013/03/chien-patent-trolls.html.

4 The White House, “Patent Assertion and U.S. Innovation,” (June 4, 2013), p 5, (emphasis added), available at www.whitehouse.gov/sites/default/files/docs/patent_report.pdf.

5 Annual Report of the Director of the Administrative Office of the United States Courts (1950; 2010). 6 Id. 7 Id. 8 Database available from the Inter-University Consortium for Political and Social Research (ICPSR) in

Ann Arbor, MI. www.icpsr.umich.edu. 9 Studies of patent litigation statistics covering periods beginning on or later than 1970 include: Alan C.

Marco, Shawn P. Miller, and Ted M. Sichelman, “Do Economic Downturns Dampen Patent Litigation?” (October 27, 2014). 5th Annual Conference on Empirical Legal Studies paper. Available at SSRN: http://ssrn.com/abstract=1641425; John L. Turner, “In Defense of the Patent Friendly Court Hypothesis: Theory and Evidence,” (April 2005). Available at http://ssrn.com/abstract=713601; Deepak Somaya, “Firm Strategies and Trends in Patent Litigation in the United States.” In: Gary D.

November 17, 2014

3

most of these studies report the absolute number of filed patent cases without

normalization to relevant economic activity measures.

A recent study by Zorina Khan10 covered the period 1790 to 2000 with a temporal

granularity of a decade, using patent lawsuit data from Lexis and from published

volumes of reports of patent cases. Although published judicial opinions may be a

good way to track the development of case law, they are deficient proxies for the actual

number of historical patent disputes. Cases litigated to a decision were only a small

subset of all lawsuits filed, and reported decisions, even a smaller subset of that, with

reporting coverage that varied over time.11 Indeed, Khan acknowledged the

limitations of reported decisions count as a source, but her results are nevertheless

qualitatively highly instructive.

Khan’s study of the number of reported patent lawsuits over two centuries reveals

that historically, as a fraction of issued patents, reported litigation rates in the first

part of the 19th century exceeded that of the last two decades by about a factor of five.12

She finds that high litigation rates in given technological fields were correlated with

the advent of disruptive technologies13 and observes that “vexatious and costly

litigation about all areas of law—patents, property, contracts, and torts alike—were

inevitably associated with the advent of important disruptive innovations.”14

Importantly, Khan also documents the robust NPE activity and patent litigation,

finding that it is not new and tracing such activity to the patent “wars” of the 19th and

early 20th centuries.15

This present study aims to fill the void in studies on the national aggregate statistics

of all patent lawsuit filings before 1970. The contributions of this study are twofold:

first, sources for such litigation information are identified and data compiled for the

fiscal years beginning in 1923 through 2013. It is recognized with virtual certainty

that, short of collecting individual archival records prior to 1923 from each district

court, no national compilation of patent lawsuit filings earlier than 1923 likely exists.

Second, in order to derive the important normalizing measure of the total number of

patents in force each year, a detailed analysis of this measure is provided which

factors for each year (a) the number of issued patents, (b) the number of patents

Libecap (Ed.), Advances in the Study of Entrepreneurship, Innovation and Economic Growth. (G. Libecap, ed), JAI Press, Greenwich, CT. (2004) pp. 103–148; William Landes, "An Empirical Analysis of Intellectual Property Litigation: Some Preliminary Results,” 41 Houston Law Review 749 (2004); Gauri.Prakash-Canjels, "Trends in Patent Cases: 1990-2000." 41 Idea 283 (2001); .Jon F. Merz, and Nicholas M. Pace, “Trends in Patent Litigation: The Apparent Influence of Strengthened Patents Attributable to the Court of Appeals for the Federal Circuit,” 76 J. Patent and Trademark Office Society, 579 (1994).

10 Zorina B. Khan, “Trolls and Other Patent Inventions: Economic History and The Patent Controversy in the Twenty-First Century,” 21 Geo. Mason L. Rev. 825 (2014).

11 Matthew Henry, Thomas P. McGahee, and John L. Turner, “Dynamics of Patent Precedent and Enforcement: An Introduction to the UGA Patent Litigation Datafile,” (March 4, 2013). Available at http://ssrn.com/abstract=2228103

12 Id., at 861, Figure 3. 13 Id., at 862-863 (showing the effects on litigation rates in the telegraph, telephone and automobile

industries). 14 Id., at 842. 15 Id., at 833-835, 839-842.

November 17, 2014

4

expired at the end of their term, (c) the number of patents expired earlier than their

term for failure to pay maintenance fees, (d) the distribution of post-GATT patents

having patent term greater than 17 years (but no more than 20 years) based on

prosecution pendency time.

Finally, a statistical analysis of patent lawsuit filing rates normalized by the four

normalizing measures described above is presented. It is shown that, with the

exception of the AIA-caused litigation anomaly of 2011-2013 explained in the paper,

for all four normalizing metrics, patent litigation intensities in recent years had not

exceeded those experienced during the 20th century.

2 Data and methods

2.1 Patent litigation data

Beginning in 1873, national compilation and reporting of civil cases filed in Federal

district courts were published in the Annual Reports of the Attorney General of the

United States.16 Unfortunately, no breakdown of the number of civil suits by the

nature of suit was reported and thus the number of patent suits cannot be obtained

from these annual reports. It was not until January 1935 that the Justice Department

inaugurated a new system for compiling Federal court statistics. Under this system,

individual reports of civil and criminal cases were furnished monthly by court officials

by means of punch cards and a central IBM Hollerith tabulating machine was used to

compile the annual statistics.17 For fiscal years 1935-1936, the Annual Reports of the

Attorney General reported the total number of intellectual property suits (patent,

copyright, and trademark suits) in a single table entry. The first fiscal year for which

those counts are available separately is 1937, although they were not reported until

after 1940, when the Administrative Office of the United States Courts was organized

and had taken over the Department of Justice's system.18 Separate numbers for suits

filed in patent, copyright, and trademark cases are therefore available from fiscal year

1937 onwards in Table C-2 and its predecessors in the Annual Report of the Director of

the Administrative Office of the United States Courts, which this study relies upon.

No appeal cases were included in these counts.

Fortunately, patent lawsuit filing events for earlier years since fiscal year 1923 have

been reported on a weekly basis in the weekly Official Gazette of the US Patent Office,

albeit requiring laborious week-by-week and case-by-case collection and data entry. In

the Act of February 18, 1922, Congress codified in R.S. §4921, the predecessor of 35

U.S.C. §290, a Federal court reporting requirement: “it shall be the duty of the clerks

of such courts within one month after the filing of any action, suit, or proceeding

arising under the patent laws to give notice thereof in writing to the Commissioner of

Patents…”19 In turn, the US Patent Office published these notices shortly thereafter

16 Will Shafroth, “Federal Judicial Statistics,” 13 Law & Contemp. Probs. 200, 204 (1948). See also Will

Shafroth, “A Short Bibliography of Judicial Statistics,” 38 Law Libr. J. 37 (1945). 17 Annual Report of the Attorney General of the United States for the Fiscal Year 1935, Washington, D.C.

(1935) p183; Shafroth (1948) note 16 supra, at 205. 18 Id. 19 Revised Statutes § 4921, Ch. 58, § 8, 42 Stat. 392 (Feb. 18, 1922).

November 17, 2014

5

in its weekly Official Gazette,20 identifying the patents involved, the inventor(s), the

suit filing date, the parties’ names, and the court docket number.

Copies of the weekly “Patent Suits” sections appearing in the Official Gazette of the

United States Patent Office during the years 1922-1937 were obtained, from which a

total of 14,617 suit records from district courts (excluding circuit court appeals) were

manually entered in a database of filed patent cases. Of these cases, 14,155 have

filing dates between July 1, 1922 and June 30, 1937, corresponding to Fiscal Years

1923 through 1937.21 In addition, 14 cases were reported with no filing date and for

the purpose of this study, these were recorded as having been filed in the fiscal year in

which the respective notice was published in the Official Gazette.

In a check for the completeness of the Patent Office “Patent Suits” data, it was found

that the number of filed cases reported for FY 1937 (870 cases) was about 87% of that

reported by the Administrative Office of the United States Courts for that fiscal year

(1,004 cases). This under-reporting by the courts to the Patent Office appears

consistent with similar comparisons of more recent data elsewhere.22 Therefore, the

data in this study for the fiscal years 1923-1936 likely underestimate by more than

10% the actual numbers of patent suits during those years.

2.2 Derivation of the number of patents in force

The number of utility patents in force by the end of each fiscal year is derived as

follows. The number of issued patents each fiscal year was obtained from the US

Patent and Trademark Office’s (PTO) annual reports starting from 1905. To begin,

two important legal changes affecting patent term must be considered. The first is the

establishment of the patent maintenance fee system that came into effect in 1981.23

Failure to pay maintenance fees during any of the maintenance windows at 4, 8 and

12 years after issue results in early patent expiration. Initial effects of this law

occurred in 1985 for patents issued in 1981 that expired at the first maintenance

window for failure to pay maintenance fees. The second key legal change is that

Congress enacted in the Uruguay Round Agreements Act24 implementing the General

Agreement on Tariffs and Trade (GATT). Under this legislation, the 17-year patent

term was changed and 35 U.S.C. § 154(a)(2) since provides that patents filed on or

after June 8, 1995 have a term of 20 years from the priority application date.

Moreover, under § 154(c)(1) the term of a patent that was in force on or that results

from an application filed before June 8, 1995 is the greater of the 20-year term, or 17

years from grant.

Finally, relevant legal changes that have had no effect on the present analysis are as

follows: Patent term adjustments under 35 U.S.C. § 154(b) began taking effect only for

20 Official gazette of the United States Patent Office, (starting in 1922). 21 This fiscal year choice was made for want of consistency with the fiscal year reporting system of the

later data from the Annual Report of the Director of the Administrative Office of the United States Courts.

22 Somaya (2004), note 9 supra, at 108. 23 P.L. 96-517, 94 Stat 3015 (Dec. 12, 1980); 35 U.S.C. §41(b). 24 Pub. Law 103-465, 108 Stat 4809, December 8, 1994.

November 17, 2014

6

patents issued after 2000 and as such, the effects are not manifested in patent

expirations for patents issued to date. Moreover, the number of patents having terms

extended due to the FDA approval process under 35 U.S.C. § 156 is negligible for the

purposes of this analysis, as PTO data shows that the terms of only about 70 such

patents per year are extended.25

Accordingly, three separate counting treatments of issued patents were used: the first

is for all “Old Patents” issued between 1905 and June 8, 1978. These have had a

simple term of 17 years from issue as they were subject to neither maintenance fees

nor GATT effects. A simple running 17-year “leaky” accumulation establishes the

number of patents in force for this category. The second treatment is for all patents

with priority application dates later than June 7, 1975, as they are subject to GATT

effects under § 154(c)(1). The third is the maintenance effects affecting all patents

issued since 1981 and overlapping the second treatment.

* Except for 1995, where applications filed on or before June 7, 1995 are shown per

search in FPO database. Note that k(t) = 0 for all t > 1995; b(t) = 0 for all t < 1995;

c(t) = 0 for all t < 1996; d(t) = 0 for all t < 1997; e(t) = 0 for all t < 1985;

** Utility patents in force at year t are calculated recursively as follows:

For t < 1995: 𝑚(𝑡) = 𝑚(𝑡 − 1) + 𝑎(𝑡) + ℎ(𝑡) − 𝑒(𝑡) − 𝑘(𝑡) + 𝑔(𝑡 − 17);

For t 1995:

𝑚(𝑡) = 𝑚(𝑡 − 1) + 𝑎(𝑡) + ℎ(𝑡) − 𝑏(𝑡) [1 −𝑔(𝑡−17)

𝑎(𝑡−17)] − 𝑐(𝑡) [1 −

𝑔(𝑡−18)

𝑎(𝑡−18)] − 𝑑(𝑡) [1 −

𝑔(𝑡−19)

𝑎(𝑡−19)]

− 𝑒(𝑡) − 𝑘(𝑡)

The last terms involving the variable g(t) are fractional corrections to avoid double

counting of expirations included in variable e(t).

Table 1. The sources and the variables used to derive the number of patents in force.

Table 1 shows the sources for, and the count variables used to derive the count of the

patents in force. Because it has been less than 20 years since the GATT change, no

patents expired yet under § 154(a)(2) and the only two categories of patents requiring

separate analysis are those having the 17-year term which expired no later than June

7, 1995, “Old Patents” in the table, and those governed by § 154(c)(1), shown in the

25 See 78 Fed. Reg. 31886 (May 28, 2013).

Issued

Nominal

term

expired in

FY (long

pendency)

Nominal

term

expired in

FY (mid

pendency)

Nominal

term

expired in

FY (short

pendency)

Issued

Nominal

term

expired

t a b c d e f g h k m

SourceAnnual PTO

Report

Annual PTO

ReportCalculated **

FY

FPO Database (Utility Patents) Annual PTO Report*

OLD PATENTS

Issued < June-8-1978 Application date > June-7-1975

PATENTS SUBJECT TO GATT EFFECTS MAINTENANCE

Fraction

issued in

FY that

expired

any time

before

nominal

term

Total

issued in

FY that

expired

any time

before

nominal

term

Total

expired in

FY due to

failure to

renew

PTO TAF database

Utility

Patents in

Force

November 17, 2014

7

table as “patents subject to GATT effects.” Expirations from both categories were

combined. In addition, the numbers of early expirations under the maintenance

system were derived from the PTO’s Technology Assessment and Forecast (“TAF”)

database, from which count variables f and g were derived.

Patents subject to GATT effects were divided into three coarse indicator sets (count

variables a, b and c) depending on their pendency in prosecution. These were

determined by selective search on FreePatentsOnline.com (FPO) for the number of

patents filed 20 years prior to the focal fiscal year t and issued within 3-4 years (b), 2-3

years (c), and 1-2 years (d) from their filing window. These were used to estimate a

proportional correction and shorten the otherwise 20-year term of the patents as

shown in the second equation in Table 1.

Noteworthy is the fact that the results obtained in this analysis for US utility patents

in force are within 1% of the limited results published by the World Intellectual

Property Organization for the years 2004-2012.26

2.3 Civil suits data

The numbers of civil lawsuits filed over the years manifest trends in the overall scale

of all disputes taken up by the Federal district courts. As one of the normalizing

measures for patent litigation, the annual total numbers of civil suits filed in federal

district courts were obtained from Table C-2 and its predecessors in the Annual Report

of the Director of the Administrative Office of the United States Courts for fiscal years

1937-2013. The rest of the data for the years 1923-1936 were obtained from the

Annual Reports of the Attorney General, wherein all civil cases were included in the

tally with the exception of cases brought by the government under the National

Prohibition Act of 1919.27 As a control for civil suits related to patents, the numbers

for copyright and trademark suits were separately obtained from Table C-2 for the

Fiscal Years 1937-2013.

2.4 Real GDP data

The US Bureau of Economic Analysis (BEA) is the official government source for Gross

Domestic Product (GDP) information. It now reports GDP in “real” dollars based on

chain-weighted indexes. These indexes use updated weights in order to provide a

more accurate picture of the economy, to better capture changes in spending patterns

and in prices, and to eliminate the bias present in fixed-weighted indexes.28 However,

BEA’s GDP data coverage begins at 1929. Fortunately, the BEA real chained dollar

index was extended to cover US GDP starting at 1790 by MeasuringWorth.com.29 For

26 See “WIPO IP Statistics Data Center,” Patents In Force Indicator. Available at

www.wipo.int/ipstats/en/statistics/patents/. 27 Civil prohibition cases artificially inflated the count until the Act’s repeal in December of 1933.

According to the Annual Reports of the Attorney General, 1932 saw a record high of 15,455 prohibition cases brought by the government.

28 J. Steven Landefeld, Brent R. Moulton, and Cindy M. Vojtech, “Chained-Dollar Indexes, Issues, Tips on Their Use, and Upcoming Changes,” Bureau of Economic Analysis (November 2003). Available at www.bea.gov/scb/pdf/2003/11November/1103%20Chain-dollar.pdf.

29 See www.measuringworth.com/usgdp/.

November 17, 2014

8

this study, this source was used to obtain the GDP data in Real 2009 chained dollars

for 1923 to 2013.

2.5 Statistical Analysis

The annual numbers of patent lawsuits (shown in Figure 1) were normalized by (a) the

number of issued patents in the year (result shown in Figure 2); (b) the total number

of patents in force at the end of the year (result shown in Figure 3); (c) the total

number of Federal civil suits in the year (result shown in Figure 4), and (d) the Gross

National Product (GDP) in real dollars (result shown in Figure 5).

The White House assertion described in the introduction, that the increased

prevalence of patent suits “in recent years stands in contrast to the 20th century,

when suits for patent infringement were relatively rare” is tested here. As explained

above this statement is meaningless without considering economic scales and thus the

terms “prevalence” and “relatively rare” are taken to mean as the lawsuit filing rate

relative to other relevant economic activity rates – such as the four normalizing time-

series discussed above. These four normalized litigation time series as shown in

Figure 2 through Figure 5 form the set, which we define as “litigation intensities.”

Because litigation intensities are always positive (asymmetric, bounded from below)

and because the magnitudes of their temporal fluctuations are not sufficiently small

compared to their mean values, their distributions cannot be taken as normal.

Therefore, statistical inferences should rely on non-parametric methods used when the

distributions are unknown. Accordingly, the two hypotheses consistent with the

White House assertion to be tested are formulated non-parametrically as follows:

HA: Patent litigation intensity after the year 2000 had a maximum excursion that

exceeded any of the excursions in the period 1923-2000.

HB: Patent litigation intensity distribution after the year 2000 differed from the

distribution in the period 1923-2000.

It follows that the alternative to Hypothesis HB is a null hypothesis H0:

H0: The patent litigation intensity distributions of both time periods are identical,

so that there is a 50% probability that an intensity value randomly selected

from one time period exceeds an observation randomly selected from the other

time period.

Hypothesis HA and its logical complement are readily tested by a simple peak

excursion test, comparing maximum intensities in both periods. The testing of H0

against HB is accomplished by using the non-parametric Mann-Whitney test, also

called the Wilcoxon rank sum test. The results of these tests are tabulated in Table 2

below.

3 Discussion of Results

The annual numbers of patent lawsuits are shown in Figure 1. The normalization by

the number of issued patents in the year are shown in Figure 2 whereas normalization

by the total number of patents in force at the end of the year are shown in Figure 3.

November 17, 2014

9

Normalization by the total number of Federal civil suits in the year is shown in Figure

4 and the normalization by the GDP is shown in Figure 5.

Several observations can be made: substantial reduction in patent litigation intensity

occurred during World War II. Figure 5 shows that this decline also occurred in

copyright and trademark litigation. A preliminary examination of the types of patents

asserted during the relatively high patent litigation period of the 1920s and 1930s

indicates substantial rise of the chemical technologies and particularly in the new field

called electronics. These recorded cases show a large share of lawsuits involving

patents on radio and vacuum tube electronics, consistent with historical reports of

litigation surge involving such patents30 Indeed, litigation followed years of brisk

patenting and development in the vacuum tube electronics and radio fields.31 This

suggests that further studies should examine the degree to which such surges of

patent litigation activities shown in the figures are related to major shifts in

technological developments such as the development of radio and vacuum tube

electronics in the 1920s-1930s, the development of semiconductor transistor

electronics in the 1960s and the wireless communications and internet-based

technologies at the turn of the 20th century

Figure 3 shows a substantial rise in the number of patents in force after 1995. This

trend is attributable to three factors: (a) the effective extension of patent term due to

GATT for patents having prosecution pendency shorter than 3 years, (b) the steady

increase in patent maintenance rate, prolonging the average patent life,32 and (c) an

increase volume of new patent issued. All figures show that the number of patent

lawsuits surged in 2011-2013. As discussed below, this is directly attributable to the

America Invents Act (AIA) and is not reflective of an underlying increase in the

number of litigated patent disputes.

3.1 The AIA lawsuit filing surge

Several factors created by the AIA caused, and will likely continue to cause, increased

rate of lawsuit filings. The effects of these factors were building up even before

passage of the AIA in September 2011, as patentees acted in anticipation of more

adverse legal conditions at various phases of the AIA implementation. As codified by

the AIA, 35 U.S.C. § 299 substantially curtails joinder of multiple defendants in a

single case, requiring multiple lawsuits where one would have been filed before

September 2011. Thus, the AIA forced patentees to file more lawsuits, each naming

only one defendant.

30 W. Maclaurin, Radio Patent Litigation, Appendix II in Invention and Innovation in the Radio

Industry, pp. 268-287. Arno Press and the New York Times, New York (1971). 31 John Howells and Ron D. Katznelson, “The Coordination of Independently-Owned Vacuum Tube

Patents in the Early Radio Alleged Patent 'Thicket',” (June 13, 2014). Available at SSRN: http://ssrn.com/abstract=2450025 (describing the brisk development of vacuum tube inventions from 1912 to 1920 that were later litigated in the 1920s and 1930s); K. R. Thrower, “Evolution of circuit design for A.M. broadcast receivers: 1900-1935,” 56 Journal of the Institution of Electronic and Radio Engineers, pp. 325-341, (October/December 1986).

32 Dennis Crouch, “Patent Maintenance Fees,” Patently-O Blog, (September 26, 2012). At http://patentlyo.com/patent/2012/09/patent-maintenance-fees.html.

November 17, 2014

10

Figure 1. Patent lawsuit filings in US district courts, 1923-2013.

Figure 2. Patent lawsuits in US district courts as percentage of patents issued in the year.

100

1,000

10,000

Pat

en

t la

wsu

it f

ilin

gs in

fis

cal y

ear

Fiscal Year

Patent lawsuit filings, US district courts, 1923–2013

Sources: FY 1923-1936: Patent suits: Offical Gazette of the U.S. Patent Office, count of listed cases in "Patent Suits" sections (1923-1937); FY 1937-2013: Patent and civil suits: Administrative Office of the U.S. Courts, Annual Reports of the Director, Table C-2 and its predecesors (1940–2013)

0.0%

0.5%

1.0%

1.5%

2.0%

2.5%

3.0%

3.5%

4.0%

Pat

en

t su

its

file

d r

ela

tive

to

pat

en

ts is

sue

d

Fiscal Year

Patent lawsuits as a percentage of patents issued in the year, US district courts, 1923–2013

Sources: Issued patents: Commissioner of Patents and USPTO Annual Reports, (1923-2013);. Patent suits: Offical Gazette of the U.S. Patent Office, (1923-1937); FY 1937-2013: Administrative Office of the U.S. Courts, Annual Reports of the Director, Table C-2 and its predecesors (1940–2013).

November 17, 2014

11

Figure 3. Patent lawsuits filed in US district courts as a fraction of US patents in force.

Figure 4. Patent lawsuits in US district courts as fraction of Federal civil suits, 1923–2013.

0.0

0.5

1.0

1.5

2.0

2.5

0.00%

0.05%

0.10%

0.15%

0.20%

0.25%

0.30%

Pat

en

ts in

fo

rce

(m

illio

n)

Pat

en

t la

wsu

its

as p

erc

en

tage

of

pat

en

ts in

fo

rce

Fiscal Year

Patent lawsuits as a percentage of US patents in force,US district courts, 1923–2013

Sources:For FY1923-1936: Offical Gazette of the U.S. Patent Office, "Patent Suits" sections;For FY 1937-2013: Administrative Office of the U.S. Courts, Annual Reports of the Director, Table C-2 (1940–2013);Patents in force calculated based on U.S. Patent Office Annual reports and online application date data (see text)

Patents in force

0.0%

0.5%

1.0%

1.5%

2.0%

2.5%

3.0%

3.5%

4.0%

Pat

en

t su

its

file

d a

s a

pe

rce

nta

ge o

f ci

vil s

uit

s fi

led

Fiscal Year

Patent lawsuits as a percentage of federal civil suits, US district courts, 1923–2013

Sources: FY 1923-1936: Patent suits: Offical Gazette of the U.S. Patent Office, count of listed cases in "Patent Suits" sections (1923-1936); Civil suits: Attorney General Annual Reports (1923-1936); FY 1937-2013: Patent and civil suits: Administrative Office of the U.S. Courts, Annual Reports of the Director, Table C-2 and its predecesors (1940–2013)

November 17, 2014

12

Figure 5. IP lawsuit filings US district courts per billion real dollars of GDP.

That the surge in the number of patent suits in 2011-2013 is dominated by AIA-

related provisions has been widely documented. A report by the U.S. Government

Accountability Office (GAO) recognized that the surge in patent lawsuit filings was

due to the AIA joinder provisions,33 and in a detailed empirical study, Professors

Cotropia, Kesan, and Schwartz have shown that there was “no major difference

between both the number of unique patentees and the number of alleged infringers

from 2010 to 2012. While the number of cases increased, the totals for the main

players—patentees and defendants—stayed essentially constant.”34 This observation

is consistent with the lawsuit filing surge wrought by the AIA rather than by an

inexplicable surge in litigation.

Incentives to advance the filing of patent infringement suits before the AIA enactment

in September 2011 also grew substantially due to AIA’s 35 U.S.C. §315(b) which

provides that no inter partes review (IPR) can be instituted if the petition is filed more

than one year after petitioner is served a complaint alleging infringement of the

subject patent. Thus, in order to inoculate any particular existing infringed patent

against the threat of IPR challenge, patentees simply had to advance filing lawsuits

earlier than September 2011, which was one year before any party could institute an

IPR proceeding which came into effect in September 2012.

33 U.S. Government Accountability Office, “Assessing Factors That Affect Patent Infringement

Litigation Could Help Improve Patent Quality,” GAO-13-465, (Aug 22, 2013) p 15. 34 Christopher A. Cotropia, Jay P. Kesan, and David L. Schwartz, “Unpacking Patent Assertion Entities

(PAEs)” p. 28 (June 29, 2014), Minnesota Law Review, Forthcoming. Available at SSRN: http://ssrn.com/abstract=2346381

0.0

0.5

1.0

1.5

Filin

gs p

er

bill

ion

do

llars

of

GD

P (

in 2

00

9 r

eal

do

llars

)

Fiscal Year

IP lawsuit filings per billion real dollars of gross domestic product, US district courts, 1923–2013

Copyright

Trademark

Patent

Sources: FY 1923-1936: Patent suits: Offical Gazette of the U.S. Patent Office, count of listed cases in "Patent Suits" sections (1923-1936); FY 1937-2013: IP suits: Administrative Office of the U.S. Courts, Annual Reports of the Director, Table C-2 and its predecesors (1940–2013); Real GDP: Measuring Worth at www.measuringworth.com/usgdp/

November 17, 2014

13

A third provision of the AIA that contributes to the lawsuit surge is related to

incentives created by AIA-codified 35 U.S.C. §325(b) to file patent infringement suites

within 3 months of patent grant, as explained by this author elsewhere.35

3.2 The patent litigation spike of 1935

As the figures above show, a spike in patent litigation occurred in 1935, reaching more

than 1,360 suits, a level exceeding by about 45% the average intensity in proximal

years. This rise was mostly due to one patentee. After the US Supreme Court upheld

on January 7, 1935 in two separate decisions36 his patent on egg incubator (US Pat.

No. 1,262,860), Samuel B. Smith filed a few hundred patent infringement lawsuits in

district courts across the country against farmers and hatcheries during February

through May, 1935. As a result, many courts were upholding Claim 1 as valid and

infringed. It would not be until 1937 that obscure prior use antedating the invention

was discovered and proven, which the Supreme Court affirmed.37

3.3 No significant change in patent litigation intensity in recent years

compared to the 20th century

Table 2 summarizes the results of two non-parametric statistical tests for the four

patent litigation intensity measures. Because patent litigation intensities during

2011-2013 were anomalous due to the AIA – artificial reasons and mechanisms that

have little to do with the underlying number of patent disputes – data points for these

three years were removed from the analysis to prevent bias.

As can be seen in the table, Hypothesis HA that patent litigation intensity after the

year 2000 had a maximum excursion that exceeded any of the excursions in the period

1923-2000 is rejected. Rather, the fraction of samples in which litigation intensity in

1923-2000 period exceeded the maximum intensity in the later period ranged from 3%

to 45% depending on the normalizer.

Similarly, Hypothesis HB that patent litigation intensity distribution after the year

2000 differed from the distribution in the period 1923-2000 is rejected for three out of

the four normalizers. Table 2 shows that for these three normalized intensities, the

null hypothesis H0 cannot be rejected, i.e., that there is no statistically significant

difference in the intensity distributions of both time periods. However, for one

normalizer – the total number of patents in force – the distributions are significantly

different (Mann–Whitney U = 56, n1 = 78, n2 = 10, p < 10-5 one-tailed).

35 Ron D. Katznelson, “The America Invents Act at Work – The Major Cause for the Recent Rise in

Patent Litigation,” IPWatchdog, (April 15, 2013). At http://bit.ly/AIA-Litigation. (Explaining how changes in 35 U.S.C. §§ 299, 315(b), and 325(b) have changed lawsuit filing practices that caused the filing surge).

36 Smith v. Snow et al., 294 US 1 (1935); Waxham V. Smith et al 294 US 20 (1935). 37 Smith V. Hall et al. 301 U.S. 216 (1937).

November 17, 2014

14

Table 2. Results of two non-parametric statistical tests for four patent litigation intensity

measures shown in Figure 2 through Figure 5.

While it can be said that litigation intensity normalized by the total number of patents

in force observed in the two time periods are not drawn from the same probability

distribution, the peak excursion test shows that in 3% of the time such normalized

intensities in the first time period were actually higher than the maximum intensity

observed in the second period.

4 Conclusion

Newly acquired patent litigation statistics going back to 1923 provide a much-needed

perspective into patent litigation rates during the last 90 years. Four normalizing

metrics reflecting trends in economic and commercial activity scale were obtained to

derive normalized litigation intensities. Using non-parametric statistical tests, it is

shown that, with the exception of the AIA-caused litigation anomaly of 2011-2013

explained in the paper, for all four normalizing metrics, patent litigation intensities in

recent years had not exceeded those experienced during the 20th century. Rather,

patent litigation surges are consistent with major shifts in technological developments,

which introduce novel terms and uncertainty in patent claims and require

infringement analysis of novel and less understood products.

Patent lawsuits rate

measure

Mann-Whitney Test for

Two Independent Samples

Sample Period1923-

2000

2001-

2010

1923-

2000

2001-

2010

1923-

2000

2001-

2010

1923-

2000

2001-

2010

Count 78 10 78 10 78 10 78 10

Median 1.59% 1.72% 0.11% 0.16% 1.00% 1.09% 0.215 0.197

Rank sum 3387 529 3137 779 3438 478 3486 430

U 474 306 724 56 423 357 375 405

one tail two tail one tail two tail one tail two tail one tail two tail

alpha 0.05 0.05 0.05 0.05

U 306 56 357 375

Mean 390 390 390 390

Standard deviation 76.06 76.06 76.06 76.06

z-score -1.1044 -4.39132 -0.43387 -0.19721

U-critical 264.4 240.4 264.4 240.4 264.4 240.4 264.4 240.4

p-value 0.1347 0.2694 5.63E-06 1.13E-05 0.3322 0.6644 0.4218 0.8437

Significant no no yes yes no no no no

Rate Excursion Test

% of time a rate in first

period exceeded the

maximum rate in second

period

As % of patents

issued in FY

As % of patents in

forceAs % of all civil suits

24% 3% 40% 45%

Suits per $billion of

Real GDP