ArticlePDF Available

Economic Cost-Benefit Analysis of a Unified and Integrated European Patent Litigation System

Authors:
  • Max Planck Institute for Innovation and Competition
1
Economic Cost-Benefit Analysis of a
Unified and Integrated European Patent
Litigation System
Prof. Dietmar Harhoff, Ph.D.
Ludwig-Maximilians-Universität (LMU) München
Institute for Innovation Research,
Technology Management and Entrepreneurship (INNO-tec)
Final Report
26 February 2009
Tender No. MARKT/2008/06/D
2
Table of Contents
Executive Summary ...................................................................................................................5
1 Objective and Structure of the Report...............................................................................7
1.1 Objective ....................................................................................................................7
1.2 Structure of the Report...............................................................................................7
2 Results of the Literature Review.......................................................................................9
3 Status Quo of Patent Litigation in Europe and the Presidency’s Proposal .....................12
3.1 The Current System and Its Shortcomings...............................................................12
3.1.1 Patent Litigation Systems in EU Member States .............................................12
3.1.2 Features of the Most Frequently Used Systems...............................................13
3.1.3 Litigation in a Fragmented System ..................................................................14
3.2 The Presidency’s Proposal .......................................................................................19
3.2.1 Elements of the Proposal..................................................................................19
3.2.2 Relevance for the Estimation of Costs and Benefits........................................21
3.3 A Conceptual Framework for an Economic Assessment of Costs and Benefits .....22
4 Incidence and Cost of Patent Litigation in European Countries.....................................24
4.1 Survey Data from a Study carried out under German Presidency ..........................24
4.2 Data from Regional Courts in Germany .................................................................26
4.3 Data from the German Federal Patent Court............................................................27
4.4 Data from the Cremers and Harhoff Studies............................................................28
4.5 Data from the CJA Study .........................................................................................29
4.6 Practitioner Estimates...............................................................................................30
4.7 2003 Survey of the Working Party on Litigation.....................................................31
4.8 Incidence Data from the Pharmaceutical Sector Inquiry..........................................32
5 Estimates of the Operational Costs of the Unified Patent Court.....................................35
5.1 Judges’ Salaries........................................................................................................35
5.2 Staff Salaries ............................................................................................................36
5.3 Facilities, Information Technology and Other Costs ...............................................36
5.4 Total Annual Operating Costs..................................................................................36
5.5 Caveats .....................................................................................................................36
6 Benefits from Avoiding Duplicated Litigation ...............................................................37
6.1 Patents at Risk..........................................................................................................37
6.2 Estimating the Incidence of Duplication..................................................................38
6.3 Benefits from Avoiding the Costs of Duplication and from Cost Reductions.........40
6.4 The Impact of Litigation Costs.................................................................................42
6.5 Further Effects from a Unified Patent Court............................................................43
7 Additional Welfare Effects and Design Considerations .................................................47
7.1 Changing Character of Patent Litigation..................................................................47
7.2 Incentives for Innovation and Investment................................................................47
7.3 Quality Effects in the Patent System........................................................................48
7.4 Substitution between Litigation and Other Instruments of Conflict Resolution......48
3
7.5 Limiting Strategic Behavior.....................................................................................48
7.6 “Trolls” and Abuse of Litigation..............................................................................49
7.7 Reduction of Uncertainty .........................................................................................50
7.8 Benefits for SMEs and YICs....................................................................................51
7.9 Particular Design Elements of the Presidency Proposal ..........................................51
7.10 Impact on Further Harmonization of EU Patent Law..............................................52
8 Summary .........................................................................................................................53
9 Bibliography....................................................................................................................55
10 Annex I - Economic Studies of Patent Litigation and Litigation Systems......................61
10.1 Fundamental Economic Studies...............................................................................61
10.2 Extensions of Basic Models.....................................................................................63
10.3 Specific Problems Related to “Bad and Weak Patents”...........................................67
10.4 Studies on the US Court of Appeals for the Federal Circuit....................................69
10.5 Empirical Studies of Litigation in Countries Other than the USA...........................71
10.6 Specific Issues in Pharmaceuticals...........................................................................74
11 Annex II – Scenario Computations.................................................................................76
4
5
Executive Summary
The patent system in Europe is still incomplete. Appropriating returns from patented
technology is impaired by the fact that patent-holders may have to enforce their patent rights
in multiple courts. Moreover, third parties interested in showing that particular patent rights
have been granted erroneously are disadvantaged by having to initiate revocation proceedings
in multiple jurisdictions.
Given that the successful pursuit of these two objectives can generate welfare gains for the
European economy, a unified European patent litigation system has immediate appeal.
The current study seeks to provide approximate cost and benefit calculations in order to
inform policy-makers in Europe about the choices they face in this important field of public
policy. The study groups cost and benefit effects into the following categories: (i) effects from
avoiding duplication of litigation; (ii) effects from changes in the demand for litigation,
induced by changes in the cost structure; (iii) effects from changed incentives for patenting.
Towards performing an assessment of the first two effects, the report collects data from a
variety of sources in order to support the estimates and to test the plausibility of a number of
necessary assumptions. Given that no official data on the incidence, outcomes and cost of
patent litigation exist, an effort is made to triangulate data and estimates in order to
demonstrate that the approximations are justified.
The following results are particularly important. Avoiding duplication of infringement and
revocation cases is likely to generate large benefits for the European economy. The results
obtained here suggest that currently, between 146 and 311 infringement cases are being
duplicated annually in the EU Member States. By 2013, this number is likely to increase to
between 202 and 431 duplicated cases. Total private savings from having access to a unified
Patent Court in 2013 would span the interval between EUR 148 and 289 million.
An assessment of the operating costs of the proposed Patent Court is obviously subject to a
large number of caveats. Based on data from earlier efforts (in particular the Working Party
on Litigation set up under the auspices of the European Patent Organisation), an upper-bound
estimate for the operating costs of a court with a capacity of 940 cases indicates that the Court
would cause operating costs of EUR 27.5 million.
Hence, the cost-benefit assessment focusing on avoided duplication leads to a highly positive
evaluation of the proposal. Even if the low estimate of savings (EUR 148 million) is taken,
the new system would create substantial benefits and reach a benefit-cost ratio of 5.4.
However, this view may be unduly conservative, and the benefit-cost ratio could be as high as
10.5.
Additional benefits could flow in case of additional litigation activity, be it in terms of
infringement or revocation actions. The availability of a low-cost litigation path offered by a
unified Patent Court is likely to lead to additional activity from parties in countries which
currently do not use the European patent system extensively. Moreover, the cost level of
6
litigation in the unified Patent Court system is likely to be below the cost levels currently
observed in some Member States and parties in these Member States are also likely to engage
in more litigation activity in the medium-run. These effects will also contribute to generate
private and public benefits.
The above estimates and considerations are based on the assumption that the unified Patent
Court will offer litigation at roughly the same cost level as the three largest low-cost national
systems. In a robustness check, the report explores to what extent the gains from saved
duplication would be dissipated if the cost level were higher. The computations show that
even with a substantial average cost increase, benefit-cost ratios remain above one, and for
most scenarios considerably above one.
It is more difficult to predict cost-induced changes in the demand for litigation. The relevance
of such changes will depend on the level and type of costs imposed on users of the new
system. These will be mostly determined by the private costs for legal support and advice, but
also by the fees levied by the Court itself. Measures to contain the private costs to parties in
litigation are important, and the Presidency’s proposal includes a number of such measures
which are discussed in the report with respect to their impact. A particularly promising
measure is to admit representation of parties by specialized European Patent Attorneys.
Another measure of importance is the contribution from the Community budget and from
States which are not EU Member States to the Court’s budget, in order to keep fees at low
levels.
The report also discusses in a qualitative manner effects which emerge from changes in
patenting and litigation incentives. It is argued that effects will be beneficial if the unified
patent litigation system puts emphasis on fast and low-cost proceedings, high quality
judgement, and a fair balance between the legitimate interests of patent holders and alleged
infringers. It is emphasized that particularly strong positive welfare contributions can be
expected if an effective and rapid, low-cost revocation procedure is available. The latter
feature should provide an effective means against strategic and (possibly) frivolous litigation
activity which could be mounted in the future by “patent trolls”. The report also comments on
particular design choices in the Presidency’s proposal.
To summarize, this report recommends strongly that the Presidency should proceed in its
efforts to establish a unified and integrated patent litigation system for European patents and
future Community patents. For conservative estimates of the relevant parameters, the
economic benefits from such a system are likely to exceed the costs of the establishment and
operation of the new court by a large multiple of between 5.4 and 10.5. Moreover, with
prudent design choices it should be possible to implement a litigation system that will be
balanced and supportive of overall efforts to improve the quality of patents in Europe.
7
1 Objective and Structure of the Report
1.1 Objective
Over the last four decades, Europe has achieved a considerable degree of harmonization in its
patent systems. Despite some problems, the examination and grant system operated by the
European Patent Office (EPO), based upon the European Patent Convention (EPC), represents
a high-quality system that is currently not available in other regions of the world. With the
opposition and appeal proceedings at the EPO, a first level of litigation-type system for
validity issues has been established. Yet, despite a number of promising efforts undertaken in
the past at Community or intergovernmental level, a unified and integrated patent litigation
system is still not available in Europe.
At this point it is not fully clear what the overall costs and benefits of the creation of such a
system would be, and how they would be impacted by the choice of various crucial design
parameters for a unified litigation system. In order to inform the policy debate, it is of
considerable importance to estimate the economic impact of a unified patent litigation system
in Europe. Such a study should seek to give policy-makers guidance in their choice of policy
parameters and present evidence concerning the economic impact of such a system. Within its
tender MARKT/2008/06/D, the European Commission has asked for such an assessment. The
current report summarizes the analysis and efforts undertaken in the course of this project.
The overall objectives of this report are according to the tender as follows:
The report is to lay out a systematic framework for the assessment of economic effects
from a unified patent litigation system.
It will discuss qualitatively the comparative advantages of various system designs,
based on the experience of various European and non-European countries;
It will identify in a systematic way the different costs and benefits associated with
such a unified system (and with particular parameters of the system’s design).
It is to quantify (to the best extent possible) the costs and benefits from the
introduction of a unified European patent litigation system.
1.2 Structure of the Report
The remainder of this report is structured in seven sections which are followed by a detailed
bibliography and two annexes.
Section 2 discusses the results from a detailed literature review which captures important
theoretical and empirical insights on patent litigation as well as salient results of the studies
that have been undertaken in various attempts to propose unified European patent litigation
systems. The review itself is attached in Annex I of the report.
Section 3 provides an assessment of the shortcomings of the current litigation systems in
Europe. The discussion takes up arguments from the academic literature as well as
8
assessments put forth by practitioners and users of the system. The section then briefly
summarizes the Presidency’s proposal for the establishment of a unified Patent Court.
1
Finally, the section describes a conceptual framework for the simplified welfare assessment
pursued in the report.
Section 4 collects information about the frequency and cost of patent litigation in Europe
from various sources. While some survey data are available for the most important European
countries, most surveys do not distinguish between litigation activity concerning European or
national patents. Moreover, there is no data that would allow the computation of the extent of
duplication of litigation cases. The section describes the data sources in order to provide a
consistent basis for quantitative assessments.
Section 5 provides estimates of the operational costs of the unified Patent Court. The
estimates should be taken as a first approximation. They are based on assumptions, data and
cost structures already employed in similar calculations for the EPLA Patent Court.
Section 6 seeks to establish conservative estimates on the extent to which the unified
litigation system would reduce duplication. The salient feature of the unified system would be
to offer users of the system a non-duplicating litigation path – for both revocation and
infringement procedures. Several scenarios are developed in order to provide reasonable
estimates of the duplicated share of patent litigation cases, both for 2008 and for 2013. The
immediate effects from avoiding duplications are then estimated.
In Section 7, the more complex selection and incentive effects are addressed. As the cost of
litigation may change (relative to the cost level now present in many Member States) after the
establishment of the unified Patent Court, the impact of these changes needs to be considered.
Moreover, the establishment of the Court may have important incentive effects which are hard
to quantify but ought to be discussed at least on a qualitative level in the policy debate.
In Section 8, the results are summarized, and various caveats, which are unavoidable given
the incomplete database and lack of structural studies, are discussed. The section concludes
that with proper attention being paid to a number of important design decisions the future
unified system is likely to generate large benefits for Europe’s inventors, businesses and
society at large.
Section 9 contains the bibliography. Annex I contains the detailed literature survey. Annex II
contains the documentation of scenario computations summarized and discussed in section 6.
1
The terminology adopted in this report will abbreviate the official term “European and Community Patents
Court” with “unified Patent Court”. The court system proposed in the European Patent Litigation Agreement
(EPLA) will be referred to as the “EPLA Patent Court”.
9
2 Results of the Literature Review
This section of the report summarizes the most pertinent results from a detailed literature
review which captures important theoretical and empirical insights on patent litigation as well
as salient results of the studies that have been undertaken in various attempts to propose
unified European patent litigation systems. The review itself is attached in Annex I of the
report.
A comprehensive theoretical or empirical welfare balance of the patent system and of
patent litigation is not available at this point. Hence, an assessment of the impact of a
unified litigation system cannot be conducted from first principles, e.g. by first computing
costs and benefits of the current system and of a unified one, and then assessing the
differences.
A large-scale simulation analysis (which would be a conceivable alternative) has not been
attempted either (and would be beyond the scope of this study).
It is a well-established result in the litigation literature that private and social incentives
for litigation deviate. This is due to the fact that when one party engages in litigation it
takes its own private costs and benefits into account, but not the costs and benefits
incurred by other parties. Yet, litigation is beset with external effects, some of which are
listed here:
o Patent litigation can resolve uncertainty about the extent of patentability and the scope
of protection conferred by patents. Since valuable patents are most likely to the
subject-matter of disputes, the economic impact of uncertainty is potentially large in
these cases.
o Patent litigation judgments may provide patent offices with precedence information
and help to consolidate divergent patent office practices.
o Patent litigation limits the scope of strategic patenting by imposing limits on
opportunistic behavior while safeguarding the rights of owners of valid patent rights.
o Patent litigation corrects distortions in incentives for research and development (R&D)
by providing patent-holders whose rights have been harmed with proper recourse and
compensation or damages.
But patent litigation can also cause negative effects, such as delay for the other party and
uncertainty. To the extent that a court will not compensate the parties for these effects, a
complex mix of positive and negative effects will be present.
The theoretical literature has emphasized three different mechanisms that may lead parties
to fight a dispute through in court. These are:
o divergent expectations, which arise when uncertainty leads parties to different
expectations about facts of the case or the law;
o asymmetric information, which arises when one party has superior information on
particular aspects of the controversy and seeks to exploit this information in order to
extract rents;
10
o asymmetric stakes, which arise when the defendant in a suit is unable to adequately
compensate the patent-holder or the stakes are contingent on the outcome of
litigation.
2
Patent litigation will become more prevalent (i) as costs of litigation decrease, (ii) as the
stakes become higher and (iii) as divergence of expectations and asymmetry of
information increase.
Patent litigation can generate considerable externalities – negative as well as positive. This
is particularly true for revocation suits because the outcome determines the extent of entry
barriers.
3
While there has been little empirical research on patent litigation in Europe, the US
system has been scrutinized in some detail.
o The main finding of recent empirical studies is that, on aggregate, litigation in the
USA has increased in line with patenting.
o Another important finding is that in the USA small firms are more likely to be
involved in either the defence of their own patents (relative to patents) or as alleged
infringers of others’ patents (relative to R&D expenditure) than large firms.
4
The
literature points to problems for SMEs in settling disputes out of court.
o There is a host of studies which show again mostly for the USA that patent
litigation presents a serious and elevated cost to firms with small patent portfolios
relative to firms with large patent portfolios. Small firms appear to face substantially
higher marginal costs when protecting their patents than larger firms, and these costs
seem to have increased recently.
5
Small biotechnology firms even appear to avoid
patenting in certain patent classes strongly affected by patent litigation.
6
Thus, the
concerns of SMEs need to be taken into account when considering design options for
the unified Patent Court
o There is evidence that repeated interaction reduces the likelihood of litigation.
7
o Several studies have pointed out that new business models involving patents as assets
or as legal threats are of importance, and that the litigation system is likely to play an
important role in how these models will develop further. One such business model is
sometimes referred to as “patent troll” activities. A patent troll is a company that
acquires patents of failed companies or independent innovators and uses these to
threaten suit against alleged infringers, without having the intention of actively using
the patent they assert.
8
Patent trolls can obtain high quasi-rents in the USA because
many of those they attempt to hold up are unwilling or unable to fight a patent
infringement case through to a judgement in order to have the patent invalidated. Cost
2
This case will arise, for example, if the patent in dispute has a value to the patent-holder which goes beyond the
market in dispute. Somaya (2003) provides a good discussion of such cases.
3
See Graham and Harhoff (2006), Shapiro (2003), Farrell and Shapiro (2008).
4
See Lanjouw and Schankerman (2004), Meurer and Bessen (2005).
5
See Lanjouw and Schankerman (2004), Meurer and Bessen (2005).
6
See Lerner (1995).
7
See Lanjouw and Schankerman (2004).
8
This definition is quite slippery as patents are created in order to allow inventors to recoup fixed outlays on
R&D. Typically a suit will be classified as being brought by a patent troll if the patent being asserted is of
dubious quality; this often means that the patent is also very broad.
11
allocation rules and the cost of litigation are important aspects in this realm. The
extent of “troll activity” is uncertain in Europe, but recently patent funds have
acquired several thousand patents and may seek to enforce them.
o A large number of studies have focused on the creation of the CAFC (Court of
Appeals First Circuit) in the USA which was established in 1982 in order to counteract
the fragmentation of appeals courts. Some assessments are quite critical and point to
the CAFC as being responsible for the extension of patentable subject-matter and for
the overly pro-patent court posture in the USA. However, the analogy to the situation
in Europe is limited, since the USA never faced a problem of duplication of patent
litigations in various national courts. Nonetheless, important lessons can be learned
from the US case, and these are discussed in detail in section 7 of the report.
Taken together the literature gives some qualitative suggestions for the design of an unified
Patent Court system. The ideal patent litigation system operates at low cost levels for the
parties involved and generates precise and reliable judgements quickly. It is clear that these
objectives may have to be traded off against each other, but it is worthwhile repeating why
these aspects are important.
First, patent litigation can exert positive external effects these will not come about if
settlement is chosen instead of adjudication. But the choice between settlement and
adjudication is largely determined by the cost of litigation.
Second, the cost level of litigation determines to what extent a potential for hold-up exists.
As a US judge noted, high litigation costs distort patent trade and the patent system.
9
A third dimension of hold-up emerges from delayed proceedings literally, “time is
money”. Delays in proceedings translate into advantages for some of the parties involved.
Therefore, proceedings may be delayed for strategic purposes.
In a fast and low-cost litigation system with precise judgements, almost no potential for hold-
up and strategic behavior is left. These considerations have a number of important
implications which are considered in the section 7.
For the quantitative assessment of the impact of a unified patent litigation system in Europe, a
suitable framework has not been presented as yet. The next section first considers the status
quo of patent litigation in Europe and studies the features of particular systems from an
economic perspective. It then presents a concept for an approximate assessment of costs and
benefits of a unified system.
9
Ellis, T.S. (2000): “(…) It is, simply put, that the escalating, indeed skyrocketing litigation costs of the 1970’s
and 1980’s have distorted patent markets and patent economics.” This comment concerns the development in the
United States.
12
3 Status Quo of Patent Litigation in Europe and the Presidency’s Proposal
In section 3.1, the current system of patent litigation in Europe is briefly described. Strengths
and weaknesses of selected national systems are summarized. In section 3.2, the most recent
proposal made by the Presidency for the establishment of a European and Community Patents
Court is briefly summarized. Then, in section 3.3, the conceptual approach of the report is
summarized which will be used to assess the costs and benefits of a unified patent litigation
system.
3.1 The Current System and Its Shortcomings
3.1.1 Patent Litigation Systems in EU Member States
Patent litigation systems in EU Member States have evolved over several decades. During this
process, they have developed their own characteristics and features. The heterogeneity in the
utilization of patents, in the number of actions initiated before national courts and in system
designs (single vs. dual system regarding infringement and validity issues) is striking and
reflects the fact that, in the field of patent litigation, Europe is still lacking an integrated
jurisdiction taking full account of the single market.
Some countries have developed refined, specialized systems which attract a large number of
cases; among these are France, Germany, the Netherlands and the United Kingdom .
Typically, these are also the countries which have developed a strong propensity to generate
European (or national) patents. Specialized patent litigation courts have also been set up in
countries like Austria, Finland, Italy, and Sweden, inter alia. Specialized litigation systems
are favourable, since they allow for relatively fast court proceedings and for low error rates in
first instances, as well as fast feed-back from the jurisdictional to the administrative part of
the patent system, in particular patent examination. Heterogeneity in litigation systems is also
apparent in the extent to which different systems make use of technical expertise. While some
systems involve technical judges, others draw on extensive use of technical experts without
bringing the dedicated technical expertise “into judges’ chambers”.
As this report will document in the following sections, one crucial aspect of heterogeneity
concerns the costs of litigation. Costs will differ according to type, complexity and technical
field of the case, but may also differ significantly by jurisdiction, with particularly high costs
in common law countries. While a case with value of EUR 250,000 may cost each party in
first (second) instance proceedings EUR 50,000 (90,000) in Germany (and similar amounts in
France and the Netherlands), the costs of litigation can be at between EUR 150,000 and EUR
1,150,000 (150,000 to 1,000,000) in the United Kingdom.
10
Since procedures differ across
countries, the potential for economies of scale in duplicated proceedings is limited.
10
See Table 4.6 below for details.
13
3.1.2 Features of the Most Frequently Used Systems
No detailed comparison of all national systems has been undertaken in the literature.
However, both academics and practitioners have engaged in studies and assessments of the
most frequently used patent litigation systems in Europe.
11
This report draws on these
assessments in order to identify system components that offer particular advantages in terms
of efficiency, cost effectiveness and precision. For the purpose of this report, four national
court systems are particularly interesting since litigation in these courts accounts for about
90% of all patent litigation activity in the EU.
12
The German system
13
is considered particularly appealing for a large number of users.
Germany is said to attract anywhere between 50 and 70% of all patent litigation activity in
Europe. While it sets relatively high thresholds for injunctions and (up to the recent past) for
the collection of evidence by the plaintiff, practitioners emphasize the following advantages:
(i) the fast resolution of cases; (ii) the relatively low costs of litigation which allows SMEs to
participate in litigation; (iii) the high level of technical competence, and the "technical
quality" of decisions; (iv) concentration on a few, highly specialized courts; (v) the
parsimonious use of expert opinions (which elsewhere often lengthen procedures); infrequent
settlements and thus frequent adjudication of cases; rules that allow the winning party to
recover costs and fees; and an adequate level of damages. The presence of technically
qualified judges in revocation proceedings is seen as a positive feature. Clearly, as Leroux and
Bourguet point out, Germany also profits from stringent national standards in patent
examination. The bifurcation principle (i.e., the split between infringement and revocation
actions which are dealt with by different courts) is considered as both a shortcoming in some
respects, and as an advantage in others.
The Dutch system
14
also has a rather positive image regarding patent litigation. However, it
is noted that it mainly accommodates large firms and that it offers less opportunity to the
winning party of cost recovery. Moreover, securing evidence is more burdensome to the
plaintiff than – for example – in France. On the positive side, Dutch proceedings work
quickly, there is a rather strict time framework for the cases, efficient summary proceedings
are available, and the level of damages to compensate patent-holders for actual infringement
is adequate.
The UK system
15
is the most costly one, and this aspect is generally noted as negative. Costs
are also considered to be a decisive factor in generating a large number of settlements in the
UK system. The rarity of preliminary injunctions is also noted as a drawback by practitioners.
On the positive side, the UK courts are considered highly competent and experienced,
11
In the following sections, the report draws in particular on a comparative assessment prepared by Leroux and
Bourguet (2006) at Bird & Bird and on seminar materials authored by Dr. Sabine Rojahn and colleagues for
Taylor Wessing.
12
See Annex 1, paragraph 2 of WPL/11/05 "Assessment of the impact of the European Patent Litigation
Agreement (EPLA) on the litigation of European patents", dated 1.12.2005.
13
Cf. Leroux and Bourguet (2006, pp. 17-25)
14
Cf. Leroux and Bourguet (2006, pp. 26-30)
15
Cf. Leroux and Bourguet (2006, pp. 31-35)
14
proceedings are very fast, the timetable is organized very strictly, there are satisfactory means
of cost recovery and adequate damage awards. To some parties, the availability of a coercive
method of securing evidence (“disclosure”) is attractive.
In France
16
, another “popular” litigation forum in Europe, the main comparative advantage
was for many years the saisie-contrefaçon which allowed plaintiffs to secure evidence in a
highly effective manner. Enforcement measures like the saisie-contrefaçon are now included
in Directive 2004/48/EC on the enforcement of intellectual property rights
17
and should by
now be available in all EU Member States. Moreover, the relatively low costs, the unitary
design of the litigation proceedings (validity and infringement are dealt with in one
proceeding), and the experience of judges are emphasized as contributing to the status of the
French courts. Potential weaknesses are, according to Leroux and Bourguet, the duration of
proceedings and the relative lack of technical expertise in courts.
Several national systems have seen reforms lately, some of these in the context of the
implementation of Directive 2004/48/EC which covers remedies available to owners of
intellectual property rights in civil courts. Examples are the introduction of a disclosure-style
instrument for obtaining evidence in Germany, of a similar instrument in the Netherlands
since 2002, and of a streamlined procedure in the UK. These reforms reflect efforts to
increase the efficiency of patent litigation, but they are also due to the fact that there is
competition among the various systems to attract “business” into the respective national
forum.
The next section turns to patent litigation in the overall system. The discussed features of the
four most prominent litigation forums play a role in these considerations, but the most
important aspect is the impact of fragmentation on the decision-making of parties engaging in
litigation and on economic outcomes and welfare.
3.1.3 Litigation in a Fragmented System
Patent litigation in a fragmented system with large institutional and cost differences leads to a
proliferation of litigation tactics and strategies which may cause hold-up problems and
wasteful duplication. Moreover, the overall working of the patent system is affected by these
aspects. Despite the infrequent occurrence of patent litigation (in particular at appeal and
supreme court level), the importance of cases can be considerable. Patent litigation cases
occur in two basic forms: either as revocation proceedings challenging the validity of patents
granted by the respective patent authority or as infringement proceedings seeking to enforce
patent rights. The likelihood of a patent being involved in litigation
18
at some point during its
term is estimated as between 1% and roughly 3% in most patent systems, with some variation
across technical domains, industries and countries. Patent litigation is known to occur
16
Cf. Leroux and Bourguet (2006, p. 49ff.)
17
Directive 2004/48/EC of 29.4.2004, OJ EU L 157, 30.4.2004 (corr. in OJ EU L 195, 2.6.2004, p. 16).
18
Statistical statements regarding the likelihood of patent litigation can be expressed in two ways: i) as the
likelihood of a patent being subject to litigation proceedings during its term (from application to lapse date), and
ii) as the number of patents litigated in a given year divided by the total number of patents in force in that year.
15
particularly frequently (i) for valuable patents, (ii) in the presence of divergent assessments of
case quality and (iii) in the presence of asymmetric information. Patent litigation is thus “the
tail that wags the dog of the patent system” litigated cases provide legal precedence and
important signals to patent holders, potential infringers and third parties seeking to steer free
of patent conflicts. A well-designed litigation system is therefore the capstone of any patent
system, and conversely, a flawed litigation system may effectively counteract any welfare
gains from such a system or cause welfare losses of its own.
Duplication.
19
Since infringement and validity of European patents fall under the jurisdiction
of national courts, patent-holders and parties seeking to revoke granted patents may have to
enter into litigation in multiple countries. The exact extent of duplication is unknown. While
there are a number of high-profile cases with extreme duplication and heterogeneous
outcomes
20
, there are currently no reliable statistics that would allow us to compute with
precision the incidence and costs of duplication.
Divergent outcomes. Case duplication may lead to divergent outcomes as has been observed
in practice. In a fragmented court system, the divergences may never be consolidated (e.g., in
a second instance ruling). Instead, they may persist. The impact of divergent outcomes is
complex. Per se, fragmentation does not necessarily lead to uncertainty – unless the rulings in
the courts duplicating the case are subject to enhanced uncertainty themselves (which may
very well be the case if non-specialized courts are involved, as is currently the case in
Europe). The interpretation of claims and the assessments in “product clearing” or in private
settlement negotiations may thus become ambivalent and uncertain if conflicting legal
precedents co-exist. Moreover, divergent outcomes contribute to a fragmentation of the patent
system since the geographic scope of a European patent now depends on the divergent
national interpretations of patent law. Leaving aside the increased cost of litigation, the non-IP
cost of doing business in the EU is raised significantly, since investments as well as
production and distribution decisions may have to be tailored to the respective national extent
of patent protection. Thus, divergent outcomes of national patent litigation proceedings
clearly hamper a smooth operation of the EU's Single Market.
Private Consolidation of Court Proceedings. Duplication is costly, and the parties to legal
disputes regarding patent rights tend to seek reductions of the cost of conflict resolution. The
parties may agree on one court location and on abiding by that court’s decision. In this case,
an efficient solution to the legal controversy can be found. Indeed, if all of the disputes would
19
Duplication as referred to in this report does not require that exactly the same legal matter is brought by
identical parties into different national courts. For the purpose of the computations below, we can speak of
duplicated cases if the introduction of the unified Court would render one or several of the cases unnecessary, i.e.
if the different national cases are substitutes in a legal and economic sense.
20
In the Epilady case (EP0101656), infringement suits of the patent-holder were successful in Belgium,
Germany, Italy and the Netherlands, but not successful in Austria, France and the United Kingdom. In Securities
System Inc. vs. ECB (EP0455750), the German and Dutch courts upheld the patent, while it was revoked in
France and the UK. In the Senseo case (EP0404717), initial divergent rulings have been issued by Belgian and
Dutch courts, but several other national cases are pending. In the Monsanto case (EP0546090), the District Court
The Hague gave an interim judgment on March 19th, 2008 and referred the case to the European Court of Justice
for an interpretation of Directive 98/44/EC of 6.7.1998 on the legal protection of biotechnological inventions;
several parallel cases are pending in different Member States.
16
be guided by a court judgement (in one Member State) and followed by settlement (in other
Member States), then this would constitute a reasonable solution without the cost of
duplication. In terms of economic theory, one court judgement can be taken as an indicator of
future rulings; and thus the extent of diverging expectations or informational asymmetries
would be reduced. It is unclear to what extent such settlements take place in the current
fragmented system.
However, it would be unrealistic to assume that parties will always want to forego the
strategic manoeuvring space that the existence of multiple forums gives them. When the
stakes are high, a patent holder will always want to enforce the patent right in at least some of
the multiple jurisdictions in Europe; similarly, an alleged infringer will always seek to revoke
the patent in at least some jurisdictions. Nonetheless, it is important to keep in mind that in
some cases both parties will try to find a cost-efficient solution to a controversy. Such
tendencies will also persist in a unified system. In other cases, litigation in multiple forums
may not be necessary. In industries with relatively concentrated production, an infringement
suit in one country may suffice to lead to resolution of the conflict. However, to the extent
that commerce in Europe becomes increasingly border-crossing, such cases may become less
important in the future. To summarize, not all cases will automatically be duplicated, and the
extent of duplication becomes an important empirical parameter for the analysis in this
report.
21
Patent Revocation for Market Access. The converse is true in some sectors where stakes are
high and where patent protection has a major impact on the entry decisions of producers. An
example of considerable economic importance is the pharmaceuticals sector where two types
of firms are active: originator firms with extensive R&D operations and generics producers
who rely mostly on efficient manufacturing capabilities in order to produce generic
pharmaceuticals which are no longer protected by patents. The European Commission (DG
Competition) presented on 29.11.2008 a preliminary report with results of its Pharmaceutical
Sector Inquiry (DG Competition 2008
22
). The preliminary results suggest inter alia that
there is a high degree of duplicated litigation in this sector. Frequently, revocation actions are
successfully resorted to by generics firms in order to obtain access to national health markets,
but market entry by generics firms is impeded by the need to enter into revocation
proceedings in multiple jurisdictions.
Strategic Litigation and “Forum Shopping”. The efficiency of private settlements is also
reduced by strategic behavior. The fragmentation of the litigation system has led to the
development of refined strategies where the attacking party may file a case in a strategically
selected court system first and enter into settlement negotiations afterwards. Naturally, the
21
It is important to realize that in cases where a dispute is resolved by an adjudication in one court, considerable
costs of private settlement may still be felt by the parties if they seek to settle in other jurisdictions. Ideally, a
cost-benefit assessment should take such costs into account as well, but the costs of private settlements are even
harder to assess than the costs of court proceedings. Note that accounting for these costs would affect the cost-
benefit computation largely in favour of the unified system.
22
The preliminary report is available at
http://ec.europa.eu/competition/sectors/pharmaceuticals/inquiry/preliminary_report.pdf
17
plaintiff will usually choose the most convenient and privately beneficial forum for first
litigation actions. Forum shopping” occurs not only where courts in different countries have
jurisdiction but within countries with multiple entry points for litigation, such as Germany and
France. But European “forum shopping” can clearly exploit a much wider heterogeneity of
systems than “forum shopping” between, for example, the Munich and the Düsseldorf courts.
Cross-Border Injunctions and Litigation. In the late 1990s, some courts began to issue
injunctions which reached beyond the territorial boundaries of the respective jurisdiction. This
“legal innovation” started in the Netherlands and was subsequently picked up by courts in
various other countries. Patent-holders were allowed to start infringement proceedings in a
Dutch court, not only based on a European patent validated in the Netherlands but also on
other national patents derived from the same European patent. The Dutch court would assume
jurisdiction in cases in which the infringer was domiciled in the Netherlands or when the
Dutch patent was being infringed. The court would then apply the respective law of the
country where the patent was in force and where the plaintiff sought to obtain an injunction
and would possibly grant a cross-border injunction. The approach was based on Article 5(3)
of the Brussels Convention (now Regulation 44/2001 on jurisdiction and the recognition and
enforcement of judgments in civil and commercial matters
23
) which allows for proceedings in
the country of infringement. A subsequent development was the “spider in the web” concept
which led Dutch courts to assume jurisdiction where the main defendant (the “spider”) was
located in the Netherlands and other defendants (e.g. subsidiaries) were part of a group of
enterprises (the “web”) acting in concert. This judicial practice soon became highly
controversial, with some national courts following the Dutch example (e.g. in Germany) while
other courts (e.g. in the UK) refused to assume jurisdiction over foreign patents. From the
perspective of patent-holders, cross-border injunctions were highly attractive since they
lowered the cost of litigation and opened new avenues in terms of litigation strategy.
On July 13, 2006 the European Court of Justice (ECJ) ruled in two important cases,
effectively putting an end to the practice of cross-border injunctions.
In GAT v. LuK,
24
the ECJ ruled on the interpretation of Article 22(4) of Regulation
44/2001. The ECJ declared that the national courts of the State of registration of a
patent have exclusive jurisdiction over all proceedings relating to the validity of that
patent, and that courts of other Member States have no jurisdiction.
The decision of the ECJ in Roche v. Primus and Goldberg
25
effectively ended the
possibility of national courts granting cross-border injunctive relief for infringement of
European patents, even against related companies acting in concert. The legal
innovation of cross-border injunctions which from the perspective of many patent-
holders and practitioners – had held the promise of reducing the costs of patent
infringement litigation was thus removed, and the need to duplicate proceedings may
23
Regulation (EC) 44/2001 of the Council, 22.12.2000, OJ EC L 12, 16.1.2001.
24
Case C-04/03, Gesellschaft für Antriebstechnik mbH & Co. KG v. Lamellen- und Kupplungsbau Beteiligungs
KG, ECJ Report 2006, I, 6509.
25
Case C-539/03, ECJ Report 2006, I, 6535.
18
in future be felt more strongly than prior to these ECJ rulings. These rulings therefore
inevitably raise the cost of litigation because they force the party seeking to enforce a
patent to initiate actions in multiple countries.
26
Delay Strategies and “Torpedoes”. The fragmentation of the system has also created
opportunities for a number of delay strategies. The best-known among these are “torpedo
motions” actions for declaration of non-infringement in court systems which are known or
alleged to work very slowly. According to Article 27 of Regulation 44/2001, any court not
first seized with an action must decline jurisdiction or stay the proceedings when another
action is filed subsequently. Thus, while the action for declaratory statement is pending in the
“slow” court, an infringement action in other courts is essentially blocked. Delays of this kind
can conceivably be turned into settlement conditions favourable to the party that initiated the
delay. In the Member States where “torpedo tactics” were employed, there have been a
number of reform measures attempting to make “torpedoes” less effective. Moreover, some
courts have become reluctant to accept motions that are clearly meant to merely delay an
infringement case. But some potential for these and other forms of strategic litigation remain
present in a fragmented system.
Summary. The main aspects of the discussion can be summarized in three points:
- wasteful duplication: whenever multiple litigation is undertaken, resources are wasted on
duplication without generating concomitant benefits;
- raising the cost for appropriating returns from patented inventions in Europe: leaving
aside the cost of litigation, cross-border commerce is made more difficult and costly when
diverging outcomes (patent protection in some, no patent protection in other EU Members
States) prevail;
- delay and hold-up: the system can be used to delay decisions in infringement cases or to
raise the costs of entrants seeking access to the market; this may either reduce innovation
incentives or the level of competition in a way that is welfare-reducing.
The creation of a unified court system could lead to improvements along all three dimensions.
Of these three, the first lends itself most easily to a quantitative treatment. The next subsection
describes the basic features of the most recent proposal made by the Czech Presidency of the
European Council. This summary is helpful since unification and integration are abstract
terms the actual outcomes of policy measures depend crucially on the institutional design
that is used to implement those abstract concepts.
26
Cf. Luginbühl and Stauder (2006).
19
3.2 The Presidency’s Proposal
3.2.1 Elements of the Proposal
The Czech Presidency, in Working Document 5072/09 (Draft Agreement on the European
and Community Patents Court and Draft Statute, dated January 8
th
, 2009) has put forth a new
proposal for a unified Patent Court, to be named "European and Community Patents Court".
27
Essentially, the proposal foresees the establishment of a unified patent litigation system the
unified Patent Court - which will cover both European patents as well as future Community
patents. This proposal has revived the policy discussion after the EPLA proposal did not find
sufficient support in the Council.
Jurisdiction. The proposed Court will have jurisdiction over Community patents and
European patents which are in force when the Agreement enters into force or granted
thereafter, as well as any applications pending at that date. Article 58 of the draft Agreement
provides for two opt-out clauses. Most importantly, during a transitional phase of seven years
after entry into force, proceedings concerning European patents may still be initiated before
the respective national courts.
Design of the Court System. The Court will consist of a Court of First Instance, a Court of
Appeal and a Registry. The Court of First Instance will consist of a central division as well as
local and regional divisions. Local divisions may be set up in any Member State. A Member
State may decide to have up to three local divisions if the annual number of litigation cases is
sufficiently high.
28
The seats of local divisions are determined by the respective Member
States which also provide the facilities. Regional divisions allow Member States not hosting a
local division to participate in the court system via regional divisions which may sit at
multiple locations. The seat of the central division and the seat of the Court of Appeal are still
to be determined.
Composition of Panels. All first instance court panels will have a multinational composition
and will consist of three judges. In local divisions, two of the three judges will be permanent
judges, a third one will be from a Pool of Judges. In certain situations, the third judge will be
permanent as well. Two of the three judges at a regional division will be nationals of the
participating Member States, the third will be seconded from the Pool of Judges. The panels
of the Court of First Instance will include technically qualified judges who are specialized in
the relevant technical field. The panels of the Court of Appeal will consist of five judges,
three legally qualified ones and two technically qualified ones.
27
The following overview focuses on particular aspects and does not intend to present a full summary of the
draft Agreement. Various elements of the proposal are still under discussion. A recent progress report notes that
these include “(…) the composition of the judicial panels, the language arrangements, jurisdiction on validity,
control exercised by the Court of Justice, the financing of the judicial system and the transitional arrangements”
(see Council Progress Report 15674/08 of 14.11.2008, available at http://register.consilium.europa.eu/pdf/en/
08/st15/st15674.en08.pdf).
28
The proposal requires that more than one hundred litigation cases have been initiated during three successive
years in the Member State in order to set up an additional local division.
20
Registrar and Advisory Committee (Articles 8-9). The Registry of the Court will be set up at
the location of the Court of Appeal and will have the usual “house-keeping” obligations. It is
supported by sub-registries at each court divisions of the first instance. The registry is also
charged with publishing reports containing statistical data. The proposal foresees the setup of
an Advisory Committee comprised of experienced judges and practitioners.
Selection of Judges (Articles 10-12) and Pool of Judges (Articles 13-14). Appointment is a
multi-stage process by which a list of candidates is established by the Advisory Committee,
the Council then selects nationals of Member States who it proposes to the Mixed Committee,
and the Mixed Committee decides by common accord on the appointment. The proposal
foresees the setting up of a Pool of Judges, comprised of all appointed judges from the Court
of First Instance. Allocation of judges from the Pool to local or regional divisions is made by
the Court’s President, based on the judges’ legal, technical and language skills. A training
framework is to be implemented in order to maintain and build expertise in technical and legal
aspects as well as in civil procedures among appointed judges and candidate-judges.
Substantive Law (Article 14). As in the case of courts which had assumed jurisdiction in cases
with cross-border litigation aspects, the unified Patent Court will have to apply the EPC,
Community law, national law implementing Community law and national law. The Court will
also have jurisdiction over Community patents and will thus have to apply the future
Regulation on the Community patent.
Jurisdiction and Effects of Decisions (Articles 15-16). Actions for revocation or declaration of
non-infringement have to be brought before the central division. Such actions do not have to
exhaust opposition at the EPO first. Actions for actual or threatened infringement shall be
brought before either the local or regional division of the Member State or Member States
where the infringement occurs, or the respective regional division or the local division hosted
by the Member State where the defendant resides. Cases are heard by the central division if no
local or regional division exists in the respective location.
Patent Mediation and Arbitration (Article 17). The proposal includes provisions for setting up
a mediation and arbitration center. This can be seen as a measure towards the reduction of the
cost of litigation for the parties involved. For good reasons
29
, validity issues are exempt from
the mediation or arbitration process.
Financing of the Court (Articles 18-21). The budget of the court is to be financed from fees
and other incomes as well as contributions from the Community budget and from non EU
Member States which are Parties to the Agreement.
Organization and Procedures (Part III, Articles 21-34). For the purposes of this report, it is
noteworthy that the Presidency proposal includes various measures that attempt to allow for
relatively low litigations costs. Among these are provisions for case management (Article 24)
29
If one party has private evidence that may lead to the revocation of the patent, a private settlement may be an
attractive outcome in which a duopoly of the two parties could be maintained – albeit at some cost to the public.
A similar comment applies to Article 52 of the proposed agreement.
21
and the representation of parties by European Patent Attorneys who have specialised in patent
litigation (Article 28). Experts may be appointed by the Court, and the Court may impose
provisional or permanent injunctions. Moreover, the Court can order corrective measures. The
Court can revoke patents, fully or in part, or amend them following a direct action for
revocation or a counterclaim for revocation. The award of damages is supposed to restore the
infringed party to the position it would have been in without infringement. Damages are not
supposed to be punitive. Costs are generally borne by the unsuccessful party, unnecessary
costs caused by one party shall be borne by that party. Legal aid shall be available to parties
unable to meet the costs of the proceedings.
Mixed Committee (Article 57). A central governance body is the Mixed Committee in which
each Contracting Party has one vote and the Community is represented by the Commission.
Decisions are generally made by a three quarter majority of the represented Contracting
Parties. The Mixed Committee has influence on the setup of the Advisory Committee, the
selection of judges, amendments of the Statute, the determination of court fees, the rules for
legal aid and the determination of financial regulations and the budget.
Transitional Period (Article 58). During seven years after the establishment of the Patent
Court, infringement and revocation proceedings may still be initiated before the national court
which currently has jurisdiction.
3.2.2 Relevance for the Estimation of Costs and Benefits
Several elements of the proposal have important implications for the level of litigation costs.
From the proposal, it becomes clear that the Presidency is seeking to set up a system with
relatively low costs of litigation. Among the cost-reducing elements in the proposal are: (i) the
utilization of existing infrastructure and human resources, in particular experienced judges;
(ii) mediation and arbitration schemes; (iii) contributions from the Community and from
States which are not EU Member States; (iv) active case management; (v) representation by
experienced and specialized European Patent Attorneys.
It is beyond the scope of this report to consider the efficacy of these measures. Yet, in order to
anchor the scenario computations in section 6 around reasonable priors, it is important to
relate the future level of litigation costs to those in the current fragmented system.
Discussions in the Working Party on Litigation in 2005 yielded the result that litigation costs
in proceedings before the EPLA Court would be 2 to 3 times higher (due to high court fees
and high representation fees) than costs in the national systems in Germany, France and the
Netherlands. The Presidency's proposal for a unified Patent Court clearly attempts to install a
system with lower litigation costs, and two of the most important steps towards this end
would be (i) contributions from public budgets to financing of the Court and (ii)
representation of parties by specialized patent attorneys. Hence, the initial benchmark in the
cost-benefit estimations will be based on the assumption that the unified Court will offer
patent litigation at the relatively low cost levels comparable to the current German system.
Possible implications of a deviation from this baseline will then be explored in separate
22
calculations in order to demonstrate the impact of the level of litigation costs on the cost-
benefit balance of the unified court system.
30
3.3 A Conceptual Framework for an Economic Assessment of Costs and
Benefits
This report seeks to shed light on the costs and benefits of the establishment of a unified and
integrated Patent Court system in Europe. To do so, it is helpful to consider the impact of the
proposed Court on different types of parties engaging in litigation. The immediate impact of
the new system would be threefold:
1. The establishment of a unified Patent Court would effectively eliminate the need for
engaging in litigation regarding the same legal matter in multiple jurisdictions. The benefit
from the new system is then the savings from avoiding duplication. Duplication occurs
particularly frequently in industries such as the pharmaceutical industry where stakes are
high and where access to markets is highly regulated and contested.
2. The private costs of litigation under a unified Patent Court will differ at least in some
countries from the currently prevailing cost level. Since the Presidency’s proposal seeks
to establish a court system with low private costs, this report will initially assume that
private costs of litigation in the unified Patent Court will correspond to the relatively low
costs of patent litigation in the current German, French and Dutch systems. Effectively,
this means that with the introduction of the unified Patent Court, the cost of litigation
would become lower for parties now undertaking litigation in high-cost countries such as
the United Kingdom. Under the stated assumption, there would thus be cost savings from
the establishment of the Patent Court.
3. Moreover, various demand effects might occur due to (i) changes in the geographic reach
of the rulings of the Court, (ii) changes in the quality of the court system and (iii) “price”
changes.
(i) The geographic reach of court rulings will increase outcomes do no longer apply to
one country only, but to all the Member States where the patent at stake is in force.
(ii) For several countries, the establishment of the unified Patent Court will mean
improvements in the quality of the court system. At given costs of litigation, these effects
will make litigation more attractive for some parties.
(iii) Changes in the costs of litigation will also impact the demand for litigation: lower
costs will induce parties to prefer litigation in court over settlements or no action; higher
costs will have the inverse effect.
30
Moreover, beyond the negative effect that relatively costly litigation has on the computed savings from
avoided duplication, there is also a strong incentive effect best captured by Ellis (2000, 24): “In sum, then, the
pernicious effect of the escalating expense of patent litigation is that it artificially discourages court challenges to
patent validity and thereby contributes to the risk that invalid patents will pollute the market.”
23
The cost savings from the first two effects will be estimated in section 6 of the report. Due to
the lack of precise data, it will be necessary to employ scenario techniques and parameter
estimates to establish an estimate of the benefits from avoiding duplicated litigation. The
demand effects described above are even more difficult to quantify. A qualitative assessment
is again provided in section 6, following the derivation of the data framework in section 4 and
estimates of the operational costs of the unified Patent Court in section 5. Additional
qualitative aspects are then dealt with in section 7.
24
4 Incidence and Cost of Patent Litigation in Europe
Currently, no official statistical data on incidence and cost of patent litigation in Europe exist.
In this section the report collects and describes data which together with a number of
simplifying assumptions – will permit the estimation of the overall litigation activity in
Europe as well as the degree of duplication that is currently present in the system. Towards
that objective, various partial datasets are described and compared below, before settling on a
number of plausible scenarios.
4.1 Survey Data from a Study carried out under German Presidency
A particularly interesting and comprehensive source for the calibration and triangulation of
the scenarios to be developed below is the statistical data contained in a survey undertaken
under the German Presidency
31
. This survey delivers the most comprehensive data in terms of
Member States surveyed.
The data are summarized in Table 4.1 below.
An estimate of the number of litigation cases is given in the second column.
The third column indicates if the respective court system makes use of technically
qualified judges.
The extent to which inventors and companies in the particular country generated European
patent applications in 2005 is indicated in the forth column.
While Table 4.1 is highly informative about the heterogeneity of national patent litigation
systems, its incidence estimates include both national patents and European patents, but no
data on the extent of duplication.
31
See Council document WD 11622/07 PI 135, dated 12.7.2007, available at
http://register.consilium.europa.eu/pdf/en/07/st11/st11622.en07.pdf.
25
Table 4.1 – Characteristics of National Patent Litigation Systems in EU Countries
Country Number of Cases (first instance only,
European and national patents)
Specific Comments EP Patent Applications from
Country in 2005
Austria (AT)
2005: 25 (V), 17 (I), 2006: 12 (V), 19 (I) dual system, technical judges 1,728
Belgium (BE)
30 cases p.a.
single system, no technical judges 1,760
Bulgaria (BG)
<5 p.a. dual system, no technical judges 22
Cyprus (CY)
n.a. (very few) single system, no technical judges 13
Czech Republic (CZ)
2005: 4 (V), 5 (I) dual system, no technical judges 167
Denmark (DK)
10-15 cases p.a. single system, technical judges (1
st
instance) 1,321
Estonia (EE)
n.a. (very few) single system, no technical judges 16
Finland (FI)
15-20 cases p.a.
single system, no technical judges 1,761
France (FR)
2005: 459 cases, 2006: 487 cases single system, no technical judges 9,891
Germany (DE)
220 (V), 600-700 (I) p.a. dual system, technical judges 27,584
Greece (GR)
n.a. single system, no technical judges 101
Hungary (HU)
<10 (V), 15-20 (I) dual system, technical judges 218
Ireland (IE)
1-2 cases p.a.
single system, no technical judges 338
Italy (IT)
270-300 cases p.a. single system, no technical judges 5,429
Latvia (LV)
1-2 cases p.a.
single system, no technical judges 18
Lithuania (LT)
n.a. (very few) single system, no technical judges 11
Luxembourg (LU)
n.a. (very few) single system, no technical judges 121
Malta (MT)
n.a. (very few) single system, no technical judges 7
Netherlands (NL)
50 cases p.a.
single system, no technical judges 4,268
Poland (PL)
20 cases p.a.
dual system, no technical judges 160
Portugal (PT)
20 cases p.a.
dual system, no technical judges 89
Romania (RO)
2005: 20 cases, 2006: 13 cases dual system, no technical judges 36
Slovak Republic (SK)
5 (V) p.a., 5 (I) p.a. dual system, no technical judges 46
Slovenia (SI)
3-6 p.a. single system, no technical judges 138
Spain (ES)
n.a. single system, no technical judges 1,752
Sweden (SE)
30-50 cases p.a.
single system, technical judges 2,949
United Kingdom (GB)
2004: 153 cases, 2005: 54 cases single system, no technical judges 7,353
Source: Annex of WD 11622/07 PI35 – For Italy: personal communication with Prof. Mario Franzosi. Patent application counts are based on own computations using PATSTAT
2008-10 and EPASYS databases. A total of 67,297 EPO applications were filed by inventors and companies located in the tabulated countries.
Legend: V - validity, I - infringement.
26
4.2 Data from Regional Courts in Germany (infringement actions)
In Germany, 12 Regional Courts (Landgerichte, LG) have jurisdiction to deal with the
infringement of German patents and of European patents validated in Germany. In practice,
only three of them – the Regional Courts in Düsseldorf, Mannheim and Munich – are of major
importance and attract between 80% and 90% of all patent infringement cases in Germany.
While the Courts do not provide any official statistics, judges at the courts have been
collecting data
32
and an estimation of the actual number of patent cases can be made. Table
4.2 presents the distribution of cases at the Mannheim Regional Court according to subject-
matter. Between 84.9% and 91.6% of the cases are patent cases, the remainder being
concerned with utility models and inventor compensation issues. These data, while not
originating from the largest patent court in Germany, will allow us to estimate the share of
genuine patent cases at other Regional Courts.
Table 4.2 – Cases at the Mannheim Court by Subject-matter
Year
Subject-matter 2005 % 2006 % 2007 %
patent cases 197 84.9% 207 90.0% 273 91.6%
utility model cases 23 9.9% 20 8.7% 18 6.0%
inventor compensation cases 12 5.2% 3 1.3% 7 2.3%
Total 232 100.0%
230 100.0%
298 100.0%
Source: Prof. Dr. Christoph Ann - personal communication
These data allow us to correct the actual number of patent infringement actions initiated at the
three Regional Courts of Düsseldorf, Mannheim and Munich (assuming that the above
distribution holds for the two other Courts as well). Moreover, we will assume that these three
Courts attract 5/6 (83.3%) of all infringement actions initiated in Germany. This figure is an
estimate that was discussed with a number of practitioners.
Under these assumptions the number of infringement cases can be estimated to be as
displayed in Table 4.3.
32
I would like to thank Prof. Dr. Christoph Ann (Technische Universität München), former judge of the
Mannheim Court, for making these data available.
27
Table 4.3 – Estimated Patent Infringement Cases by Court and Year
Year
Court 2005 2006 2007
LG Düsseldorf 530 444 555
LG Mannheim 232 230 298
LG München I 89 95 117
Total 851 769 970
Estimated Patent Cases (90%) 766 692 873
Total - Patent Cases in Germany 919 831 1048
Source: Prof. Dr. Christoph Ann - personal communication, own estimates for
estimated patent cases and total patent cases in Germany
Note that these cases include actions relating to national patents granted by the German Patent
and Trademark Office (DPMA) as well as actions relating to EPO-granted patents.
4.3 Data from the German Federal Patent Court (revocation actions)
Official statistics are available for the number of revocation cases at the Federal Patent Court
in Germany (Bundespatentgericht) (see Table 4.4).
Table 4.4 – Revocation Cases at the Federal Patent Court - 1998-2007
Year Cases Decisions
Pending
Cases
1998 172 138 211
1999 146 156 201
2000 189 177 213
2001 166 147 232
2002 163 168 227
2003 181 174 234
2004 200 147 287
2005 225 183 329
2006 221 199 351
2007 234 235 350
Source: Annual Reports of the Federal Patent Court (2000-2007)
Again, the filing numbers do not distinguish between DPMA- and EPO-patents, but additional
micro-data (see below) will be used to derive the distribution. It is evident from Table 4.4 that
the incidence of revocation cases has been increasing over the time period from 1998 to 2007.
Figure 4.1 displays the upward trend in filings and in the number of pending cases and
decisions.
28
Figure 4.1 – Revocation Cases, Decisions and Pending Cases at the Federal Patent Court
0
50
100
150
200
250
300
350
400
1999 2000 2001 2002 2003 2004 2005 2006 2007
Filings Decisions Pending Cases
Source: Annual Reports of the Federal Patent Court (2000-2007)
4.4 Data from the Cremers and Harhoff Studies
As described in the literature survey below (Annex II), Cremers (2006) obtained data on a
sample of litigation cases at the Düsseldorf and the Mannheim Court. Harhoff (2004) used
data provided by the Federal Patent Court to study outcomes of revocation proceedings and
their relationship to opposition proceedings. These data can be used to estimate the following
parameters: (i) the share of EPO-granted patents among patents under litigation and (ii) the
distribution of cases over technical fields.
Cremers (2006) used information on all litigation cases filed between 1993 and 1995 at the
Mannheim and Düsseldorf Courts. Of the 824 patents involved in litigation, 26% had been
granted by the EPO and 77.1% were owned by German proprietors. Moreover, in 25% of the
cases, a revocation action had been initiated. In the case of EPO-granted patents, revocation
actions were initiated in 33% of all litigation cases. Since the Cremers study deals with cases
filed in the early 1990s, the data are likely to be non-representative for the first decade of the
21
st
century. In particular, one would expect that the share of EPO-granted patents in patent
litigation proceedings has risen steadily over the years.
This is indeed the pattern that the data from the Federal Patent Court suggest. Using two
micro-level datasets
33
on revocation proceedings, one for DPMA-granted patents and one for
EPO-granted patents, we can show that by 2003, 60% of revocation actions initiated at the
Federal Patent Court related to European patents. Assuming that 33%of all European patent
litigation cases trigger revocation cases (while only 25% of litigation cases against DPMA
granted patent do), we estimate that approximately 50% of all litigation cases in German
33
The datasets were kindly provided by the Federal Patent Court in 2004. I would like to thank Lutz von Raden
and Thomas Baumgärtner for their support.
29
courts are concerned with EPO-granted patents. Note that together with the “hard data” from
Table 4.3, this implies that we should expect a total of about 525 infringement cases relating
to European patents in Germany. We will use this implication later on in order to demonstrate
that the various partial datasets can be combined to give a coherent picture.
4.5 Data from the CJA Study
In 2003, CJA Consultants was commissioned by the European Commission to undertake a
study on the potential of patent litigation insurance. In the course of the investigation, CJA
constructed a set of estimates regarding the incidence of patent litigation in several European
countries. For countries with revocation and infringement procedures in two separate
proceedings
34
, separate estimates were presented. Table 4.5 summarizes this information.
Table 4.5 – CJA Estimates of Case Filings Related to EPO-granted Patents
Country
Patent Stock
at Risk
CJA Estimates
for 2004
Germany* 307,488 500+500
France 252,798 50
UK 257,600 120
Netherlands 121,337 50
Spain 97,146 50
Belgium 84,621 30
Austria 83,636 10
Sweden 82,125 15
Denmark 45,067 15
Greece 27,963 20
Finland 18,239 10
Poland* 12,457 15+15
Total (Filings)* 323,672 1,228
Source: Appendix 1 and Appendix 2 in CJA Consultants (2003). Patent estimates refer to
2004.
* Note: In Germany and Poland, the first number refers to infringement, the second to
revocation cases. In the aggregate (TOTAL), 833 German and 25 Polish cases were
assumed to have been initiated, taking into account that some revocation cases would not
cause separate litigation in a court system without separation of infringement and
revocation proceedings.
Clearly, for Germany, some of these estimates are in contradiction to the statistics from the
German Courts, summarized in Table 4.3 and Table 4.4. German patent infringement cases
are underestimated in the CJA data, while revocation cases are apparently overestimated.
Moreover, the data for France are highly inconsistent with those of the Survey carried out
under German Presidency (referred to in section 4.1) and of the Working Party on Litigation
(see Table 4.6 below). The overall likelihood of litigation in the CJA data appears
34
The separation of infringement and litigation proceedings is often referred to as bifurcation of proceedings.
30
underestimated for some, and overestimated for other countries.
35
Leaving aside the
distribution of cases by country, the CJA data imply that about 68% of all actions are initiated
in Germany.
4.6 Practitioner Estimates
In several discussions, estimates of the incidence of patent litigation were elicited from a
number of practitioners, both judges and attorneys. Moreover, some estimates are published in
internet sources.
Véron (2001) provides a detailed study of litigation activity in France over the time period
1990 to 1999. The share of European patents involved grows from 17.5% to 30.8% from 1990
to 1999. Véron reports that 24% of the plaintiffs (mostly from DE and US) and 30% of the
defendants are foreign entities (mostly from IT, DE and US). Comparing the case filings, he
has 455 cases in Germany, 350 in France, 120 in the Netherlands and 100 in the UK. Italy
leads in terms of case filings (650), but supposedly very few of these concern European
patents.
Holzer (2005) estimates the following annual filing number for 2003: 85 (UK), 700 (DE), 300
(FR), 70 (NL), 15 (DK), 40 (SE), 20 (FI) 30 (CH), 40 (AT). For Germany, he suggests that
60% of all cases pertain to EPO-granted patents, for the Netherlands 80%, for Switzerland
90% and for Sweden 30%. He estimates that 500 to 600 cases (in 2004) pertained to European
patents. This assessment is broadly in line with the detailed estimates used below. Implicitly,
the high concentration of infringement cases in the German judicial system provides us with a
first estimate as to how many cases are duplicated. If as Holzer’s estimates suggest
between 70% and 80% of all infringement cases in Europe are dealt with in German courts,
then at a maximum 20% to 30% of all cases can be duplications.
36
In personal discussions, other practitioners confirmed the order of magnitude of the above
estimates. Estimates ranged between 550 and 650 for infringement cases in Germany in the
years 2003 and 2004, with an additional 180 revocation cases. Moreover, it was estimated that
between 75% and 80% of all infringement cases were located in Germany.
35
The data actually imply that four out of thousand patents in the stock will be involved in litigation. Note that
this also leads to an estimate of a patent having a likelihood of 4% of being involved in litigation over an average
life time of ten years. This implicit estimate is not consistent with most of the literature or the data.
36
In some practitioner publications, both a high concentration and a high incidence of duplication are mentioned.
That is logically infeasible. Let x (in %) be the share of litigation cases going to court in Germany (or the
country with the largest case number). Then the maximum share of duplicated cases is 100-x. As the subsequent
sections show, duplication costs are however a first-order problem even if German courts were to account for
70% of all infringement cases.
31
4.7 2003 Survey of the Working Party on Litigation
The Working Party on Litigation undertook a survey in 2003 in order to obtain an estimate of
the number of patent litigation cases related to European patents.
37
The survey yielded the
following results for the year 2000:
700 cases in Germany (of which 420 related to European patents);
300 cases in France;
105 cases in the United Kingdom; and
70 cases in the Netherlands (of which 56 related to European patents).
The survey assesses that litigation in these four countries represent 90% of all patent litigation
activity in Europe in 2000. The implied share of cases tried in Germany would be 59.5%.
Given that over the last decade, concentration of litigation on Germany has increased
according to practitioner statements, this statistic is not inconsistent with the more recent data
and estimates.
Another study
38
the 2006 EPLA Impact Assessment prepared by the EPO as Secretariat of
the Working party on Litigation includes data on average costs of patent litigation in the
four countries considered. Table 4.6 summarizes the data from both sources.
Table 4.6 – Incidence and Costs of Patent Litigation of European Patents in 2000
First Instance Proceedings Second Instance Proceedings
Country
Litigation
Cases
(EP)
"Small-
Medium
Case"
"Large
Case"
"Small-
Medium Case"
"Large Case"
DE 420 50,000 €
250,000 €
150,000 €
190,000 €
FR 210 50,000 €
200,000 €
40,000 €
150,000 €
GB 105 150,000 €
1,500,000 €
150,000 €
1,000,000 €
NL 56 60,000 €
200,000 €
40,000 €
150,000 €
Total 791 63,982 €
399,115 €
113,009 €
284,071 €
Sources: WPL/4/03 and 2006 EPLA Impact Assessment.
The estimates in Table 4.6 are subject to a number of uncertainties as the WPL study notes
“they must be viewed with due circumspection.” The row labelled “Total” contains the
incidence-weighted averages across the four countries. The costs reflect the costs per party
(including court fees, fees for hearing witnesses, and attorney costs). However, they do not
include a mark-up for private, internal costs of the parties which accrue in addition to cost of
representation and fees (e.g., costs for providing documents, securing evidence, own
37
See WPL/4/03 “Workload and Cost of the European Patent Judiciary”, 31.10.2003.
38
Available at http://www.epo.org/patents/law/legislative-initiatives/epla/assessment.html. See in particular
Annex 1.
32
personnel involved in settlement negotiations etc.). These data will be used later to assess the
cost savings from avoided duplication and from cost reductions.
4.8 Incidence Data from the Pharmaceutical Sector Inquiry
The pharmaceutical sector is of special interest to the current study. First, it is well-known
that the impact of patent protection is particularly pronounced in the field of pharmaceuticals.
Patent protection is highly valuable to originator pharmaceutical firms. At the same time,
generics producers are likely to see high value in a court institution that allows them to
effectively challenge patents and to see those rights revoked that have been granted
erroneously. Given the high private value of patents and patent revocation, litigation is also
likely to have strong effects. Both types of litigation cases – infringement and revocation – are
also likely to be duplicated in this sector since access to national markets will be strongly
contested. Moreover, companies are likely to see strategic value in a multiplication of
litigation.
The Commission's preliminary report on the Pharmaceutical Sector Inquiry (DG Competition
2008)
39
confirms these expectations. Patent litigation in this sector is frequent, cases are taken
to multiple jurisdictions, and costs of litigation are high. Given the particular importance that
a unified patent litigation system may have for this sector, pharmaceuticals will be treated as a
separate technical field in the estimation exercise in section 6 of this report.
The Sector Inquiry generates the following results of interest to this report:
1. The survey data collected in the course of the Sector Inquiry cover approximately 80% of
the relevant EPO-granted patents in the field of pharmaceuticals.
40
2. In the period between 2000 and 2007, a total of 698 cases of patent litigation were
initiated in the EU between originator companies and generics makers.
41
3. 54% (378) of the 698 cases were initiated by originator companies, 46% (320) were
initiated by generic companies.
42
4. The incidence of litigation is growing, from 36 cases in 2000 to 132 in 2007. The
development over time is not monotonic, but clearly systematic.
43
5. Most of the cases are brought to German courts (90), but the country distribution does not
confirm the expectation that Germany accounts for the lion’s share of cases in
pharmaceuticals. UK (71) and Spain (70) are close followers. Other countries also see a
high incidence of patent litigation: Italy (59), Austria (59), Sweden (54), Portugal (43) and
Denmark (40). The Netherlands and France, in many studies named as important countries
39
Available at http://ec.europa.eu/competition/sectors/pharmaceuticals/inquiry/preliminary_report.pdf
40
See DG Competition (2008), paragraph (338). The survey collected data on 28,750 patents (applications and
grants), compared to a total of 30,010 filings (as in Table 16).
41
See DG Competition (2008), paragraph (468).
42
See DG Competition (2008), Figure 62.
43
See DG Competition (2008), paragraph (471) and Figure 63.
33
for patent litigation do not play the pronounced role that one would expect based on
aggregate statistics and estimates (33 cases in NL, 31 in FR).
44
6. Originator companies typically initiate infringement actions, while generics firms initiate
revocation actions or actions for declaratory judgements of non-infringement.
45
7. The 698 cases concerned 68 pharmaceutical substances (as classified under their
International Non-Proprietary Name, INN). There appears to be considerable duplication
or potential for duplication. All of the 20 most litigated INNs were subject to litigation in
at least 3 Member States, the top 6 INNs were subject to litigation in at least 5, some even
in 8 and 9 Member States.
46
8. The top six INNs accounted for 49% of all reported litigation.
47
9. Of the final outcomes reported in 149 litigation cases, generics companies won a large
share of cases. In cases which they had initiated, they won 73% of the cases. When
originator firms had initiated a case (and a final judgement was reported), they were
successful in 51% of these cases and not successful in 49%.
48
10. The Sector Inquiry states explicitly: “(…) generic companies won overall nearly two
thirds of all patent litigations initiated in the EU from 2000 to 2007 in which a final
judgment was given. However, this outcome was achieved at the expense of the
multiplication of costly and often lengthy litigation before different national jurisdictions,
thus entailing a significant burden and legal uncertainty for generic companies.”
49
11. Average costs of originator companies amounted to EUR 230,000 per case. It is estimated
that the total cost of litigation amounted to EUR 420 million for all 698 cases.
50
12. The Sector Inquiry also contains an assessment to what extent court rulings were
contradictory. In 11% of the final judgements reported by the surveyed companies, two or
more different courts in EU Member States gave conflicting final judgements.
51
13. As to litigation between originator companies, the incidence figures are far smaller. The
study only report 66 cases in the time period between 2000 and 2007. 39% of these
actions were initiated in Germany. Two thirds of the cases were revocation actions. Only
13 cases went to a final judgement.
52
The report does not give cost estimates for these
cases.
44
See DG Competition (2008), paragraph (474) ff.
45
See DG Competition (2008), paragraph (478) and Figure 65.
46
See DG Competition (2008), paragraph (482) ff. and Figure 67.
47
See DG Competition (2008), paragraph (483).
48
See DG Competition (2008), paragraph (502) ff.
49
See DG Competition (2008), paragraph (505).
50
See DG Competition (2008), paragraph (532) ff. Total costs per case are estimated to amount to EUR 600,000
on average. This figure include internal costs and fees.
51
See DG Competition (2008), paragraph (536).
52
See DG Competition (2008), paragraph (1006).
34
Returning to the dominant group of cases between originators and generics companies, the
Sector Inquiry does not state the number of infringement cases that are actually duplicated.
Some cases may concern the same INN, but different patents. Some may concern the same
patent, but cases may be initiated by multiple parties and concern different aspects or claims
of the patents. These cases may be filed in a harmonized system as well as in a fragmented
one. However, it is also clear that there is substantial duplication. For the estimates reported
below, a range of duplication parameters will be used. Given the country distribution by INN
litigation
53
, it seems justified to assume that between 25% and 50% of all patent litigation
cases in the field of pharmaceuticals would not be filed under a unified patent litigation
system. Furthermore, the incidence of litigation (132 cases in 2007) will be used later on in
the cost-benefit estimates. Taken together, these results and estimates suggest that legal costs
on the order of between EUR 19.9 million and EUR 39.7 million
54
were spent in the year
2007 in this sector alone on duplication (respectively multiplication) of patent litigation. The
legal cost borne by the public (costs of court operations not covered by fees) would have to be
added to this figure, but the study does not report data on the public costs of these litigation
cases.
53
See DG Competition (2008), Figure 67.
54
Computed as (132/698) x EUR 420 million x 0.25 (respectively 0.5).
35
5 Estimates of the Operational Costs of the Unified Patent Court
This section provides estimates of the operational costs of the unified Patent Court. The
computations build largely on assumptions developed by the Working Party on Litigation in
various publications. Moreover, it uses the caseload data from Annex II, Table II.2d as the
capacity requirement for the Court in year 2013. Thus, for the purpose of this section, the
number of (non-duplicated) cases is assumed to be 940 cases.
55
The cost structure of the Court
is composed of three main components: (i) judges’ salaries, (ii) staff salaries, and (iii)
additional costs for facilities, IT and other expenses.
It is important to keep in mind that none of the cost estimates below have normative content:
they are not meant as recommendations. They merely serve as conservative approximations in
order to obtain an upper-bound estimate of the operational costs of the proposed unified
Patent Court.
5.1 Judges’ Salaries
The estimation of total judges’ salaries is based on a number of assumptions regarding
capacity, salary levels and work organization.
For the Court of First Instance, the following assumption are made:
1. A judge rapporteur can deal with 25 cases per annum.
56
2. The average annual gross wage costs per judge are EUR 225.000.
57
3. Regional capacity adjustment can be managed by employing judges partially, e.g. in
part-time positions while they also serve on national courts.
4. A full-time equivalent of 39 judges (13 panels) will handle first-instance cases
(maximum case capacity: 975 cases).
Furthermore, for the Court of Appeal, it is assumed:
1) The appeal rate is 25% of all first instance cases, calling for an appeal capacity of 235
cases per annum.
58
2) A judge rapporteur can deal with 20 cases per annum.
59
3) The average annual gross wage costs per judge are EUR 250.000.
60
4) Three panels will be set up for a maximum case capacity of 300 cases.
The total costs in terms of judges’ wages would then be 39 x EUR 225.000 plus 15 x EUR
250.000, i.e. equal to EUR 12.5 million.
55
Rounded from 941 cases in Table 6.2.
56
See WPL/4/03 "Workload and cost of the European Patent Judiciary", of 31.10.2003, p. 17.
57
The amount would be roughly equivalent to the cost of a judge of the ECJ’s Civil Servant Tribunal.
58
COM(2003) 828 final, p.51 gives 25% as an estimate for the appeal rate.
59
See WPL/4/03 of 31.10.2003, p. 17
60
COM(2003) 828 final, p. 58 gives EUR 275.000 as the cost of a judge of the ECJ’s Court of First Instance.
36
5.2 Staff Salaries
This assessment is based on the estimates provided in the 2003 Legislative Financial
Statement
61
which applied to a fully centralized court system without arbitration centre and
training program. The 2003 assessment called for a staff of about 100 FTE, with annual gross
wage costs of EUR 108.000 per person on average. The staff includes registrars, legal and
clerical secretaries, translators and interpreters, staff for training and arbitration.
For the unified Patent Court, a staff requirement of 120 is estimated to be required, since
decentralization and additional task definitions require more resources. The total costs of staff
are thus EUR 13.0 million.
5.3 Facilities, Information Technology and Other Costs
The costs of facilities (court rooms, offices, …), IT, telecommunications etc. can be
approximated as between 6 and 10 % of wages.
62
It is assumed that this cost category can be
estimated at 8 % of total wage costs. Hence, total costs relating to facilities etc. will be EUR
2.0 million.
5.4 Total Annual Operating Costs
In total, the operating costs for the proposed Patent Court with capacity for 940 cases per
annum is estimated to be at EUR 27.5 million. Thus the average operational costs per case are
estimated to be at EUR 29.280.
5.5 Caveats
The following costs have not been taken into account: costs for experts according to
Article 36 and costs for legal aid according to Article 44 of the draft Agreement.
61
See COM(2003) 828 final, p. 52.
62
Assumption: rented premises. See for example Report of the Working party to the Interim Committee for the
Community patent, Brussels Oct. 10, 1985 - CIBC/835/85, p. 10.
37
6 Benefits from Avoiding Duplicated Litigation
6.1 Patents at Risk
In order to quantify the potential for patent litigation and its duplication across Europe, it is
necessary to estimate the actual stock of European patents in force. Since incidence of patent
litigation differs considerably by technical field, the calculations are undertaken by main
technical field.
63
Moreover, we will break out calculations and estimates for Germany, since
this country has attracted most of the European patent litigation cases in the past, and since
the data on Germany allow us to arrive at more precise overall estimates of the cost-benefit
balance of the proposed court system.
A calculation of the patent stock as of March 2008 can be based on factual data, rather than
estimates. Using the March 2008 EPASYS database of the EPO and the PATSTAT database
of April 2008, the patent stock data summarized in Table 6.1 (column (1a)) were computed.
Moreover, while more than 90 % of EPO-granted patents within a technical field are usually
validated in Germany, there are slight differences across technical fields. Column (2b)
tabulates the patent stock for Germany only.
A prediction of patent stocks for future dates, such as March of 2013 as chosen here, must be
based on estimates. While there are some uncertainties, the number of patent applications and
and grants can be predicted relatively well. The estimates in Table 6.1, columns (1b) and (2b)
were obtained using the following approach. Starting from the factual data describing the
patent stock and stock of filings in March of 2008, we (i) added an estimate of newly granted
patents for each of the following years up to 2013, (ii) subtracted patents that would
(factually) have expired by that time after reaching the statutory limit of 20 years, and (iii)
subtracted patents that were predicted to have lapsed by virtue of non-renewal by that date. A
proportional factor (specific to the technical field) was applied to arrive at an estimated
number of patents valid in Germany in 2013.
64
Long-term developments in the patent system are currently difficult to predict. Several aspects
need to be considered in this context.
First, while filings at the EPO are still increasing, the EPO has initiated a policy of
“raising the bar”. The impact of this new policy is likely to be felt in the medium- to long-
run, but not immediately. The impact could be a reduction of the grant rate, but it is
unclear how pronounced the reduction will be. A similar comment applies to new policies
regarding the use of divisional filings at the EPO. A less generous granting policy and
stronger sanctions on patent filing tactics will ultimately lower incentives to file marginal
patent applications.
63
The report assumes no differences between technical fields with the exception of the pharmaceutical field. The
data by technical field should nonetheless give a rough indication of case incidences by technology.
64
The predicted number of grants is derived from the actual number of applications filed at and published by the
EPO by applying the historical grant rate. Details of the calculations are available upon request.
38
Second, countering the first effect, the entry into force of the London Protocol on May 1
st
,
2008 will effectively lower the cost of patent protection in Europe. This will lead to an
increase in the number of validations per granted patent, and it will also lead to an
increase in the number of patent filings at the EPO.
65
Finally, the proposed unified Patent Court will also have jurisdiction to deal with litigation
relating to Community patents. But an estimation of the incidence of Community patent
applications and the subsequent impact of the Community patent on the use of the EPO
system is well beyond the scope of this study.
Table 6.1 Estimates of the Stock of Valid Patents by Technical Area – 2008 and 2013
Technical Area
(1a)
EPO Granted
Patents in Force
in At Least One
EPC State -
March 2008
(1b)
EPO Granted
Patents in Force in
At Least One EPC
State -
March 2013
(2a)
EPO Granted
Patents in Force in
Germany -
March 2008
(2b)
EPO Granted
Patents in Force in
Germany -
March 2013
Electrical
Engineering
98,337 147,900 92,677 139,400
Instruments
67,294 92,600 63,293 87,100
Chemistry (excl.
Pharmaceuticals)
82,751 93,000 78,888 88,600
Pharmaceuticals
17,583 36,500 16,648 34,500
Process
Engineering
66,883 85,200 63,324 80,600
Mechanical
Engineering
82,411 129,200 77,795 122,000
Consumption and
Construction
30,537 43,100 28,089 39,600
All Technical
Areas
445,796 627,500 420,714 591,800
Source: Own computations based on EPASYS data (March 2008) and PATSTAT (April 2008).
6.2 Estimating the Incidence of Duplication
In this section, four scenarios are being developed and the stock of patents in force is
computed for 2008 and for 2013. The purpose of this step is to describe reasonably supported
data constellations and to use them to derive an estimate of the benefits from introducing the
unified court system. In all scenarios, we perform the following operations.
65
See Harhoff, Hoisl and van Pottelsberghe (2008) for a study of the impact of the London Protocol on
validations. Mejer and van Pottelsberghe (2008) have studied the impact of the London Protocol on the total
number of patent filings and conclude that these are likely to rise by 12%. This development will not impact the
predictions for 2013 in a major way, but it will lead to even longer backlogs at the EPO.
39
1. We first compute the stock of European patents “at risk” for 2008 by broadly defined
technical groups.
2. We then focus on German data since the available estimates and actual statistics for
litigation are best developed for Germany. This stock is computed from first taking all
granted patents within a technology group and then subtracting all patents that have
expired or have not been renewed in Germany.
3. We then estimate the extent of German patent litigation. We use the available statistics to
estimate the number of infringement actions by broadly defined technical field. These
should be consistent with the data discussed previously.
4. We then estimate the number of cases likely to emerge in other countries in Europe and
the share of cases which duplicate other national litigation proceedings. The “duplication
share” is quintessential for the estimation of the direct benefits of a unified system of
patent litigation.
5. The parameters chosen for the scenarios are as follows. It is assumed that
per annum, 0.125% of the patents in the stock (and at risk) will be involved in
infringement litigation in Germany. It generates the best fit with the data in Table 4.3
and is therefore used to calibrate the scenarios. This parameter is maintained
throughout in the remainder of the report.
the second parameter assesses the concentration of patent infringement litigation and
measures the share of cases prosecuted in Germany. This parameter ranges between
60% and 70%. There is no clear evidence favoring one or the other. The data in Table
4.6 suggest that this share was as low as 50% in 2000, but subsequent developments
have led to greater concentration, possibly as high as 70%. Since the recent rulings of
the ECJ may have increased the need for multiple litigation again, a plausible estimate
is 60%.
the third parameter measures the extent to which the cases not prosecuted in Germany
reflect outright duplication. This parameter ranges between 60% and 70%, with the
former being again the preferred parameter value.
6. For pharmaceuticals, we assume a set of particular parameters.
We assume that in 2008, 150 cases are filed, that 30 of these are prosecuted in
Germany, and that the duplication rate of the remaining cases ranges between 25%
(Scenario 1a) and 50% (Scenario 1d) and is 37.5% for the intermediate cases (Scenario
1b and 1c). The scenarios for 2013 follow the same assumptions, but they use the
projections for the future patent stock as tabulated in Table 6.1. In pharmaceuticals, we
replicate the sector specific litigation rate from 2008, albeit at the higher incidence
numbers of 2013.
40
The results of the computations are documented in Annex II. A summary is included in Table
6.2.
Table 6.2 – Scenario Analysis of Duplication Incidence
Year
a -
Litigation
Incidence
b –
Concen-
tration in
DE
c –
Dupli-
cation
Rate
Total
Number of
Cases
Dupli-
cated
Cases
Pharma
Dupli-
cated
Cases
Other
Fields
Dupli-
cated
Cases -
Total
Non-
Duplicated
Cases
2008 1a 0.125% 0.7 0.5 872 38 108 146
726
2008 1b 0.125% 0.7 0.6 872 56 130 186
686
2008 1c 0.125% 0.6 0.6 992 56 202 258
734
2008 1d 0.125% 0.6 0.7 992 75 236 311
681
2013 2a 0.125% 0.7 0.5 1,206 53 149 202 1,004
2013 2b 0.125% 0.7 0.6 1,206 79 179 258 948
2013 2c 0.125% 0.6 0.6 1,372 79 279 358 1,014
2013 2d 0.125% 0.6 0.7 1,372 106 325 431 941
Source: Own computations based on Table I and II (see Annex II). Case filings and duplications are annual
figures.
The initial setup of scenarios is consistent with the relatively hard evidence of the number of
cases in Germany. Table 4.3 implies that roughly 1,050 infringement actions would be
initiated in Germany in 2008 of which 52% would concern European patents. That yields a
total number of infringement cases of roughly 525 while Annex II (Table II.1a) specifies 535
cases. The assumptions in the three basic scenarios reflect this data constellation for Germany,
but allow for variation in concentration and duplication (and thus in the number of cases in
other countries). Table 6.2 summarizes the parameters, the total number of litigation cases and
the number of duplicated cases, separately for pharmaceuticals and for other technical fields.
For 2008, the estimated number of duplicated cases ranges between 146 and 311; for 2013
(the year in which a steady state of the unified Patent Court is assumed to be reached) the
range is between 202 and 431 cases.
6.3 Benefits from Avoiding the Costs of Duplication and from Cost
Reductions
Towards a monetary assessment, the duplicated cases need to be valued at their cost. Given
that duplicated cases are likely to be linked to particularly valuable patents, it does not appear
reasonable to apply a representative value and cost distribution. Estimates in the 2006 EPLA
Impact Assessment are based on the assumption that two thirds of all litigation cases (in the
German, French and Dutch courts) are “small to medium” and that one third is “large”. Due to
self-selection, the distribution is likely to be reversed for cases in the United Kingdom. For
the assessment here we will assume that 50% of the duplicated cases are “large” and 50% are
“small to medium”. Moreover, we assume that appeals occur not in 25% of the cases (which
would reflect the overall population of cases) but in 30% of the cases. Using these parameters
and the overall estimates in Table 4.6, the cost per duplicated case per party is estimated at
41
EUR 291,000.
66
For both parties, the total private costs are then EUR 582,000 per case. We
do not add a mark-up here for internal costs to the party, since we also do not take costs
savings from duplication into account. For pharmaceutical cases, we assume that cases that
are duplicated are again more costly to the parties, in this case by 15% in excess of the
average.
67
The results of the assessment are tabulated in Table 6.3.
68
Currently, the parties to patent
litigation are estimated to spend between EUR 89 and 189 million on duplication of litigation.
Due to the increase in the stock of patents “at risk”, the private costs of duplication would
range between EUR 148 and 289 million in 2013. The b and c scenarios reflect particularly
likely constellations, but even at the relatively conservative estimates (with high concentration
of litigation in Germany and low duplication rates of the remaining cases), duplication clearly
is very costly to the European economy.
Table 6.3 also includes an estimate of additional cost savings arising from the assumption that
the unified Patent Court will offer litigation at the cost level currently observed in Germany,
France and the Netherlands. In effect, the implementation of such a cost level across the EU
would profit parties in the United Kingdom and in other countries with relatively costly
systems. To approximate this benefit figure conservatively, we assume that for each of the
remaining (non-duplicated) cases in the United Kingdom, there is a total cost saving of EUR
250,000, i.e., of EUR 125,000 per party. In the second but last column of Table 6.3 the
aggregate benefits are then tabulated.
69
Table 6.3 – Estimates of Private Benefits from Avoided Duplication and Cost Reduction
Year
Sce-
nario
Total
Number of
Cases
Dupli-
cated
Cases
Pharma
Dupli-
cated
Cases
Other
Fields
Dupli-
cated
Cases -
Total
Private
Benefits
from
Avoided
Duplication
(EUR Mill.)
Additional
Private
Benefits
from Cost
Reduction
(EUR Mill.)
Total
Private
Benefits
(EUR Mill.)
Benefit-
Cost
Ratio
2008 1a 872 38 108 146 89.1 17.5 106.6 3.9
2008 1b 872 56 130 186 114.3 16.6 130.9 4.8
2008 1c 992 56 202 258 156.2 21.1 177.3 6.4
2008 1d 992 75 236 311 189.1 19.6 208.7 7.6
2013 2a 1,206 53 149 202 123.3 24.3 147.6 5.4
2013 2b 1,206 79 179 258 158.7 22.9 181.6 6.6
2013 2c 1,372 79 279 358 216.9 29.2 246.1 8.9
2013 2d 1,372 106 325 431 262.3 27.1 289.4 10.5
66
That is, 0.5*EUR 63,982 + 0.5*EUR 399,115 + 0.3*(0.5*EUR 113,009 + 0.5*EUR 284,071).
67
EUR 300,000*1.15 per party per case = EUR 690,000 per case.
68
The figures for 2013 should be interpreted as estimates in nominal 2008 Euro. Attorneys’ fees and other costs
components have been increasing above the rate of inflation, but since the estimate is meant to be conservative
no projected price changes are included in the computation.
69
Cases after duplication were estimated by assuming that case filings in France, United Kingdom, the
Netherlands and other countries are distributed in the ratio of 4:2:1:1. This is roughly commensurate with the
distribution in Table 4.6. The number of cases in Great Britain was between 70 and 78 in the 2008 scenarios, and
between 97 and 109 in the 2013 scenarios. These case numbers were multiplied with the stated cost reduction.
42
Source: Own computations based on Table I and II (see Annex II). Benefits and case filings are given as annual
figures.
To compute a benefit-cost ratio, the above results need to be compared to the estimates of
operating costs computed in section 5 of this report which puts the operating costs of the
unified Patent Court with capacity for 940 cases per annum at EUR 27.5 million. Thus the
average total costs per case are estimated to be at EUR 29,280. The computed benefit-cost
ratio is shown in the final column of Table 6.3.
Clearly, the total savings from avoiding duplication and from cost reduction are considerably
larger than the operating costs, even for the most conservative scenarios. For 2013, the
benefit-cost ratios range between 5.4 and 10.5. In other words, duplication of litigation
combined with high costs of litigation in some countries costs firms about 5.4 to 10.5 times
more than the establishment and annual operation of the unified Patent Court. Since the Court
itself substitutes some of the existing national court capacity (or, to put it differently, not all
court capacity will have to be financed anew), the benefit-cost ratios would be even more
favourable if the substitution were taken into account.
6.4 The Impact of Litigation Costs
The preceding considerations were based on the assumption that the unified Patent Court will
allow parties to litigate at the relatively low cost levels which characterize the national
systems in low-cost litigation countries such as Germany, France and the Netherlands. A close
reading of the Presidency’s proposal shows that several cost-reducing measures are planned.
As has been pointed out in the discussion already, the cost of litigation is a central parameter
which affects not only the extent of litigation, but also settlement activities and the potential
for strategic behavior in the patent system. To strive for a low-cost system is therefore
important. At the center of the positive results in the preceding section were savings from
avoided duplication and from cost reduction. It is simple to see that these savings could be
reduced substantially if the unified Patent Court system were to come with an increase in the
cost of litigation beyond the level now present in Germany, France and the Netherlands.
In order to explore how strongly the benefit-cost ratios are affected by litigation cost levels,
Table 6.4 performs the following experiments. First, a critical cost increase is computed, i.e.,
the level of costs which would dissipate the total benefits computed in Table 6.3 completely.
As the results show, under the assumptions of scenario 2a, an increase of average litigation
costs of EUR 147,012 (EUR 73,506 per party) would dissipate the benefits of EUR 147.6
million. Moreover, an increase of average litigation costs (cumulative to both parties) of EUR
(147,012 29,280) = EUR 117,732 would yield a benefit-cost ratio of unity. Table 6.4 also
tabulates a revised benefit-cost ratio after a EUR 50,000 increase in average litigation costs
per party. Such an increase can be called drastic the average cost level per party in the low-
cost countries Germany, France and the Netherlands is EUR 145,000.
70
This means that even
70
Computed from Table 4.6 as a weighted average if two thirds of cases are “small-medium” and one third is
“large”, and if one quarter of all cases enter the appeal stage.
43
in this robustness check, the Presidency's proposal performs well. It would take roughly a
50% increase in average litigation costs to dissipate the private benefits computed above
completely. However, it needs to be noted that this thought-experiment disregards demand
effects clearly, at the increased cost levels, many of the parties may no longer want to
undertake litigation and are forced into potentially more costly private settlements.
Table 6.4 –Testing the Impact of Litigation Costs
Year Scenario
Cases without
Duplication
Total Private
Benefits
(EUR Mill.)
Critical Cost
Increase
(EUR)
Revised Benefit-
Cost Ratio after
EUR 50,000 Cost
Increase (per
Party)
2013 2a 1,004 147.6 147,012 1.7
2013 2b 948 181.6 191,561 3.2
2013 2c 1,014 246.1 242,702 5.3
2013 2d 941 289.4 307,545 7.1
Source: Own computations based on Table I and II (see Annex II). Benefits and case filings are given as annual
figures.
To conclude this discussion, the estimates in Table 6.3 form the core of the results. For 2013,
the benefit-cost ratios range between 5.4 and 10.5. The computation summarized in Table 6.4
affirms the robustness of this result. It also becomes clear that offering litigation at the new
Patent Court at levels close to the ones in the current low-cost countries is advantageous. Cost
increases lead to dissipation of benefits.
6.5 Further Effects from a Unified Patent Court
The discussion in this section assumes again that the unified Patent Court system will allow
parties litigate at the relatively low cost levels of the three Continental European systems. The
purpose is to discuss medium- to long-term consequences of the introduction of the unified
court system.
Impact on infringement suits. The purpose of the unified Patent Court is inter alia to
reduce the variety of interpretations of patent scope and claim interpretation in Europe,
especially at non-specialized courts. Economic theory suggests that this will ultimately reduce
the extent of divergences and asymmetric information regarding patents.
71
In the course of
this process, parties that would decide in favor of litigation in a fragmented system will again
revert to settlement because settlement costs will be reduced over time
72
and parties’
expectations and information will become more homogeneous. In the long-run, it will then be
a hallmark of success if (relative) litigation activity will decrease despite of a low-cost
litigation path. Based on the currently available data, it is not possible to predict when these
effects would set in.
71
See the discussion in Lanjouw and Lerner (1998) and Priest and Klein (1984).
72
See Galasso and Schankerman (2008) for empirical evidence that settlement lags were reduced substantially
after the introduction of the CAFC in the USA.
44
Moreover, the above development may not become fully visible due to other effects of the
unified system. The unified Patent Court system will be more attractive than the fragmented
one (for most users) because it offers the resolution of legal conflicts for the whole EU.
Utilizing the Court will also become necessary for an increasing number of parties which may
in future make more frequent use of the European patent or the Community patent. Currently
(as Table 4.1 has documented), firms in the new EU Member States use European patents
infrequently. As these economies grow and as their commerce gets more intertwined with
other Member States, there will be increasing demand for litigation involving parties in these
States. Finally, in countries in which the current national system is very expensive, a shift
from settlements to litigation may occur, since litigation offers the more cost-effective way of
dealing with the legal conflict once the unified system has been introduced.
The development of the demand for litigation will therefore reflect three aspects: (i) a
potential long-term reduction in demand for litigation due to the Court’s impact on divergent
expectations and asymmetric information; (ii) increasing use of European patents in some
Member States which are not particularly active in patenting at this point; and (iii) increasing
use of litigation in Member States whose national litigation system is more costly than the one
offered by the future unified Patent Court. All three developments are positive. Under (i),
parties will resort to settlement, but under more homogeneous expectations. In most cases,
this shift to private negotiations will be positive since these occur under greater certainty
regarding the scope of patent rights. Tendency (ii) simply reflects economic growth in new
Member States, and the ensuing need for an efficient patent litigation system. Fast and low-
cost resolution of legal controversy should support that process. And (iii) is a reaction to the
implementation of the low-cost litigation system (from the perspective of parties in formerly
high-cost national systems). The first two tendencies will be observed in the long run. They
cannot be quantified easily, but they would affect the overall analysis positively.
The third aspect could be felt relatively soon after the establishment of the Patent Court. The
effect is not taken into account in Table 6.3, since the computations there are guided by the
assumption that all other litigation activity beside duplications will be maintained at the
original levels (inelastic demand for litigation). A quantitative benefit-cost framework cannot
be developed easily here, but as a robustness check for the previous calculations, it is assumed
that 200 additional cases are filed as a result of the (relative) cost reduction perceived by
parties in currently high-cost litigation systems.
73
These cases are initiated because, with the
establishment of the unified system, litigation has become more appealing than settlement or
no action. Hence, parties must have the expectation of private benefits from switching to
litigation. As an order of magnitude, consider the case in which each infringement case
generates an additional private surplus to both parties of half of the cost reduction achieved by
implementing the unified system. For “small to medium” cases in the UK, the reduction is on
the order of EUR 100,000 per party
74
and thus EUR 200,000 per case. Hence, for our assumed
73
This figure can be roughly thought of as the current difference between the level of cases in the United
Kingdom and in France.
74
See Table 4.6 and the above discussion. We take as the relevant difference EUR 150,000 for a small case in
the United Kingdom vs. EUR 50,000 for a small case in Germany, both times in first instance.
45
200 additional cases, a private gain of EUR 10 million
75
could be expected. The additional
court costs would have to be held against this effect. Given that the cases were assumed to be
“small to medium”, operational costs would consume on the order of half of the gains (about
EUR 5 million). Hence, the overall private benefit from additional infringement suits is likely
to be small in comparison to the overall cost savings tabulated in Table 6.3. In any case, the
impact of these cases would raise the benefit-cost ratio further. The above estimates are thus
robust to the inclusion of this additional effect.
Revocation proceedings. Somewhat different results emerge from a discussion of revocation
proceedings. These have found relatively little attention in the literature, presumably because
the US system does not offer a particularly effective way of revocation in cases of erroneously
granted patents.
76
The results of the Commission’s Pharmaceutical Sector Inquiry indicate
that revocation has considerable public benefits because, if successful, it allows firms to enter
the market. The Presidency’s proposal foresees a highly attractive revocation path which does
not have to exhaust opposition at the EPO first. Opposition at the EPO used to involve more
than 10% of granted patents in the early 1980s, but has declined to a level around 5%.
77
One
reason for the declining attractiveness may be the long delays in resolving opposition and any
subsequent appeal cases. The unified Patent Court would offer an interesting alternative.
Suppose that the unified Patent Court would attract new additional invalidity suits
78
and that
the average private value of patents attacked in invalidity suits is EUR 2,0 million if
maintained in the proceedings. Assume further that roughly 25% of patents would be revoked
or partially revoked in revocation proceedings (as at the German Federal Patent Court). Each
case leading to revocation could then cause total benefits of approximately EUR 1,0 million
79
,
since the erroneously granted patents can no longer impede competition. Suppose that a case
with these stakes would cause total litigation costs of EUR 50,000. Per case, a net benefit of
EUR 200,000 accrues to society.
80
Thus, 200 revocation cases would lead (under these
assumptions, without appeal stage) to a net benefit of EUR 40 million in addition to the
benefits computed before. The effect is clearly larger for infringement cases (see above), but
there is again a positive contribution. It is important to keep in mind that the societal gain
from revocation will in all likelihood not be fully internalized by the party or parties initiating
a revocation action, while the gains from a successful infringement action will be internalized
to a much greater degree.
Fast and low-cost revocation proceedings are also a good defense line against “patent trolls”
seeking to extort licensing fees from other parties based on weak or questionable patent rights.
75
200 cases multiplied with half of the reduction in private costs to both parties (EUR 200.000).
76
A revocation action can be initiated