Private Liability Rule Regimes

Munich, Germany


In Chap. 2, several indications were found that there is basically an inefficient overuse of property rules in current German law and hypothetically also in the regulation on the European Patent with Unitary Effect; relying so intensively on property rules is also not mandatory from other perspectives such as distributional preferences and higher-ranking law. Another question, which has not been answered until now and should be addressed specifically in the following discussion, is whether private liability rules sufficiently reduce the overuse of property rules. An additional question will be whether it is possible to support even more efficient use of those mechanisms than is currently the case.

As already mentioned above, an argument for the implementation of property rules is that right holders are in the best position to switch to protection by liability rules, especially to reduce transaction costs. Although the patentee will only make this switch if it is advantageous to him, private liability rules are also basically positive for society because some of the other above-mentioned disadvantages and risks could also be reduced. Moreover, the incentives to innovate are normally not reduced: since there is a lower risk of under-compensation and under-use than generally assumed under liability rules, coordination may become easier and access to the patented invention improved. The legislature should normally rely on private liability rules before considering the use of compulsory liability rules because they basically mean a lower degree of state intervention, and administrative costs may be lower, too. However, one must be aware of the circumstance that private liability rules do not always enhance welfare. Their effects are often ambivalent with respect to economic efficiency and may also differ between the variants of them. This will be investigated in the following analysis. The legislature should provide measures to eliminate the disadvantages. The evidence from the German and European patent law systems has already suggested that there is an overuse of exclusivity and presumably it is currently not sufficiently reduced by private liability rules.

Private liability rules may be relevant in different contexts, but, as explained above,1 three major constellations can be identified (see also Fig. 3.1). Therefore, this chapter is divided into the following sections: declaration of willingness to license which is equivalent to the term licence of right (Sect. A.), patent pools with liability rule mechanisms (Sect. B.) and royalty collection clearinghouses (Sect. C.). All these mechanisms are explained separately in the following discussion. Private liability rule mechanisms could basically coexist; however, each variant poses specific economic and legal questions because each could have its own advantages and disadvantages. Therefore, one variant may not be beneficial in every constellation. Sometimes one may be preferable to the other2; in certain cases, a combination of the different mechanisms may also be advantageous. The most detailed view is provided on the first variant because researchers have not focused very intensively on this mechanism so far.


Fig. 3.1
Property and liability rules in patent law (with variants of private liability rule regimes)

In the first two sections below, the second research question will be addressed, followed by the third question. In the third section solely the third question can be addressed because only the possibility to declare the willingness to license and patent pools with liability rule mechanisms are currently used—in contrast to royalty collection clearinghouses. The chapter ends with the results (Sect. D.).

A. Declaration of Willingness to License and Licences of Right

The following discussion investigates whether the possibility to declare the willingness to license according to Section 23 Patent Act reduces the overuse of exclusivity efficiently and, if not, whether the legal environment should be altered.

The possibility to declare the willingness to license is a mechanism implemented in the German patent law system and it is one that is constantly used. Therefore, empirical research is suited to gathering as much information as possible about the facts regarding the use of Section 23 Patent Act. A specific empirical study of the willingness to license had not been conducted until recently. An economist and the author of this work attempted to fill this gap by combining detailed legal and economic research.3 For the empirical analysis, inter alia, data from the so-called Patent Utility Model-Band, or PU-Band, provided by the DPMA, merged with data from the EPO Worldwide Patent Statistical Database (PATSTAT Database) were analysed. The former data contain information on all legal events from 1983 to 2008 published in the German Patent Gazette; the latter contain data on patents from more than 80 countries and around 70 million records.4 Additionally, some variables from the PatVal-EU dataset were used.

This analysis not only explains the main empirical considerations and results, but also provides a detailed legal analysis. Therefore, the analysis starts with some introductory considerations regarding Section 23 Patent Act, including the debate among legal scholars about the advantages and disadvantages of this provision (Sect. 1.). It continues with theoretical and mainly economic considerations regarding the use of Section 23 Patent Act (Sect. 2.) and empirical specifications and results (Sect. 3.). It then makes a comparison between German and other licence of right provisions, especially that of Regulation (EU) No. 1257/2012, the British and French provisions (Sect. 4.). Subsequently, it evaluates the new findings (Sect. 5.) and what can be concluded for the proposal of the so-called European Interoperability Patent (Sect. 6.) which has attracted some attention in the greater debate about open innovation (understood in the wider sense). Finally, the results of this chapter are presented in summary form (Sect. 7.).

1. Introductory Consideration Regarding Section 23 Patent Act

In the first section of the following analysis, this work introduces considerations regarding the declaration of willingness to license according to Section 23 Patent Act. The analysis starts by providing the most important information about the function and effects of the elements of this provision and their interpretation and application in practice (Sect. a.). Theoretical considerations are then explained, especially the legislature’s intention and comments by legal scholars (Sect. b.).

a. Explanation of Section 23 Patent Act

(1) The Declaration of Willingness to License and the Withdrawal

According to Section 23(1), first sentence, Patent Act, the patent owner or applicant is allowed to declare the willingness to license, but only as long as “no entry concerning the grant of an exclusive license [Section 30(4)] is recorded in the register or a request for the recording of such entry has been filed with the patent office”5 [Section 23(2)]. It is concluded from this stipulation that the formal allowance that normally becomes apparent with the entry in the patent register is sufficient for being allowed to declare the willingness to license; the real entitlement is not important.6 Some commentators claim that the declaration is valid only if the real and registered patentee issues it, which may be justified by the non-legal force of the German Patent Register.7 However, this last opinion would be irrelevant in practice because it is contrary to the wording of Section 23 Patent Act and because the users of an invention may already have had expenses. In general, it is also of no relevance whether a non-exclusive licence or, based on the above considerations, an exclusive licence exists for which no request for recording in the register is pending or which is not recorded in the register.8

When the patent owner or applicant can declare his willingness to license depends on the circumstance of where the application was filed. For German patent applications, a declaration is possible from the day of the application [Section 23(1), first sentence, Patent Act], for PCT applications from the day Germany is a designated state and for EPO patent applications from the day of publication of the mention of the grant [Arts. 2(2) and 97(3) EPC]. The patent owner must declare his willingness to license in writing [Section 23(1), first sentence, Patent Act], which means that it becomes effective at the point in time when it reaches the DPMA [Section 130(1) Civil Code], and the paperwork must contain an original signature of the patent owner or an agent who is entitled to act for the patent owner [Section 126(1) Civil Code]; pursuant to the jurisdiction of the Federal Patent Court, a facsimile is not allowed.9 According to Section 23(7), first and second sentences, Patent Act, the willingness to license can be withdrawn, which takes effect upon filing a written communication to the DPMA as long as no intent to use the invention has been notified to the patent owner.10

(2) The Effects of the Declaration and the Withdrawal

The declaration of willingness to license has the effect that the patent owner’s entitlement is no longer protected by a property rule, but by a liability rule.11 In other words, the patent owner abandons his rights as the sole user of the patented invention [Section 23(3), fourth sentence, Patent Act], as well as of granting exclusive licences [Section 30(4), second sentence] and forbidding use even by someone willing to pay reasonable compensation [Section 23(3), sixth sentence, e contrario].12 The declaration is published in the Patent Register and the Patent Gazette [Section 23(1), third sentence] and—very importantly—the future annual patent fees are reduced by half [Section 23(1), first sentence]. This reduction can only take effect from the third year following the date of filing because every owner of a German patent, or every patent applicant if the patent office has not yet decided about the grant, has to pay annual fees from that point in time onward to keep the patent (application) alive [Section 17(1)].13 The amount of fees is determined in the German Patent Costs Act (PatKostG) and its appendix [Section 2(1)]. Figure 3.2 illustrates the German (but not the European!) patent fees for patents and supplementary certificates of protection (Section 16a Patent Act) according to the Patent Costs Act.14


Fig. 3.2
Annual patent fees in euros according to the German Patent Costs Act

If the patentee withdraws his declaration of willingness to license, protection of the entitlement changes again from a liability rule to a property rule. Additionally, the patent owner is obliged to pay the amount by which the annual fees have been reduced to the DPMA [Section 23(7), third sentence, Patent Act] and the withdrawal is published in the Patent Register and Patent Gazette.15

(3) The Legal Situation Between the Patent Owner and the User

Anyone who wishes to exploit the invention must notify the patentee of his intention via registered mail; the notification must contain a statement of how the invention will be exploited [Section 23(3), first to third sentences, Patent Act]. After notification, the notifying party is allowed to use the invention [Article 23(3), fourth sentence].16 Normally the user can produce and sell the products in Germany, but this could also take place in the following way: the notifying party produces the products abroad which contain the patented invention, then imports and sells them to German customers; since exhaustion takes place with the importation, the customers are allowed to resell them.17 The user is also obliged to report the details of use at the end of every calendar quarter and pay remuneration [Article 23(3), fifth and sixth sentences]. Only if the user does not fulfil his obligations is the patent owner allowed to apply for an injunction [Act. 23(3), seventh sentence].18 The parties may agree upon the amount of remuneration; alternatively, as will be explained in detail in the next section, remuneration can be assessed by the patent division of the DPMA at the written request of a party [Article 23(4), first sentence].

Legally, the declaration of willingness to license may be categorised as an offer and the notification by the user as an acceptance of a contract.19 Alternatively, the declaration may be considered as a disposal of the patent owner over his right and the indication as allowing use of the patented invention. The disposal and allowance constitutes a kind of licensing relationship but, at least in the first instance, it is not a contractual relationship.20 The second alternative seems to be the prevailing opinion, although (according to German law) the parties need not necessarily agree on the amount of remuneration when concluding a licensing agreement.21 This also seems to be the consequence when entitlement theory is taken as a basis. The reason is that the willingness to license or licence of right is protected by a private liability rule and such mechanisms do not necessarily lead to contractual relationships between the patent owner and user.

(4) Determination of the Amount of Remuneration

There are two first instances that can determine compensation or make an alteration [Section 23(4–5) Patent Act] if the parties do not reach an agreement themselves. First, the patent division of the DPMA is explicitly mentioned in Section 23(4), first sentence, Patent Act. It decides if one party files a written request and the parties already agree that an obligation for paying remuneration exists.22 The fee must be paid by the applicant to the patent office and may be reimbursed by the opponent [Section 23(4), fourth sentence]. The amount is currently €60 [Section 3(1), Patent Costs Act, Fee No. 313 200 in the table of costs]. The second first instance may be a Lower Regional Court (LG) [Section 142(1), Patent Act]. The court seems to be generally accepted although it is not explicitly mentioned in the law.23 The court may also be allowed to decide about other issues that are in dispute. The amount of court fees is not fixed and depends on the value of the claim. Therefore, the fee is certainly higher than €60. Consequently, parties normally file a request to the patent division if the amount of remuneration is the only point of contention.

Section 23 Patent Act does not contain any information regarding the meaning of the term reasonable compensation. Indications can be found in the literature24 and, more important in practice, in published administrative and court decisions. Since the (re)foundation of the German Patent Office in 1949, the patent office has only decided three times about a reasonable amount of remuneration. In each of these cases an appeal was filed to the Federal Patent Court, and once a further appeal to the Federal Supreme Court.25 According to the decisions and the literature, the amount of remuneration should depend on the circumstances of each individual case.26 The aim should be to strike a balance between the interests of the patent owner and users. The patent owner should be compensated for having made and disclosed the invention, but users should not be burdened unreasonably, certainly if the amount of compensation has the effect that (potential) users decide not to use the invention.27 Other commonly known and accepted principles to determine remuneration or compensation are also applicable. These include principles developed for compulsory licensing provisions,28 for contractual licences, and for the calculation of damages according to the licence analogy.29 So far, normally as a first step, the courts have made comparisons with licence agreements that were concluded before the willingness to license was declared.30 Additionally, courts consider as relevant, inter alia, the following: the importance of the invention for the user and the value derived from the current state of the art31; doubts regarding the validity of the patent32; the volume of the user’s sales33; and what is reasonably agreed under comparable conditions.34 However, the declaration of willingness to license itself should not, in principle, have relevance for the amount of remuneration, especially if there are already a significant number of licensees35 and if non-exclusive licensing contracts were used as an indication before.36

b. Theoretical Considerations Regarding Section 23 Patent Act

(1) The Aim of the Legislature

Section 23 Patent Act37 came into force on 1 October 1936 after a decision by the German government of the Reich.38 Although Section 2339 was introduced by the German government of the Reich and not a parliament, an explanatory memorandum exists for the act. It is important to keep in mind that the idea for such a provision originated not primarily from the Nazis, since introducing such a mechanism was not their idea. Licence of right provisions were first introduced in 1903 in Canada, which was at that time a British Crown Colony, and in 1919 in England.40 However, according to the introductory remarks of the memorandum, the main purpose of the act was to exert the ideas of National Socialism in the field of patent and utility model law.41 On the one hand, the evolution of the creative personality was to be promoted and protected against exploitation by, inter alia, facilitating access to patent protection for inventors who were short of funds. On the other hand, the inventor was not to be aided in reaching selfish goals; instead he was to make the value of his invention available to the public because he owed his achievements to society.42

In concrete terms, the reasons for introducing Section 23 were to serve the interests of inventors and the general public by promoting technology transfer, as well as the use and commercialisation of inventions, i.e., innovations. Technology transfer and use of the invention should take place due to the possibility that third parties can easily become aware of the declaration of willingness to licence, since the patent owner can no longer prevent the use of his invention, and since every declaration is published in the German Patent Register and the Patent Gazette.43 Section 23 serves the interest of the general public because an invention can be used by third parties before the patent expires; the interests of patent owners are served in that they receive remuneration (potentially even more than would normally be the case).44 This mechanism leads to more innovations because individual inventors who lack the necessary assets for commercialisation are then able to pay the annual fees for a longer time and receive remuneration if the invention is useful.45 The introduction of Section 23 and a coincidental increase in patent fees should also have hampered the maintenance of—economically often not desirable—blocking patents.46

Section 23 Patent Act was slightly amended several times.47 However, there was one important amendment that went into force on 1 June 1992, when the possibility was added to withdraw the willingness to license as long as no one had indicated an intention to use the patented invention [Section 23(7)].48 According to the act’s explanatory memorandum on the amendment in 1992, the aim of the legislature was to improve the focus on the goal of promoting innovation.49 The disadvantage of not being able to grant exclusive licences was the main reason the declaration of willingness to license did not have, as was reported by specialists, the desired effects of disseminating inventions and encouraging more licensing activity.50 The amendment was intended to alleviate this disadvantage. Additionally, the legislature considered the interests of society to be less important than those of the patent owner, as long as no third party indicated his intention to make use of the patented invention.51

(2) Criticism of Section 23 Patent Act and Its Explanatory Memoranda

Section 23 Patent Act and its explanatory memoranda have been criticised in many ways; in most cases, however, without referring to any empirical evidence. Only one empirical study, which was published by the Institute for Economic Research in 1974 and in which the effects of the patent system on the innovation process were investigated based on a survey, came, inter alia, to the result that the declaration of willingness to license was irrelevant for the conclusion of licensing agreements.52 The disadvantages of Section 23 are analysed below.

One critic, Eggert, argued that the aim of the legislature cannot be achieved because costs can only be partially reduced and remuneration under a willingness to license regime is supposedly lower than under “normal” conditions and cannot be compensated by a saving in fees.53 The reasons for lower compensation are inter alia—which may also be a general argument against the current provision—that the patent owner who has declared the willingness to license is no longer allowed to limit the use of his invention and is obliged to accept every user.54 Moreover, the provision is not suited to serving the interests of an individual inventor who is short of funds, as the annual fees in the first 6–10 years are only a small part of the high overall costs. Small and medium-sized enterprises do not profit because they need an exclusive right or the possibility of exclusive licensing to be able to shoulder the investment risks for further development, production and marketing.55 Additionally, the system under Section 23 does not lead to more licensing activity since no one who is seeking a licence searches the Patent Register or Gazette for patents for which the owner has declared the willingness to license.56 A further disadvantage is the situation that was partially improved by the amendment in 1992, that the patent owner loses his right to grant exclusive licences.57 As a consequence of all these arguments, the declaration may be perceived, contrary to the intention of the legislature, as a devaluation of the patent and may only be used if the patent owner does not want to decide whether he wants to keep the patent alive or let it expire.58

(3) Additional Theoretical Considerations

(a) About the Construction of Section 23 Patent Act

One could also theoretically consider the following and add the argument that Section 23 Patent Act may benefit wilful infringers, who use a patented technology without notifying the patent owner although they are aware of the legal protection. In the event that they are caught, they normally risk a claim for damages and, perhaps even more fatal, the patent owner may close down the infringer’s business by applying for and enforcing an injunction. But this risk would not exist if the patent owner declared his willingness to license. Then, an infringer could continue to use the patented invention after getting caught by formally indicating his use according to Section 23(3), first sentence, Patent Act. The infringer may not even have to fear criminal sanctions according to Section 142 Patent Act because this provision has no practical relevance, as mentioned above.59 Whether this view is correct depends crucially on one question: is the user of an invention obliged to indicate use before the first use or does a later indication have the same consequences for the future? Mes concluded from the wording of Section 23(3) that only a user who makes his intention clear before the first use is allowed to use the invention; otherwise the patent owner can successfully apply for an injunction.60 The Düsseldorf Lower Regional Court did not agree with this argument in a decision in 200161 and denied an injunction after the infringer indicated the use to the patent owner. It is, however, important to note that the defendant claimed not to have knowingly infringed the patent which was not challenged by the claimant. Therefore, the decision could have been different had the infringer acted wilfully.

One should agree with the court’s decision since no support for Mes’ view can be found in the wording of Section 23(3). Moreover, a wilful infringer should have no more consequences to fear than someone who was unaware of the patent, because there is no indication in the wording to make such a differentiation. Additionally, the principles concerning the misuse of law62 seem not to be applicable because the patent owner made a deliberate decision to allow the use of his invention if someone indicates it and is willing to pay a reasonable remuneration.

(b) The Procedural Rules of Section 23 Patent Act

All the arguments above refer mainly to the construction and the effects of Section 23 Patent Act. The only criticism of the procedural rules is that the declaration of willingness to license is published solely in the Patent Register and Gazette. The rest of the procedural rules, i.e., how to declare the willingness to license and indicate the use and determination of compensation by the patent office or a court are, however, not in the focus of this debate. Therefore, one can agree with other commentators who do not see any problems here. In theory, the process seems very simple because a message from the patent owner to the patent office is enough to declare the willingness to license, and likewise from a potential user to the patent owner to be allowed to use the invention. Additionally, the patent office should be competent to determine the compensation at an efficient point without overly high assessment costs. One reason for this view is that the German patent office, besides its familiarity with patent law per se, has an arbitration board to settle disputes arising between employers and employed inventors on the basis of the German Employees’ Invention Act (ArbNErfG) [Section 29(1)], in which the amount of remuneration [Secs. 9(1), 10(1), first sentence] is often the subject of debate and around 50–75 requests are presented each year.63 The principles for determining compensation are similar to those in Section 23 Patent Act because both consider as applicable what is normally agreed on in licensing contracts.64 Though the arbitration boards are not identical to the patent divisions that decide disputes arising under Section 23 Patent Act [Section 27(1)(1) Patent Act; Section 29(1) Employees’ Invention Act], an exchange of experience could take place. This procedure should also lead to lower transaction costs as the parties have the possibility to ask the patent office to determine compensation and need not necessarily come to an agreement by themselves.

In conclusion, both procedural rules and the publication mechanism prevent a patent owner from declaring the willingness to license or a potential user from using the patented invention.

2. Theoretical Economic Considerations Regarding the Use of Section 23 Patent Act65

The following theoretical analysis considers economic explanations for declaring the willingness to license, i.e., for switching voluntarily from protection by a property rule to a liability rule. Therefore, factors that have to be taken into account are the legal consequences of the declaration and the voluntary nature of the patent owner’s decision. In principle, the patent owner always has not only two, but three options, particularly before the annual fees become due: pay the full annual fees and keep the patent as before; declare the willingness to license and pay half; or let the patent expire.

Since this is a voluntary decision by which the patent owner will, in all probability, try to maximise his profit, the focus must lie in the private and not in the objective value of the patent. There are many different ways in which patent owners realise (private) value, but the dimensions may be reduced to the following five: practice, licensing, blocking, offensive or litigation and defensive value.66 The first way, which is also the most common way of exploiting a patent right, is to apply the patented technology to one’s own products or processes, whereby success depends on the producer’s ability to stay ahead and to exclude competitors. The second way is a function of the value to licensees of using the patented invention, which normally also depends on whether it is an exclusive licence or not. Blocking value exists—as opposed to the value of working the patent—if the patent holder is able to prevent competitors from using an equivalent technology or technology that builds upon the patented subject matter. The offensive value of a patent is attributed to the possibility of suing rivals and claiming injunctive relief; then the patent owner derives value from settlements, licences or damages. The defensive value refers to patenting for the purpose of providing a basis for counter-infringement claims in patent infringement litigation.67

There are currently three different concepts for measuring the value of patents.68 The first is a concept in which the value of the patent right is calculated for a “stand-alone” inventor who compares his profit in the case of technical leadership to the profit gained in some ex ante state of the industry.69 The second is a method that is called the asset value of a patent right; here the value is defined as the difference in profits if the patent owner keeps his patent or if he transfers his right to one of his competitors. The last concept measures the value of renewed patent protection by calculating the difference between the expected profits of a patent owner who renews the patent, i.e., keeps the patent alive by paying the next annual fee, and an owner who lets the patent expire. The first concept is considered unsuitable because it neglects strategic aspects completely.70 By evaluating and comparing the second and third concept, Harhoff et al. argue that the asset value should be especially useful in a constellation where the blocking power of a patent is an important phenomenon.71 Since the renewed value should include all the above-mentioned forms of private value under a liability rule regime, which in this case means that, after a declaration of willingness to license, blocking power no longer exists or is at least very limited, this concept seems to be a better starting point than the method in which the asset value is calculated.

Abstractly, it could be said that, without taking the possibility to declare the willingness to license into account, the patent owner should let his patent expire if the renewal value is lower than the next annual fee. If the patent owner declares the willingness to license, the renewal value will amount to zero since the patented invention falls into the public domain. However, since the patent still exists as a non-exclusive right, he can gain what should be called declaration value from compensation paid by the notifying parties or non-exclusive licensing and from positive externalities, namely network effects, commitment, publicity through publication in the Patent Gazette and Register, etc. Additionally, the owner saves half of the annual fees. Thus, the patent owner should declare his willingness to license instead of letting the patent expire if the sum calculated from the declaration value and half the annual fee is higher than the renewal value. Thus, the declaration value must be more than half of the normal annual fee. In other words, the patent owner who has perfect knowledge will choose the option by virtue of which the combination of private value and obligation to pay annual fees is the most lucrative.

The above analysis leads to a very general hypothesis: The probability of observing a patent with declared willingness to license (1) increases with expected revenue from compensation or non-exclusive licensing, (2) increases with expected profits from positive externalities, (3) increases with expected renewal fees and age given that the patent owner wants to keep the patent, and (4) decreases in the value that results from the difference in expected profits if the patentee keeps his patent and the expected profits if he lets it expire.

But, to be more concrete, the following analysis explains three possible motivations for declaring the willingness to license, namely, non-exclusive licensing strategies (Sect. a.), cost-saving strategies (Sect. b.) and patent litigation and settlement strategies (Sect. c.).

a. Non-exclusive Licensing Strategies

A constellation in which the patent owner declares his willingness to license is comparable with a situation where he tries to maximise the private value of his patent right by concluding several non-exclusive licensing agreements for the following main reasons. In both cases there are potentially infinite users or licensees who pay some remuneration or licensing fees to the patent owner. As a consequence, potentially infinite users are allowed to use the patented invention and may compete with each other. Licensing fees should be low because, basically, no one can earn monopoly profits. Therefore, the motivation of the patent owner should be similar in both cases and normal theoretical considerations regarding strategies for non-exclusive licensing should be transferable to the situation described here.

(1) Complementary Assets Theory

Complementary assets are assets or capabilities that are always needed for the successful commercialisation of an innovation. These assets include competitive manufacturing facilities, distribution channels, different services and complementary technologies.72 According to an article by Teece, these assets can usually be classified in three different categories, depending on the relationship between the asset and the innovation. In the first group are generic assets that need not be tailored to the innovation. In the second category are specialised assets in which a unilateral dependence between the innovation and the complementary asset exists. Co-specialised assets, which comprise the third group, are those with a bilateral dependence.73 The latter two are assets that need to be tailored to the innovation; as a consequence, there is a strong dependence of the asset on the innovation and the other way round.74 If the innovation and the complementary assets are not owned by the same person—especially the case for individual inventors and small and medium-sized enterprises—both parties must commit capital to certain irreversible investments that will be valueless if their relationship breaks down. Thus, an agreement, which is a licensing agreement in the case of technology transfer, between these two parties will tend to be an exclusive one to overcome the problem of possible opportunism by one party. Of course, a cross-licensing agreement in some constellations, especially in the case of complementary technologies, might also be an option, but will not be debated here. A patent owner who lacks generic complementary assets may simply license his technology non-exclusively to firms that own them, mainly for the following two reasons: Firstly, the licensees need not invest in their assets and are not dependent on the relationship with the patentee. Secondly, the patent owner may be interested in licensing his technology to several parties because if one relationship breaks down, e.g., due to insolvency, he would not face big disadvantages.

Assuming that the patent owner does not care who the licensees are, this is a constellation in which the declaration of willingness to license would make sense, even at an early stage in the patent term, because the inventor would be interested in finding several licensees and would profit additionally from the reduction in the annual patent fees. As a consequence, patent owners who lack complementary assets, particularly generic assets, and who are probably individual inventors and small or medium-sized enterprises, but possibly also inventors engaged in the field of research, like universities, would be most likely to declare their willingness to license.

(2) Diffusion of Innovation

There may be constellations in which innovators have sufficient incentive to freely reveal an innovation by letting a patent expire or not keeping the invention secret. By doing so, the innovator would increase diffusion of the innovation and could benefit from several positive externalities, such as network effects, reputational gains and related innovations induced and revealed by other users. Additionally, an innovation that is freely revealed and adopted by others could become an informal standard that would pre-empt the development and/or commercialisation of other versions of the innovation.75 As a consequence, a free innovation would reduce research and development expenditures in the whole industry, but might also be advantageous for the innovator if the innovation is designed in a way that is especially appropriate to conditions unique to the innovator.76 Of course, it could also lead to competitive disadvantages and losses; therefore this strategy might not always be advantageous. But in some constellations, for example in industries where standards play a crucial role, the patent owner might weigh the advantages and disadvantages against each other. The patent owner might decide to license non-exclusively and/or declare his willingness to license. In this case he has basically revealed his invention, but could still profit from payment of licensing fees or compensation.

The situation here is also comparable with what was described as a specific open-innovation process,77 or what from the patent owner’s viewpoint is more precisely described as an “inside-out process”. This is used to reduce fixed costs for research and development and share with other companies the risks involved in establishing a standard and for spillovers into other industries.78 The term open innovation (in contrast to closed innovation if understood in a narrow sense) describes the strategy in which the innovation process is opened to other participants and their participation is used to increase the innovation potential by reducing costs and risks, as well as by shortening innovation cycles.79 The processes may be differentiated in the event that external knowledge is internalised in the innovation process (“outside-in process”), internal knowledge is externalised (“inside-out process”) and cooperation in innovation and commercialisation results.80

As a consequence of the above situations, willingness to license should be observed more often in sectors in which technology is more frequently revealed than in those industries in which secrecy and exclusivity are the prevailing strategies.

(3) Licensing as a Commitment Device

As the willingness to license can only be withdrawn as long as no one has indicated use of the innovation [Section 23(7), first sentence, Patent Act], this mechanism can also be used as a credible commitment device to allow competition. If the setup costs to use a product are high in a market, buyers of a new product may be reluctant to incur these costs if they fear that the patent owner will raise the price later. As a consequence, this might reduce demand and therefore the profits of a monopolist. But if the seller declares his willingness to license and invites competitors into the market, he commits to competitive instead of monopoly prices in the future. Additionally, he ensures a higher demand and earns royalties or compensation from his competitors. Farrell and Gallini present a theoretical model which shows that the described strategy should be profitable as long as the product-specific setup costs that technology users would incur are high enough.81

Assuming that a patent does not confer monopoly power, it can even be argued that, if there are no product-specific setup costs, committing to non-exclusive licensing can be a profitable strategy.82 Especially in complex fields of technology, in which one product consists of several patented inventions and substitutes are frequently available, this assumption might often be correct. In such a constellation, the revenue effect from licensing might be higher than the rent dissipation effect, i.e., losses from increased competition in an already competitive market.83 Presumably, the incentives to license non-exclusively and therefore to declare the willingness to license will increase with the degree of technological homogeneity. Additionally, a firm with a small market share is often also a small firm that might profit more from revenue than from the rent-dissipation effect.

To summarise, even at an early stage in the patent term, a declaration of willingness to license should occur more frequently the higher the competition is in the technology market, and the higher the degree of substitution is between them. But this hypothesis will not hold if the patent owner has a weak position; rather, it might be the case, inter alia, when the market for technology is extremely competitive or the number of potential licensees is limited.

b. Cost-Saving Strategies

One important element of Section 23 Patent Act is the possibility to save half the future renewal fees [Section 23(1), first sentence]. As illustrated in Fig. 3.2,84 the fees increase from year to year, becoming especially relevant from approximately the tenth year on. Before that time, fees should be of minor importance because they are normally very low compared to the costs, i.e., attorney and application fees, already incurred for the patent application.

According to the PatVal-EU study, 17.4 % of all patents are sleeping (unused),85 which suggests that the owners of these patents are likely to be interested in cost-saving opportunities. The following discussion presents three constellations in which the declaration of willingness to license could become especially relevant because of the possibility to save costs.

(1) Large Patent Portfolio

Larger firms in the field of complex technologies are reported to own large patent portfolios and frequently set up patent thickets, therefore facing enormous costs for the annual maintenance of the patents. These firms have to decide repeatedly between paying the next annual fee and letting a patent expire. The willingness to license regime offers such firms a compromise.

There are mainly two constellations in which the declaration of willingness to license may become interesting if a firm has a large patent portfolio and wishes to cut costs. Firstly, if a firm developed a technology and made a later improvement for which it has several patents for both stages, it may decide to keep all patents but declare the willingness to license for the patents for the first stage to save costs. These patents may no longer be of value for internal use, but may be to others, e.g., low-quality producers from which the patent owner could earn licensing fees. Secondly, a firm may have several for patents for a certain technology, but only consider some to be essential, i.e., using the technology or producing a certain product may only make sense or, in the case of blocking patents, may only be possible with access to the essential patents. In this case, the firm could decide to save costs by declaring the willingness to license for the non-essential patents. Contrarily, firms will almost never declare the willingness to license for patents that are used internally and are essential; otherwise they would risk losing a competitive advantage.

(2) Cross-Licensing

Another constellation in which the declaration of willingness to license may be motivated due to cost constraints is in connection with cross-licensing agreements. Such contracts are typically concluded between two or more patent owners and consist of several licences. In the PatVal-EU survey, around 3 % of the interviewees indicated that their patents are cross-licensed.86 In another study, companies in the electronic industry were reported to have the highest share of cross-licensing agreements.87 A declaration of willingness to license would not normally be made before the conclusion of a cross-licensing agreement because a patent owner would not be in a strong bargaining position if everyone is already allowed to use his patented invention for a reasonable remuneration. But after the conclusion of such an agreement, especially when several competitors are involved, the patent owner may decide to declare his willingness to license to save costs. The patent owner may even conclude some type of cross-license by deciding to agree to declare his willingness to license for all relevant patents.

(3) Non-core Technology

Large firms that operate in fields in which products and processes are complex and involve multiple technologies may own patents that do not cover their core business. These patents are spin-offs from innovation processes and are considered as having some value, but which cannot or will not be applied and exploited internally by the firm, e.g., because of the lack of complementary assets.88 The situation can also be defined as an open-innovation, inside-out process, as described above.89 Declaring the willingness to license for these patents to save costs and to find someone who is willing to pay a reasonable remuneration is a viable option, but basically only if these patents are not of value to direct competitors. The declaration would normally be made in the advanced term of a patent, since the annual fees are low in the first years and the patent owner might not be aware of the full strategic value of the patent at that stage.

c. Patent Litigation and Settlement Outcome Strategies

The declaration of willingness to license could also be used strategically in connection with opposition (Section 59 et seq. Patent Act) and patent litigation proceedings, especially regarding the validity of a patent (Section 81 et seq. Patent Act). Three cases can be distinguished, each of which depends on the point in time when the patent owner makes his declaration and on different motivations: before, during or after the legal proceedings. A patent owner may declare the willingness to license to reduce the risk of litigation for the following reasons: A potential plaintiff who wants to challenge the validity of the patent may fear the costs incurred by the proceedings and decide to pay a reasonable remuneration instead of litigating. Additionally, a patent owner may abandon the option in order to deter potential plaintiffs. This strategy may be especially relevant for owners of patents that have a broad scope; the broader a patent, the more likely is the involvement in legal proceedings. During validity proceedings the patent owner may declare his willingness to license to signal his readiness to negotiate licensing terms.90 After the proceedings, if the patent owner and another party have settled a dispute with a licensing or cross-licensing agreement, a patent owner who is no longer entitled to use the patented invention exclusively may declare his willingness to license to allow all competitors to use his invention against remuneration.

All these considerations should be more relevant for certain smaller firms than for large firms. It was shown above that the former may be more likely to be involved in patent litigation than the latter,91 which typically have broader patent portfolios for cross-licensing, interact repeatedly with licensing partners and have bigger legal departments.

3. Empirical Results Regarding Section 23 Patent Act92

After an overview (Sect. a.), further empirical results regarding Section 23 Patent Act are explained and analysed in the following discussion. The topic includes who the users are (Sect. b.), at what point in time the declaration of willingness to license is made (Sect. c.), results about the notification of use and the determination of compensation (Sect. d.), the importance of the possibility to withdraw the willingness to license (Sect. e.) and finally some empirical specifications (Sect. f.). Whereas the first five sections contain statistics, the last section describes the results of regression. Finally, some interim results will be presented (Sect. g.).

a. Overview

Between 1983 and 2008,93 76,094 declarations of willingness to license were made, of which 22,722 were patent applications and 53,372 took place after publication of the grant [Section 58(1), first sentence, Patent Act; Arts. 2(2), 97(3) EPC]. As explained above, the declaration of willingness to license for an EPO patent is only possible after publication of the mention of the grant, i.e., the first statistic only includes applications filed with the DPMA. In 383 cases an opposition was filed against a patent for which the willingness to license was already declared, and in 16 cases invalidity proceedings were initiated. In contrast, for 702 patents and therefore almost twice as many, the willingness to license was declared during opposition proceedings, and for 19 during invalidity proceedings. These data indicate the relevance of the possibility to declare the willingness to license as a settlement outcome strategy.

In Fig. 3.3, it can be seen that the number of declarations averaged around 2,500 each year until 1992 and the ratio of declarations occurring before and after grant was about 0.6. In the following years, more patent owners and applicants declared the willingness to license and the ratio fell to 0.16 in 1999, which is also the year in which the number of declarations reached its maximum. Since 2003, the number of declarations has risen to more than 4,500 per year, whereas the pre- or post-grant ratio has returned to about 0.5.


Fig. 3.3
Declaration of the willingness to license in 1983–2008

The rise that began in 1992 may be explained by the increase in existing patents and the statutory amendment that went into force on 1 June 1992. The amendment made it possible to withdraw the willingness to license under certain conditions. The high number of declarations in 1999 may be explained by another statutory amendment, in which the legislature increased the annual patent fees by around 15 % with effect from 1 January 2000.94 Since a draft for this amendment was already published on 27 August 1999,95 before it was enacted on 22 December 1999 and published in the Federal Law Gazette on 28 December 1999,96 several patent owners might have revised their patent portfolios and declared the willingness to license for some patents with the aim of reducing the overall amount of patent fees. This is a first indication that the possibility to declare the willingness to license is used to save costs.

In Fig. 3.4 the number of declarations is illustrated as a fraction of applications, of totally granted patents and of granted patents by the DPMA and the EPO. The applications or patents were grouped according to the year the application was filed (beginning with the year 1983 and ending with the year 200697). In these groups are 1,845,343 patent applications of which 1,074,627 were granted (342,585 by the DPMA and 732,042 by the EPO), and the willingness to license was declared for 57,045 (18,252 before and 38,793 after grant). One can consider the groups from 1983 to 1991 as completed because, for the vast majority, the patent never reached the full term. For these groups, 4–5 % of all applicants declared the willingness to license. For granted patents as a whole the fraction remained at around 5.7 % for all groups, whereas for patents granted by the DPMA the number amounts to 9 % and for EPO patents only around 4 %, albeit with an upward trend. The difference between DPMA and EPO patents might be because EPO patents are considered to be more valuable as the procedure is normally associated with higher costs. Thus, the declaration of willingness to license might be more likely for less valuable patents, such as non-core technologies or certain patents in a large patent portfolio. However, this statistical difference could also have other reasons. Firstly, EPO patent applicants are, in contrast to DPMA patentees, not allowed to declare the willingness to license, a restriction that reduces the time period for such a declaration. Secondly, applicants who only file their application with the DPMA might have better knowledge of German law. Thirdly, holders of EPO patents usually own patent families that consist of several national patent rights with the same claims in different countries, and they may not consider the declaration as attractive if this possibility does not exist or does not have the same legal consequences in all legislations.


Fig. 3.4
Fraction of declarations for applications and patents in 1983–2006

b. The Users

(1) Users Across Technology Classes

In Fig. 3.5 the distribution of the fraction of declarations (Fig. 3.4) is split into the following six technological classes: electric engineering; instruments; chemicals and pharmaceuticals; process engineering; mechanical engineering; and consumer goods and construction.98 A cursory glance already shows that the possibility to declare the willingness to license is used significantly more in the area of electrical, mechanical engineering and instruments than in the other three classes. For all granted patents from the years 1983 to 1991, which can be considered as complete, the willingness to license was declared for around 11 % in the area of electrical engineering, for 8 % in the area of mechanical engineering and 6 % in the area of instruments. One finds lower numbers in the area of consumer goods and construction (4 %) and process engineering (3 %). The lowest percentage is found in the area of chemistry and pharmaceuticals (2 %). The first two categories can be considered as especially complex technologies and the latter two as discrete, in which patents are very valuable and important to protect innovations.99 This confirms the theory that declaring the willingness to license would be more likely in industries in which rather large patent portfolios exist and where not only non-exclusive licensing but also cross-licensing occurs.


Fig. 3.5
Fraction of declarations for patents by technology class in 1983–2006

(2) Users Across Applicant Types

Another question is what types of patentees declare the willingness to license. As mentioned above, the initial intention of the legislature was to promote innovation by small and medium-sized enterprises as well as by individual inventors. It is possible to distinguish five types: non-profit organisation or university, individual inventor, small company, medium company and large company. The latter three are defined by the number of granted patent applications by the same firm in 1 year: a company is called small if less than five applications are granted and large if there are more than 50. In Fig. 3.6 the distribution of the fraction of declarations (Fig. 3.4) is split up according to this categorisation. It can already be observed at first glance that individual inventors do not significantly use the possibility to declare the willingness to license and that the larger a company is, the more likely it is to declare the willingness to license. This confirms that owners of large patent portfolios more often declare the willingness to license. The number for large companies was 13.2 % in 1983 and reached the same figure in the years 1984–1988. For medium-sized companies the proportion was 7.3 % in 1983, decreasing to around 5 % in the following groups. The number for small companies as well as individual inventors was 2.25 % in 1983, decreasing slightly for the former type of owners and increasing to 3.5 % in 1997 for the latter. The importance of the possibility to declare the willingness to license appears to be increasing for individual inventors, albeit from a very low level. Moreover, individual inventors declare their willingness to license earlier than patent owners from the other categories. Judging by the groups, universities and other non-profit organisations are the least likely to declare the willingness to license. These data do not support the notion that patentees who usually lack complementary assets tend to declare the willingness to license, but this situation may change to a certain degree in the future, at least for individual inventors.


Fig. 3.6
Fraction of declarations for patents by applicant type in 1983–2006

c. Time of Declaration and Patent Lifetime

Figure 3.2 illustrates that annual patent fees are increasing. As a consequence, the possibility to save costs also grows every year. Thus, one would expect applicants and patent owners to declare the willingness to license, especially for older patents. Figure 3.7 confirms this assumption: in the years 1983–1988 the declaration occurred on average 9.0–9.5 years after the application; in more recent years this number has decreased because there are potentially more patents that still exist. For patentees who submitted their patent application in 1983, 75 % of all declarations occurred before the 12th year; the tendency in the following years was towards declarations earlier in the life of the patent.


Fig. 3.7
Time between application and declaration by year of application (1983–2006)

In Fig. 3.8, the years 1983–1988 show that after the declaration of willingness to license for a patent continues to exist for 5.5–6.5 years on average. The legislature initially intended that, due to the possibility to save costs with the declaration of willingness to license, patents should be maintained for a longer time. Patents with declared willingness to license do exist for a longer time than other patents. While “normal” patents from the years 1983 to 1988 continued for around 13 years, a declaration of willingness to license prolonged the life of a patent by around 15.5 years. Therefore, Section 23 Patent Act seems to have had the effect that the lawmakers intended. Two additional important aspects of patent duration should be noted. Firstly, while “normal” EPO patents exist on average 0.66 years longer than DPMA patents, the difference for patents with a declared willingness to license is higher (1.6 years). The reason may again be that EPO-granted patents are more valuable than those granted by the DPMA. Secondly, patents of large companies and not those of individual inventors and small companies are kept longer after the declaration of willingness to license, which is a contradiction to the initial intention of the legislature.


Fig. 3.8
Time between declaration and expiration by year of application (1983–2005)

Another question is whether the income of the DPMA and the state is lower because of the existence of Section 23 Patent Act. The above figures support an analysis with the following results: Patent owners who filed their application from 1983 to 1987 and declared the willingness to license received a subsidy of around €29.2 million. But because they kept the patents longer the loss of income may only amount to about €4.7 million. As a consequence, large companies receive a subsidy to keep their patents longer without the possibility to exclude others from using the protected invention.

d. Notification, Licensing and Determination of the Amount of Remuneration

Section 23 Patent Act is intended to facilitate licensing. Since information on licensing is confidential in Germany and notification of use according to Section 23(3), first sentence, Patent Act is only sent from the potential user to the patent owner, i.e., the patent office is not involved, no concrete data is available on the licensing of patents for which the patent owner declared the willingness to license. However, there are several reasons why there have been very few notifications and licensing agreements so far regarding these patents. Although there are a significant number of declarations each year, the DPMA has only assessed the amount of remuneration in three (published) cases. In the areas of mechanical and electrical engineering where this is very common,100 licensing in Germany usually only occurs within cross-licensing agreements or if a patent is infringed. Generally, firms do not look for patented technologies but try to develop their own. Moreover, in several fields involving complex technology, knowing what the patent is about is often not sufficient, because one needs the know-how to implement it as well. The PatVal-EU survey confirms this assumption: inventors indicated that only 3.5 % of the patents for which the willingness to license was declared are in fact licensed (licensing and use: 2.5 %). Remarkably, this percentage is even lower than when the willingness to license was not declared (5.86 %; licensing and use: 3.84 %). Interestingly, the opposite is the case for cross-licensing (willingness to license declared: 7.5 %; not declared: 2.92 %). Therefore the assumption that the declaration has some importance in connection with cross-licensing might be true. That a share of patent owners who declared their willingness to license may be interested in licensing (and also saving costs) can be seen by the higher percentage of sleeping patents for which the willingness to license was declared (21 %) in contrast to those where there was no declaration (17.08 %).

e. Withdrawal of the Willingness to License

As already explained above, the 1992 introduction of the possibility to withdraw the willingness to license might be the reason for slightly more intensive use. Another question is, however, whether some patent owners who declared their willingness to license also withdrew it later. The aim of the legislature was that the patent owner should be able to optimally exploit his patented invention. He should still be allowed to switch back to protection by a property rule and grant an exclusive licence or use the patented invention internally if no one indicated interest in using the patent. While it might be assumed that this amendment would have led to an intensive use of the possibility to withdraw, this was not the case until 2000. In that year, almost 8 years after the amendment came into force, a withdrawal was submitted to the DPMA for the first time. From 2000 to 2008 identifiable withdrawals totalled 386, which is still low compared to the number of declarations. Interestingly, individual inventors make most use of withdrawals compared to the other above-defined groups and based on the fraction of all declarations. Unfortunately, it is not possible to find out from the data available what the motivations for withdrawal were.

f. Empirical Specifications (Regression Analysis)

Up to this point, the theoretical considerations and results of the statistics might lead to the assumption that certain independent variables are positively or negatively correlated with the dependent variable declaration of willingness to license. An empirical model (regression analysis) lends more certainty to the analysis when testing the correlations and examining which effects are statistically significant, i.e., where there is a very low probability that the result is a coincidence. Such a model was used to test the influence of the characteristics of the patent, the applicant, the technological environment and the invention on the probability of declaring the willingness to license. Because of the availability of more variables, not the entire DPMA and PATSTAT dataset was used. Rather the model used that dataset merged with the data from the PatVal-EU survey. Therefore, the sample for the regression analysis contained only 6,277 patents granted by the EPO with effect in Germany and 210 declarations of willingness to license. The results that are summarised in the following discussion therefore show not only a trend, but also additional information on what the reality is:

  • The following variables were used to test whether the willingness to license is declared for less valuable patents: whether a patent was opposed at the EPO, the number of countries for which the patent was granted (family size), the number of citations a patent application received within 5 years (forward citations) and the number of other patents the patent cited.101 Although all these variables are negatively correlated to the variable declaration of willingness to license, only the variable for family size had a statistically significant effect. Again, one can only assume that the willingness to license is declared for less valuable patents, possibly in large patent portfolios or non-core technologies. Contrarily, the patent scope and its generality seem to have no importance.

  • Another question is whether certain types of motivation for patenting the invention have an effect on the declaration of willingness to license. These can be categorised into how important commercial exploitation, licensing, cross-licensing, prevention from imitation, blocking other patents and reputation are for patenting the invention. Prevention from imitation and cross-licensing had a significant effect: the former had a negative effect and the latter positive. Both findings were expected since there is a liability regime after the declaration and exclusivity cannot reasonably be the dominant strategy. Additionally, cost-saving especially makes sense after the conclusion of cross-licensing agreements, as argued above. Interestingly, licensing as a motivation for patenting again does not seem to play a role.

  • The age of the patent also seems to have an effect, as assumed above.

  • The data analysis also indicates that small and medium-sized enterprises make less use of the possibility to declare the willingness to license than individual inventors and especially large firms. Interestingly, the portfolio size and the previous use of the possibilities of Section 23 Patent Act seem to have an effect on the probability to declare the willingness to license.

  • As seen above, declarations for patents are less likely in the chemical and pharmaceutical industry than in other areas of technology. Competition has a negative effect on the probability of declaring the willingness to license. Whether the willingness to license is declared more often for non-core patents, unfortunately, remains unclear.

  • Moreover, a declaration of willingness to license seems to be less likely for science-based patents. Since science-based technologies usually require further technological development and are easier to transfer as they are codified, strategic licensing considerations normally play a role when deciding about the declaration of willingness to license. Additionally, there is a positive correlation between the number of declarations and the fact that a patent has built on other innovations. Licensing is also unlikely here because the implementation of cumulative technologies normally requires support.

g. Interim Results

The most important results follow. Section 23 Patent Act is a provision that is used more often by large firms with large patent portfolios for less valuable and/or older patents in order to save costs. The patents are predominantly in the area of complex technologies and less likely in fields where a firm directly competes with others. Contrary to theoretical considerations, the declaration of willingness to license does not usually play a role in connection with simple licensing, but rather with cross-licensing agreements; settling disputes in certain constellations may play a role. Additionally, the two statutory amendments (the possibility to withdraw the willingness to license, although seldom used, and the rise in patent fees) seem to have had a positive effect on the number of declarations. Importantly, the implementation of Section 23 has led to a longer patent lifetime, especially for patents owned by large companies, but to lower revenue for the state. As a consequence, Section 23 may currently reduce the overuse of property rules, but whether this is efficient and has a significant impact cannot be answered with certainty and is very doubtful.

4. Licences of Right in the Proposal for a European Patent with Unitary Effect, France and the United Kingdom

Besides Germany, licence of right provisions exist or have existed in several other countries, for example in Bulgaria (Article 30 Bulgarian Patent Act), Italy (Article 50 Italian Patent Act), France (former Article L613-10 French Patent Act102), Luxembourg (Article 56 Luxembourgian Patent Act 1992/1998), Slovakia (Section 25 et seq. Slovakian Patent Act), Spain (Section 81 et seq. Spanish Patent Act), UK [Section 46 et seq. United Kingdom Patents Act (UK-PA)], and supposedly in some developing countries, since a licence of right provision is part of the Model Law for Developing Countries for the Protection of Inventions.103 However, they do not exist in two of the most important patent systems: the United States and Japan. For the latter, a study aimed at helping the Japanese government to implement a proper patent fee system and an effective financial management policy recommended the introduction of a licence of right scheme to facilitate open innovation.104

A licence of right provision is also part of the regulation on the European Patent with Unitary Effect; it will be explained in the following discussion (Sect. a.). Moreover, it is worth looking at other legislations for alternatives and possible improvements to Section 23 Patent, as the provisions on the willingness to license and licences of right have not been harmonised.105 This work takes a closer look at the situation in France where the provision was introduced in 1978, but abolished in 2005 (Sect. b.), and in the UK, where the provision was introduced in 1919 (Sect. c.). Both legislations vary in certain respects from Germany.

a. Licences of Right Under Regulation (EU) No. 1257/2012

A licence of right provision, which is only slightly different to Section 8 Regulation (EU) No. 1257/2012, was already a part of all proposals for a European Community/Union patent beginning with the Convention for the European Patent for the Common Market in 1975, which was signed by nine member states but never entered into force because it was not ratified by enough states. Article 8 Regulation (EU) No. 1257/2012 provides that the proprietor of a European patent with unitary effect may file a statement with the EPO that he/she is prepared to allow any person to use the invention as a licensee in return for appropriate compensation (para. 1). A licence obtained under this Regulation shall be treated as a contractual licence (para. 2). Article 8 does not contain a possibility to withdraw from the declaration of willingness to license; however, this should be possible pursuant to Article 9(1)(c) Regulation (EU) No. 1257/2012.

Clearly, Article 8 Regulation (EU) No. 1257/2012 is very similar to Section 23 Patent Act. Regarding the amount of remuneration, Article 8(1) Regulation (EU) No. 1257/2012 contains the term appropriate compensation. The words reasonable and appropriate are supposedly interchangeable [reasonable is used here as a translation of the German word angemessen; see Section 23(1), first sentence, Patent Act]. Therefore, the German and European provisions are identical in this respect (but neither contains any further information as to what this term means) and the same factors could be relevant to determining compensation.

The differences between Section 23 Patent Act and Article 8 Regulation (EU) No. 1257/2012 and other issues that are currently unclear are the following. Firstly, in Article 8 Regulation (EU) No. 1257/2012, the percentage by which the annual fees will be reduced has not yet been determined: Article 14(3) Regulation (EU) No. 1257/2012 specifies solely that the renewal fees for patents that fall due after receipt of the statement will be reduced. Secondly, Article 8(2) Regulation (EU) No. 1257/2012 determines explicitly that a licence obtained under this Regulation will be treated as a contractual licence; however, this should not lead to any differences in application compared with Section 23 Patent Act. Beyond these dissimilarities, Article 8 Regulation (EU) No. 1257/2012 is less concrete than Section 23 Patent Act regarding the legal situation between the patent owner and the user. Regulation (EU) No. 1257/2012 does not contain any information about the notification process; it is therefore currently unclear how this should take place. Additionally, there are no conditions specified under which the patentee may still be allowed to apply for an injunction. A clarification regarding these two points would make sense. Thirdly, it is also uncertain at the moment whether the EPO, a court or another body will be the first instance that will be competent to determine the amount of compensation. According to Article 20(5) of the Proposal for a Council Regulation on the Community Patent,106 which originated from the European Commission in 2000, the Commission itself should be that instance. However, this changed to the Community Patent Court in the Proposal for a Council Regulation on the Community Patent in 2003.107 Regulation (EU) No. 1257/2012 does not contain any information about this question. Likewise, it is currently unknown how much the fee for determining the amount of compensation will be.

b. Licence de Droit in France

(1) Comparison with Section 23 Patent Act

Since Article L613-10 French Patent Act was repealed in 2005, the following discussion refers to the former version of this provision. The mechanism provided by that provision could be viewed as being in the middle of compulsory and voluntary licensing.108 It is similar to Section 23 Patent Act, but differs in the following essential points.

Firstly, the former Article L613-10 did not contain the formal allowance necessary to switch to a licence of right regime, but the Director of the French National Institute of Industrial Property (Institut National de la Propriété Intellectuel), who decided on the application of the licence of right system, was only allowed to accept it “if a documentary report has been drawn up not showing any anticipation manifestly affecting the patentability of the invention”.109 Secondly, whereas in Germany only the granted patent is an exclusive right [Secs. 58(1), third sentence, and 9 German Patent Act], in France “the exclusive right of exploitation […] shall take […] effect as of the filing of the application” (Article L613-1 French Patent Act).110 Therefore, in France protection by a property rule changed to a liability rule but in Germany it only changes if the declaration takes place after the grant. Thirdly, according to the former Article L613-10(3), the annual fees were also reduced like in Section 23(1), first sentence; but the former did not contain the exact percentage of reduction to be determined “by Conseil d’Etat decree” [Article L612-19(1)],111 which was 40 % in 2005.112 It is also important to mention at this point that the annual fees for a French patent are significantly lower than for a German patent.113 Therefore, the percentage of savings was not only lower, but also the absolute amount. Fourthly, if the patent owner and the (potential) user did not reach agreement on the amount of compensation, the patent office was not allowed to determine it [Section 23(4), first sentence, German Patent Act], but rather a certain court [former Article L613-10(2), third sentence, French Patent Act]. Fifthly, it was explicitly determined that the licensee was allowed to surrender the licence at any time. Sixthly, contrary to the German provision, which only allowed a patent owner to switch back to protection by a property rule if there was no notification of use [Section 23(7), first sentence, German Patent Act], in France the patent owner was allowed at any time to withdraw from a licence of right regime which, however, had “no effect on the licences of right already obtained or requested” [former Article L613-10(4) French Patent Act].114 Seventhly, whereas the non-exclusive licensee in Germany cannot, in principle, institute infringement proceedings,115 the holder of a licence of right is allowed to do so if, “after a formal notice, the owner of the patent” does “not institute them” (former Article L615-2(2), second sentence, French Patent Act).116

(2) Reasons for the Repeal

In connection with the repeal of Article L613-10 French Patent Act, Article L612-20 was amended, according to which the amount of fees charged at the time of application, the examination and the delivery of the patent as well as its renewal could be reduced in certain cases: provided the applicant is a natural person, small or medium-sized business or non-profit institution in the teaching and research sector (para. 1) and makes a simple declaration regarding these requirements to the patent office (para. 2, first sentence).117 Similarly the former Article L612-20 only contained a fee reduction for natural persons whose resources were insufficient to be liable to income tax, on condition that non-patentability was not obvious. The political reason for the amendment was predominantly to support small and medium-sized enterprises, but also the other mentioned groups.118 The practical reason was that the former provisions caused several difficulties, especially in the examination of the requirement of patentability and the conditions under which a person may be entitled to receive a subsidy.119 As a consequence and with regard to the new direct subsidisation of several special groups, the French legislature regarded a licence of right regime as no longer necessary or justified.120

c. Licences of Right in the United Kingdom

The British licence of right provision, now in existence for more than 90 years without any major changes, is also comparable to Section 23 Patent Act. However, there are important differences.121 Firstly, applying for a licence of right is not possible for patent applications [cf. Section 23(1), first sentence, German Patent Act] until after the patent grant [Section 46(1) UK Patents Act]. Secondly, although future renewal fees are halved under a licence of right regime similar to Germany’s [Section 46(3)(d) UK Patents Act and Section 23(1), first sentence, German Patent Act], the cost-saving opportunities in absolute terms are lower because of lower renewal fees. Depending on exchange rates, the renewal fees are very similar to the French fees.122 Thirdly, the British law contains another variant under which conditions the patent owner is allowed to withdraw from the licence of right regime. According to Section 47(2) UK Patents Act, the comptroller may cancel the entry made under Section 46 if he is satisfied that there is no existing licence under the patent or that all licensees under the patent consent to the application. Fourthly, the British law explicitly stipulates in Section 46(3)(c) that, if in proceedings for infringement of the patent (other than by importation of any article from a country that is not a member state of the European Economic Community), the defendant or defender undertakes to take a licence, no injunction shall be granted against him. This matter is still the subject of debate in Germany.123 Fifthly, similar to the situation in France, the licensee can request the proprietor of the patent to initiate proceedings to prevent any infringement of the patent; if the proprietor refuses or neglects to do so, the licensee can institute proceedings to prevent infringement in his own name as if he were the proprietor [Section 46(4)]. Sixthly and most importantly, whereas in Germany the possibility of using a patented invention is basically unlimited after the declaration of willingness to license, the situation in the UK is different. According to Section 46(3)(a) UK Patents Act, any person is entitled to a right to a licence under the patent on such terms as may be settled by agreement or, in default of agreement, by the comptroller. The relevant aspect is that the comptroller can determine not only the amount of compensation but also other licensing terms between the patent owner and the licensee. The licensee is not allowed to work the invention until the terms are settled.124

Several decisions concern the determination of licensing terms under a licence of right regime; some of them originate, however, from the time after the patent term was prolonged in 1978 from 16 years to a maximum of 20; it was possible to keep certain already existing patents alive with licences of right endorsed from the beginning of the 17th year until the end of the 20th year.125 The basic principle is that the licensing terms should be those that could be agreed between a willing licensor and a willing licensee.126 Basically, limitations and conditions may be imposed by the comptroller at his discretion, but there may be constraints due to EU or World Trade Organisation law.127 So far, under certain conditions regarding the concrete situation, inter alia, licensing terms on security for royalty,128 import and export prohibitions,129 prohibition on assigning or sub-licensing,130 termination in the event of a take-over situation,131 quality control,132 termination on the insolvency of the licensee, the appointment of a receiver or administrator of its assets133 or if the licensee challenges the validity of the patent134 have been accepted.135 Inter alia, clauses on termination without prior warning for even trivial or accidental breaches136 or requiring a patent attribution137 have been refused.138

d. Interim Results and Additional Information

A comparison of Section 23 German Patent Act, Article 8 Regulation (EU) No. 1257/2012, ex-Article L613-10 French Patent Act and Section 46 UK Patents Act shows that there is a basic idea regarding willingness to license and licence of right provisions. Most importantly, all of these provisions transform an exclusive right into a non-exclusive one, which depends on the decision of the patent owner, and it is also the patent owner who may benefit from a reduction in the annual fees. Several variations exist concerning the absolute and relative amount of cost-saving opportunities, the instance that determines compensation, whether only the amount of compensation or additionally any further licensing terms can be settled and the conditions under which the licence of right may be withdrawn.

Since cost-saving opportunities are fewer in France and the UK than in Germany, the importance of reducing costs might be less relevant there. Moreover, an advantage of the British provision may be that the comptroller can balance the interests of the licensor and the licensee by settling licensing terms whereby the licence is not necessarily unlimited. Both the British and the former French provisions present the advantage for the patent owner that the possibilities for withdrawing from the licence of right regime are more flexible.

However, regarding the situation in the UK it has been argued that the provision might only be used by a patent owner who has been unable to exploit his patent to good effect and wants to economise on the renewal fees.139 Whether making an entry in the register that licences of right are available makes any difference would be doubtful because, if the invention were commercially attractive in the first place, the patent owner would have had no difficulty in finding a licensee.140 According to the Gowers Review of Intellectual Property provided by the UK’s economics and finance ministry, the take up of licences has been low141 due to a lack of awareness and the unattractiveness of negotiating licensing terms.142 In the latter case, the possibility of settling all licensing conditions and not solely the amount of compensation might be one reason. Additionally, the French provision was abolished for two reasons: it did not seem to have the desired effects and there were doubts as to whether individual inventors or small and medium-sized enterprises would have profited from it. As a consequence, it is doubtful whether the differences between the provisions have any relevant effects, especially on the number of licences.

5. Evaluation

Willingness to license and licence of right schemes in general, i.e., without referring to a concrete provision, have recently attracted attention in the debate about how to improve the patent system.143 For example, in the Scientific and Technological Options Assessment Report of the European Parliament on the Improvement of the European Patent System in 2008, the declaration of willingness to license is considered to be attractive for many applicants and especially for three groups. The first group comprises independent inventors who have no clear picture of which use their inventions might have and cannot, therefore, easily promote their inventions. The second group includes small and medium-sized enterprises that do not have the financial stamina to defend their rights. Finally the third group comprises universities.144 It was shown above that the current German provision and possibly other licence of right mechanisms do not predominantly serve the interests of the second and third group. Whether the legislature should implement provisions to promote innovation by these groups is a question that cannot be answered here. If the legislature decides to do so, however, a provision such as Article L612-20 French Patent Act seems preferable to the willingness to license provision in the current Section 23 German Patent Act, because it is easier to name and identify the groups directly than to create a provision that is supposed to be used by them.

Another question is whether a licence of right provision in general should be part of a patent law system because the economic advantages are at least very doubtful. One can argue the following against such a provision. Firstly, it may make the problem of patent thickets even worse despite the fact that there is no exclusivity, because patent owners receive a subsidy to renew their patents for a longer time (which large firms, the main users, actually seem to do). Secondly, if the main purpose is cost-saving (which currently seems to be the case), there is an easier solution: lowering the patent fees in general. Thirdly, current provisions fail to promote the use and commercialisation of inventions and open innovation in general. By extension, this might be the case for all licence of right mechanisms.

The following arguments favour such a provision. Firstly, although the willingness to license seems to be declared for less valuable patents in a field where there is less competition, the potential for blocking is reduced since injunctions are not available. Secondly, dynamic effects may result because, although the patents are kept for a longer time, there is still a subsidising effect. As a consequence, innovators may have more funds available for future innovations. Lastly and very importantly, although this is currently not the case, in theory the provision potentially promotes patent owners who want to license their patented inventions non-exclusively and who want to open their innovations to others, especially in a so-called inside-out process. As a result, having a licence of right provision as part of a patent law system seems preferable to not having it.

Considering the alternative elements that a licence of right provision could include, this study concludes the following. First and foremost, the major concern is that the current provisions—which may also be true for Article 8 Regulation (EU) No. 1257/2012—seem not to promote licensing activity, although this may theoretically be possible, and they are used only in connection with cross-licensing. So what are the alternatives to promote licensing? One reason for the current situation may be that the declaration of willingness to license is normally only published in the Patent Register and Gazette.145 An alternative would be for the patent office or another entity to institute some kind of innovation pool or clearinghouse where interested parties could find at least a better categorisation of the patents for which the willingness to license is declared. This idea will be pursued further below under the heading “Royalty Collection Clearinghouses”.146

In order to increase the number of declarations, the following alternatives to Section 23 Patent Act are proposed. Firstly, the patent office should be allowed to settle other licensing terms than just compensation because this should lead to a better balance of the interests of both parties and—what is often necessary—the conditions under which a transfer of know-how may also be settled. To solve the problem caused by the unattractiveness of negotiating licensing terms, the patent office or the legislature should provide a model licence.147 The patent office and not a court should be the first instance to settle the terms because it can be considered as competent to make fast and cheap decisions. Secondly, the introduction of the system with the European Patent with Unitary Effect, including Article 8 Regulation (EU) No. 1257/2012, may prove useful because its provisions would lead to harmonisation in the common market; whereas patent systems are still limited to national borders, the market is not limited. This work considers the current limited possibility to withdraw the willingness to license as optimal because it leads to a larger number of declarations. Caution must be exercised, however, since extending this possibility could augment the hold-up potential in certain constellations.

Whether the percentage of the annual patent fees should be reduced, for example by making the reduction dependent on the age of the patent, requires further economic research and cannot be answered here, but should be investigated.148

6. European Interoperability Patent

A very prominent proposal that fosters a licence of right provision can be found in the European Interoperability Patent, originating from IBM and based on the Soft-IP Approach (the Soft-IP Approach will be explained in more detail in Sect. B.6 in Chap. 4).149 It is important to note that, whereas the Soft-IP Approach is based on a mechanism in which a liability rule applies by default, the European Interoperability Patent is a regime in which the patent owner can decide if he prefers protection by a property rule or a liability rule. Because the proposal is similar, but not directly comparable with the above-explained licence of right regimes, it makes sense to explain the concept, its advantages as mentioned by its advocates (Sect. a.) and evaluate it (Sect. b.) in the following analysis in its own section.

a. The Concept and Advantages Mentioned by Its Advocates

The main commentaries by advocates are summarised as follows. The existing national patent systems should remain in effect. Additionally there should be an EU patent system in which patents are automatically endorsed to allow licences of right. The patentee should still file his application with the EPO. After the patent grant the patentee can decide whether to go the national or the European route. Whereas national legislations normally require a translation into the national language, this would not be the case if the patentee makes a decision for the EU patent. If parties do not reach an agreement on the licensing terms, a European patent court could settle the issues.150 If a licensee does not pay the agreed or settled compensation, the patent owner may request an injunction in a national court.151

The main legal and economic advantages of such a system are the following. Firstly and generally, this type of system should lead to IP protection that balances open and proprietary protection because society benefits from both models.152 Secondly, there should be strong incentives to use the EU patent system because firms can save translation and other costs.153 Thirdly, assuming that there are strong incentives for many patent owners to use the future system with the European Patent with Unitary Effect, the problem of the so-called innocent infringer could be reduced.154 An innocent infringer may be defined as a person who uses the patented invention but does not know or could not reasonably have been expected to know of the patent’s existence. An innocent infringement typically occurs when someone uses a patented invention for which a translation in the user’s national language does not exist. Such an infringer could rest assured of being allowed to use the patented invention and would not be faced with the prospect of his/her business being disrupted or closed down. Fourthly, new business models should be expected but with no disruption to existing businesses since these firms could still use the present patent system.155 Fifthly, the risk of encouraging patent trolls should be lower compared with the official proposal of the EU patent.156

b. Evaluation

While the two approaches, that of the European Interoperability Patent versus that of the possibility of declaring the willingness to license in a national system where patents are normally exclusive rights, are not directly comparable, some of the lessons from the above analysis of Section 23 Patent Act can be applied to the evaluation of this proposal. Major differences exist between two such systems. The former should typically present more cost-savings and the patentee cannot make a decision whether he prefers a non-exclusive right up until the patent expires, but only after the patent grant; additionally, a withdrawal is not possible. As a consequence, there should be no connection between the age of the patent, the (supposed) use in cross-licensing and the choice of the European Interoperability Patent. However, as the patentee expects lower value, it will probably be used for patents from bigger firms in the field of complex technologies because cost is of special concern to them and the potential of exclusion may be of lesser importance. In a final overall comparison, the proposal seems likely to be an improvement over the current patent system. Admittedly, some patentees like IBM would benefit and, in some cases, stronger incentives would promote the choice of a liability rule system, leading in turn to more technology transfers. However, the major disadvantage is that the proposal would only be attractive to certain patentees. Possibly more efficient will be the situation under the regulation on the European Patent with Unitary Effect, which is attractive to almost all patentees acting at a European or international level. Therefore, this work considers the European Interoperability Patent as a step in the right direction, but not perfectly suited for implementation.

7. Results

The most important results are the following. It was shown that Section 23 Patent Act contains a mechanism that, theoretically, should attract patentees who follow a non-exclusive licensing, cost-saving or patent litigation and settlement strategy, which involves switching from protection by a property rule to a liability rule. Although Section 23 is considered as having disadvantages, the mechanism is used by numerous patent owners. Most importantly, it affects the licensing relationship in very few cases. Therefore, if it affects the overuse of property rules at all, it should be to a lower extent, where it is really relevant. A comparison with other licence of right provisions shows that there are better alternatives than the German provision. As a result, this work argues that licence of right provisions should be part of a patent law system, in which patents basically confer exclusive rights, although the positive effects may not have a greater impact. Most importantly, a mechanism that promotes licensing relationships should be incorporated into this regime. This proposal will be explained in more detail below at C.2.c in the section “Royalty Collection Clearinghouses”. Additionally, it seems preferable if the patent office rather than a court is allowed to settle not only the amount of compensation, but also other licensing terms. Moreover, although the European Interoperability Patent appears to be an interesting suggestion, it should not be pursued further.

B. Patent Pools with Liability Rule Mechanisms

As explained above,157 patent pools may be defined as “an agreement between two or more patent owners to license one or more of their patents to one another, or to license them as a package to third parties who are willing to pay the royalties that are associated with the licence”.158 As a general guideline, patent pools with private liability rule mechanisms are only those in which “patent holders accept to submit the use of the patent to the rules of the pool allowing for the use of the patent by parties who are members of the pool and possibly also by third parties”.159 In the following discussion it is explained to what extent these pools exist and what variants they have (Sect. 1.), what the general advantages and disadvantages of patent pooling are (Sect. 2.) and what the current legal environment for patent pools is (Sect. 3.). The legal environment will then be evaluated, followed by some suggestions (Sect. 4.). The analysis aims at answering two questions, namely whether patent pools as currently employed are efficient mechanisms to reduce the overuse of exclusivity and whether the legal environment should be altered.

1. Existence of Patent Pools and Important Variants

As explained above, patent pools with a liability rule structure are basically only those containing a great number of patents formed by many patent owners. Large patent pools can currently be found in the area of electrical engineering and typically establish a technical standard, such as MPEG-2, Bluetooth, OpenCable, DVD3C, G.729 Audio Data Compression, MPEG-4, IEEE 1394/FireWire, 3G,160 DVD6C, DVB-MHP and AVC/H.264.161 Important patent pools can additionally be found in three technological sectors: pharmaceuticals, biomedical and agricultural technologies. There are, inter alia, the Golden Rice and the AvGFP pool (green fluorescent protein).162 Several pools in this area of technology are quite different from those with the purpose of standardisation because the motivation may not be purely economic—indeed, many do not aim to make profits—and issues such as public health may be the driving force. One example is the patent pool of the organisation UNITAID, which bundles patents to produce medicaments for developing countries.163 Non-economic aspects such as public health are however not taken into consideration here.

The existing pools differ in the reasons for their creation, in the way they are managed and in their licensing conditions, especially how royalties are determined.164 The purposes that are mentioned are, for example: offering “one-stop-access” for licences necessary to produce products according to a standard (e.g., MPEG4165), improving the availability and quality of products (e.g., G.729 Audio Data Compression166) and assisting in clearing a patent thicket (e.g., AvGFP167). A pool may be managed by one member company (e.g., the DVD3C pool168), an entity founded by several companies (e.g., the Bluetooth pool169), or by a third party (e.g., the OpenCable pool170). A pool designed around the 3G Platform Arrangement provides more flexibility than the others and may form its own category of pools due to there being one overall umbrella organisation, as well as multiple entities, each of which develops licensing programmes for specific standards. Beyond having a standard on offer, the aim is that licensors and licensees may agree upon other arrangements such as cross-licensing and licensing of non-essential patents.171

For some patent pools, independent experts determine the essential patents that should be part of the pool (e.g., OpenCable172). Determination of royalties may take place according to FRAND conditions (e.g., MPEG-2173). Alternatively, the pool may determine an ex-ante fixed fee for each end-product that uses the technology (e.g., OpenCable174). In other examples, all pool members, but only members, may be allowed to use the technology on a royalty-free basis (e.g., Bluetooth175). In these cases, the patent owners’ reward for submitting or licensing patents to the pool lies solely in being able to use the entire technology in their own products; the value of the individual patents has no importance. In other cases, the division of the royalty is based on numeric proportion rules (e.g., MPEG-2) or on rules in which the value matters (e.g., DVD3C).176 In some cases only pool members can obtain a licence (e.g., Bluetooth177) and in other cases third parties also may (e.g., G.729 Audio Data Compression178), but sometimes only specific parties are accepted. Some patent pools have grant-back clauses in their agreements, requiring all licensors to incorporate their new essential patents into the pool (e.g., MPEG-4179).

2. General Economic Advantages, Disadvantages and Risks

As the previous paragraph suggests, large patent pools with a structure that could be characterised as a liability rule mechanism currently exist to reduce the overuse of exclusivity, which might, at first glance, enhance economic efficiency. However, it must be said that the effects of patent pooling are ambivalent with respect to allocative, productive and dynamic efficiency.180 What the exact effects of the existing pools are would require specific empirical research, which so far has only been conducted to a very low degree and might be complicated by the difficulty of access to available data and the dependence on several factors. One aim alone, an investigation of the effects of contemporary patent pools, would require data over a much longer time period than is currently available.181 However, the general and most important economic advantages (Sect. a.), along with the disadvantages and risks (Sect. b.) of patent pooling, irrespective of any legal regulation and the situation in a concrete industry sector, can be determined based on what existing pools demonstrate and will be explained below.