Overuse of Exclusivity in Patent Law

Munich, Germany


Nowadays, the main purpose of a patent law system is predominantly, while balancing private and public interests, to govern economic decisions, to encourage innovation and, in general, to reach economic efficiency. On the one hand, the question is, therefore, whether the current German patent law system (including German and European antitrust laws), as it is currently applied, and the future system under the regulation on the European Patent with Unitary Effect, as it is designed, contain too many exclusive elements to be efficient. On the other hand, because the choice between the two options may not be purely economic, it must also be investigated whether other factors could be taken into account or—more importantly for an analysis that focuses on the economic effects of the patent system—make it necessary to use either exclusive or non-exclusive elements. Two additional perspectives may be relevant: distributional preferences and other justice considerations.1 Although these two categories and economic efficiency may not be strictly distinguished from each other, the additional perspectives should be understood here as limitations on economic efficiency. Also, due to the present focus on a concrete legal system, at least shortsighted legal reasons (i.e., higher-ranking law and multilateral contracts) may play a role.

As a first step, an analytical framework is provided to explain the terms property rules and liability rules, then the terms exclusivity and non-exclusivity, as they are used here, and finally to analyse the extent to which German law and the regulation on the future European Patent with Unitary Effect make use of them (Sect. A.). The next step investigates economic efficiency, generally and with respect to the field of patent law, and its significance for the choice between property rules and liability rules. Next, it is analysed whether this type of economic consideration is consistent with the choice the German legislature made and the European legislature wants to make (Sect. B.). The third step of the analysis examines whether there could be or—more importantly here—whether there are other binding and relevant perspectives that must be taken into account and with which the choices inherent in the above legislation could be explained (Sect. C.). Fourthly, the analysis investigates whether some of the relevant factors found in the previous paragraphs, concluding that property rules are not always preferable, can actually be identified in the German and European systems. Facts and typical situations are presented which the current statutes cause and which may result in inefficiencies (Sect. D.). The chapter concludes with the presentation of results (Sect. E.).

Very importantly, it must be noted that, in answering the first research question, in this chapter, the so-called private liability rule regimes are not taken into account. As far as a separation is possible, however, these private liability rules are explained below in terms of their mechanisms and where they could have relevance. Perhaps a separation is not possible, but the negative effects of the current patent system can be identified despite the existence of private liability rules. There are two reasons for conducting the analysis in the manner explained: Firstly, presentation of the analysis would otherwise be too complicated. Secondly, private liability rules can sometimes be used to solve the problems outlined in this chapter. Therefore, it makes sense to combine their analysis with the suggestions in Chap. 4.

A. Analytical Framework and Status Quo

Before starting a concrete analysis, it is necessary to answer the following introductory questions: What are the issues concerning exclusivity and non-exclusivity, i.e., questions about the exercise of the right, and what distinguishes them from issues regarding the emergence of a right? What are exclusivity and non-exclusivity? Is it necessary or possible to create certain sub-categories of the liability rule? What are the most relevant provisions and regimes regarding these aspects in current German law and the future regulation on the European Patent with Unitary Effect, and what is the balance between exclusivity and non-exclusivity? One way to answer these questions is by using the renowned law and economic framework of Calabresi and Melamed,2 which originates from an article published in 1972, and which is viewed as having provided new and important conceptual, descriptive and normative contributions to legal theory.3 This framework categorises different elements of law and provides an assessment of the conditions under which each category would see improvement. In the first step, the relevant aspects of categorisation to answer the posed questions are explained (Sect. 1.). Then the categories are modified where necessary to analyse patent law, the most relevant German and European provisions are identified and the balance between exclusivity and non-exclusivity is explained both as it stands in current law and as it will stand in the regulation on the European Patent with Unitary Effect (Sect. 2.). This paragraph ends with some concluding remarks (Sect. 3.). The findings outlined in this paragraph form the basis for the following analysis.

1. Calabresi and Melamed’s Categorisation and Critique

a. Calabresi and Melamed’s Categorisation

Calabresi and Melamed’s concept, which was first applied in the field of nuisance disputes and tort law,4 divided into two the issues that must be faced by any legal system. The first issue is the problem of entitlement and the second issue is the problem of enforcement, that is, how to enforce while maintaining the initial choice of entitling someone.5 Both choices are necessary if a situation of “might makes right” is to be avoided.6

Regarding the first step, Calabresi and Melamed stated that “whenever a state is presented with conflicting interests, it must decide which side to favour […] hence, the fundamental thing that law does is to decide which of the conflicting parties will be entitled to prevail.”7 Therefore, an entitlement could be defined as a choice made by the state—mostly by the legislature and sometimes by a court or another instance—to judge someone’s interests higher than those of someone else.8 The question of enforcement and how to protect the initial entitlement follows the decision on which side should prevail and is the second step. The legislature or a court is allowed to choose from three possibilities: property rules, liability rules and inalienability.9 Categorisation should be useful since it reveals some of the reasons that lead to the protection of certain entitlements in certain ways.10 It is important to note that most entitlements to most goods are mixed,11 i.e., one single entitlement can be protected by property rules under certain conditions and by liability rules in a different situation.

An entitlement is protected by “a property rule to the extent that someone who wishes to remove the entitlement from its holder must buy it from him in a voluntary transaction in which the value of the entitlement is agreed upon by the seller.”12 Thus, what is indicated and commonly understood by this definition, protection by a property rule, should have the consequence that an exclusive right is created,13 by virtue of which the entitlement owner is allowed to exclude others from using it. Setting a property rule has the consequence that the state does not decide its value and the level of state intervention is therefore, in principle, the lowest possible.14

An entitlement shall be protected by a liability rule “whenever someone may destroy the initial entitlement if he is willing to pay an objectively determined value for it”.15 Thus, protection by a liability rule means here that a non-exclusive right is created. The terms liability rule and non-exclusive right are used synonymously because within this framework the person whose interests are judged higher prevails, and whenever an entitlement exists in a certain constellation it must have positive consequences for its owner. In other words, whenever that person is not allowed to demand something, he cannot have an entitlement, i.e., a right.

The major difference between a liability rule and a property rule is that someone may destroy the initial entitlement, but only if he is willing to pay a certain amount of compensation.16 The value needs not necessarily be determined objectively; but it will be determined objectively if the parties do not agree, for example in a settlement. That the value can be determined objectively has two important consequences. Firstly, the value can be, but need not essentially be, the same as what the owner of the entitlement would have sold it for under property rule protection.17 This value can be a hypothetical market price or can be fixed at average damages or some other (non-punitive) level.18 Secondly, in principle, there is an additional stage of state intervention since an organ of the state (normally) determines the value.

Inalienable entitlements describe a situation in which the entitlement may not be separated from its owner.19 In other words, the owner is not allowed to sell it and a third party is not allowed to buy or destroy it. Since this possibility is almost irrelevant in the field of patent law, it is not discussed here.20 As a consequence, in a constellation of two parties or groups, it can basically be concluded that there is a four-rule framework in which the legislature can entitle one party or the other and protect the entitlement under property or liability rules.21

b. Critique Regarding the Categorisation

Whereas Calabresi and Melamed limited the categories of enforcing an entitlement to three, or to a four-rule framework as explained above, several other commentators have argued that there must generally be more. The following contributions can be considered as the most important:

  • Krier and Schwab added one additional rule to the original four-rule framework, which may be called “rule 5”.22 The rule focuses on a situation in which a polluter causes a nuisance to a resident; the former may stop at his discretion and demand damages from the latter, which are determined objectively, or continue to pollute and receive nothing.23 This rule is identical to what was discussed as a “liability rule with put options”,24 or a “reverse liability rule”.25 Whereas liability rules were first defined as an option to take an entitlement from the owner with the obligation to pay an objectively determined remuneration, this variant describes an option to give an entitlement away and to receive objectively determined compensation.

  • Levmore focused on the Calabresi and Melamed contributions as a basis for his own work and derived 16 rules from the original three. He especially suggested that a differentiation could be made, inter alia, as to whether a party must only pay compensation for the past or the future, or whether they acted wilfully or negligently.26

  • Bell and Parchomovsky viewed Calabresi and Melamed’s model as static and added a dimension that could be seen as dynamic.27 They identified so-called pliability rules, i.e., rules that consist of a combination of property and liability rules whose identity as such becomes apparent at a certain point in time or under certain conditions, but could be different at another point in time or under other conditions.28 These pliability rules should be viewed as ontologically distinct from pure property rules or liability rules because each contains within itself its own conditions of change, and it would be wrong to look at them at only one point in time and therefore neglect the dynamism.29

c. Evaluation

What all these proposals have in common is that a legal and economic analysis could be more detailed and suggest more possibilities of how to protect an entitlement than does Calabresi and Melamed’s basic framework. However, the new proposals make an analysis more complex. They could also differentiate between certain categories in an unhelpful manner if the differences are only marginal. It is also arguable that some rules are only “created” for academic reasons and that their application would not make sense in reality. This could, for example, be true for reverse liability rules because if a potential polluter anticipated the application of this rule, he would have an incentive to pollute even more.30 At the least, only a very narrow area would remain in which these rules could be applied.31 It is also often ignored that Calabresi and Melamed anticipated this debate, saying that most entitlements to most goods are mixed32 and that the model should be kept as simple as possible; therefore, a more detailed distinction was omitted.33 The strength of this basic model lies in its simplicity, and everything that was proposed afterwards can be reduced back to the three rules: property rules, liability rules and inalienability. On the other hand, the reality of and the options for protecting an entitlement are admittedly more complex than proposed in the simple model, and it might not make sense to simplify where the case itself is not simple. This work suggests that the middle way is optimal when analysing a concrete area of law: In principle, one should stick to the simple model as long as it provides as much insight as a more complex one. However, the model must be extended if certain options would otherwise be ignored or if different cases would be put in one category although they vary distinctly. Additionally and most importantly, an extension is useful if it reveals new insights for legal and economic analysis. Of course, its usefulness depends on the question of which field of law or even which part of the field is analysed; providing a unified method would neglect several peculiarities.

2. Modification of the Categorisation and Identification of Its Elements

The advantages of taking the Calabresi and Melamed concept as a starting point in this analysis are the following: Firstly, their model seems to fit perfectly to answer the three posed research questions. The question of entitlement is equivalent to the issue of a right coming into existence; enforcement of the entitlement is about the exercise of the right and the terms exclusive right and non-exclusive right can be used synonymously to entitlement protected by property rules and entitlement protected by liability rules. Secondly, with regard to its analytical structure, this framework is widely accepted and takes a very general law and economic approach which perfectly suits the categorisation and analysis of the different ways to enforce entitlements. Due to its generality and the possibility of applying it in many different areas of law, especially in the broad field of business law, the concept itself does not lead to one predetermined assessment. Thirdly, a lot of articles have been published that deal with the differentiation between property rules and liability rules. These insights—as far as they are applicable to the field of patent law—can also be used here. Fourthly, the concept has only recently attracted attention in the patent law area34 and was never extensively used to analyse the German patent law system (including the German and European antitrust laws) and should, therefore, provide new insights, help to understand the whole system and provide a new way of comparing the German system with others. It might also help to identify areas in which further research, especially from the economic perspective, might be possible and necessary.

Of course, the concept may also have a downside—as does every framework and model. For example, it could be argued that the classification therein might be overly simplified and strict.35 This reservation is possibly true, but should not lead to a rejection of the whole concept, which has not been debated yet. Additionally, the concept does not fit perfectly, as will be shown below, for use in the field of IP law, particularly for patent law. However, it can be modified and concretised to take the peculiarities of this field of law into account. After the concept has been altered where necessary, it should fit perfectly to analyse and improve the patent law system.

a. The Question of Entitlement and Emergence of a Right

Whereas there is no need to modify the definition of the first step, in which the legislature’s decision to favour one side leads to an emergence of a right and differentiation from the second step, it is important to note—especially in the context of patent law—that there is a possibility not to grant an entitlement at all36 or to restrict the initial entitlement. Why and where this is the case is shown in the following discussion.

In patent law, essentially all national legislatures made a decision—which is widely accepted37—to entitle inventors to apply for and receive a patent for an invention, but under certain conditions and for a certain period of time.38 According to Section 1(1) Patent Act [the European Patent with Unitary Effect will be granted under the provisions of the EPC (Article 2(c) Regulation (EU) No. 1257/2012) and Article 52 EPC corresponds to Section 1(1) Patent Act], “patents shall be granted for inventions that are novel, involve an inventive step and are susceptible of industrial application”.39 But excluded from patentability are, for example, “the human body” [Section 1a(1)40] and situations in which the “commercial exploitation” of an invention would be “contrary to public order or morality” [Section 2(1)41].42 In contrast to the US patent system,43 for example, business methods are not patentable in Germany (due to the restriction on the technical area44), whereas these methods are the subject of debate in Europe with respect to Article 52 EPC.45 Because there is no answer to the question whether a patent owner can exclude others from using his invention, all these provisions undoubtedly determine the entitlement, as well as its scope. From these examples it is clear that only under certain conditions does a patent right emerge and that the legislature decides here on conferring an entitlement or not.

Whether the patent duration46—which, according to German and European law shall be a maximum of “20 years, beginning on the day following the filing of the application for the invention” [Section 16(1) Patent Act47]48—is subject to the problem of entitlements or to the protection of them is more complicated. It may be argued that this is an issue of the second step, because after the patent expires anyone is allowed to use the formerly patented invention upon payment of zero compensation to the inventor; this could be called a “zero order liability rule”.49 But this construction seems very artificial and unnecessarily complicated. After expiration of the patent there is no change in the way the entitlement is enforced. The entitlement is, rather, removed from its owner. The legislature decided that the interests of competitors and the general public outweigh the interest of the inventor. However, it could be argued that this must consequently be a special case in which the entitlement switches to the general public and is not enforced. But this interpretation would be as complicated as a zero order liability rule. Therefore, it would be most convincing to argue that no one is entitled after the patent expires, although different interests exist, because an entitlement is useless without enforcement. This interpretation would also be closest to the system of law, because the legislature does not positively determine what happens after expiration of the patent. It should also be noted that a situation with no entitlement can even be effected by the entitlement owner himself: a patent owner may decide to abandon the patent. A situation with no entitlement is, by the way, basically a solution in which no transaction of a right is necessary and therefore a solution without transaction costs, i.e., “costs incurred during the process of buying or selling (on top of the price of what is changing hands)”.50

The classification between entitlement and enforcement becomes more difficult with respect to such provisions as Secs. 11–12 Patent Act (cf. Article 27 Agreement on a Unified Patent Court). Section 11 determines that “the effect of a patent shall not extend to” certain acts, for example to “acts done privately and for non-commercial purposes” (No. 1).51 Section 12 determines the “limitation of the effect of patents vis-à-vis users”, those “who had already begun to use the invention” at the time the application was filed.52 These provisions could be classified as part of the mechanism of how entitlements are protected and could therefore be characterised as a limitation to the exclusivity of a patent,53 because they allow someone to use the patented invention. This is obviously different to the situation in which the patent expires, because the entitlement, i.e., the patent, still exists. Nevertheless, it is important to note that these provisions limit the patent in such a way that the patent owner is not entitled to exclude others or claim compensation in these particular constellations; they should therefore be viewed like issues regarding the scope of the patent.54 Additionally, a distinction must be drawn between the invention and the patent. The latter is the legal entitlement, whereas the former is only a fact to which the law refers. Therefore, in a case such as that described here, it is true that the patented invention is used, but the patent per se—as an artificial legal construct—is not affected. Thus it becomes apparent that this is a restriction to the entitlement and that there should be no difference made regarding the categorisation between Section 11 et seq. Patent Act and Section 16 Patent Act which determines the patent duration.

To sum up, all questions about patent duration, patent scope and what should be patentable are questions about the entitlement and the emergence of a right. Additionally, all these examples show that a strict classification is possible between the first and second step in analysing a legal system according to the framework that is used here. However, the decision between a limitation of an entitlement and an entitlement that is enforced as a non-exclusive right, where compensation has to be paid, i.e., with liability rules, may be very close and intertwined.

b. Enforcement of Entitlements by Property Rules

The definition of a property rule, considering that a voluntary transaction is necessary to remove the entitlement, is also useful in the field of patent law. Property consisting of tangible goods, as recognised in almost all democratic states, is, in principle, an entitlement protected by a property rule for two reasons: first, because the right owner is normally allowed to exclude others from using his goods55 and, second, because a voluntary transaction is necessary to transfer a property right. Also, patents are, basically, entitlements that are enforced by property rules. According to Section 9 Patent Act (cf. Article 25 Agreement on a Unified Patent Court), “a patent shall have the effect that the patentee alone shall be authorised to use the patented invention” and an agreement between the patent owner and the buyer and the licensee is necessary for a transfer of the patent or for a licence.56 These examples also illustrate that the terms property right and property rule are not identical: whereas the property right is a name for an entitlement, property rules are one way of protecting an entitlement.

Preliminary and permanent injunctions are important elements of a property rule system to make sure that the holder of an entitlement is able to exclude others from using it. An example of injunctive relief in the field of property law is Section 1004(1) German Civil Code (BGB), in which “the owner may seek a prohibitory injunction … if the ownership is interfered with by removal or retention of possession and further interferences are to be feared”.57 The Patent Act contains a similar provision in Section 139(1): “The party, whose patent is infringed, may, in case of a risk of recurrence, assert a claim for injunctive relief against any person who exploits a patented invention.”58 This means that the patent owner or an exclusive licensee may file for injunctive relief in any case in which the patent has been infringed. It is of no importance whether the infringer acted wilfully or negligently.59 The risk of recurrence is generally assumed if an infringement has already taken place; the assumption can be refuted with a cease and desist order or a final judgement granting injunctive relief.60 As a consequence, basically the only condition for successfully filing for injunctive relief is an infringement.

An injunction is enforced in such a way that if the defendant infringes again the court can determine an administrative fine up to a maximum amount of €250,000 or the defendant can be jailed for a maximum of 2 years [Section 890(1) German Code of Civil Procedure (Zivilprozessordnung)]. In principle, administrative fines take precedence over imprisonment.61 Therefore a competitor who only takes civil law into account may decide to infringe and risk having to pay damages and an administrative fine—although this is rather theoretical, because the described measures are very effective and are frequently used successfully. This example illustrates that it might be difficult to have strict exclusivity without criminal sanctions.62 Such criminal sanctions are also part of the Patent Act. In Section 142 Patent Act,63 the legislature determined that wilful infringement of a patent is an offence. Notably, however, Section 142 currently has very little relevance64; one reason for this might be that wilfulness is difficult to prove because of the mass of existing patents.

As a result, it can be concluded that patents are basically exclusive rights under German law, to which the regulation on the European Patent with Unitary Effect is effectively identical; therefore, injunctions should be granted and possibly enforced in basically every case of infringement if the patent owner or exclusive licensee applies for it. Such practice is actually the rule in German courts—theoretical exceptions are, at the moment, the only exceptions and are mentioned below.

c. Enforcement of Entitlements by Liability Rules

The term liability rule can also be used in the field of patent law. It is argued here, however, that a concretisation and sub-categorisation would be useful. Why and how this should be the case is explained in the following section.

(1) Concretisation of the Definition Under Patent Law

Other authors use the term liability rule in the field of patent law and—although considering the differences between IP and tangible property rights—do not explicitly alter the definition provided by Calabresi and Melamed.65 This study argues that the definition above does not fit perfectly in the context of IP and patent law. Under property law, the owner of an entitlement that is protected by a liability rule may lose the whole entitlement or, at least, the unrestricted possibility to use the entitlement. By contrast, the situation under patent law is more complex because the entitlement is an intangible right which protects an invention that can be used by more than one person at the same time. Therefore, the patent owner (if he has not granted an exclusive license) is normally still allowed to use the patented invention even if someone else uses it and pays a certain amount of compensation. Additionally, implementing a system that leads to an automatic transfer of a patent from one private party to another—even only under certain conditions—has never been discussed and would make no sense as this would not convert the patent from an exclusive to a non-exclusive right. Moreover, a system could exist in which the state buys the patent and places it in the public domain; such a purchase would represent a liability system that would remove and not simply transfer the initial entitlement.66

Thus, protection by a liability rule in the field of patent law and the existence of a non-exclusive right normally mean that others can use the patented invention if they are willing to pay a remuneration. However, the patent itself, i.e., in this context the right to be compensated, should still only be transferred to another person if the patentee and seller are in agreement. Therefore, a liability rule could be defined as a legal instrument that protects an entitlement if the patented invention can be used by any user (without consent of the owner of the entitlement) as long as that user is willing to pay an objectively determined value for it. Removal of the entitlement may take place, but is rather an exceptional case. Transfer of the initial entitlement does not take place unless the patent owner and seller agree upon the value in a voluntary transaction.

(2) Sub-categorisation

This study makes the following argument: Three sub-categories that take into account the particularities of IP should be added to the category of entitlements protected by liability rules. These new sub-categories should prove very helpful in promoting both understanding and insights into the possibilities of exercising patent rights.

Up to this point, the differentiation between liability and property rules seems to be sufficient when analysing whether patents should be exclusive or non-exclusive rights. However, concluding the categorisation at this point would mean neglecting the fact that there are three variants that differ significantly from one another. Until now, no clear sub-categorisation has been proposed, especially not into three categories. Other authors have only noted, without providing clear definitions, that certain liability rule regimes are based on legal measures and some on a combination of legal and private measures,67 with the result that there may be private liability rules into which the patent owner may contract.68 The three proposed categories are the following:

  • Firstly, the legislature may decide that an entitlement should be protected by default by a property rule and that only under certain conditions a person or instance other than the patent owner may decide that a liability rule applies (compulsory liability rule).

  • Secondly, under certain conditions or even in general, the legislature may let the patent owner decide if he prefers protection by a property or a liability rule. In these cases, the legislature normally protects an entitlement by a property rule by default, but allows, or at least does not prohibit, the owner of the entitlement to switch to protection by a liability rule. If the patent owner makes that choice, the liability rule should be termed a private liability rule. The legislature may set different incentives to choose one or the other rule. The legislature may even allow or prohibit switching back to a property rule. It must also be noted that this definition is broader than what was defined before as a private liability rule because it would not only cover arrangements that the patent owner contracts into them, but also those where a contract is not necessary.

  • Thirdly, a liability rule may apply by default, i.e., independently of the acts of the parties (liability rule by default) which could but need not take place directly after the patent grant. For example, application of a liability rule may also depend on the age of the patent. It is important to note that the last mechanism could be combined with mechanisms that are inverse to those in the second group, i.e., the patent owner may be allowed to switch to a property rule.69 Therefore, one could basically also categorise certain property rules as private property rules if a liability rule applies by default. This is, however, not the case here because there is only a very narrow area, if at all, where liability rules by default could be useful or beneficial.70

Grouping certain legal constructions in these sub-categories and not only as liability rules has several reasons. Firstly, all these subtypes have in common that they all limit exclusivity, but they differ in how liability rules are applied: in the first case the legislature in addition to a third party, an administrative body and/or a court is the driving force; in the second, where the legislature may not even be involved, the patent owner is predominantly involved; and in the third, solely the legislature is involved. Secondly, a liability rule by default must be separated from a compulsory liability rule because, as explained, such a mechanism could theoretically be combined with a mechanism that is inverse to private liability rules. Thirdly, as will be shown below, different constructions exist, or are conceivable, in every category. Fourthly, the legislature has to consider different incentives, either deciding itself or letting a third party or even the patent owner determine whether there should be a liability rule. Last but not least, consideration of these sub-categories should lead to a legal and economic analysis that not only focuses on results but also on the options of the legislature; it could also facilitate an analysis of the behaviour of the relevant parties.

(3) Application of the Modified Definition and the Sub-categorisation

(a) Compulsory Liability Rule

In Germany, provisions that constitute the basis for the application of a compulsory liability rule can be found in patent law, but also in antitrust law. These two areas of law are viewed nowadays as complementary and not as opposites, because both aim at promoting competition and welfare, as well as an efficiency-enhancing allocation of resources.71 In patent law there are two very similar legal institutions: the system that is commonly called compulsory licensing (Section 24 Patent Act) and what may be called a limitation of the effect of a patent in the interest of public welfare and national security (Section 13 Patent Act).72 The main difference between them is that the former involves a private party and the latter involves the state that wants to use the patented invention or wants it to be used by some other private party.73 In antitrust law, Secs. 19–20 of the German Act against Restraints of Competition (GWB)74 and Article 102 TFEU (ex-Article 82 EC) must be taken into consideration.75 All liability rule regimes, which follow directly from these provisions or have been derived from them by the courts, function in such a way that, if the legal conditions are fulfilled, there may be several conditions, such as the application for a compulsory licence. Protection by property rules is partially replaced by protection by liability rules. That is, the patent owner cannot successfully request an injunction or claim for damages, but for a reasonable remuneration paid by the beneficiary or the state. The remuneration is normally determined by a court and, therefore, objectively determined. This is the case, for instance, for a compulsory licence according to Section 24(6), fourth sentence, Patent Act.76 It is also true, at least when a dispute arises, in the case of Section 13(3) Patent Act.77

It is sufficient only to mention these provisions at this point because all mechanisms, including those at the European level, currently have very little importance in practice and there are no initiatives by the legislature, courts or the administration to radically change that. This lack of legal action is somewhat surprising because in Germany there is constant academic debate about them. The practice in Germany and at the European level is extreme compared to other legislations; application of compulsory liability rules also has relatively low relevance internationally, but at least some.78 Regarding Section 24 Patent Act, from 1950 to 1979 there were 37 applications for compulsory licensing but not one grant.79 Similarly, from 1980 to 2003 there were eight applications before the German Federal Patent Court (BPatG)80 but only one grant in 1991,81 which was annulled in 1995 by the German Federal Supreme Court (BGH).82 83 From 2004 to 2011, neither the Federal Patent Court nor the Federal Supreme Court decided a single case concerning Section 24 Patent Act. Since 1945 there has been no case in which Section 13 Patent Act was relevant.84 Regarding German antitrust law, there has not been a direct grant of a compulsory licence. Additionally, one cannot find any cases in which the Federal Supreme Court denied the grant of an injunction based on antitrust or any other legal construction, even though the conditions laid down in Section 139(1) Patent Act were fulfilled.85 Compulsory liability rules derived from Article 102 TFEU have been applied only with respect to copyright,86 but not patent law. These figures all show that they have very low relevance; even the number of court proceedings in such cases seems to have decreased. Perhaps the existing provisions have such a psychological effect that patent owners grant licences to avoid the application of these provisions.87 Or perhaps patent owners especially try to prevent court proceedings regarding the grant of compulsory licences because they prefer to negotiate the terms of a licensing agreement themselves and do not wish to disclose internal matters such as research and development costs.88 Such conjectures seem erroneous, however, because normally where the possibility for a dispute exists, disputes will take place.89 It is more likely that compulsory licensing is considered to have no place and, therefore, no practical importance in industrialised countries.90 Moreover, as long as no compulsory liability rule has been applied, the threat will remain relatively low. In some cases, patent owners may conclude anticompetitive settlement agreements in which they simply try to circumvent the application of compulsory licensing provisions inside the patent law system.

The regulation on the European Patent with Unitary Effect itself does not contain provisions on the application of compulsory liability rules. It is solely determined in Article 15 Regulation (EU) No. 1257/2012 that the regulation should be without prejudice regarding the application of competition law and the law relating to unfair competition. Details about the German provisions and more specific considerations are explained in Chap. 4.

(b) Private Liability Rule Regimes

Theoretically, several different forms of private liability rules can be contemplated. There are, however, only a few major mechanisms either already used or, at least in this study, considered as potentially suitable for broad application.91 The following section presents an idea, at least briefly, of the two main existing constructions involved when property rule protection is replaced by liability rule protection in which the patent owner is the driving force. These represent the two options available to declare the willingness to license, which is also called licence of right, and the possibility to form or join certain forms of patent pools. A construction named a royalty collection clearinghouse will be explained shortly, and the question will be answered of whether a fair, reasonable and non-discriminatory (FRAND) declaration leads to protection by a private liability rule. Details about the major private liability rule regimes are not given here, but rather in Chap. 3, because they are not relevant to answer the first research question.

i. Option to Declare the Willingness to License (Licence of Right)

According to Section 23 Patent Act92 “the patent owner” (or the patent applicant) can “declare to the Patent Office […] that he is prepared to allow anyone to use the invention in return for reasonable remuneration”. As a consequence “the annual patent fees […] shall be reduced by half” (Subsection 1, first sentence) and any person who is willing to pay a reasonable remuneration “shall be entitled to use the invention” (Subsection 3, fourth sentence). In principle, the patent owner loses his right to exclude others from using his invention and to grant exclusive licences [Section 30(4)]. If the patent owner and the user of the invention do not agree upon the amount of remuneration, each party is allowed to request that the Patent Division of the DPMA assesses the standoff [Section 23(4), first sentence]. “The willingness to license may be withdrawn […] as long as no intent to use the invention has been notified to the patentee” [Section 23(7), first sentence].93

In principle, after the grant, the patent is protected by a property rule as described above,94 but if the patent owner declares his willingness to license, the entitlement is protected by a liability rule, since whoever is willing to pay a remuneration, which can be determined objectively, is allowed to use the invention. Therefore, it is the patent owner’s decision whether the patent is protected by a property rule, which was the initial choice of the legislature, or by a liability rule.95 The patent owner is allowed to switch back to a property rule as long as no one indicates his intention to use the invention. Details about this mechanism are explained below.96

ii. Possibility to Form or Join Patent Pools

Also, the possibility to form or join patent pools with a certain legal structure should be grouped into this category. A legal definition does not exist for the term patent pool, either in the Patent Act or in the regulation on the European Patent with Unitary Effect. Moreover, one must be aware that there may be many different forms of patent pools. Commonly, a patent pool 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”.97 It is debatable whether single cross-licensing agreements can already be defined as a patent pool. This shall not be discussed here, however, because a construction in which only cross-licensing agreements98 are concluded, and in which the involved parties agree upon the value of the patents before allowing their use, in fact constitutes protection by a property rule because individual bargaining takes place. In contrast, the structure of some larger patent pools may be characterised as liability rule protection.99 An exact categorisation of when a liability rule applies is difficult and would most likely not lead to a better economic analysis. However, in general, a patent pool with liability rule mechanisms should have the following structure: “[M]ultiple patent holders assign or license their individual rights to a central entity, which in turn exploits the collective rights by licensing, manufacturing, or both”100 and, most importantly, which “regularises the valuation of individual patents, making […] a division of royalties according to the value attributed by the parties to their respective claims”.101 One may also call the mechanism after the pool agreement was concluded an “internal liability rule”102 that applies between the patent owners and the pool. It is possible that something close to a liability rule also applies between the pool and the licensees. This is the case if the pool offers standard licensing terms to (potential) licensees—which usually happens103—and commits itself to accepting every licensee.104 The value of the patent and, therefore, the amount of the royalty fee may be determined, for instance, by a pool committee105 or the members of the pool.

In principle, depending on the actual circumstances and legal constraints, any patent owner has the possibility to form or join such a patent pool. However, if he is free to join and makes use of this option, he is no longer allowed to exclude specific other parties, especially not other pool members or maybe—depending on the pool agreement and also on competition law constraints—other third party licensees from using his patented invention.

Normally, under a liability rule as defined above, there is a state organ that can determine the amount of compensation objectively, but a pool committee may also act very similarly. A patent pool committee is supposedly appointed by the pool members and acts, in principle, in the interest of its members when it grants a licence to a third party and distributes the licensing revenue among its members. Such an arrangement would almost always involve extensive negotiations.106 However, it should result in the amount of remuneration being determined objectively—especially in pools with many members—and offering a simple, coherent menu of prices and other terms to licensees,107 thus replacing individual bargaining.108 This result can also ensue if patent owners manage the pool, as long as objectivity is ensured. Of course, the degree of objectivity may also depend on legal requirements and the specific circumstances. Patent pools with liability rule mechanisms will be explained below in more detail.109

iii. Royalty Collection Clearinghouses

Clearinghouses may be defined as a “mechanism by which providers and users of goods, services and/or information are matched”.110 At present, basically three types of clearinghouses can be distinguished, depending on their extent of involvement in the licensing process of IP rights.111 This threefold distinction also makes sense here because this work deals with liability rule protection. The relevant types are “information clearinghouses”, “technology-exchange clearinghouses” and “royalty collection clearinghouses”. Apart from that, the “open source clearinghouse” is another model, which does not, however, involve patented technology112 and will therefore not be taken into account here. The first variant describes “a mechanism for exchanging technical information and/or information that is related to the IP status of that information”.113 The second type, which “is inspired by the internet based business-to-business […] model”, “provides an information service that lists the available technologies to allow technology owners and/or buyers to initiate negotiations for a licence and possibly also comprehensive mediating and managing services”.114 The final variant “would comprise major aspects of the technology-exchange scheme”, but “would cash” additionally “licence fees from users on behalf of the patent holder in return for the use of certain technologies or services”. Under this scheme, “the patent holder would be reimbursed by the clearinghouse pursuant to set allocation formula”.115 The patent owners license their rights to the clearinghouse, which then issues sublicences.116

Only the last variant can be categorised as a private liability rule mechanism because only in that group the patent owner deliberately gives up his exclusive right. The clearinghouse then objectively determines the amount of remuneration according to set allocation formulas. Therefore, a liability rule applies between patent owner and clearinghouse. This could possibly also be the case between the clearinghouse and licensees, depending on the legal regulation. Although there are currently no active royalty collection clearinghouses, the concept will be discussed below in more detail.117

iv. Fair, Reasonable and Non-discriminatory Declaration?

A FRAND declaration means that the patent owner declares that he is willing to license his right under fair, reasonable and non-discriminatory terms. The patent owner may issue such a declaration to a standard-setting organisation, such as the European Telecommunications Standards Institute (ETSI), which then ensures that the owner does not misuse his position after his patented invention becomes part of a standard.118 It is doubtful whether a patent is protected by a private liability rule after such a declaration. This is arguably the case because the patent owner is obliged to license his patent on reasonable terms, meaning that some objective considerations may become relevant and his right to apply successfully for an injunction may also be limited. It is argued here, however, that these rights cannot be grouped into the category of patents protected by a private liability rule because there is an important difference: the patent owner is only obliged to license his patent; no one is basically allowed to use the patented invention without the patent owner’s consent or the issuance of a compulsory licence. Additionally, the FRAND declaration is of no importance to the buyer of the patent if only the seller issued the declaration; therefore the declaration only affects the patent owner, but not the patent itself.119 This situation therefore differs to that in which the patent owner declared his willingness to license and joined a patent pool or a royalty collection clearinghouse as described above, because the declaration or licensing affects the patent itself and not only the patent holder [cf. Secs. 23 and 15(3) Patent Act, according to which the assignment of rights or the grant of a licence shall not affect licences previously granted to other parties]. As a consequence, patents for which a FRAND declaration exists remain protected by a property rule. A patent pool with a liability rule mechanism may be the next step after the patent owners have issued a FRAND declaration120 and may, therefore, result in private liability protection. Alternatively, a property rule may be transformed into a liability rule after the FRAND declaration by a compulsory liability rule, possibly under certain additional circumstances. This example shows, however, that the situation in which a patent is protected by a property rule, but in which the patent owner is willing to license or actually licenses his patent, may (at least regarding the economic consequences) be very similar to a situation in which the patent is protected by a private liability rule.

(c) Liability Rule by Default

German law and the regulation on the European Patent with Unitary Effect basically do not contain any elements in which there is a liability rule by default or which are of importance to the analysis here. Nevertheless, two such constructions do exist. Firstly, compensation for filed applications according to Section 33 Patent Act121 fits into this category.122 Under this provision, the patent “applicant can demand from that person who has used the subject matter of the application, although he knew or should have known that the invention used by him was the subject matter of the application, compensation appropriate to the circumstances.”123 The use of an invention that could be patented is not unlawful124 and compensation is determined by a court if a dispute arises.125 Both elements are typical of a liability rule regime, which applies by default.

Secondly, a mechanism whereby the patent is protected by liability rules by default, but the right holder is allowed to switch to property rule protection, is also part of German patent law (but should no longer have relevance because all these patents should now have expired). This mechanism is, therefore, a combination of a liability rule by default with an inverse licence of right regime. It refers, however, to a very special context. After German reunification in 1990, when the former German Democratic Republic joined the German Federal Republic,126 the patents granted by the DPMA and those granted by the eastern German patent office (Amt für Erfindungs- und Patentwesen) were ruled to coexist and especially to extend to the entire federal territory according to Article 8 and the appendix of the Unification Treaty. Besides several other problems caused by this solution,127 the German Federal Republic had to decide how to deal with two different patent categories in the GDR: Ausschließungspatente and Wirtschaftspatente. Whereas the former were exclusive rights comparable with patents in the German Federal Republic, the latter were non-exclusive patents on inventions that were made, for example, in a socialist firm and for which the inventor was only entitled to receive remuneration.128 The German Federal Republic decided that Wirtschaftspatente should be transferred into patents for which the willingness to license is declared [Section 7(1) German Extension Code (Erstreckungsgesetz)]; however, the patent owner was allowed to revoke the willingness to license (Subsection 2) and could therefore choose between protection by a liability or property rule.

3. Results

The most important points in this paragraph can be summarised as follows. One can distinguish two steps when analysing law. In the first step, the legislature determines the emergence of a right. In the context of patent law, this means, inter alia, that it decides whether there should be a patent or not, how long it should exist and what scope it should have. In the second step, it is determined how patent rights should be exercised, especially whether the entitlement should be protected by property or liability rules or, in other words, whether there should be an exclusive or non-exclusive right. There are basically three different possibilities how a non-exclusive right can be designed. The first is by default, since the legislature decides on the application of liability rules (liability rule by default; but there may be an option for the patent owner to switch to a property rule). The second and third possibilities are situations in which protection by a property rule is replaced by a liability rule. With the former, not only the legislature but also a person other than the patent owner is the driving force (compulsory liability rule). With the latter, the patent owner himself can make a decision (private liability rule). The results are illustrated in Fig. 2.1, where one can also see which actors can basically be the driving forces of the decision.129


Fig. 2.1
Entitlement, property and liability rules in patent law

It is again important to note that the state need not choose one option for the whole patent law system; rather it can decide which option should apply under what conditions. The options of how to implement one of the mentioned rules are also important and theoretically infinite. Although all these decisions are possibly intertwined, the focus lies in the following discussion of the second step.

While leaving out the possibility of applying private liability rules, German and European regulators clearly prefer, as the analysis above shows, a patent system with property rules. Liability rules never apply by default where this is of major relevance and, at least in current practice, basically also never apply as compulsory liability rules.

B. Property Rules, Liability Rules and Economic Efficiency

In the following discussion, this study uses the analytical framework in the first paragraph to explain theoretically what reaching economic efficiency means for the choice between property and liability rules. The assumption has been made that an economically efficient system must basically only contain property rules because German and European legislators, as well as the bodies that apply the law, prefer such a system, as shown above. To show that this is not necessarily the case is the aim of this paragraph.

The term economic efficiency generally means—however it is defined exactly130—that the aim is to get “the most out of the resources used”.131 Different factors, which should be minimised, maximised or internalised, but are also often interdependent, are measures of economic efficiency. These are, inter alia, the amount of transaction and enforcement costs, the deadweight loss, spillover effects and externalities, and—very importantly here—the consequences for investment in the entitlement itself, or even in other assets or entitlements.132 Transaction costs are, according to Williamson, the economic equivalent of friction in physical systems,133 which were already defined above as costs incurred during the process of buying and selling, on top of the price of what is changing hands. Typically, a distinction is made between ex ante and ex post transaction costs,134 despite the fact that they may be interdependent.135 The former are costs incurred in drafting, negotiating and safeguarding an agreement136; the latter are control and adaption costs.137 The term enforcement costs includes all costs, especially administrative costs, that result from the enforcement of a property or liability rule. A deadweight loss occurs if a market is not perfectly competitive, i.e., where buyers and sellers are all price takers and the former earn normal profits (the bare minimum profit necessary to keep them in business), where products are homogeneous and where information is perfect.138 Spillover effects and externalities may be viewed as economic side-effects and, more specifically, as costs or benefits resulting from economic activity affecting persons other than those directly engaged in a particular economic activity,139 which are therefore not captured by the price system.140 Whereas, generally speaking, the different costs and the deadweight loss should be minimised and the investment in the entitlement should be maximised, spillover effects should be internalised.

The following discussion first investigates whether there is a general approach (outside the area of patent law and independent of the characteristics of the entitlement) that could be preferable.141 This discussion should demonstrate the strengths or weaknesses of some general arguments. Some opinions have been transferred to the area of patent law without evaluation. However, this step may have been necessary, since doubtful and challengeable opinions can be found both in the field of legal research and economic research.142 Therefore, the conclusions of Calabresi and Melamed will be summarised (Sect. 1.); challenges and alterations to that approach will be discussed (Sect. 2.) and evaluated (Sect. 3.); and attention will be drawn to the patent law area (Sect. 4.).

1. The Contribution of Calabresi and Melamed

Calabresi and Melamed, who do not explicitly refer to the field of IP law but take a general approach, use Pareto’s definition of economic efficiency, which describes “a situation in which nobody can be made better off without making somebody else worse off […] if economy’s resources are being used inefficiently, it ought to be possible to change the allocation in such a way that somebody is made better off without anybody else becoming worse off.”143 Then, the authors refer to Coase who, however, did not preliminarily focus on the assignment of legal entitlements. Rather, on the problem of externalities, Coase concluded that under certain assumptions (usually termed the absence of transaction costs, which should be understood extremely broadly and contain perfect knowledge and the absence of any impediments or costs of negotiation) Pareto efficiency will occur regardless of the initial entitlement.144 Calabresi and Melamed point out, as do others who have commented on the Coase theorem, that nobody would assume that there would be no transaction costs in practice, but it can be derived from his analysis that the distribution of wealth and economic efficiency may affect a society’s choice of entitlements.145

Calabresi and Melamed mention that minimising the administrative costs of enforcement may be one reason for a particular entitlement,146 and they come to the result that property rules are preferable if transaction costs are low, if only a few parties are involved and if the parties are readily identifiable. Under these conditions, state intervention to determine the amount of compensation and achieve an efficient allocation of resources should not be necessary.147 Liability rules should, therefore, take precedent over property rules if transaction costs are high, many parties are involved and the parties are difficult to identify. In other words, a reason for employing liability rules would be that market valuation of the entitlement is deemed inefficient because of its unavailability or its expense.148 It can also be concluded that Calabresi and Melamed view private bargaining, in principle, as better than state intervention to reach an efficient outcome.

2. Challenges and Extensions to Calabresi and Melamed’s Concept

Since Calabresi and Melamed conceived this very simple differentiation to determine which rule would be preferable under what conditions, several authors have challenged and commented on this approach.149 The following analysis explains which general challenges and extensions should be considered as most important and should have—at least some—influence on the debate in the area of patent law.

a. Importance of Assessment Costs Under Liability Rule Protection

Calabresi and Melamed concluded that liability rules are better where transaction costs are high. Few scholars have questioned this approach.150 However, any proposal should also encompass, as some authors have pointed out, another consideration: when transaction costs for the involved parties are high, information costs151 for the body that would determine compensation are low and a third instance would always be able to set compensation efficiently.152 However, such an assumption is not always realistic. For example, the strategy of the parties may lead to inadequate information about the infringer’s benefits and would prevent the instance from setting compensation efficiently.153 Another constellation that could involve high information costs is if the owner of the entitlement has an idiosyncratic value or consumer surplus154 that could be difficult to discover. Krier and Schwab even suggest that there is often a positive correlation between the circumstances that lead to high or low transaction costs and those that generate high or low assessment costs.155

As a result, by rejecting the assumption that assessment costs are low when transaction costs are high, it cannot be abstractly asserted that liability rules are always better when transaction costs are high. Rather, one has to distinguish between four different constellations: (1) Where both transaction and assessment costs are high, neither rule is likely to be efficient, because bargaining will break down and an efficient assessment is not possible; (2) Where transaction costs are high and assessment costs low, liability rules are preferable; (3) Where transaction costs are low and assessment costs high, property rules are preferable; (4) Where transaction and assessment costs are low, both rules are likely to be efficient.156

b. Importance of Enforcement Costs Under Property and Liability Rule Protection

Taking only transaction and assessment costs into account, however, might not be sufficient. So far, it could be concluded that the decision to enforce entitlements either by property or liability rules should mainly depend on the extent of transaction and assessment costs. Additionally, there is an implicit assumption that property rules would entail lower enforcement costs than liability rules, since the degree of state intervention is perceived to be lower. This assumption is, however, not necessarily correct and must be weighed against the entire enforcement costs of each option.157

Under liability rules, there are not only assessment costs of determining remuneration, but there may also be additional enforcement costs,158 including institutional expenses such as for personnel and training. Even this may not be enough. A good example can be derived from German constitutional law: any time an administration decides an amount of remuneration, a court must be available159 to examine the calculation of the remuneration if one of the parties involved appeals.

On the other hand, the decision for property rules means first and foremost that the parties involved in a transaction can perform it without intervention by the state, because remuneration is determined by the market. In order to guarantee exclusivity under property rules, however, whenever infringement is involved the legislature must establish not only the necessary substantive and procedural law, but also—to prevent enforcement by the right owners themselves—the institutions, particularly courts and bailiffs, for enforcement. These institutions must be financed—possibly by the infringer—and must be able to work effectively. Moreover, criminal law provisions, such as those against damage of property160 and theft,161 with the corresponding prosecution service and courts, might be necessary or at least helpful.

One cannot abstractly say which system would be more efficient and cause less enforcement costs, including litigation costs, because in constellations with imperfect information—which is normally the case—it is hard to predict which situations will result in less litigation.162 Therefore, a clear differentiation between liability and property rules is not possible in theory, but could be easier in practice regarding a concrete case.

c. Liability Rules in Low Transaction Cost Settings?

Whereas Calabresi and Melamed state that property rules are preferable in low transaction costs settings, there are several arguments in favour of liability rules. Four important arguments (1) and also some counterarguments (2) are provided in the following analysis.

(1) Arguments in Favour of Liability Rules in Low Transaction Cost Settings

(a) More Efficient Contracting Under Liability Rules?

Firstly, several authors have argued, especially Ayres and Talley, that liability rules should not only be viewed as “market-mimicking substitutes for consensual trade, which induce non-consensual takings and uses, subsequent litigation, and judicially determined prices”.163 Rather, they are instruments that could generate more efficient contracting than property rules.164 Where an entitlement is protected by liability rules, the owner may wish to engage in bargaining if he wishes to bribe a potential taker or user not to take or use the entitlement. The owner may do the same if, contrarily, he wishes to sell or let his entitlement be used at a price that is lowered to the amount that a third instance would supposedly determine.165 The first alternative will happen if the entitlement owner values his entitlement higher than the third instance. The second alternative becomes relevant if the owner does not expect the buyer to pay the objectively determined compensation but is, nevertheless, interested in selling his entitlement. Both cases should also show that the owner has, in contrast to property rule regimes, incentives to foreclose his true valuation. Such results can reduce the degree of asymmetric information and bias towards strategic behaviour.166 However, Ayres and Talley also mention that infringers may have strategic incentives to misrepresent their true valuation under a liability rule because the owner’s behaviour depends on how he assesses the buyer’s valuation and not on what the valuation is in reality.167

(b) Lesser Chance of Failure for Efficient Bargains Under Liability Rules?

Secondly, Kaplow and Shavell have especially argued that under liability rules there may be a lesser need for bargaining and therefore a lesser chance of failure to conclude potentially efficient bargains.168 Assuming that courts do not systematically and substantially underestimate compensation awards to entitlement owners and that bargaining under liability rules is not substantially more successful than under property rules, one could conclude the following. Property and liability rules are, according to Coase, equally efficient if bargaining is perfect and all negotiations are successful. Liability rules are better than property rules if bargaining is impossible,169 because under a property rule regime inefficiencies cannot be corrected by bargaining. As a result, if liability rules are as efficient as property rules whenever bargaining is possible and are superior in situations whenever bargaining is impossible, then they must also be better in situations in which only some negotiations fail.170

(c) Lower Endowment Effects Under Liability Rules?

Thirdly, it has been argued that under liability rules the endowment effect can be expected to be lower than under property rules. This term describes—in contrast to the Coase theorem171—the fact, which has been proven in several experiments, that people value an entitlement they already own more than a similar entitlement they have an opportunity to acquire.172 That is, the willingness to pay is normally lower than the willingness to accept.173 The endowment effect may reduce the range of potential transactions that, according to Lewinsohn-Zamir and Jourdan and Rachlinski, should be negative from an efficiency point of view.174 One reason for the reduction under a liability rule regime should be, according to Ayres and Talley, that the roles of buyer and seller are obscured.175 Another argument is, according to Jourdan and Rachlinski, that the lack of power to refuse to sell a good seems to have psychological importance; an owner of an entitlement does not feel truly an owner if someone can wilfully appropriate it upon payment of a fee.176 As a consequence, an endowment effect can still be present under a liability rule, but the effect is lower than under a property rule.177

(d) Inefficient Hold-Out Situations Under Property Rules?

Fourthly, property rules may cause hold-out situations which can cause inefficiencies that would not normally occur under liability rules. Under a property rule, a hold-out problem may especially arise whenever the circumstances limit each side to a single trading partner.178 In a competitive market, i.e., where equivalents or substitutes are available, potential buyers can play one seller off against another until the price equals marginal costs, and no deadweight loss will result from the transaction. But if there is no alternative, with only one buyer and different independent sellers owning complementary goods, each seller will have an incentive to be the last one to sell and thus to capture a greater share of the surplus created by the exchange.179 Whereas this is presumably the case in which transaction costs are high and liability rules are favoured because of the costs, it could also be the case in which only one seller and several buyers exist and transaction costs are low. In such a constellation the seller can charge monopoly prices and inefficiencies normally result in a deadweight loss. Liability rules could lead to a more efficient outcome here if assessment costs are low.

(2) Arguments Against Liability Rules in Low Transaction Cost Settings

A conclusion from the arguments in favour of liability rules could be that the main concern about property rules is the emergence of strategic behaviour and the lack of efficient bargaining, which would allegedly not happen under liability rules. Admittedly, property rules might lead to strategic behaviour, but there is no assurance that this threat is as serious as described above because one can also make a case that the arguments above are based partly on questionable assumptions and conceal certain disadvantages of liability rules.

(a) Do Liability Rules Lead to Inefficient Bargaining?

One concern weakens the first argument in favour of liability rules. When there is awareness of a positive attribute with respect to economic efficiency, the owner may tend to bribe potential users or buyers if he values his entitlement higher than the expected objectively determined compensation. In such a constellation, bargaining only becomes necessary to avoid inefficiencies that would not come into play under a property rule.180 These deals are themselves potentially inefficient and cause deadweight losses since potential users or takers spend resources to find such assets and owners spend resources to protect them.181 Additionally, it is questionable whether the information-forcing effect of liability rules facilitates efficient bargaining. Kaplow and Shavell have used the same example as Ayres and Talley to show that welfare might be greater and there might be more value-increasing trades with a higher increase in welfare per trade under property rules than under liability rules.182 The outcome may, however, depend on other factors such as litigation costs.183

(b) Risk of Under-Compensation Under Liability Rules?

Regarding the second argument in favour of liability rules, the assumption that under-compensation does not occur is questionable and, as a consequence, the whole argument can be turned into a pro-property-rule statement. Courts and other third parties normally follow legal provisions to determine compensation at the point they believe is the objective market value. Therefore, there could often be a gap between the objective and idiosyncratic or subjective valuation of the entitlement.184 The reason could be, inter alia, a strong endowment effect185 and the consequence that concerns a more dynamic approach to efficiency, as well as a lower investment in the entitlement.186

The reason that courts have to determine an objective value is the difficulty, or in most cases the impossibility, of figuring out what the idiosyncratic or subjective value is, due to the expected strategic behaviour of the entitlement owner. This statement is similar to the above argument that assessment costs should be taken into account to determine an efficient remuneration and that liability rules should not be considered as the optimal choice if the costs for assessing efficient compensation are high.187 However, the difference is that the statement above is more general and static, whereas what are highlighted here are the risk of under-compensation and, therefore, the risk of (too) high assessment costs of determining the idiosyncratic value.

For clarification, it is worth mentioning at this point that under- or over-compensation could be used as a mechanism to achieve distributional goals.188

(c) Endowment Effects Lower Under Property Rules?

Concerning the third argument in favour of liability rules, it may be wrong to argue that endowment effects are more likely to emerge under property rules than under liability rules. Lewinsohn-Zamir came to the opposite conclusion: the use of liability rules produces a stronger endowment effect than the use of property rules.189 The endowment effect represents the pain of giving up one’s entitlement; therefore, it is reasonable to believe that the pain will be greater if such parting is or can be forced upon an unwilling owner than with his consent.190 Property rules should induce a frame of mind that focuses on the profits of exchange rather than on the good that is given up. By contrast, the power of coercion under liability rules should create an attitude of resistance.191 However, Korobkin’s critique was that no empirical findings support these hypotheses and that Lewinsohn-Zamir’s arguments are based on wrong assumptions.192 He concluded, while also questioning the assumption that the “stickiness” of an entitlement is always undesirable, that based on current knowledge the endowment effect will either be smaller under a liability rule regime than under a property rule regime, or the size of the endowment effect will be unaffected by the choice of how to protect an entitlement.193

(d) Are Parties Acting Fairly Under Property Rules?

The assumption may be wrong that the parties involved in a voluntary transaction under property rules will act strategically and opportunistically whenever possible, as indicated above, especially in arguments one and four. Lewinsohn-Zamir, who argued in contrast that liability rules discourage potential takers from offering owners acceptable compromises,194 referred to the literature that challenged the game-theory prediction by which all actors will behave rationally and try to maximise their profits. Because of such experiments as the Ultimatum game,195 it can be assumed that not only pure economic motives influence people’s behaviour, but also fairness and willingness to sacrifice some of their well-being in order to punish people who act unfairly.196 As a consequence, property rules may facilitate bargaining by encouraging potential buyers to act fairly and offer some of the gains from the trade to sellers.197 Additionally, the assumption that parties act strategically and opportunistically whenever possible would also imply that entitlement owners would be risk-seeking by trying to extract the whole surplus of a trade. However, since Tversky and Kahneman have shown in their so-called Prospect Theory that people are relatively risk-averse when confronting potential gains and relatively risk-seeking when confronting potential losses,198 it can be concluded that sellers will be more willing to compromise than buyers. Property rules where the roles of buyer and seller are naturally invoked should therefore lead to successful and efficient bargaining in low transaction cost settings.199

3. Interim Results and Evaluation

As an interim result, one can say first that Calabresi and Melamed’s conclusion seems to be too simple to make the right choice between liability and property rules. There are definitely more factors that are relevant than transaction costs, the number of parties involved and whether the parties are readily identifiable. Additionally, the hypothesis that liability rules would be better than property rules when market valuation is deemed inefficient on account of its unavailability or expense seems to be quite idealistic, because the market price is often one of the best indicators of efficient compensation. Likewise, taking assessment costs into account is a very important enhancement of the initial analysis, but it is not sufficient alone, since enforcement costs and other factors such as predictable strategic behaviour and endowment effects may also be relevant.

Concluding the debate whether liability or property rules would be better in low transaction cost settings is difficult because most of the arguments brought forward are based on disputable assumptions and are, therefore, relatively weak. For example, whether bribing a potential taker or user from taking or using an entitlement promotes efficiency is questionable because there could often be situations involving many potential takers and users. The owner of the entitlement could then face high costs to secure his position. Whether the endowment effect is higher and whether the parties act more strategically and opportunistically under liability or property rules definitely has relevance, but a general conclusion cannot be drawn because each case greatly depends on the concrete constellation, the circumstances and the exact implementation of the legal rules. However, two risks should be emphasised and weighed against each other: first the risk of under-compensation possibly resulting in a lower degree of investment in an entitlement with liability rules and, second, the risk of hold-out situations with property rules. Additionally, one can conclude that although transaction costs were the starting point for distinguishing between situations in which property or liability rules would be optimal, they cannot be the only decisive factor. Assessment and enforcement costs, the potential of strategic behaviour, the occurrence of endowment effects and deadweight losses caused by hold-out situations must also be taken into account. This should be true for almost all areas of law and, therefore, also for the field of IP law.

To sum up, negatively speaking, all these arguments show that it is difficult or even impossible to determine abstractly which rule is preferable under what conditions without focussing on a concrete constellation with certain legal rulings. Additionally, care must be taken in transferring the above arguments undigested to the field of patent law because they are already somewhat disputable per se. But the positive aspect is that one rule cannot generally be considered superior to the other and the whole debate shows that the main risk is generalisation. The legislature must find the right mixture of rules while considering the different risks and determinants mentioned in this paragraph.

4. The Particularities of Patent Law

The general debate about property rules and liability rules shows that there is, in theory, no clear preference for one or the other regime, and that different factors have to be weighed against each other. The following analysis focuses on the area of patent law and its particularities, and draws conclusions from the above arguments. The result should be a clearer statement on whether and under what conditions property or liability rules would be suitable for that field of law.

The speciality in the field of patent law is that efficiency has one particularly important dimension that should be emphasised here: dynamic efficiency. Whereas static efficiency relates to an economy that appropriately balances short-term concerns, i.e., optimisation of existing products, processes and capabilities, dynamic efficiency also takes long-term concerns into account and focuses especially on encouraging research and development.200 Since there is a trade-off between the two, the aim is to find the optimal balance. Society is interested in profiting from innovation, including the generation of it, as well as access to it. The conventional wisdom is that protecting entitlements by property rules in patent law would be optimal.201 This opinion may originate from the assumption that IP is a special form of property,202 although the latter must be viewed more as natural and the former as intentionally granted, man-made exclusive rights because tangible goods can normally be used best if only one person is allowed ownership, which is not the case with IP.203 As a first step, the most important arguments are stated for property rules (Sect. a.), followed by those for liability rules (Sect. B.); all are then evaluated (Sect. c.). It should again be emphasised that, for the most part, no differentiation is made at this point between certain variants of liability rule regimes. Liability rules basically only mean those for which the patent owner is not the driving force (liability rule by default and compulsory liability rule). However, the analysis does include the third category of liability rules (private liability rule), which may also play a particular role.

a. Arguments for Property Rules

(1) Risk of Under-Compensation Under Liability Rules

A major concern about liability rules is, as was already mentioned above,204 the risk of under-compensation. This concern is of special interest in the area of patent law because of the dynamic effects, i.e., incentives are very important to motivate the (potential) patent owner to invest in the entitlement, commercialise his invention and also invest in future research and development. Additionally, it is in the public interest that the inventor has an incentive to disclose his invention instead of keeping it secret. If the patent owner expects or receives too little compensation for his efforts, he will not make them or, alternatively, he will make them but will keep his invention secret. Accordingly, promoting disclosure is one of the principal aims of the patent system.205 Of course, many other factors are important for the question whether these incentives exist; whether patents are treated as exclusive rights is one consideration. Under property rule protection, the patent owner may rarely earn pure monopoly profits, but such profits can frequently be attributed to the exclusive right. The effects may be lower and less relevant in areas of complex (cumulative), rather than discrete, technologies. Likewise, the effects may be less relevant where innovations are only incremental and more important where development and product cycles are long. The key difference between complex and discrete technology is whether a new, commercialisable product or process is comprised of numerous separately patentable elements versus relatively few.206 For example, the innovation type is mainly discrete in the pharmaceutical industry, discrete and complex in the biotechnology industry, and complex in the computer hardware and semiconductor industry.207

Additionally, the assumption that the individual right holder is basically in the best position himself to negotiate and determine efficient compensation under a property rule regime208 should also be true in patent law, especially in the relationship between patentees and follow-up innovators.209 By contrast, under a liability rule regime the (expected) remuneration due to the lack of exclusivity and—not only a question of potentially overly high assessment costs—the possibility of court intervention to accurately identify damages may lead to a systematic undervaluation of innovation210 and, therefore, to the inefficiency of low incentives.

(2) Risk of Under-Use Under Liability Rules

Another argument against a system with liability rules is that it would not only lead to lower incentives for innovators because of the risk of under-compensation, but also to lower use of the whole system. This is true for several reasons. Complete transferability of a right and the exclusive use of that right, possibilities that exist only under property rules, are both important to ensure that commercialisation is conducted by the lowest-cost provider.211 Additionally, varying degrees of licensing, a possibility less likely under liability rules, must occur to facilitate price discrimination in order to reduce a deadweight loss.212 The potential infringements caused by liability rules would reduce the incentive to invest in inventions.213 Moreover, as already indicated above, unlike property rules, liability rules may lead to strategic behaviour,214 which should also be true in the context of patent law.215 Only property rules increase such types of competition and access as coordination among entrepreneurs, inventors and venture capitalists to facilitate the commercialisation of new ideas.216 In contrast, liability rules and acts of infringement lead to a destruction of value.

The question remains whether taking these arguments into account in a system with primarily liability rules to prolong the patent term or geographical scope actually increases incentives again.217 It could, instead, lead to a larger number and increasing fragmentation of patent rights, making it potentially more difficult to commercialise a patented invention.218 Faced with these complexities, actors may even decide not to engage in innovation.

(3) Risk of Free-Riding and Inefficient Entries Under Liability Rules

The possibility of easily accessed patented inventions could have negative effects. Liability rules could, in contrast to property rules, lead to free-riding. Companies may prefer to profit from new developments that are accessible at pre-established prices instead of taking the risk of investing in the pursuit of new, untested ideas and concepts.219 Risk-averse actors may not engage in research and development. As a consequence, liability rules especially reduce the positive effects on dynamic efficiency. Liability rules may also reduce economic efficiency, especially in the short run, by facilitating the entry of inefficient producers.220 This may, however, only be the case under specific conditions, particularly if the royalty is not determined per unit of output and with a zero fixed fee.221

(4) Right Holders in Best Position to Lower Transaction Costs

Overdue for mention is the following argument which takes account of the possibility of using private liability rules. Calabresi and Melamed contend that transaction costs are especially important for making the choice between property and liability rules.222 In the field of patents, transaction costs can be quite substantial and are basically rising with the number of patents and right holders. Therefore, liability rules instituted by the legislature may be necessary and more efficient than property rules in these cases. However, as explained in the first chapter, the patent owner himself can replace liability rule protection by property rule protection (private liability rule).223 It was argued that a patentee would have an incentive to invest in institutions with a liability rule structure that lower transaction costs if he faces complex and expensive individual bargaining costs.224 One could even argue that, in some cases, such an investment would not even be necessary, for example, where a certain institution is unnecessary or already exists. Moreover, patent owners may be perfectly knowledgeable of their situation and their possibilities. Therefore, there need not necessarily be a liability rule in the first place due to high transaction costs. It could be argued that the incentive to invest in such institutions would also exist under liability rules and that they may also have the effects described above, but they would also limit the possibilities of patent owners to decide which protection of entitlements is optimal for them.

b. Arguments for Liability Rules

Recent arguments in favour of property rules in patent law have been challenged, and various scholars have argued for more liability rule elements in patent law. Generally speaking, the argument is that exclusivity should only be dominant where and to the extent that other non-exclusive schemes cannot achieve the same or even better results and/or generate more beneficial effects for society as a whole.225 Not only liability rules come with some costs, as property rules can also have a price.226 The basic and most general concern, a concern that especially relates to dynamic efficiency, is that strict property rules involve a risk of over-compensation that would result in innovators receiving more than what is necessary to create an incentive to make and market an invention. Moreover, patents protected by property rules may lead to a deadweight loss because the owner can exclude others from using his patented invention, or in rare cases restrict output or engage in some kind of uncompetitive behaviour. These actions could also prevent the positive spill-over effects that would occur if the innovation were open to other market participants and could lead to faster and better innovation from which society would benefit. It is noteworthy that it is generally impossible to always assign an entitlement to the most efficient user in the field of IP when, as a consequence, multiple transactions are required.227

Another general approach to this issue is based on the differentiation among three market stages for patents: the innovation stage, the use stage and the production stage. Whereas exclusive rights are normally beneficial at the first stage, they are questionable at the second stage and often disadvantageous at the last stage.228 Although the differentiation may be tentative because of the interdependency of the stages,229 it shows that it is important to focus on all stages and the effects of IP rights, as they are currently exclusive rights that might not be optimal in all constellations. The major arguments for liability rules and against property rules, which are also partly interdependent, are laid out in more detail below. Most of them also take into account the effects on the last two stages.

(1) Uncertainty About Boundaries of a Patent

As indicated above, the conclusions of Calabresi and Melamed are especially appropriate if the rights are well defined.230 Patents typically involve, however, a degree of uncertainty about boundaries.231 Whereas it is easy to define the scope of a property right regarding a certain parcel of land, patents are intangible rights and the difficulty lies in attempting to define inventions in terms of words: patent claims normally require interpretation. The scope of a patent is ascertained at different stages by different decision-makers, specifically by the patentee, by the patent office and possibly by a court.232 Whereas the patentee formulates the patent specification subjectively, judges must interpret it objectively.233 Moreover, answering difficult questions of fact and interpretation of law is often required in the field of patent law.234 As a consequence, with regard to real property it is almost always possible to tailor the scope of injunctive relief to the violation of the right. In contrast, with patents it is often difficult to determine in advance whether a particular use is infringing.235 An injunction based on an unclear right will sometimes under-serve, but probably more often over-serve, the goal of protecting the legal right because courts tend to make sure that the right is fully protected.236 Moreover, it is often impossible to limit the scope of an injunction to cover only the infringement of the right, basically for two reasons. Firstly, a patent often covers only a small part of a product and it is not possible to sell it without the infringing part. Secondly, due to information failure and significant costs to acquire such information, a defendant may find it difficult to know in advance whether a particular use will or will not violate an injunction. As a result, a defendant often refrains from participating in a market.237 The problem increases with the number of right holders and patents involved.

As a result, the use of property rules to provide protection may lead to systematic over-compensation of plaintiffs and may overly deter defendants. This should not be the case under liability rules.238 Additionally, costs of private bargaining under a property rule regime are often higher than assumed (with a resulting complication of the licensing process); because of uncertainty about the boundaries of a patent and about its value, transaction costs in private bargaining are substantial.239 Costs are deemed inefficient when the acquisition of information about the scope of a right exceeds the social value of that information.240 In contrast, determination of the amount of remuneration under liability rules is not usually impossible, and there may be at least some cases in which the costs under liability rules, especially involving assessment costs, are not higher than those under property rules.241

(2) Uncertainty About Validity of a Patent

Another important factor regarding the risk of over-compensation under property rules may be the uncertainty about the validity of a patent242 in connection with the presumption of validity.243 Whereas it is relatively easy to investigate whether a property right exists regarding tangible goods (without answering the question to whom it belongs), this determination is often more complex in the area of patent law because it requires interpretation of often vague law. If the patent office grants a patent, the patent is regarded as existent, but it may later be revoked. As a consequence, patents are regarded as probabilistic rights: i.e., rights that are valid with a certain degree of probability.244 Whether it is revoked depends in principle, however, not only on the validity itself but also on several other factors, such as the motivation and funds of someone other than the patent owner and, again, on the patent office or a court. Thus, if no one opposes an invalid patent, the patent owner can extract profits from it, which would not be the case if the patent did not exist. In the case of licensing, licensees would not only conclude needless contracts, but would also waste time and money on negotiating them.245 Of course, this problem would also be relevant under a liability rule regime, but it could be mitigated because the risk of (over-) compensation would be lower than under a property rule regime. Additionally, in a situation that could lead to wilful infringement, fewer (potential) users of the invention would enter into negotiations because the penalty for infringement would potentially not be greater than the price of the licence.246

(3) Risk of Inefficient Hold-Out, Hold-Up and Anticompetitive Effects Under Property Rules

Whereas uncertainty about the scope and validity of a patent may per se lead to over-compensation of the patent owner under a property rule regime, this problem may also arise or even be made worse if the patentee acts strategically, thereby misusing his position. As explained above, protection by property rules may also involve the risk of hold-out situations,247 and actors may not always act fairly under property rules.248 Although a hold-out situation is basically always statically inefficient due to the potential of deadweight losses, it may initially be beneficial with respect to dynamic efficiency. Even if the patentee does not use the patented invention, optimal forward-looking behaviour could turn occasional cases into an advantageous situation. A liability rule could inefficiently reduce incentives for investment in research and development in such a case.249 Where these positive effects for dynamic efficiency do not exist, however, hold-out situations are not desirable. Moreover, it was shown theoretically that welfare could be increased if a liability rule is implemented instead of a property rule.250

Hold-out situations involve blocking patents; blocking patents should be determined relative to other patents and be defined as patents covering features that are essential for a certain activity or function.251 A typical example may be the following situation: given two patentees, the first owns a (blocking) patent that provides essential infrastructural input or crucial components and the second is a follow-up inventor who has made an improvement. The second invention can, however, only be commercialised if there is access to the first patented invention.252 If the first patent owner does not consent for strategic reasons, he may inefficiently hold out against the second one because society would not profit from the withheld improvement and no positive spill-over effects would occur.253 In other words, a property rule creates a setting in which any subsequent innovator is at risk of being (inefficiently) enjoined from using already patented innovations,254 a situation that is especially negative if the innovator has already invested in the innovation or in complementary assets. Consequently, not only competitors would suffer from this behaviour, but also and especially consumers and society because they may not benefit as long as the situation remains unclear. Another example in which property rules may favour inefficient hold-out situations is where the rules make it difficult or impossible to comply with de jure or de facto industrial standards. Likewise, competition may be preserved in markets with network effects or a strong need for compatibility and interoperability255 because certain patent owners—especially those who are dominant in the market and profit from these effects, for instance, because they hold the essential patents and entered the market earlier than others—may block others from using their patented inventions and reach a monopolistic position.

Of particular concern in the area of patent law are so-called hold-up situations and other anticompetitive behaviour. The term hold-up in the area of patent law foremost describes situations in which the patent owner does not want or is not able to work the patent and intends solely to demand higher licensing fees than would usually be the case. A specific hold-up situation in which the patent owner may also work his patent is known under the term patent ambush. There is no uniform definition of the term,256 but it normally describes, according to the widest definition, constellations in which an owner of several patents participates in a standard-setting procedure but does not disclose all the patents that are essential for the standard to the standard-setting organisation. After the standard is established and applied, the patent owner is thus able to hold up the users of the standard.257 In a hold-up situation, the licence fees can significantly differ from the inherent value or utility of the achievement258; therefore, it is no longer possible to rely on the market to lead to the right outcome.259 Typically, the (potential260) infringer is not aware of using the patented invention in commercialising his own product. If he has made a significant investment, the patent holder shows up, claims licensing fees and threatens the infringer with an injunction. In this situation, as pointed out by Shapiro, the (potential) infringer basically has seven different options (not all of which may be available in a concrete constellation). He may (1) pay the licence fees; (2) exit the market; (3) wait and challenge the patent; (4) wait, redesign and avoid litigation; (5) wait, redesign and challenge the patent; (6) sell and redesign; or (7) sell and not redesign the product.261 With the last two options, the (potential) infringer risks that the patentee successfully applies for an injunction. Of course, no option is perfect and, normally, the more the infringer has already invested, the more disadvantageous it is for him and consequently also for consumers and society. Where a standard is concerned, competitors who disclose all their patents may incur losses if their technology is not accepted, potentially even solely because of the anticompetitive behaviour.262

The described problems would not arise under liability rules, since there is no possibility of excluding someone; additionally, coordination should be easier. Problematic, however, in these constellations is the fact that the agency that determines compensation has to attempt to calculate a fair price. This calculation necessarily involves a certain amount of guesswork since a market price may not exist or may be inefficient.263 But liability rules may still look better than property rules: if the legal system itself creates an opportunity for a hold-up, it may have the special duty to prevent such an act.264

(4) Risk of Negative Effects due to Tragedy of the Anticommons and Patent Thickets

The number of patents in a certain technological area must be taken into account, and whether these patents may be overlapping. If this number is immense,265 it is arguable whether liability rules are better than property rules.266 Because exclusive rights are involved, such situations may lead to the tragedy of the anticommons and be characterised as patent thickets. Whereas the focus in the first situation with liability rules is on the fragmentation of rights, in the latter it is on the potential of overlapping. The anticommons problem was first described by Heller267 and mirrors the problem of the commons where coordination breaks down because multiple users are allowed to use a scarce resource but no one has the right to exclude the others. The resource is therefore prone to over-use.268 The neologism, in contrast, describes a situation in which multiple owners are each endowed with the right to exclude others from a scarce resource that is overpriced and, therefore, prone to under-use.269 In a world of costless transactions, people are always able to avoid commons or anticommons; in actual practice, however, avoiding tragedy requires overcoming transaction costs, strategic behaviour and the cognitive bias of participants.270 It was shown that overpricing might occur because the price charged by complementary monopolists may be higher than that of a single agent monopolist.271 The result of under-utilisation of joint property increases monotonically, in both the extent of fragmentation and forgone synergies and complementarities between the property fragments.272

A patent thicket may be defined as “an overlapping set of patent rights requiring that those seeking to commercialise new technology obtain licences from multiple patentees”.273 High fragmentation is required between essential, blocking patents.274 Thickets are viewed as being bad for the exploitation of knowledge and dynamic competition and innovation because they lead to substantial costly searching and licensing, activities that discourage the use of existing technology and hamper the emergence of new technologies.275 Statistics suggest an impact on the amount of transaction costs.276 Additionally, thickets may make the emergence of inefficient hold-out and hold-up situations even more likely.277 Thickets can also lead to bottlenecks and royalty-stacking.278 The latter term refers to situations in which a single product potentially infringes on many patents and thus may bear multiple royalty burdens,279 resulting in increased consumer costs.

As already explained, the employment of liability rules makes hold-out and hold-up situations less likely and problematic than does the use of property rules. Moreover, transaction costs are lower because the (potential) infringer needs not search as intensively for other relevant rights.

(5) Endowment Effects with Property Rules

The previous argument illustrates that endowment effects may, in general, be relevant when making a decision between liability and property rules.280 These effects are important in the area of patent law, where the subject-matter of the entitlement is created instead of simply existing.281 Contrary to the general debate, the effects especially apply to property rules because researchers have the tendency, commonly referred to as attribution bias, to overvalue the contribution of their own research compared to that of others.282 This assumption was confirmed in experiments, the results of which show that substantial endowment effects are likely to be present in transactions involving exclusive IP rights. Additionally, optimism bias and regret aversion were identified as reasons for the emergence of these effects.283 Therefore, the market price of IP may not only be affected by strategic behaviour, but also by endowment effects resulting in an inefficient reduction of the bargaining range in patent licence negotiations and in transaction costs. Moreover, in contrast to property rules, liability rules appear to be better with regard to the emergence of endowment effects because errors created in valuation (under- or over-compensation) are more likely to be distributed symmetrically on both sides of the optimal price. If this assumption is correct, disincentives to create inventions should be prevented.284 However, one must be aware that endowment effects under property rules may be mitigated by such other measures as specific terms in licensing agreements pertaining to the determination of royalties. This should especially be the case where valuable entitlements are concerned, as is normally the case in the area of patent law, because basically only those patents are licensed. If this is the case, then the situation is unclear and requires further research.285

c. Evaluation

The above general discussion and that focussing on patent law show that neither a system with solely property rules, nor one with solely liability rules is optimal. Regarding the general results, it is important to reiterate that transaction and assessment costs, the potential of strategic behaviour, the occurrence of endowment effects and deadweight losses caused by hold-out situations must be taken into account. These factors are also of particular relevance in the area of IP law. Specific to the area of patent law, it is emphasised how important innovation is to the economic interests of society, but also that it is commercialised and new knowledge disseminated. Two arguments must again be highlighted: on the one hand, the risk of under-compensation and, on the other, of taking the characteristics of intellectual rights into account along with the risk of inefficient hold-out situations. These arguments seem to be conclusive although they lead in different directions. It must be noted, however, that because of great uncertainty about the boundaries and validity of a patent, implementation of property or liability rules should not be the first choice to improve the situation; rather, these aspects should be generally taken into account as additional factors. Also in doubt is whether the risk of under-use only exists under liability rules. Consider the case of blocking patents under property rules: when an inventor and an improver both exist, coordination is necessary before society can benefit from the innovations. Bargaining breakdown may frequently occur, however, because a failure may result not only from strategic behaviour or other above-mentioned factors, but also from difficulties involved in placing separate values on the contributions of the inventor and the improver. Additionally, a classic situation involving the breakdown of negotiations results from immense uncertainty about the future development and profitability of each technology.286 Moreover, arguments regarding the risk of free-riding under liability rules and the risk of endowment effects under property rules are not entirely convincing because the former may simply not be the case287 and the latter may occur—as the general debate showed—independently of the chosen regime. It should not be forgotten that contracting is also possible under liability rules.

In conclusion, in order to establish the best-suited system, the pros and cons must be weighed in view of the specific task to be resolved. Some conditions, such as uncertainty about the boundaries of a patent, may be taken as given but also depend on the practice in the relevant patent office. At the same time, uncertainty about validity or the growth of patent thickets may, inter alia, depend on several factors. Whether these problems really exist is an empirical question that cannot be addressed in the abstract. Nevertheless, the result should be a system that embraces both types of rules and a certain combination of them in which the relationship between them could be dynamic.288 What this means for the different types of liability rules is that a combination between property rules and compulsory liability rules is most likely optimal (when leaving out private liability rules). With liability rules being the default in most regimes, it is difficult to balance interests perfectly because inverse compulsory liability rules do not make sense, except in special cases.

That said, however, the following considerations should be weighed. Since the optimal choice depends on concrete circumstances, it may often be difficult or even impossible for the legislature to make the optimal choice in advance. It has already been discussed that patent owners are in the best position and have a special interest in lower transaction costs: one could simply let the patentees decide which regime is best suited, while basically implementing property rules (private liability rules). However, a patent owner will only switch to liability rules if he profits from them and not because society would benefit. The legislature may be able to set incentives in such a way that the patent owner will choose the optimal regime. In other words, the legislature should establish a system that mirrors what firms do if they engage in indirect price discrimination. The term indirect price discrimination describes a situation in which a firm sets different prices, but because it cannot directly identify consumers’ willingness to pay, it will provide incentives for consumers to choose themselves according to preference.289 By way of comparison, the legislature may not know in advance whether property or liability rules are better, but it can set certain incentives to make a choice between the options.

However, a statement that private liability rules are generally beneficial would be wrong, as is shown in the next chapter. Additionally, it is not difficult to predict the impossibility of setting incentives in such a way as to always achieve the optimal result. Therefore, the legislature’s first step should focus on mechanisms that contain private liability rules, but the next step should be to identify situations in which compulsory liability rules might be optimal. In rare cases, liability rules by default could be taken into account. Regarding compulsory liability rules, the argument that focuses on inefficient hold-out and hold-up situations should be of special concern because these are the typical two situations in which the patent owner would be unwilling to switch to a liability rule regime, irrespective of what incentives were set.

5. Results

The analysis has shown that there are, from the standpoint of economic efficiency, a multitude of relevant factors that have to be taken into account when determining the optimal regime. The general evaluation, as well as the arguments for property and liability rules in patent law, does not lead to a clear preference for one regime. Therefore, contrary to conventional wisdom, it is at least questionable whether property rules should prevail. Rather, the evidence suggests that the optimal regime would consist of a combination of property and liability rules. When comparing the findings of this section with those of the first, in which it was shown that the German law and the regulation on the European Patent with Unitary Effect clearly prefer patents to be exclusive rights, it can be concluded that there is an inefficient overuse of property rules and that the legislature’s considerations were presumably based on conventional wisdom. However, it has already been noted that private liability rules could be of greater importance.

C. Property Rules or Liability Rules: Other Perspectives

The decision on whether an initial entitlement should be protected by property or liability rules needs not necessarily be (in fact, commentators agree that it should not or even must not be) purely economic. According to Calabresi and Melamed’s general concept, distributive and other justice reasons may also be relevant when deciding between property and liability rules.290 All categories may be considered equivalent or, as suggested here, according to the foremost economic view on the patent system, the latter as limitations to the focus on economic efficiency. To what extent and in which constellations this should be the case is the subject of debate (Sect. 1.). Additionally, since this work focuses on patent law in Germany and on European legal aspects, constraints due to international, European and constitutional law may currently be relevant (Sect. 2.). The regulators who decided about these higher-ranking norms may have already taken economic aspects into account, but the effect can be perceived as a short-sighted limitation of new considerations. All aspects are explained below; the aim of this discussion is to analyse what these aspects mean for the choice between property and liability rules in theory and to show that neither German nor European regulators are obliged to prefer property rules in patent law to the current extent.