Ringnalda.qxd 1 Orphan Works, Mass Allard Ringnalda Rights Clearance, and Online Libraries: The Flaws of the Draft Orphan Works Directive and Extended Collective Licensing as a Solution 1. Introduction European lawmakers have recently shown a profound interest in copyright law as an obstacle to the creation of Allard Ringnalda, Project researcher, Centre for online digital libraries. Unfortunately, a new directive that is Intellectual Property Law, Molengraaff Institute soon to be proposed to remove these obstacles suggests for Private Law, Utrecht University; PhD resear- that they have focused on the wrong problem. It provides a cher, Willem Pompe Institute for Criminal Law solution for the problem of orphan works: the many copy- and Criminology and the Department of legal righted works whose rightsholders are unknown or unloca- theory, Utrecht University. table. These orphan works cannot be used in a manner that The author would like to thank Rebecka Zinser and requires the rightsholder’s consent.1) With a new Orphan Willem Grosheide for their comments on earlier drafts. Works Directive, the Commission aims to facilitate the ex- The usual disclaimer applies. pansion of Europeana, the non-profit online library that should disseminate the digitised collections of all European national libraries. The problem of orphan works obviously stands in the way of a successful online library: if copyrights cannot be cleared, copyrighted works cannot be digitised and made available online. However, in this article, I shall propose and defend two claims. First, that the issue of orphan works is not the main hurdle on the way to a successful Europeana. Instead, the orphan works problem is only a symptom of a much larger issue: the inability to clear copyrights for the mass digitization and online dissemination of entire library collections. Second, that the legal technique of extended collective licensing seems to be appropriate to address the problem of mass rights clearance, and that it should there- fore be preferred over a solution that focuses on the orphan works problem only. In what follows, in section 2 I will first give a brief description of the Commission’s plans as they presently appear, and of the two general approaches to the orphan works problem: statutory licensing and statutory excepti- ons and limitations. In section 3, I will argue that because these approaches require the user to prove reasonably dili- gent efforts in trying to locate rightsholders, they do not benefit the digitisation and online dissemination of entire library collections. Section 4 details a solution to that pro- blem: extended collective licensing. While the system of ex- tended collective licensing appears to be well suited for facilitating online digital libraries, a number of important questions remain. These are discussed briefly in section 5. 2. Background and Contents of the Orphan Works Directive 2.1 Orphan Works and the i2010 Digital Libraries Pro- ject name.qxd 2 In 2005, the European Commission launched its initiati- ve to create an online digital library by 2010.2) This online library, called Europeana, is more properly described as a platform through which all European national libraries can disseminate digitised works from their collections. Europea- na is intended to make available to the general public a wide variety of freely accessible digitised content, including books, newspapers, articles, music, video, and photo- graphs. The advantages are obvious. All material of cultural significance is preserved, archived and made available to as wide an audience as possible. Instead of having to get hold of a physical copy of a work, anyone in the world can enjoy or study European cultural heritage by simply visiting a website. However, if we consult Europeana today, we find that it almost exclusively contains older material that is in the public domain. Recent works that are still subject to copyright are only rarely included. That is due to the fact that copyright law forms a major obstacle to digitising and disseminating library collections. Libraries need to obtain permission from copyright owners both for digitising a work (making a reproduction) and publishing it on a website (communication to the public, making available). This requires libraries to find the copyright owners of all works in their collections in order to negotiate a licence. Dealing with copyright law in this way is new for libraries. In their traditional role, libraries hardly, if at all, had to deal with copyright law and copyright owners. The copyrights in the physical copies of a work are exhau- sted, which means that they can be further distributed without the permission of the copyright owner.3) Therefo- re, no copyright issues emerge if libraries allow the public to consult their collections on site. Lending rights are not exhausted, but as they are subjected to mandatory collec- tive management, libraries do not have to deal with rights- holders directly.4) Libraries also benefit from exceptions and limitations for reproducing works for preservation pur- poses or making digitised content available on on-site ter- minals.5) Traditionally, they did not, therefore, have to be aware of the identity of the copyright owners of the works in their collections. That changes if libraries want to make their collections available online. Orphan works occur in the process of identifying and locating copyright owners. As there are no mandatory copyright registries, libraries have to attempt to find copy- right owners by other means.6) This can be a laborious pro- cess.7) Very often, no copyright owner can be identified or located, making the work orphaned.8) Orphan works occur for a number of reasons. While most works include some information about the author, it is not always possible to identify who the current copyright owner is or how he can be contacted. While famous writers, photographers, jour- nalists or scientists will not usually pose any difficulties, les- ser-known authors for whom only a name is available can be very difficult to find. Moreover, with copyright protec- tion in the EU lasting for 70 years after the death of the aut- hor, it will often be necessary to identify and trace all the heirs, requiring family relations to be traced and wills to be examined. Copyrights are also frequently (partly) transfer- red and retransferred, and usually transfer contracts are not publically accessible. Furthermore, not all copyright owners name.qxd 3 are represented by collective management organisations, and even if they are, these organisations may not manage the rights that are required for digitisation and online disse- mination. Finally, for some works, the author may be totally unknown, yielding no clues where to start searching for the copyright owner; photographs, for instance, are often un- marked. While estimates of the size of the orphan works problem vary, it is generally held to be extensive. Libraries estimate that between 10 and 40 per cent of written mate- rial held in their collections is orphaned.9) Another study estimates the overall number of orphan works in the UK to be between 13 and 50 million.10) The European Commission was quick to recognise that orphan works would form a major obstacle to creating a comprehensive European online library.11) A High Level Expert Group was set up with a subgroup on copyright, and it considered various suggestions on how to solve the problem.12) It is important to distinguish between solutions that aim to prevent the occurrence of future orphan works, and those that seek to redress the problem as it currently stands. The first type may consist of enhanced voluntary registration of rights management information, creating a database for orphan works,13) or encouraging the use of Creative Commons-like licensing.14) Such approaches do not, however, solve the impossibility of clearing the rights in works that are already orphaned. Regarding the use of such orphan works, an important question emerges. While a statutory-based solution would seem appropriate, it could also be argued that users might simply start using works if attempts to find the copyright owner do not meet with success; no legislative intervention would then be required.15) Users could, for instance, inclu- de a disclaimer stating that reasonable efforts were made to locate all rightsholders; and that those who could not be found are invited to contact the user to agree on further use and a licence fee. Mass users can allocate funds for such claims or even set up an insurance scheme together with collecting societies to distribute the costs and risks.16) However, it is important to note that copyright law is not merely an economic affair that can be settled ex post facto. Many continental European countries protect copyright by marking infringement as a criminal offence. Inducing pu- blic libraries to wilfully violate criminal law by having them use orphaned works without permission would therefore clearly violate public order and policy. Self-regulation can- not suffice. A legal solution is required. 2.2 Legal Solutions to the Orphan Works Problem The European expert group considered various ways of providing for an exemption from the copyright protection of orphan works. Even if the beneficiaries of these solutions are non-profit libraries, it is obvious that a balance should be struck between the need for digitising and dissemina- ting cultural heritage on the one hand, and the interests of authors and copyright owners on the other. In particular, rightsholders have economic interests that should be res- pected. One important way of doing so is by limiting the scope of an exemption. Firstly, requiring that reasonably diligent search efforts be proved before a work is conside- name.qxd 4 red orphaned creates an important threshold. Furthermore, balancing can be achieved by setting up a system to com- pensate rightsholders in the event of their surfacing after the use of the orphan work has commenced. Thirdly, the exemption could be limited to works that are no longer commercially available (i.e. works that are out of print), so as to prevent interference with commercial exploitation; after all, works that are exploited hardly ever become or- phans, as the rightsholder or his representative is involved in the exploitation process. The possible solutions to the orphan works problem fall into two general categories: licensing systems, and excepti- ons and limitations. A licensing system requires the user of an orphan work to obtain a licence prior to using the work from a licensing body designated by law, such as a court, a copyright tribunal, or a collective management organisa- tion.17) Evidence of the reasonably diligent efforts needs to be offered, and a licence fee is set and may be collected immediately. Under an exception or limitation system, no such prior administrative efforts are required. The law pro- vides that works whose right owner cannot be found despi- te reasonably diligent efforts may be used without permissi- on, provided that an equitable remuneration is paid to the copyright owner in case a claim is brought.18) Generally speaking, the advantage of licensing systems is that they offer full legal certainty due to the prior authorisation they require, whereas an exception or limitation is more effi- cient, since a procedure is only needed if the rightsholder appears. 2.3 New European Legislation on Orphan Works Whatever the type of solution, the final report of the experts’ copyright subgroup identified four criteria that should be met before an orphan work can legitimately be used. I quote these in full: – ‘A user wishes to make good faith use of a work with an unclear copyright status; – Due diligence has been performed in trying to identify the right holders and/or locate them; – The user wishes to use the work in a clearly defined manner; – The user has a duty to seek authority before exploiting the orphan work […]’19) After some years of consideration, the Commission has now indicated that it is prepared to legislate on the issue of orphan works, taking into account the four criteria explica- ted by the expert group. While no official preparatory documents have been issued so far, various sources allow us to obtain a view of the Commission’s plans. Firstly, Tilman Lüder, head of the Commission’s Copyright Unit, has spo- ken and written on the possible approaches to the orphan works issue and has mapped the course for the future.20) Secondly, and more importantly, a draft version of the soon to be proposed Directive on the mutual recognition of or- phan works in the print sector has been leaked in Novem- ber 2010. Since none of these sources are official, I shall dis- cuss only the general policy outlines as they presently appe- ar without going into any details. name.qxd 5 Lüder explains that there are five options to address the issue of orphan works in the context of a European digital library. Four of these would prescribe modalities of either an exception or limitation, or of a statutory licensing sche- me. The fifth, however, is the one that he thinks should be preferred. It leaves it up to the Member States each to devi- se a legal technique with which libraries can include orphan works in their online collections, provided that some form of prior authorisation has been granted for each work.21) However, to ensure that orphan works can be made availa- ble throughout Europe, it includes a rule of mutual recogni- tion – a novelty in copyright law.22) According to Lüder, this rule has two aspects. First, all Member States should recognise a work as orphaned once it has been marked as such. This prevents the need to conduct and prove reason- able search efforts in all countries of use.23) Second, if a work is made available under the laws of one of the Mem- ber States, the legality of that use should be recognised everywhere else. That ensures that a user does not have to comply with the divergent rules on orphan works that would come to exist throughout the EU.24) The fact that the prior approval of the orphaned status of works is required has to do with this very important pro- blem of enabling the European-wide use of orphan works. European copyright law is not unified. It consists of 27 sepa- rate copyrights each covered by the copyright laws of one of the Member States. Under the international and Europe- an copyright law principles of territoriality and national treatment, the legality of making works available on the Internet should be judged according to the laws of the countries where the work is made available. The act of making available is not localized in the country of transmis- sion, but in all the countries of reception, and their respec- tive laws apply.25) All of these should therefore legalize the use of orphan works.26) The problem, then, is that a user will be required to conform to 27 different laws on the use of orphan works. However, solving this problem is more dif- ficult than claims of a harmonised European copyright law would lead one to suppose. The most straightforward solution would be to introdu- ce a unified solution to the orphan works problem by means of a regulation. One and the same approach would then apply in all Member States. However, European copy- right is traditionally subject to minimum harmonisation only. With a few exceptions, Member States are free to se- lect exceptions and limitations from an exhaustive list.27) There is no true tradition of creating mandatory exceptions and limitations, because copyright policy is held to be strongly contingent on domestic cultural circumstances.28) As the political support and the legal basis for copyright unification are presently rather weak, alternatives had to be sought with more respect for the diversity between natio- nal copyright systems and the principles of proportionality and subsidiarity.29) This has resulted in the draft Orphan Works Directive. The contents of this Directive, with its five substantive Articles, are as follows. The first article limits its scope to printed works that have been published with the consent of the author (books, journals, newspapers, maga- zines or other writings).30) Audiovisual works are not cover- ed, nor are other visual works, except if they are included in name.qxd 6 printed matter (e.g. illustrations in books or press photo- graphy). The second article defines orphan works as works whose rightsholders could not be identified or located despite diligent efforts. Article three circumscribes the de- gree of diligence required by referring to an annex that lists the sources that should be consulted to find the copyright owner. The provision also states that if a work was first published in one of the Member States, the search need only be conducted in that country.31) Article five requires Member States to allow libraries, educational establish- ments, and museums to reproduce (digitise) orphan works and make them available online, but only for non-commer- cial purposes. States are free to choose a statutory licence or exception or limitation.32) The directive includes no provi- sions on remuneration, stating only that a rightsholder in an orphan work should have the possibility to ‘exercise’ his exclusive rights. It would seem that the delicate issue of remuneration – whether or not users should pay for the use of orphan works, and if so, how the fees should be determi- ned and when and to whom they should be paid – is left to the Member States to decide. In the draft Orphan Works Directive, the principle of mutual recognition has been slimmed down considerably. Article four provides that Member States shall mutually recognise as orphaned those works identified as such in the member state where they were first published.33) This ensu- res that the search efforts will only have to be made once, namely in the country of first publication (provided that it is an EU member state). However, the other aspect of mutual recognition, ensuring that the making available of orphan works in one member state is recognised as legal in others, has not been implemented – even if the 13th preamble sta- tes that ‘[i]n order to foster the Union’s citizens’ access to Europe’s cultural heritage, it is also necessary to ensure that orphan works which have been digitised and made availa- ble to the public in one Member State are also available in other Member States’. Without any rules of mutual recogni- tion on this point, it seems that libraries simply need to comply with the orphan works legislation of all 27 Member States. Although they need not duplicate their search efforts, they will still be required to apply for various territo- rial licences and comply with divergent national remunera- tion schemes, as these are all issues that are not harmonised by the proposed Directive.34) 3. Reasonable Search or Mass Digitisation? We have seen that the European legislature obliges the prior scrutiny of works being marked as orphaned. A user will always have to apply to an authority to determine whether his search efforts have been sufficiently diligent for the work to be considered orphaned. By implication, the proposed solution requires that libraries conduct reasonab- ly diligent search efforts for every work in their collection which they wish to digitise. While that approach may be adequate for solving the orphan works issue in itself, it does not truly facilitate the digitisation and online dissemination of entire library collections. Obtaining a licence for each work in the collection of a library takes a lot of time, effort, and money.35) There are name.qxd 7 estimates that the average costs for clearing the copyrights required to include a book in an online collection are as much as $ 1,000, regardless of whether the attempts are successful (and exclusive royalties).36) Another indication of the scale of the costs comes from the Dutch National Li- brary, which is currently working on an online collection of history books, which besides text contain a large volume of images, maps, photographs, etc., each of which may have its own copyright owner. One full-time employee manages to clear the copyrights of only 10 such books per month. The need to search for the copyright owner, moreover, covers a vast amount of books and texts. In 2010, research showed that the oldest copyrighted book in the UK dated from 1859.37) Consequently, extensive and expensive search efforts have to be made for all books published from that time onwards.38) Clearing copyrights on an individual scale would thus be an extremely costly and lengthy affair. Therefore, the true problem that has to be addressed is not the inability to use works whose copyright owners cannot be found despi- te reasonably diligent search efforts. The true problem is that libraries have to make such efforts to find copyright owners, even if the copyright owner can eventually be found. Licensing on an individual scale seems to be too much to ask for. If we have to wait until the copyright owners for all the works in library collections have been found or looked for, we will not see a digital library in the near future, with or without orphan works. Waiting for all those works to pass into the public domain would be more efficient, and probably quicker, too. The proposed Orphan Works Directive tries to ameliora- te the burden of diligent search efforts by referring to a number of sources that should be consulted before a search is considered to have been sufficiently diligent. These sour- ces include publishers’ associations, publishing houses, public lending right authorities, databases of ISBNs and ISSNs, library indexes and catalogues, databases of relevant collecting societies, national bibliographical indexes of published material, personal directories and search engi- nes, and biographical resources regarding authors. While the inclusion of such a list has the benefit of making the vague standard of diligence somewhat more practicable, it does little to alleviate the costs and time of searches. The sources still require manual scrutiny on a work-by-work basis. However, a lower threshold would run the risk of vio- lating the three-step test, as the exception would not be sufficiently limited to certain special cases. A solution based on a diligent search fits well with the problem of orphan works. But for online libraries, orphan works are not the true problem. Rather, they are a symp- tom of the much larger problem of clearing copyrights for entire collections such as those of libraries and archives. The proposed Directive, based as it is on diligent searches for each copyright owner, does nothing to solve the inability of mass rights clearance. It may provide for a legal technique that allows orphan works to be used, but if libraries will be required to clear copyrights for each individual work, we shall not have to worry about the orphans before their copyrights have already expired. name.qxd 8 What is needed, instead, is a legal technique that would allow libraries to make their entire collections available onli- ne without having to obtain permission from each copy- right owner individually. It goes without saying that simply allowing libraries to digitise and publish their collections, including orphans and non-orphans alike, by introducing an exception or limitation to that effect is impossible and undesirable. It would violate the three-step test: the excep- tion would apply to all works rather than certain special cases of orphans, and it would undoubtedly conflict with the normal exploitation of works that are not orphans: if a work is freely available online, it is unlikely to sell very well. Even if mandatory collective management were applied, the copyright owners of non-orphaned works would be unduly restricted in their possibilities to exploit according to their own wishes. Besides, the right of making available (as distinct from the broadcasting right) is formulated in the WIPO copyright treaty as an exclusive right of authori- sation, and therefore it may not be possible to reduce it to a mere right of remuneration or to subject it to mandatory collective management.39) 4. Extended Collective Licensing One promising technique to facilitate mass rights clea- rance is extended collective licensing (ECL), which is cur- rently employed in various Nordic countries.40) It is based on voluntary collective management, and works as follows. If a collective management organisation (CMO) repre- sents a substantial amount of copyright owners in one country, it is allowed by law to issue licences on behalf of all copyright owners worldwide. In essence, its repertoire is extended to cover all possible works. This opens up the pos- sibility of issuing blanket licences: licences with which the entire world repertoire may be used, without the need to obtain permission from individual copyright owners. No reasonably diligent searches are required. That makes it an attractive solution for facilitating online libraries. It would not only solve the orphan works problem, but also, and more importantly, the broader problem of mass rights clea- rance. Entire collections could readily and easily be licensed for online use. For instance, in Norway, extended collective licensing is currently being employed to allow the Norwegi- an National Library (NLB) to digitise up to 50,000 books, including copyrighted material.41) The NLB does not, the- refore, have to contact individual copyright owners: the transaction costs are limited and no orphans occur. It can legally make copyrighted works available by virtue of a con- tract with a collecting society, even if not all of the copy- right owners involved are represented. Extended collective licensing might seem to place a sig- nificant restriction on the exclusive nature of the exploitati- on rights. There are, however, various guarantees to ensure that the restrictions are not too great. The first is that a CMO needs to be sufficiently representative among copy- right owners before it is allowed to issue extended collec- tive licences. This ensures that the interests of all copyright owners are sufficiently respected. In fact, the system of extended collective licensing is defended by arguing that it is in the interest of all copyright owners, whether represen- name.qxd 9 ted or not, that their rights are adequately managed by a professional organisation. Second, there is the possibility to opt out from the extended collective licensing system. And, finally, there is a right of individual remuneration for non- represented right owners, so that they can receive compen- sation at commercial tariffs, rather than having to participa- te in any collective distribution schemes. The collective management organisations have a res- ponsibility to try to pay collected fees to copyright owners that are unrepresented, and thus have to search for them.42) While this would seem to shift the problem of or- phan works from before use (licensing) to after (the paym- ent of royalties), it is important to note that under ECL the legitimacy of using works does not depend on royalties being paid, but on the licence being negotiated with a suf- ficiently representative party. Making the collecting society responsible for distributing unclaimed fees also implies that the costs of searching for unrepresented right holders are for the copyright owners collectively, rather than for users (though the costs may be calculated into the licensing fees). In fact, ECL is not all that different from the technique used in the Google Books Settlement to allow Google to digitise entire library collections, including copyrighted material, and make them available online. In brief, Google intents to settle all past and future copyright infringements by striking a deal with an organisation representing a num- ber of copyright owners (the Authors’ Guild). Because the Settlement is a class action suit, the court can determine that it should apply to all possible and future claimants, and therefore grant it legal effect that extends beyond the par- ties involved. Google will pay royalties for the works that it disseminates, and an independent authority has to manage and distribute these, searching for unknown copyright owners where necessary. The difference between ECL and the Google Books Settlement is in its legitimacy: whereas extended collective licensing is an instrument firmly rooted in copyright law and policy, class action suits are intended to settle existing claims rather than reaching agreement on future ones. As a consequence, extended collective licen- sing includes guarantees tailored to copyright that are mis- sing from the Settlement. For instance, ECL requires that copyright societies are sufficiently representative, whereas the Authors’ Guild is not (it being a society of some 8,000 literary writers, publishers, and lawyers, thus excluding, for instance, scientific authors and journalists). Extended collective licensing may also be subject to specific govern- ment supervision on how funds are managed and distribu- ted. And there is an issue of principle: while Google provo- ked the class action by starting to infringe copyright – and thereby arguably gaining an advantage over other parties who might also wish to create an online library – ECL provi- des an instrument with a sound legal basis that is equally available to all sorts of different cultural heritage instituti- ons. 5. Issues of Extended Collective Licensing Before extended collective licensing can be hailed as a promising solution for the problem of mass rights clear- name.qxd 10 ance, there are at least five significant issues that need to be resolved. I will briefly discuss the most prominent of these. Of course, they are in need of further exploration and re- search. The first issue is how ECL can be used to allow Europe- an-wide use. Just as with exceptions and limitations, the effect of an extended collective licence is limited to the ter- ritory of the country under whose laws it was granted. For that reason, the Norwegian digital library that employs extended collective licensing is only available within Nor- way (due to the localisation of visitors’ IP addresses).43) But if all Member States introduce extended collective licen- sing, one could rather easily create organisations to take care of multi-territorial licensing by arranging and admini- stering extended collective licences in all Member States (one-stop shops).44) That would, however, require the full harmonisation of ECL provisions. Secondly, it can be doubted whether ECL is sufficiently legitimate. After all, it allows copyrights to be managed by a CMO even if the owner has not consented. Legitimacy is very much a matter of legal culture. In the Scandinavian countries, collective management is very popular and its benefits are widely agreed upon. It is accepted that exten- ded collective management, under the guarantees provi- ded, is to the benefit of all: the interests of society (as a user) and copyright owners coincide. That is, of course, a rather communal system with socialist origins that may not be via- ble in the legal cultures of other Member States. Compara- tive work on the regulation of and views on collective management would be required to provide an impression of the possibilities of extended collective licensing through- out the European Union. Furthermore, there is a practical side to this issue as well. Collective management may not be popular in all countries and for all types of works. As a consequence, there may not everywhere be collective management organisations that are sufficiently representa- tive to be allowed to issue extended collective licences. Thirdly, ECL should conform to international and Euro- pean copyright law. Given that it is applied in the Scandina- vian countries, and that this application is recognised as legal in the considerations of the Copyright Directive,45) one might argue that there are at least no obvious ob- stacles. ECL is only a light restriction on the exclusive nature of copyrights.46) Extended collective licensing does not preclude non-represented copyright owners from exerci- sing their rights, and copyright owners can opt out. Howe- ver, a requirement to opt out before one can fully enjoy the exclusive nature of one’s copyright may be a forbidden for- mality under the Berne Convention. Much depends on the interpretation of the Convention’s prohibition. One might say that the restriction on formalities was intended to do away with constitutive formalities: the registration of a work before copyright protection was granted. That can be distinguished from a formality that is required before the exclusive nature of copyrights can be fully enjoyed.47) Then, extended collective licensing may very well be legal. This further depends on the precise details of the opt-out process. There is a difference between simply sending an e- mail requesting to opt out and having to supply a great deal of detailed information. An easy opt-out method name.qxd 11 would seem preferable, but at the same time would increa- se the burden on libraries to verify for each work whether the copyright owner has opted out. From a library’s per- spective, it would be easier if the copyright owners would be required to indicate for which works precisely they want to opt out. The fourth issue is that ECL is, in essence, a commercial exploitation system geared towards decreasing transaction costs and resolving market failure. This means that it aims to facilitate negotiations between the collective of copy- right owners and users to allow them to reach an agree- ment on the licensing conditions and license fees. These conditions and fees will reflect market conditions, and they are not, therefore, necessarily very attractive for non-com- mercial libraries. In the Norwegian digital library project, for instance, the library has to pay an average fee of € 13 per book per year.48) That might be feasible for small or short-term projects, but it would be much too expensive to make entire collections available for the full duration of copyright protection. The essence of this issue is that extended collective management is not a policy instrument. It is not intended to be used by governments to control copyright and achie- ve certain policy goals. It is merely a facility to support the market, to lower transaction costs, and to resolve market failure. Its legitimacy depends on its protecting the interests of all copyright owners. That means that, in principle, there is no room to use extended collective licensing to force other interests, such as online accessibility subject to cheap licence fees. If governments were to intervene with their own interests, ECL would lose its legitimacy. So, by implica- tion, governments can neither compel that extended col- lective licences are concluded, nor that they can be affor- ded by libraries. How extended collective licensing could be regulated to benefit online libraries remains an open question. Fifthly and finally, the costs of administering a system of extended collective licensing could be very high. That applies in particular to the costs incurred by CMOs for loca- ting rightsholders and paying them the licensing fees that they are due. Because of uncertainty about the success rate of finding unrepresented rightsholders, large sums of licen- ce fees would have to be reserved. Furthermore, the costs of the searches and payments are probably significant. It is important to note this, because ECL has up until now been implemented on a small scale only. In the Scandinavian countries, it applied to a limited number of users and a comparatively small audience; also, it was traditionally re- stricted to narrowly circumscribed types of use.49) With its application to online libraries, the amount of fees collected and the administration involved would be brought to an entirely new level. A related problem is that many of the copyright owners whose works are being used will never claim or be given their fees, particularly if the works are orphaned. However, in a system of extended collective licensing, the CMO is deemed to represent all copyright owners, and therefore can legitimately claim their fees and manage these to the benefit of all copyright owners. As a consequence, a large amount of unclaimed fees, paid for by the public, will be ac- name.qxd 12 crued. A prominent policy question is who should benefit from these unclaimed fees of mainly orphaned works – libraries and thereby society as a whole, or collective mana- gement organisations which can distribute the money among their members? Given the nature of ECL, one would probably have to opt for the latter – confirming the finding that it is a market-facilitating tool, not an instrument for po- licy implementation. Some of these issues could be addressed if the scope of an extended collective licence were limited only to include works that are no longer commercially available; licences for commercially exploited works should then be negotia- ted on an individual basis with or via publishers. Such a limi- tation would imply that the scope of the ECL approach is de facto similar to that of the Orphan Works Directive: works that are no longer exploited are often orphaned, for after abandonment it becomes increasingly difficult to determi- ne who the current copyright owner is. However, contrary to the directive, it would allow the use of such works with- out having to make expensive and time-consuming efforts to try to identify and locate the copyright owner. Such an application of ECL would ensure better compliance with international obligations, in particular the three-step test, as commercial exploitation is not interfered with. It would enhance its legitimacy for the same reason. And, finally, with the works being without commercial interest, the fees can be set at a level that is attractive for non-profit cultural heritage institutions. 6. To Conclude The proposed Orphan Works Directive is fundamentally flawed. It will not help to make entire library collections available online. While orphan works are a real problem, they are only a part of a much larger obstacle to creating an online library: the inability to clear copyrights on a very large scale. Had the Commission addressed that obstacle, it would have killed two birds with one stone: facilitating mass rights clearance would have taken care of the orphan works issue as well. But by focusing exclusively on the latter, libraries will only be able to disseminate small parts of their collections online. It would take too long and be too costly to obtain prior authorisation for all works in their collec- tions. The problem that stands in the way of online libraries, in essence, is that copyrights have to be cleared prior to use and on an individual scale. Therefore, collective manage- ment and licensing would have been a more fruitful avenue to explore. Extended collective management seems parti- cularly promising, as it allows collective licences to be con- cluded even on behalf of rightsholders who are not repre- sented by a CMO. In effect, the world repertoire – probably best limited to those works that are no longer commercially available – can be licensed for inclusion in an online library. At the same time, the technique provides important gua- rantees for copyright owners, such as an opt-out option and a right to individual remuneration. Although there are a number of problems with ECL – its legitimacy, its compli- ance with international copyright law, its use as a policy instrument, the costs of its administration, and how it can name.qxd 13 be used to enable European-wide availability – it definitely warrants further attention in the quest for a true European digital library. 1) See the general literature on orphan works: M. van Eechoud, Har- monizing European Copyright Law: the Challenges of Better Law Making, The Hague, Kluwer Law International 2009, p. 263 – 296; S. van Gompel, ‘Unlocking the potential of pre-existing con- tent: how to address the issue of orphan works in Europe?’, IIC (International Review of Intellectual Property and Competition Law) 2007, p. 669 – 702; G. Spindler & J. Heckmann, ‘Retrodigita- liserung verwaister Printpublikationen; Die Nutzungsmöglichkei- ten von “orphan works” de lege lata und ferenda’, GRUR Int 2008, p. 271 et seq; F-M. Piriou, ‘Les “oeuvres orphelines”: en quête de solutions juridiques’, RIDA (2008). This article is based in part on research carried out for the Dutch government in 2008: M. Elferink & A. Ringnalda, Digitale ontsluiting van histori- sche archieven en verweesde werken, Amsterdam, Amsterdam, deLex 2009. 2) Communication from the Commission to the European Parlia- ment, the Council, the European Economic and Social Commit- tee and the Committee of the Regions, ‘i2010 Digital Libraries’, 30 September 2005, COM(2005) 465 final. 3) Art. 4(2) Directive 2001/29/EC on the harmonisation of certain aspects of copyright and related rights in the information society (hereinafter: Copyright Directive). 4) On exhaustion: Art. 1(2) Directive 2006/115/EC on rental right and lending right and on certain rights related to copyright in the field of intellectual property (codified version) (hereinafter: Rental Rights Directive). On the optional mandatory collective manage- ment: Art. 6 Rental Rights Directive. 5) Art. 5(2)(c) and 5(3)(n) Copyright Directive. 6) Formalities such as registration are forbidden by Art. 5(2) of the Berne Convention (BC). 7) A good view on the problems faced by libraries can be obtained by reading the comments on the pending U.S. Orphan Works Act collected by the U.S. Copyright Office (www.copyright.gov/orp- han/), in particular those from (non-profit) libraries. 8) On the causes of orphan works, see generally Van Gompel (2007), supra note 1. 9) British Library, ‘Intellectual Property: a balance’, British Library Manifesto, September 2006, www.bl.uk/news/pdf/ipmanifesto. pdf (point 5). 10) JISC, In from the Cold: An Assessment of the Scope of ‘Orphan Works’ and its Impact on the Delivery of Services to the Public, April 2009, available from www.jisc.ac.uk 11) Commission Recommendation on the digitisation and online ac- cessibility of cultural material and digital preservation, COM (2006) 3808 final. 12) Commission Decision of 27 February 2006 setting up a High Level Expert Group on Digital Libraries 2006/178/EC. See the reports of the Expert Group and the subgroup on http://ec.euro- pa.eu/information_society/activities/digital_libraries/other_ expert_groups/hleg/index_en.htm. See in particular the Final Re- port on Digital Preservation, Orphan Works, and Out-of-Print Works of 4 June 2008, available at http://ec.europa.eu/informa- tion_society/activities/digital_libraries/doc/hleg/reports/copy- right/copyright_subgroup_final_report_26508-clean 171. pdf. On the work of the group, see M. Ricolfi, ‘Digital libra- ries in the current legal and educational environment: a European perspective’, in: L. Bently, U. Suthersanen & P. Torremans, Global Copyright, Cheltenham, Edward Elgar 2010, p. 216 – 229. 13) This is the aim of the European Arrow database; see www.arrow- net.eu. 14) Van Eechoud (2009), supra note 1, p. 273 et seq. See for a similar discussion, Van Gompel (2007), supra note 1. 15) Cf. e.g. Van Eechoud (2009), supra note 1, p. 272. 16) An example of the latter is the Dutch organisation FotoAnoniem (a foundation set up by the Dutch photographers’ collecting society and the federation of photographers), which offers ‘licen- name.qxd 14 ces’ for orphaned (anonymous) photographs, stipulating that it will pay all costs incurred as a result of claims by rightsholders (see www.fotoanoniem.nl). The Dutch National Archive has recently contracted a licence with the collective management organisation for photographs – Pictoright – with similar terms (www.nationaalarchief.nl/nieuws/nieuws/overeenkomst_natio- naal_archief_en_pictoright.asp, last viewed 17 December 2010). Obviously, such licences are without legal effect vis-à-vis copy- right owners of orphan works, as these owners are by definition not represented by a collecting society. A similar initiative has been launched in Germany, where the Deutsche Nationalbiblio- thek has concluded an agreement with the publishers’ associati- on and the collecting societies VG WORT and VG BILD-KUNST. See T. Koskinen-Olsson, ‘Digital libraries: collective administration for online libraries – a rightsholders’ dream or an outdated illusi- on?’, in: Bently, Suthersanen & Torremans (2010), supra note 12, p. 252 – 264. 17) The system can either be one of statutory licensing, wherein a court or copyright tribunal issues the licence, or one of manda- tory collective licensing of orphan works, wherein collective management organisations can issue licences for orphan works. The statutory system is operative i.a. in Canada (s. 77 Canadian Copyright Act). A French expert group has suggested a system of mandatory collective management of orphan works: Conseil supérieur de la propriété littèrarire et artistique, Commission sur les oeuvres orphelines, Avis de la commission spécialisée du CSPLA sur les oeuvres orphelines, 10 April 2008, www.culture.gouv.fr/culture/cspla/avisoo08.pdf. 18) In the USA, a limitation on remedies has been proposed for orp- han works (Orphan Works Act, pending). 19) See the 2008 Final Report, supra note 12, p. 12. Similar ideas have been expressed in a 2008 Memorandum of Understanding that was signed by various EU rightsholders and cultural heritage organisations: http://ec.europa.eu/information_society/activities /digital_libraries/doc/hleg/orphan/mou.pdf. 20) See T. Lüder, ‘The “orphan works” challenge’, 2010 GRUR Int., p. 677 – 685. This paper was also presented at the 2010 Fordham IP conference (see http://fordhamipconference.com/papers/). Neelie Kroes, Commissioner for the Digital Agenda, has mentio- ned the issue of orphan works and digital libraries as a topic for future copyright reform in her speech of 5 November 2010 in Avignon (SPEECH/10/619). 21) With such a prior procedure being required, the difference bet- ween licensing systems and exceptions and limitations is slight. 22) Mutual recognition derives from the harmonisation of European private law and, from a later date, European criminal procedure. It serves the purpose of allowing minimum harmonisation with respect for national legal diversity while ensuring interoperability. However, in international copyright law (particularly the Berne Convention), the mechanism is unknown. International copy- right law proceeds on the very different principle of assimilation, or national treatment. This means that rather than referring to foreign law, the copyright protection of a work is judged by the laws of the country where the work is being used and for whose copyright protection is being sought. Foreign works are treated in the same manner as works of domestic origin. 23) The Arrow database (www.arrow-net.eu) could be used to iden- tify works as being orphaned. Other users then know that such works can be used pursuant to the applicable law, without their having to conduct a reasonably diligent search. The mutual reco- gnition principle applies within the EU only. 24) A similar approach can be found in the EU report of the ‘Comité des Sages’, The New Renaissance, Brussels, 10 January 2011, para. 5.3, available on http://ec.europa.eu/information_society/ activities/digital_libraries/index_en.htm. 25) S. Ricketson and J.C. Ginsburg, International copyright and neigh- bouring rights: the Berne Convention and beyond, Oxford Uni- versity Press, Oxford 2006 (2 vols), vol. 2, p. 1297 et seq.; J. Gins- burg & P. Goldstein, ‘Reply Comments on “Orphan Works” in- quiry’, 9 May 2005, www.copyright.gov/orphan/comments/reply/OWR0107Gins- burgGoldstein.pdf. The matter is open to debate, however. See name.qxd 15 also J.C. Ginsburg, ‘The (New?) Right of Making Available to the Public’ (2004). Columbia Public Law & Legal Theory Working Papers. Paper 0478, http://lsr.nellco.org/columbia_pllt/0478, p. 2 – 3. 26) To matters of infringement, the lex protectionis applies (art. 8 Regulation EC/864-2007 (Rome II Regulation)). If making availa- ble is localized in all countries of reception, each of these will the- refore apply its own copyright law, including its exceptions and limitations, to determine the legality of the transmission. See generally A. Ringnalda, ‘National and International Dimensions of Copyright Law in the Internet Age. Harmonizing Exemptions: The Case of Orphan Works’, 17 [2009] European Review of Priva- te Law, p. 895 – 923. To remedy the problem of ‘ubiquitous in- fringements’ (i.e. infringements occurring in all countries of re- ception) the Max Planck Group for Conflict of Laws in Intellectual Property (CLIP) has suggested that courts apply the law of the state with the closest connection to the infringements, to be jud- ged on factors such as the place of residence or business of the infringer or the greatest harm (Principles for Conflict of Laws in Intellectual Property, 3rd draft, September 2010, Art. 3:603). 27) Art. 5 Copyright Directive. Mandatory exceptions and limitations are found in a number of the vertical directives and in the Copy- right Directive with regard to caching. 28) For a more elaborate background, see Ringnalda (2009), supra note 26. The Wittem European Copyright Code, the product of a scholarly initiative, is an example of a proposition to create a fully unified, pan-European copyright governed by a single set of rules. It does not, however, contain any provisions on orphan works. See www.copyrightcode.eu. 29) This point is expressly mentioned in the 7th preamble to the draft Orphan Works Directive. 30) It is important to note that the Directive does not cover unpub- lished writings. However, archives contain large quantities of un- published and often orphaned material, such as diaries and let- ters. 31) Consequently, the Orphan Works Directive is not well suited for works originating from outside the EU. It should be recalled, however, that its main purpose is to facilitate the online dissemi- nation of European cultural heritage. 32) Thus, a new exception to the exclusive rights is introduced. The closed list of exceptions and limitations in the Copyright Directive has not been amended, but the Orphan Works Directive has prio- rity because of the lex posterior principle. 33) Cf. preamble 12 to the draft directive. 34) Applying mutual recognition to require Member States to reco- gnise the legality of works made available under the law of the country of first publication might violate the principle of national treatment (requiring states to apply one and the same rule of applicable law, without discriminating between national and for- eign works). Another point is that mutual recognition requires countries of reception to apply the lex originis to questions of infringements of their own territorially granted copyright. One can doubt whether any law other than the lex protectionis can logically apply to infringements, because that law obviously has the closest connection to the territorially granted copyright that it seeks to protect: R. Fentiman, ‘Choice of Law and Intellectual Property’, in J. Drexl and A. Kur (eds.), Intellectual Property and Private International Law – Heading for the Future, Hart Publis- hing, Oxford 2005, p. 129 – 148. 35) An overview of estimates is given in A. Vuopala, ‘Assessment of the Orphan works issue and Costs for Rights Clearance’, Europe- an Commission, DG Information Society and Media, February 2010, available at http://ec.europa.eu/information_society/ activities/digital_libraries/doc/reports_orphan/anna_report.pdf. 36) J. Band, ‘The Long and Winding Road to the Google Books Settle- ment’, 9(2) [2009] The John Marshall Review of Intellectual Pro- perty Law, p. 227 – 329, at p. 229. 37) Vuopala (2010), supra note 35. 38) A. Beunen, ‘De Google Book Settlement nader beschouwd en bekeken vanuit bibliotheken’ 2010 AMI (Tijdschrift voor Auteurs-, Media- & Informatierecht, p. 38 – 49, p. 48; Lüder (2010), supra note 20. name.qxd 16 39) M. Ficsor, ‘Collective management of copyright and related rights in the Digital Networked Environment: voluntary, presumption- based, extended, mandatory, possible, inevitable?, in: D. Gervais (ed.), Collective management of copyright and related rights, The Hague, Kluwer Law International 2006, p. 37 – 83, at p. 50 – 59. See also M. Ficsor, ‘Collective Management of Copyrights and Related Rights from the Viewpoint of International Norms and the Acquis Communautaire’, in: D. Gervais (ed.), Collective Management of Copyright and Related Rights, Alphen aan den Rijn, Kluwer Law International 2010, p. 29 – 74. 40) For Norway, see ss. 16a (extended collective licences for libraries, archives and museums) and 36 (extended collective licensing in general) of the Norwegian Copyright Act. Denmark has introdu- ced s. 50(2) into its Copyright Act, allowing extended collective licences for any subject parties might agree on. The Swedish law provides for extended collective licensing in s. 42a, but its appli- cation is currently limited to specific types of use, such as digital lending by libraries (s. 42d). However, a new extended collective licensing provision for online libraries is being considered: Beunen (2010), supra note 38. There is also a generally applicable system of extended collective management in Hungary. Article 3 of the Satellite and Cable Directive (93/83/EEC) provides for extended collective management of re-broadcasting rights. 41) This is the Bokhylla library project (http://www.nb.no/bokhylla). Nowegian libraries can obtain extended collective licences under ss. 16a and 36 of the Norwegian Copyright Act. The Bokhylla contract between the National Library of Norway and Kopinor (the literary collecting society), dated 23 April 2009, can be found at www.nb.no/pressebilder/Contract_NationalLibraryand- Kopinor.pdf. 42) Elferink & Ringnalda (2009), supra note 1, p. 79 (the information reported in this book is partly based on questionnaires completed by various copyright experts). 43) Para. 3 of the contract, supra note 41. 44) On multi-territorial licensing, cf. T. Woods, ‘Multi-Territorial Licensing and the Evolving Role of Collective Management Orga- nizations’, in Gervais (2010), supra note 39, p. 105 – 134. 45) Preamble 18. 46) D. Gervais, ‘Collective Management of Copyright: Theory and Practice in the Digital Age’, in Gervais (2010), surpa note 40, p. 1 – 28. Cf. also S. von Lewinsky, ‘Mandatory collective administrati- on of exclusive rights – a case study on its compatibility with international and EC copyright law’, e-Copyright Bulletin, January-March 2004, arguing that mandatory collective manage- ment should not be considered to be an exception or limitation under international or European copyright law. 47) D. Gervais, ‘The Changing Role of Copyright Collectives’, in: Ger- vais (2006), supra note 40, p. 3 – 36. On formalities and their function in the digital era, see S. van Gompel, ‘Formalities in the digital era: an obstacle or opportunity?’, in: L. Bently, U. Suthersa- nen & P. Torremans (eds.), Global Copyright: Three Hundred Years Since the Statute of Anne, from 1709 to Cyberspace, Chel- tenham: Edward Elgar 2010, p. 395 – 424; S. van Gompel, ‘Les formalités sont mortes, vive les formalités! Copyright formalities and the reasons for their decline in nineteenth century Europe’, in: R. Deazley, M. Kretschmer & L. Bently (eds.), Privilege and Pro- perty: Essays on the History of Copyright, Cambridge: Open Book Publishers 2010, p. 157 – 206. 48) Para. 7 of the contract, supra note 41; see also Lüder (2010), supra note 20 p. 684. 49) See supra note 40.