Data, and copyright in its teenage years
This article appeared in the European Intellectual Property Review (I think) sometime in 2019. It takes a look at the EU Database Directive and argues that the law is incoherent. Its reasoning applies to all copyright and pseudo-copyright protections of data.
Footnotes are at the end.
Data is big, and getting bigger. The OECD estimates that in 2015, the global volume of data stood at 8 zettabytes (8 trillion gigabytes), an eight-fold increase on 2010.[1] By 2020, that volume is forecast to increase up to 40 times over, as technologies including the Internet of Things create vast new data sets. Across the developed economies, governments are focusing on the importance of data and re-assessing whether the existing structures are optimally aligned for a digital age.[2]
Why is it then, that in the European Union, we have a law that discriminates against efficient collectors of data? The EU Database Directive[3] provides that the sui generis right will attach to those databases where there has been “qualitatively and/or quantitatively a substantial investment in either the obtaining, verification or presentation of the contents”. That means for a database (call it Database A), there is a minimum level of investment (call that X) below which Database A will not qualify for the sui generis right.
If I create Database A, and my level of investment is X or greater, my database is protected by the sui generis right. However, if my rival creates exactly the same database (call it Database B) but does so with an investment that is below X (X being, ex hypothesi, the minimum level of investment), her database will not qualify for the sui generis right. Databases A and B are exactly the same, but B was produced at half the cost of A: Database A is protected, Database B is not, and the only reason that Database B fails to qualify is because it was more efficiently produced.[4]
In other words, the more efficient the creator of the database, the less likely it is that her database will qualify for sui generis right. How could the EU legislators have produced a result that is so palpably absurd?
The answer is that copyright is still in its teenage years. Unlike its older brother, property in tangibles, which has had much longer to iron out its kinks and inconsistencies, copyright is not coherent, is not fully understood, and is spotty. As a result, a number of the concepts underlying our present understanding of copyright do not stand up to scrutiny and need to be revisited.
This article looks at a number of rules which commonly exist in copyright and, more interestingly, at the reasons behind some of those rules – the tectonic plates of copyright – and proposes a number of changes to the rules of copyright and the sui generis right which will put both on a footing that is more appropriate for a digital age.
For the purposes of this article I will not differentiate much between the copyright and the sui generis: the sui generis right is a form of copyright, but just wearing a different set of clothes.
PART A DATA – THE SUI GENERIS RIGHT
The investment threshold
I have set out above how an investment threshold discriminates against the more efficient producer. Clearly, that is not a sustainable position for any society in a digital age. But there is another, equally fundamental reason, to object to an investment threshold. The investment threshold functions as proxy for the merit or value in the database. An investment at or above the threshold suggests that the database has merit, and is therefore worthy of being propertised. Similarly, an investment below the threshold suggests that the database has insufficient merit, and is therefore not worthy of being propertised.
In other words, we are relying on the legal system to make an assessment of the value or merit of the database. But this makes no sense. Why should we, as a society, require the legal system to make this assessment, when we have a market which can do this much more effectively? We do not take this approach with tangible property: we would not expect a legal system to decide whether car Y is better than that car X, and award propertisation (or vary propertisation) on that basis; we rely on the market to differentiate between the two cars.
Asking the legal system to make this assessment usurps the role of the market, and requires the legal system to perform a function where it has no useful role to play. To put it another way, when the legal system decides whether or not to grant the status of property (i.e. the sui generis right) to a database, the merit or value of that database should not be a relevant consideration. Why? Because markets assess merit and value far more efficiently than courts ever can. This is picked up again below.
Corrective: no investment threshold.
However, it is always sensible to guard against unintended consequences. What unintended consequences could flow from removing the investment threshold from the sui generis right?
Unintended consequence – increased number of worthless database. This argument runs as follows: if there is no qualifying threshold, then this a) will lead to an excessive number of worthless databases, and b) an excessive number of worthless databases is an undesirable outcome. And the answer to a) is: yes, quite possibly. But there is a good reason not to be concerned about b) because this exactly what we already have in copyright, and no major ill effects result from this. Every month, millions of emails are produced which are of no use or interest except to those sending and receiving: the bulk of these are copyright works. For society at large, these are trivial works, but no ill effect results from their huge number.
Unintended consequence – monopolisation of facts. This argument runs as follows: copyright in facts is inherently monopolistic, and therefore only by having a qualifying threshold is there a suitable balance between the monopolisation effect and the benefit of the property right. One of the classic statements of the monopolisation of facts is in Feist, where the US Supreme Court stated that “to accord copyright protection [to facts] distorts basic copyright principles in that it creates a monopoly in public domain materials”[5] (emphasis added).
But this is plain wrong. Leaving aside the Supreme Court’s incorrect use of the word monopoly[6], it is the nature of most facts to exist in a source which cannot be exclusively appropriated: viz. the world about us. Mere collection and compilation of such facts will rarely establish a “monopoly” because, where the facts are propertised by copyright, it is always open to others to go direct to the original sources. However many compilations are produced then, so long as the real world source remains publicly accessible, there is, to use Locke's phrase, always enough left over. In more formal terms, what the analysis in Feist ignores is the distinction between source and version. A database of facts is a version: the facts themselves still exist in the source. London itself is the source for a list of all the companies based in London, while the database (or list, or compilation) itself is a version. Including facts in a database does not and cannot lead to the sort of “monopoly” envisaged in Feist, since the source of the facts continues to pre-exist and remains as available to second-comers as it was to first-comers. What is source for the goose is also source for the gander.
Unintended consequence – restriction on the free flow of facts
Similar to the “monopolisation” argument is the claim that property in facts (i.e. via the sui generis right or other versions of copyright) restrict the free flow of information. The weakness of this argument is that it fails to recognise the distinction between ex ante and ex post. Ex ante, a property right encourages the production of thing in question because it allows the producer to obtain a return on the market and reduces distribution costs (because most people respect property). Ex post, a property right appears to restrict the distribution of the thing in question. Except that, since production must precede distribution, the property right will generally increase the availability to all.
A simple example from the world of tangibles will illustrate this point. Will removing property from food increase the distribution of food? Yes, in the short term, because we can all go to supermarkets and fill our shopping bags without paying. In the medium and long term, clearly not, as farmers and food distributors find it increasingly difficult to make a return on their investment. This argument is doubly true with facts where the property protects the version but leaves the source accessible to all.[7]
Unintended consequence – databases created as a byproduct. In both Market Fixtures[8] and British Horseracing[9], the production of the list of forthcoming matches and forthcoming horseraces were seen as representing insufficient investment for the sui generis right to arise. Each company had to arrange the football matches and the horse races anyway: the fact that this activity generated, as a byproduct, a database was not, of itself, sufficient for the sui generis right to arise in relation to that database. The relevant investment had to be in the obtaining, verification or presentation of the contents: investment in the creation of data, however substantial, was not relevant.
On the assumption that the allocation of investment in Market Fixtures and British Horseracing was correct, the conclusion must that the producers of those databases were highly efficient: in both cases, valuable databases were produced with negligible investment.
Previous sections have argued that the investment test does not work because it penalises efficient producers. Is this conclusion any different where the database is a byproduct?[10] If we look at the world of tangibles, there are many processes where there is a primary product and also a byproduct, yet the issue of propertisation (or dis-propertisation) does not arise. Admittedly, the logistics of propertisation (or dis-propertisation) would be hard to administer to tangibles, but what is interesting (from the perspective of the reasons behind the rules) is that the property status of byproducts is never even considered in the world of tangibles. No one considers it plausible to argue that bran (a byproduct of the milling of flour), or molasses (a byproduct of refining sugar), or chicken feathers (a byproduct of raising chickens) should be treated differently, in terms of their status as property, merely because they are byproducts.
Is there anything special about databases such that a different approach is required? It is hard to see one. There is also at least one good reason for treating data byproducts and physical byproducts the same when it comes to propertisation, and that is that the classifications of primary product and byproduct are essentially arbitrary. For example, if product A produces 90% of a company’s revenues and product B produces 10% of its revenues (both A and B being outputs of the same process), A is the primary product and B is the byproduct. However, if the market conditions change such that product A now brings in 10% of the revenues and product B brings in 90% of the revenues, then product B is the primary product and product A is the byproduct. For example, in the 1950s, Arsenal’s revenues for gate and match-day income were substantial, and its revenues for TV and broadcasting were negligible. In 2016/2017, Arsenal’s gate and match-day revenues were £100m, whereas its TV and broadcasting revenues were £191m. Which is the byproduct?
Any system of property which grants propertisation on the basis of a product being a byproduct or a primary product is essentially unstable and, over time, unworkable.
Sui generis right term of protection
What other changes are required to the sui generis right to make it appropriate for a digital age? The term of protection provided by the sui generis right is 15 years.[11] In contrast, the term of protection for works of copyright is 70 years from the death of the author. Other things being equal, investors are likely to prefer to invest their resources in areas that have a longer period in which to make a return. Giving data (compared to copyright) a reduced term of protection is hardly a sensible approach in any society that recognises that data is an increasingly important part of its economy.
Assuming the sui generis term of protection was extended to 70 years pma (or its equivalent), what unintended consequences might result?
Unintended consequence – excessive number of worthless databases
A potential reason for giving the sui generis right a shorter term is that databases are unlikely to have much value after 15 years. This looks like a difficult issue to assess: after all, how can we know how society will value a particular database (or databases generally) in, say, 30 or 50 years after their creation? The good news is that we do not have to make the assessment. If databases lose all value 15 years after creation, then we are back in the “excessive number of databases” scenario set out above: i.e. we have recreated the situation we have with the vast majority of copyright works, and no harm to society results. If, on the other hand, there is still value in databases after the 15-year term, then it is only economically sensible to allow the market to continue to allocate that value.
Unintended consequence – reduction in the public domain
The standard argument for a limited term is that, once the term has expired, the work goes into the public domain and is free to all, to the overall benefit of society. This argument is examined further below.
Corrective: increase the term of the sui generis right so that it matches that of copyright.
PART B FIXING COPYRIGHT FOR A DIGITAL AGE
What other changes are required to make copyright fit for a digital age? So far we have looked at specific changes in the rules surrounding the sui generis right, but there are also changes that need to be made at a more fundamental level if copyright is to function effectively. These are not primarily changes to the rules of copyright, so much as changes to the tectonic plates that presently underpin copyright. The rest of this article looks at:
· the requirement of originality
· the reward theory of copyright
· copyright as a positive right
· the term of copyright
The requirement of originality
For a product to qualify as a copyright work it must be a work within the meaning of the Berne Convention, and original. In the EU, the definition of originality has stabilised, by means of a number of Directives, at the author’s own intellectual creation. On its own this definition does not achieve the differentiation intended by the EU: for example, databases (irrespective of their structure or arrangement) meet the author’s own intellectual creation definition. They are intellectual artefacts (unavoidably, if something is qualifying work, then the word intellectual is a redundant part of the definition), they are created, and there is an author. Therefore the author’s own intellectual creation, despite the fact that “no tests as to the qualitative or aesthetic merits of the program should be applied”[12] and “no other criteria such as merit or purpose being taken into account”[13], has been amplified by a number of cases so as to require a reflection of the author’s personality:
…. the notion of the author’s own intellectual creation refers to the criterion of originality … the criterion of originality is satisfied when … [the] … author expresses his creative ability in an original manner by making free and creative choices and thus stamps his “personal touch” … By contrast, that criterion is not satisfied when … [the work created is] … dictated by technical considerations, rules or constraints which leave no room for creative freedom.[14]
There are two problems with this approach. Firstly, it requires a human author. A work which is created by a non-human producer – whether by a computer or some other means – is not a copyright work. But why would we, as a society, make a distinction at the level of property between works produced by humans and works produced by other means? This is not a distinction that is made in the world of physical property. No one would find it sensible to make a distinction between the property in a vase made by hand and the property in a vase made by machine.
The key question is not whether a work produced by a human is different to a work not produced by a human, but whether that difference – if it exists – is relevant to propertisation. If two identical poems are created, one by a machine and one by a human, does it make any sense to allow one poem to be propertised but not the other? Given that the two works are indistinguishable from each other, the only difference is the means of production – human or non-human. If we then propertise the former over the latter, then the implicit assumption is that a human means of production has greater value or merit than a non-human means of production.[15] But, as set out above, this misunderstands the function of property and usurps the role of the market. It should not be the legal system deciding what has value, but the market.
Corrective: remove the requirement for copyright works to have a human author.
At various times and in various jurisdictions, the requirement of originality has been used to require copyright works to have above-average aesthetic merit, to be the result of skill and labour, and so on. In the EU, as set out above, qualitative or aesthetic assessments, merit or purpose no longer form part of originality (though it is worth noting that skill and labour survive in the investment requirement of the sui generis right). The surviving requirement, that the author’s personality should be visible in the work, is probably best understood as a test to differentiate works with low functional content (and therefore, ex hypothesi, high visible personality) from those with high functional content (and therefore, ex hypothesi, low visible personality). “By contrast, that criterion is not satisfied when … [the work created is] … dictated by technical considerations, rules or constraints which leave no room for creative freedom”.[16]
But why do we wish to differentiate them? If the issue is one of merit (works with high personality have more merit than those with low personality), then that usurps the function of the market, as set out above (as well as contradicting the EU’s approach that qualitative or aesthetic assessments, merit or purpose no longer form part of originality).
If the issue is types of work, then what is the policy reasoning that suggests that works of the high-functionality/low-personality type should be propertised differently to works of the low-functionality/high-personality type? It is not enough to say, “Well, that’s what copyright does”, because copyright is a form of property which can be adjusted – and which, if required, should be adjusted – to best reflect the needs of the society it serves. The preservation of this distinction is particularly odd given that, in the last thirty years, the bulk of economic growth in our societies has come from works and products that are high in functional content.
If the requirement for the author’s personality was abandoned (leaving merely that the product should be a work, i.e. an intellectual creation of the accepted type), what unintended consequences might result?
Unintended consequences resulting from abandoning the requirement of originality
Firstly, the extension of copyright to works which are low in author’s personality would have no impact on those works which are high in author’s personality: these would remain protected as they presently are.
Secondly, it is probable that a far larger number of works will qualify for copyright (though it is hard to be more certain without better data). There is unlikely to be any downside to this; this is the excessive number of works issue, discussed earlier.
Thirdly, will the extension of copyright to works with high functional content increase the amount of litigation because, being functional, they are likely to be more similar and therefore allegations of copying are likely to increase?[17]. It is hard to arrive at a firm view on this without data, but there is at least one reason to think that there would be no increase in litigation, and that is because the issue is self-regulating: the more that there is only one way, or only a limited number of ways, of writing (for example) about something, then the harder it is to draw the conclusion that any similarity between two competing texts must be the result of copying rather than independent generation. In other words, the element that favours similarity (functionality) is also the element that will limit competing claims.
Corrective: remove the requirement for copyright works to be original.
Copyright as a reward
At the heart of copyright are the notions of reward and incentive: According to Feist: “The decision should not be construed as demeaning Rural’s efforts in compiling its directory, but rather as making it clear that copyright rewards originality, not effort”[18] [emphasis added].
The notions of reward and incentive are buried so deep in copyright that they are rarely questioned. And after all, they appear to make good sense. Where the author has clearly contributed something to society, he or she should be rewarded. On the other hand, where the contribution is minor, then the author should receive a lesser form of property, if any.
But what happens if we apply the notions of reward and incentive to tangibles? For example, a baker bakes loaves of bread. If I told him that, as a reward, he could have property in his loaves, he would find me rather strange. If I then told him that, as an incentive to carry on baking, he could have property in any future loaves, he would find me stranger still.
Applying the notions of reward and incentive to tangibles produces odd results, yet it is an oddness which we tolerate quite happily when it comes to copyright. What it illustrates is that the function of property is neither incentive nor reward. Both are there, of course, occurring not at the level of propertisation, but at one removed, at the level of the market. And there they take the usual form – money. Having spent all day working, the baker does not finish with a contented sigh, thinking, “I have worked hard all day, and have reaped my reward in the property rights now vested in my bread”; he goes out and sells the loaves on the market.
Is the situation any different where the product is a work of copyright rather than a tangible? How can it be? Property is property, after all. An author may write a book in the expectation that his work will be copyright, but he certainly does not hope that the copyright will be the only reward. The primary function of copyright, like any other form of property, is not to provide a reward, but to provide a mechanism by which the reward can be appropriated. It “commodifies” the good and thereby makes it tradeable on the market (both in the primary market – i.e. the first sale or other exploitation, and on secondary markets). On its own, a property right has no value. If there is no demand for your product, then, however solid your property rights, you are not going to get a reward. On the other hand, if your product does have market value, property makes it easier to collect that value, and will usually increase the amount of value that can be collected. It may seem that the distinction between property being a reward, and property allowing a reward to be appropriated, is small but it is not: once you believe that the basis of copyright is reward, the next step is to differentiate between those works that have value and deserve the reward and those that have little value and do not. As set out earlier, this usurps the role of the market and requires the legal system to play a role for which it is poorly suited.
Although the EU’s need to harmonise copyright across a number of different jurisdictions has led the EU, as a byproduct, to the remove the most pernicious effects of the reward/merit approach, the reasoning still remains and underpins a lot of copyright and copyright-related reasoning. It underpins, for example, the investment threshold of the sui generis right.
Corrective: abandon the reward theory.
Copyright as a right
There is another, equally important reason, to abandon reward as an underpinning of copyright. Like most other intellectual property rights, copyright began as a privilege. As copyright became more common, its basis as a privilege became harder to sustain, and the justification for copyright shifted to reward and incentive. However, we have now reached a point where the creation of intellectual property is no longer exceptional but commonplace, and moreover, an activity from which a large number of people derive their livelihoods. It is time that we reconceptualised copyright (by which I include the sui generis right) as a positive right enjoyed by those that create the relevant works. After all, should producers not have the right to bring to market the work they produce? Most people would find that uncontroversial.[19]
Corrective: accept that producers of databases and Berne Convention works have a positive right to have their work propertised.
Term of copyright – an economic view
Earlier sections of this article have pointed out that investors prefer to invest their resources, other things being equal, in areas that have a longer period in which to make a return, and that giving the sui generis right a 15-year term, compared to copyright’s term of 70 years pma, is not a sensible strategy for a society that recognises data as an increasingly important part of its economy. This argument applies equally to the copyright term as compared to the perpetual term that applies to tangible property: giving copyright a limited term is not a sensible strategy for a society that hopes to get significant growth from the intellectual property sector. It simply distorts the market against intellectual property and in favour of tangible property.
Term of copyright – a rights-based view
The argument around differing terms of property also applies to a rights-based view. Why should producers of tangible goods enjoy a perpetual term to recoup their investment when producers of copyright works only benefit from 70 years pma? This hardly sounds fair. This imbalance is further compounded by the rivalrous nature of tangibles. The nature of tangibles is that, at any one time, the global stock of tangibles is finite: every tangible that is propertised involves a depletion of a finite stock of (what would have been) the public domain. If I have exclusive property in tangible X, then, by definition, you do not: the pool of tangibles is finite (for practical purposes) at any one time, and the more that is mine, the less can be yours. It is a zero-sum game. Copyright, on the other hand, is a positive sum and non-rivalrous. The creation of a copyright work or database does not deplete a finite stock: the number of potential copyright works and databases is infinite.
If we conceive as property as a subtraction from the public domain, then the greater the subtraction, the more restricted the property right should be. Applying this approach, copyright (no subtraction) should have a longer term of protection, whereas property in tangibles (subtraction) should have a shorter term of protection.
However, it is always sensible to guard against unintended consequences. What unintended consequences could flow from giving copyright a perpetual term?
Unintended consequence – a reduction in the public domain
The standard argument for a limited term of copyright is that, once the term has expired, the work goes into the public domain and is free to all. In other words, the temporary protection of copyright is a quid pro quo granted in return for the subsequent transition to the public domain and, the argument runs, this arrangement works to the overall benefit of society.
This argument is essentially an economic one. What is unusual about it is that it assumes an economic underpinning different to that which, in market economies, we generally accept as axiomatic. The standard reasoning behind market economies is Adam Smith’s invisible hand:
Every individual endeavours to employ his capital so that its produce may be of greatest value. He generally intends neither to promote the public interest, nor knows how much he is promoting it. He intends only his own security, only his own gain. And he is led by an invisible hand to promote an end that was no part of his intention. By pursuing his own interest he frequently promotes that of society more effectually than when he really intends to promote it.[20]
In other words, individual self-interest maximises the benefit to society as a whole. The key element here is, “He intends only his own security, only his own gain”: this pre-supposes that the relevant individual is able to appropriate his own security and his own gain (i.e. the return on the investment he has made). This is typically achieved by means of property.[21] The limited term followed by public domain removes the mechanism (property) that previously motivated the relevant individuals, and thereby disables Adam Smith’s invisible hand: it provides no replacement mechanism but, nevertheless, expects that an optimal allocation will result.
This seems inherently unlikely in itself, but its main weakness is that it assumes that the only investment required for getting a work from production to consumer is investment in the production of the work: in other words, that once the work has been produced and is the public domain, access to that work is free of cost to the consumer. However, if distribution (by which I include marketing, research, curation, dissemination, the creation of appropriate user interfaces, etc.) also requires investment, which it typically does, then the ex ante/ex post distinction made earlier will continue to apply. If investment in distribution is required, then, ex ante, giving property rights in the relevant works to those performing distribution will generally increase the supply and/or lower the cost of the relevant works even if, ex post, consumers have to pay to access them.[22]
Corrective: equalise the term of protection for copyright and tangibles.
PART C CONCLUSION
Copyright – and the thinking behind copyright – is still in its teenage years, and a number of fundamental changes are required to make it appropriate for a digital age. The theoretical foundation of copyright – reward – may have had a place in the past, but it is no longer fit for purpose. People who create works should have a positive right to have their works propertised so that they can bring them to market, and propertisation should not be contingent on the legal system’s assessment of elements, such as value (or proxies for value, such as investment), which are best assessed by the market. We do no need to use propertisation to differentiate between the high value and low value of works: we can let them all be property (provided they meet the criteria for being works), and the let the market filter the good from the bad.
Market distortions between different types of property should, as far as practical, be removed. Deciding on the optimal terms for different types of property is, given a lack of comparative empirical data, a difficult task, but at least two main principles are available. Firstly, equalising terms (to the extent practical) between the sui generis right, copyright, and property in tangibles, avoids distortions on the market. Secondly, other things being equal, products which are taken from an infinite pool of resources (i.e. products which are positive sum and non-rivalrous) should receive a longer term of protection than products which are taken from a finite pool.
The notion of copyright as exceptional works, which may have a had a place when copyright works really were few in number, no longer has a place, and needs to be replaced by a recognition of copyright works being ordinary and commonplace, in the same way as we accept physical property as ordinary and commonplace. This would also allow a move towards a more unified approach to databases and Berne Convention works of which the prime characteristic is that independent generation is allowed, and where the legal system is, at the level of propertisation, indifferent as to whether the main constituent of the property is expression, facts, functional description, or a mix of all three.
In 1959, CP Snow warned against the growth of two cultures (one scientific, the other literary and artistic) that did not communicate effectively with each other.[23] This distinction between author and scientist/engineer (itself a continuation of the gentleman v. tradesman distinction) continues in copyright to this day. The droit d’auteur tradition of copyright prioritises the archetypal works of copyright above the functional and the utilitarian, even though it is the functional and utilitarian that have provided us with most of the new benefits we have enjoyed in the last thirty years. It is hard to see how we can continue to maintain this flawed distinction in a digital age.
[1] OECD, ‘Data-driven Innovation: Big Data for Growth and Well-being’ OECD Publishing:2015, page 20.
[2] For example, for the UK, see The Economic Value of Data: Discussion Paper, HM Treasury, August 2018, and Unlocking Digital Competition: Report of the Digital Competition Expert Panel (Furman Report), March 2019.
[3] Directive 96/9/EC of 11 March 1996 on the legal protection of databases.
[4] This argument works in all scenarios. If there is an investment threshold. it is always possible for a producer to produce at a level of investment below that threshold. The argument holds true for all values of X except where the value of X is zero: in which case, there would be no threshold.
[5] Feist Publications Inc. v. Rural Telephone Service Co. Inc., 113 L. Ed. 2d 358 (1991).
[6] Monopoly is a conclusion which follows from an economic assessment of relevant product market, relevant geographic market, market structure, and so on. In the EU, there is a rebuttable presumption of dominance if a company has a share of the relevant market which exceeds 50%. Few copyright owners are so lucky.
[7] The approach by version and source also suggests that copyright in facts is most likely to lead to some form of “monopolisation” where the source of the facts is not equally accessible to all. See further below.
[8] Fixtures Marketing Ltd v Oy Veikkaus AB (C-46/02) EU:C:2004:694; [2004] E.C.R. I-10365; [2004] 11 WLUK 205 (ECJ).
[9] British Horseracing Board Ltd v William Hill Organisation Ltd (C-203/02) [2004] E.C.R. I-10415, [2005] 1 C.M.L.R. 15.
[10] Where privileged access to a database occurs, it can be a constituent part of a dominant position in the relevant market. E.g. Magill (Joined Cases C-241/91 P and C-242/91 P, Radio Telefis Eireann (RTE) and Independent Television Publications (ITP) v Commission, judgment of 6 April 1995). However, it is hard to see how this can justify a rule which automatically disqualifies all byproduct databases from propertisation when we have anti-trust rules that deal specifically with dominance. Equally, dominance can also occur (though less likely) where the database is not a byproduct. See, for the UK, Unlocking Digital Competition: Report of the Digital Competition Expert Panel (Furman Report), March 2019, which advocates a faster, pre-emptive approach to competition law in the digital sector.
[11] The Database Directive allows for extension of the term if a substantial new investment in the database is made. See Article 10.3, Directive 96/9/EC of 11 March 1996 on the legal protection of databases.
[12] Recital 8 of the Software Directive. Directive 2009/24/EC of 23 April 2009 on the legal protection of computer programs (codified version).
[13] Recital 16 of the Term Directive, in relation to photographs. Directive 2006/11/EC of 12 December 2006 on the term of protection of copyright and certain related rights (codified version).
Football Dataco Ltd v Yahoo! UK Ltd (C-604/10) EU:C:2012:115; [2012] Bus. L.R. 1753; [2012] 3 WLUK 1 (ECJ (3rd Chamber)).
[14] Football Dataco Ltd v Yahoo! UK Ltd (C-604/10) EU:C:2012:115; [2012] Bus. L.R. 1753; [2012] 3 WLUK 1 (ECJ (3rd Chamber)).
[15] This does not mean that we should never differentiate, at the level of property, between human-generated and non-human-generated products, but that any such differentiation should be part of an articulated policy.
[16] Football Dataco Ltd v Yahoo! UK Ltd (C-604/10) EU:C:2012:115; [2012] Bus. L.R. 1753; [2012] 3 WLUK 1 (ECJ (3rd Chamber)).
[17] This is the point made by Gary Lea in “In Defence of Originality” [1996] 1 Ent. L.R. 21 by way of a reply to Mark Sherwood-Edwards, “The Redundancy of Originality” [1995] 3 Ent. L.R. 94.
[18] Feist Publications Inc. v. Rural Telephone Service Co. Inc., 113 L. Ed. 2d 358 (1991).
[19] It may even be a legal principle in the UK. To give an example: the general common law position in the UK is that corpses and body parts cannot be property. However, “a human body, or a portion of a human body, is capable by law of becoming the subject of property. It is not necessary to give an exhaustive enumeration of the circumstances under which such a right may be acquired, but I entertain no doubt that, when a person has by the lawful exercise of work or skill so dealt with a human body or part of a human body in his lawful possession that it has acquired some attributes differentiating it from a mere corpse awaiting burial, he acquires a right to retain possession of it, at least as against any person not entitled to have it delivered to him for the purpose of burial, but subject, of course, to any positive law which forbids its retention under the particular circumstances.” [emphasis added]. Doodeward v. Spence (1908) 6 C.L.R. 406. Followed in R. v Kelly [1998] EWCA Crim 1578; (2000) 51 B.M.L.R. 142.
[20] Adam Smith, “The Wealth of Nations” (1776). Although Adam Smith's notion has subsequently been much qualified, his “basic arguments on the efficacy of free competition remain intact, a philosophical lodestar to nations relying upon a market system of economic organisation.” Scherer, “Industrial Market Structure and Economic Performance”.
[21] Though, clearly, other mechanisms are available. The importance of property is that, in addition to the convenience of a right in rem, it is generally backed by the state and so the costs of property protection to users are comparatively low.
[22] The cost to the consumer is not just the access fee the consumer may have to pay to a distributor, but also the cost of time spent looking for the thing in question. The greater the number of available works, then, other things being equal, the harder it will be to find the thing you are looking for and the greater the costs of accessing the public domain. The greater the number of available works, the more important curation and distribution becomes.
[23] CP Snow, “The Two Cultures”, Rede Lecture, 1959.