PROPERTY RIGHTS COASES THEOREM IPR DATABASE PROTECTION Yale

PROPERTY RIGHTS, COASE’S THEOREM, IPR & DATABASE PROTECTION Yale M. Braunstein School of Information Mgt. & Systems University of California, Berkeley 2004 1

Today’s Presentation 1. Economic logic for IPR & “Coase’s Theorem” 2. Economics of database protection 3. Special areas of concern 4. Conclusion 2

Economic principles that lead to the case for intellectual property rights Ø Difficulty in excluding non-payers Ø Presence of economies of scale § Alternate specification for information: problems in alienability and exclusion Ø Together, these lead to “public goods” § But this is not an all-or-nothing concept – Difficulty in exclusion impossibility – Varying degrees of economies of scale Possible digressions: lighthouses, libraries, toll roads, etc. 3

Non-economic Rationales for IPR Ø Required by the Constitution § § Article I, Section 8, Clause 8 The Constitution history and the first U. S. Copyright law can be viewed as extensions of the Statute of Anne Ø Judeo-Christian tradition of private property 4

Coase’s Theorem* provides an economic logic for a clear definition of property rights in intellectual (and all other) property. The original article actually has two conjectures: 1. Market failures arise from inadequate assignment of property rights (correct) 2. Final allocations are independent of the initial assignment of rights (probably incorrect**) * Ronald Coase, "The Problem of Social Cost, " Journal of Law and Economics, 1960 ** For a proof, see: http: //www. sims. berkeley. edu/courses/is 231/f 02/is 231 -Coase. htm Note the link between this second conjecture and the incentive effect we discuss later. 5

Motivation for Reports on DB Protection Ø Sui generis protection of databases in EU countries pursuant to Directive 96/9/EC Ø Proposed legislation in US following Feist decision Ø WTO & TRIPS Ø Developing countries seeking guidance from World Intellectual Property Organization § UN-affiliated organization in Geneva; serves as secretariat for the Berne Convention § WIPO commissioned five “expert” studies in 2001 -2002 6

Details on the Subject Ø The analysis applies to developing countries at all stages of the development process, to the least developed countries, and to countries undergoing the transition from a centrally-planned to a free-market economy. § Some of the analysis has wider applicability. Ø Focus is on non-original databases: they do not meet the originality criterion under the principles of copyright recognized in the Berne Convention and the WIPO Copyright Treaty. § As a result these databases are not protected by copyright (although the content of specific records may be). 7

Table of Contents of the Report 1. Introduction 2. Economics of database protection 3. Developing countries as producers and users of databases: data & analysis 4. The problem of “competing” protection regimes 5. Special areas of concern 6. Conclusion 8

Economics of Database Protection Two themes: 1. Economics of DB Production 2. Understanding the Economics of Pricing 9

The “incentive effect” Ø As a society we want to encourage the development of the arts and advancements in technology. Ø There are two ways to do this: § Grants to artists, scientists, and other creative people § Provide incentives via property rights [Are these the only ways? ] Ø How do we “know” the second approach will work? § The US Constitution (Art. I. , Sec. 8. , Clause 8) § Joseph Schumpeter 10

Implications of economics for structuring protection of databases Ø The applicability of principles of incremental cost pricing to the pricing of databases is severely hampered for two basic reasons: 1. 2. In most cases, the cost structure in the database industry is generally such that with prices set at the incremental cost of providing an additional user with the use of a database, the production costs will not be covered. Once a property right is granted to the producer of a database, he is vested with a certain degree of market power. A monopolist has no incentive to set prices at incremental cost. The question then arises as to the kind of pricing policies that the database producer should be allowed to set. 11

Setting prices to cover costs Ø The solution to this problem that marginal-cost pricing is economically “efficient” but not sustainable in the presence of economies of scale is to employ what are known as “quasi-optimal” or “Ramsey” prices. Ø Possibly the clearest statement of the Ramsey approach comes from the pricing rule: § For each product, the percentage deviation of quasi-optimal price from marginal cost must be inversely proportional to its price elasticity of demand. [Baumol and Bradford, 1970] 12

Application of the pricing rule - I Ø Assume either of the following: § § A database is sold both to profit-making firms as well as to nonprofit research institutes. Or it is sold to users in developed countries and those in countries whose economies are still developing. Ø For this analysis, what distinguishes those two groups is not their financial goal or the income level of their country but the responsiveness of their demands to changes in the price for the database. § In economics the measure of responsiveness of demand to price changes is known as the elasticity of demand. 13

Application of the pricing rule - II Ø A remarkable result in economic analysis is that the following “properties” hold for the optimal prices, regardless of whether one is seeking to maximize societal benefits OR maximize the profits of the database producer: § the database should have a (non-zero) price in each market § both prices should be greater than the marginal cost of providing the database to the user § the price should be lower in the market with the higher elasticity of demand 14

Implications of the pricing rule Ø The following principles then result: § Unless there are strong reasons to the contrary, the recovery of the fixed cost component of the total costs should be spread over as wide a set of customers and users as feasible. § Exempting some users from payment or arbitrarily restricting the charges that can be assessed against them leads to increased costs and charges to the less-favored users. § Compulsory licensing and other artificial restrictions on prices reduce the profits of those products that reach the market. However, the uniform prices that result may not be socially desirable. Furthermore, the total effect of such restraints may be undesirable if it impairs the flow of innovations and new databases. (Larry Lessig take note. ) 15

Special Areas of Concern One basic theme: 1. Don’t mess with the general approach unless it is really necessary 16

Special areas of concern Ø Databases with scientific and technical information (STI) or cultural content Ø Sole-source Ø Capture Ø Exclusive agreements for the commercial distribution of public data Ø Internet routing tables 17

Databases with STI or cultural content - I Ø Each of the following arguments has been applied, in one context or another, to databases: 1. “Certain types of information need to be widely disseminated and, as a result, property rights for that information should be prohibited. ” 2. “Certain information is the cultural patrimony of a nation and should not be treated as an economic commodity nor made available for commercial exploitation. ” – Variations of this argument are that the information should not be used by any commercial or for-profit entity, that it should not be used outside of the country or region from which it came, or that it should remain under the control of certain specific groups. 18

Databases with STI or cultural content - II Ø To a large degree these questions are not about economics (and often not about databases !). Ø But they do have an economic dimension: § § § In many cases, one can view them as applications of the extremes in pricing: prohibiting protection is the analog of setting the price at zero, while blocking distribution is the analog of having an infinite price. Intriguingly, both policies have the same effect on the revenues of the database producers—the revenues will be zero. In one case this is from no price; in the other it is from no sales or license fees. The long-run result of no revenues, as we have discussed previously, is little or no incentive for the production of the affected databases. 19

Databases with STI or cultural content - III Ø Some people and organizations create intellectual property for motives that are either not profit oriented or are, at most, only indirectly related to the possibility of financial reward. Others are motivated by the prospects of financial gain. It is sometimes argued that there is no need for IPR to motivate those in the first group. § While this may be true, it does not mean that it is in society’s interest to prohibit the exercise of IPR. The incentive effects lead to the creation of additional STI and cultural content by members in the profit-seeking second group, while not causing less to be produced by the first group. Ø Furthermore, even under a strict IPR regime, those in the first group are not prohibited from placing their output into the public domain. 20

Databases with STI or cultural content - IV Ø Two other arguments: 1. There are few, if any, prospects for the development of databases with STI or cultural content in developing countries, or in small countries. As a result, the only beneficiaries of rights protection will be the database producers in the developed countries that have large, rich markets. – Counter: See the data – (Possible digression: parallels with the debate over patent rights for medicines) 21

Databases with STI or cultural content - V 2. Cultural patrimony and indigenous scientific knowledge are or should be the collective property of a specific culture, group, or nation, akin to communal grazing lands. Therefore, it is reasonable that members of the community use the material freely while non-members, such as those in developed countries, pay for use. § Counter: Intriguingly, this approach can be easily seen as consistent with the lessons of optimal pricing. To the extent that cultural patrimony and scientific knowledge are organized in databases of the sort under discussion, the establishment of rights in the databases can accommodate a dual market/dual price approach. This is directly analogous to having a compulsory or blanket license at a low or zero price in the home market. 22

The case of genetic information in Iceland - I Ø Background § § The firm of de. CODE Genetics is reported to have a 12 year agreement for the exclusive marketing rights to the database of genetic information of the entire (rather homogeneous) population of Iceland (approx. 278, 000). If they relied solely on trade secrecy protection, the publication of parts of the database could result in the loss of their property rights. On the other hand, the EU sui generis approach allows for publication of records and sets of records without such a loss of rights. (Note: not a member of the EU. ) [Other, probably less controversial, examples from ECAI in the report] 23

The case of genetic information in Iceland - II Ø Policy discussion § § An interesting example of a case where the discussion of the desirability of db protection has become intertwined with other political and social issues. The creation and marketing of this database is controversial, with questions being raised about issues such as the “commodification” of genetic information, the freedom of scientific inquiry, and privacy. – The first two issues are variants of the questions raised above about the prohibition of protection and the blocking of distribution, but they are sometimes cast in moral terms, which I shall not discuss, and sometimes linked with concerns over monopoly power, which will be addressed explicitly in the sections on sole-source and capture below. 24

Sole-source - I Ø Important to distinguish between databases for which there is no substitute of any sort and those that have no close substitute. § The two defining characteristics are the closeness of the substitute, which can be measured in economic terms by the additional costs one might have to bear to use the substitute rather than the “original, ” and the importance of encouraging widespread distribution and use. 25

Sole-source - II Ø Need to distinguish between databases that, for some physical or legal reason, cannot be reproduced and those that are economically impractical to reproduce. § Databases in the former group might include geophysical data from a single monitoring station or the medical procedure codes required by government health agencies. – Databases of this sort might be viewed as being comparable to the “essential facilities” of antitrust law. 26

Sole-source - III Ø Two policy options consistent with keeping the incentives inherent in the definition of IPRs: § Compulsory licensing provisions or similar restrictions and the use of anti-monopoly laws – For the purposes of this discussion, the problem with compulsory licensing is not so much the loss of the incentive effect but, rather, the difficulty in making a statute that is narrowly targeted to affect only databases that are truly subject to the sole-source phenomenon and, therefore, have no substitutes. § The second approach is to rely on the use of antimonopoly laws. – Most appropriate in the case of irreproducible databases. – Possible digression: Rule of law in developing countries 27

Capture - I Ø Issue: § Possible “capture” of public-domain data by adding value and including these data in a commercial database – Related to the sole-source issue in that the existence of reasonable substitutes is a key factor. – If the underlying data continue to remain available and are a reasonable, albeit imperfect, substitute for the commercial database, capture is of little concern in practice. – And remember, it is the uniqueness—whether in terms of organization or ease of use—provided by the database that gives value to its users, and this perceived value is, in turn, the incentive for the creation of the database. 28

Capture - II Ø The capture problem is one of degree, as with the sole-source problem, so there is no single solution that is guaranteed to simultaneously provide a remedy and have no detrimental effect on the desirable incentive outcomes. § The best approaches seem to be measures that insure the continued availability of the content from the original sources and requirements that database producers who incorporate substantial amounts of public-domain content provide a reasonable notice of the source to potential users. 29

Exclusive agreements for the commercial distribution of public data Ø Government agencies and commercial database producers enter into agreements that grant the commercial entity the exclusive rights to produce a database containing certain sets of public data. § One can view these arrangements as combining features of both the sole-source and capture situations described above. Ø Arguments for: These agreements may be necessary and in the public interest because: § § the government is not capable of, or—for philosophical reasons—should not be in the business of marketing products with commercial possibilities, and exclusivity is required to guarantee a sufficient return to warrant the necessary investment by the private party. 30

Exclusive agreements for the commercial distribution of public data Ø Arguments against: § § § the government is giving away its “crown jewels” it is picking winners instead or relying on the market this approach creates a de facto monopoly, resulting in unnecessarily high prices to users of the database. Ø These are not arguments against property rights in databases, per se. Rather, they are criticisms about the contractual terms employed in the exercise of those rights. § The appropriate solution is to develop contractual and licensing terms in each situation that are consistent with national policy objectives. 31

General comments Ø Many of the problems raised in discussions of issues related to sole source, capture, and exclusivity situations revolve around the extent to which the underlying information is available to the public. Ø Certain information may be formally available to the public to meet various legal obligations, but the question should be whethere is sufficient public access. Ø The problem often pre-dates the exclusivity agreement and should be dealt with directly rather than via limitations on database rights. 32

Internet routing tables - I Ø A set of concerns surrounding the databases used in maintaining the Internet’s infrastructure. The specific issues are hard to pin down, but seem to include: 1. Domain name registrars will consider the data that links domain names and IP addresses proprietary 2. Backbone providers and vendors of routing hardware will …[do something bad] 3. Someone in the first group will merge with someone in the second and together they will …[do something bad] 33

Internet routing tables - II Ø Issue 1: Domain name registrars will consider the data that links domain names and IP addresses proprietary § It is difficult to imagine how this might actually occur, as those registering their domain names do so precisely so that Internet users can find their sites. § If one or more domain name registrars refrained from making the mapping information public, users would seek out other registrars who meet their needs by freely distributing this information. § If the worry is that the existing registrars would collude by only cross-licensing each other and thus effectively barring entry of competing registrars, the solution is to address that issue either directly or via the antitrust laws. 34

Internet routing tables - III Ø Issue 2: Backbone providers and vendors of routing hardware will use db protection laws to control the distribution of routing tables they develop for their own use. These tables may be optimized for the specific architecture of the networks on which they are used or designed along with other firmware. § § § This is something that should be encouraged. * One way in which the producers of routing hardware or the operators of networks compete is by offering more efficient products and services. (* “Hot potato” problem) It is also possible that effective protection is already available via trade secrecy and that db protection laws will have no effect one way or the other. In any event, so long as the original, non-optimized data remain available, this should not be a public-policy concern. 35

Internet routing tables - IV Ø Issue 3: A domain name registrar will merge with a backbone provider and/or a hardware vendor and together they will do something exclusionary. § (I have not seen this argument; it is a “straw man. ”) § This also seems difficult to imagine because of the competitive pressures described in the discussion of Issue 1. – Nevertheless, some might fear this chain of events, citing the general reluctance of antitrust authorities to attack vertical mergers. – But, if this situation were to arise, it seems likely that there would be strong calls for direct action on the tying and exclusive-dealing aspects of the arrangement, regardless of the general views on vertical mergers at the time. 36

Conclusion ! Ø Overall logic of the report in six steps: 1. 2. 3. A clear, adequate definition of property rights in databases enables markets for databases to develop and grow. Strong protection of the IPRs in databases, without too many limitations or exemptions, will encourage the growth of local production of databases in developing countries, some of which already exists. Recognition of parallel rights foreign producers of databases enables trade that is in the interest of both parties. [Not covered today. ] 37

Conclusion ! Ø Overall logic of the report in six steps: 4. The existence of differing approaches to the protection of databases across the developed counties is not likely to be a problem for the developing countries. [Not covered today. ] 5. Policies designed to deal with special circumstances such as sole-source provision or the possibility of capture should be narrowly drawn. 6. It is important to recognize throughout the discussion of database rights that there is a distinction between the unoriginal databases that are the subject of this analysis, on one hand, and their content, on the other. The content in the database fields and records may be original works already subject to copyright protection or work that is in the public domain. 38