Constitutional Bounds of Database Protection: The Role of Judicial Review in the Creation and Definition of Private Rights in Information
Abstract
The article analyzes the constitutional bounds within which Congress is empowered to regulate the production and exchange of information by creating private rights. As a test case for answering this question, it analyzes the differences between two radically different bills for protecting database providers, reported in late September of 1999 by two committees of the House of Representatives. One bill, House Bill 354, is unconstitutional. The other, House Bill 1858, probably is constitutional. The article suggests that the Intellectual Property Clause and the First Amendment create two layers of judicial review over acts of Congress that recognize exclusive private rights in information. It suggests that judges have an important role to play as a counterweight to political processes in which the value of the public domain to future generations and to users is systematically underrepresented.
Table of Contents
Imagine Congress enacted a law that stated:
(1) Whoever publishes a news report, weather report, historical or scientific article, or other statement of fact, shall have an exclusive right to the use of the facts incorporated in that report and not previously publicly reported elsewhere.
(2) For a period of fifteen years following initial publication, no person may publish or distribute in interstate commerce facts subject to the exclusive right defined in paragraph (1) without the permission of the owner of the facts.
So Matt Drudge gets to say who may, and who may not, report on the Lewinsky affair. CNN and Time get to say who reports about the existence—or non-existence—of Operation Tailwind.
What prevents Congress from enacting such a law? Is it only prudence and the good judgment of its members? Are there constitutional constraints on congressional creation of exclusive private rights in information? If so, what are they?
The answer is that there are two cumulative constitutional constraints that prevent enclosure of the public domain so extensive as to be detrimental to our information environment. First, the Intellectual Property Clause imposes a set of threshold constraints on when it is that Congress may say that one person, and no other, may control this information, this knowledge, this piece of our common culture.1 Second, the First Amendment adds another layer of judicial oversight, to assure that even legislation that meets the threshold conditions, or is exempt from them because it can properly be passed under another power, does not impose too heavy a burden on our capacity to access information and communicate to others as autonomous individuals and as constituents of a democratic society.2
The question of constitutional constraints on Congress’s power to enclose the public domain is of immediate practical import. The past few years have seen an upsurge in enclosure legislation, as those who stand to gain from stronger property rights drive their fences deeper into the heart of our information environment, staking out ever-larger holdings.3 Of most immediate concern here are the database protection bills that have been introduced in Congress over the past few years. The stark differences between the iterations of this effort currently before the House of Representatives—House Bill 354, reported from the Committee on the Judiciary by Representative Coble, creating a property right in raw information in all but name, and House Bill 1858, reported from the Committee on Commerce by Representative Bliley, that assiduously avoids creating a property right—bring to the fore both a specific and a general constitutional question. The specific question is, what are the constitutional bounds on Congress’s power to delineate property rights in facts and in compilations of facts. This article will seek to answer this question by analyzing these dueling bills. The more general question is, what are the constitutional parameters that constrain the power of Congress to regulate the production and exchange of information and knowledge through the creation of exclusive private rights. I will answer this more general question first, to frame my analysis of the database bills.
Three constitutional reference points set the parameters of the answers to both the specific and the general questions. These are the Intellectual Property Clause,4 the First Amendment speech clause, and Congress’s near-plenary power under twentieth century interpretations of the Commerce Clause. There are three basic points to understand about these parameters.
First, the Intellectual Property Clause, at least since the Trade-Mark Cases in 1879,5 has been understood to be not only an empowering clause, but also one that limits to what extent and in what manner Congress can secure exclusive rights in information. Second, the reasons for the limitation in that clause reflect a substantive concern that Congress should not create exclusive rights in information and knowledge unless these rights fall within the constraints imposed by the Intellectual Property Clause. While congressional powers usually need not be mutually exclusive, this is not the case when Congress is engaged in creating private rights to control the production and exchange of information in society. Rather, Congress cannot use a power other than the Intellectual Property Clause—most importantly, its power to regulate interstate commerce—to enact exclusive rights inconsistent with the substantive constraints imposed by that clause. Third, the Intellectual Property Clause and the First Amendment interact to contain Congress’s powers to regulate the flow of information in our information environment. Cumulatively, they seek to assure that no one will capture the legislative process to privatize that most precious of all public domains—our knowledge of the world that surrounds us. That public domain is germane to our ability to decide for ourselves and talk to each other about how we ought to live our lives as individuals and as members of a community.
Part II explains the constraints that the Intellectual Property Clause and the First Amendment impose on legislation that creates exclusive private rights in information. Part III explains why one would want a robust system of judicial review of such legislation, in terms of the centrality of free information flows to democracy and autonomy, and in terms of the political economy of legislative creation of exclusive rights in information. Part IV analyzes the pending database bills in terms of the Intellectual Property Clause, and finds House Bill 354, but not House Bill 1858, unconstitutional. A careful analysis of the provisions of House Bill 354 suggests that it is functionally an intellectual property right in data. Under the Supreme Court’s interpretation of the Intellectual Property Clause, Congress lacks the power to enact such a right. Part V analyzes the two bills in terms of the First Amendment, and again finds the former bill deficient and the latter bill more defensible. There is no serious evidence that a right of the type created in House Bill 354 is necessary to further a significant public purpose. Furthermore, there is reason to believe that it would burden valuable speech too much for too speculative a gain in database production.
In 1879, the Supreme Court was posed with the following question: can Congress enact a trademark protection statute? In 18706 and 18767 Congress had passed statutes providing for registration and enforcement of trademarks. The Trade-Mark Cases involved three prosecutions under the latter statute, and the Court considered whether Congress had the power to enact these laws. Relying on the statutory language, the Court first determined that Congress had purported to act under the Intellectual Property Clause, not the Commerce Clause of Article I.8 Then the Court held that Congress could not, consistent with Article I, Section 8, clause 8 protect trademarks because trademarks lacked the originality that is the sine qua non of protection under that clause.9 It expressly reserved the question of the appropriateness of regulating trademarks under the Commerce Clause, which indeed later became the source of authority for the Lanham Act.10
It is important to focus on what the Court did and did not do in the Trade-Mark Cases. The Court did hold that the Intellectual Property Clause is not an open-ended grant of power to Congress to create exclusive rights in information, but a specifically limited grant available only to protect original contributions to the wealth of human knowledge. This limitation on the appropriate subject matter for property rights in information is constitutionally embedded, and is germane to the power of Congress to act in this field. But the Court did not, at least in these cases, state that the limitations in the Intellectual Property Clause applied to all congressional attempts to recognize exclusive rights in information.
From a contemporary perspective, the Trade-Mark Cases establish the important principle that the Intellectual Property Clause constrains congressional power, and that it is the proper role of judges to step in and tell Congress when its zeal to enact property rights exceeds its power to do so under the Constitution. The cases also establish the principle that originality is a threshold requirement before information can become the object of exclusive rights under the Intellectual Property Clause.
These basic principles have since been upheld systematically by unanimous or near-unanimous Supreme Court opinions.11 Two clusters of opinions—in the 1960s, and later in the late 1980s and early 1990s—developed and more clearly articulated the substantive constraints imposed by the Intellectual Property Clause on Congress’s power to legislate exclusive rights in information. These cases offer a robust defense of the importance of the public domain to the “Progress of Science and the Useful Arts,” and evidence an equally robust commitment by the Court to treat the Intellectual Property Clause as a structural constraint on Congress’s power to enclose the public domain.
The first cluster of opinions involves direct statements by the Court—speaking unanimously—about the constraints imposed by the constitutional provision itself. In 1966 the Court interpreted the nonobviousness requirement for patentability in Graham v. John Deere Co.12 The Court explained that the clause “is both a grant of power and a limitation.”13 In clear language, which it has reaffirmed since,14 the Court stated that “Congress in the exercise of the patent power may not overreach the restraints imposed by the stated constitutional purpose.”15 Specifically, the Court held that the Intellectual Property Clause requires that Congress (a) act only when extending an exclusive right promotes “[i]nnovation, advancement, and . . . add[s] to the sum of useful knowledge”16 and (b) not recognize exclusive rights “whose effects are to remove existent knowledge from the public domain, or to restrict free access to materials already available.”17 The Court reasoned that this unusual express limitation on the power granted in the same clause is a reflection of its framers’ aversion to a system of government grants of monopolies in trade—a strategy used by the Crown to reward its favorites.18
Explaining why patents fall within the conceptual ambit of disfavored “monopolies” rather than the presumably beneficial category of “property,” the Court referred to the writings of Thomas Jefferson, who as Secretary of State was one of the first to implement the patent system under the Patent Act of 1790.19 Jefferson was initially skeptical about the advisability of empowering Congress to provide for patents at all,20 but he later accepted their utility within bounds.21 His skepticism is most clearly reflected in a letter to Isaac McPherson, quoted at length by the Court in Graham.22 Here Jefferson presented not a generalized abhorrence for monopoly, but a very specific sophistication about information as a good.
Poetic though his language may be, Jefferson’s reasoning is entirely comprehensible and irresistible to the economically literate late twentieth century commentator. Information is a public good. It is nonexcludable and nonrival.23 Its nonexcludability makes it unsusceptible to appropriation, except by grant of exclusive right coupled with the threat of state enforcement; consequently, this characteristic is the potential justification for recognizing and enforcing such rights. But because information is also purely nonrival, the social cost of its use by many people, once it is produced, is zero. It is therefore desirable that information be at least partially nonexcludable, because that attribute permits free access to information by those who value it at its social cost—the value of their attention to it—but not enough to pay a positive price above that social cost. Unlike land, or cows, then, information does not lend itself to the simple assumption of the presumptive beneficial effect of treating it as the subject of property. The free exchange of information, rather than its enclosure, is the presumed beneficial state, and enclosures of information are thus like sometimes-necessary “monopolies,” rather than normally-beneficial “property.”
The upshot of the Court’s interpretation of the Intellectual Property Clause in Graham is that the clause expressly constrains the power of Congress to create exclusive rights in information. This limitation is based on substantive reservations that those who drafted the Constitution had towards granting anyone exclusive rights to use and benefit from information and knowledge. That the limitation reflects a substantive concern and not a formality is important. It suggests that it should apply to laws purportedly passed under powers granted to Congress outside of the clause itself. Otherwise, simple legislative drafting contrary to the constraints imposed by the Intellectual Property Clause, enacted under a different power, could too easily undermine the substantive commitments animating the constraint. Furthermore, the constraint is not primarily the textually obvious “for limited times” requirement. Rather, it is the requirement that exclusive rights in information actually can promote the Progress of science and the useful arts; there must be at least a plausible claim that a given right can actually increase information production.24 The clause also requires that Congress recognize exclusive rights only in exchange for adding to the wealth of public knowledge, and that it not grant rights that have the effect of removing information already available in the public domain.25
A quarter of a century later, the Court again revisited the constitutional constraints imposed on Congress by the Intellectual Property Clause, this time in the context of copyright. In Feist, a unanimous Court26 held that Congress cannot, consistent with the Intellectual Property Clause, constitutionally recognize exclusive private rights in raw facts. “Originality,” stated the Court, “is a constitutionally-mandated prerequisite for copyright protection.”27 Relying on the Trade-Mark Cases, the Court made amply clear that facts could not constitutionally be made the subject of copyright protection.28 The Court relied on the concept that facts preexist their statement by a human agent; being the first to articulate a set of facts does not make that person the facts’ originator.29 And only originators, or in the terms of Graham, only those who add to human knowledge, may receive an exclusive right in what they added.30 This, in turn, is understood by the Court to create a constitutionally mandated framework not only to assure authors rights in their original expressions, but also to “encourage[] others to build freely upon the ideas and information conveyed by the work.”31 As with Graham, the concern is substantive, not formal. It is a concern, treated as going to the very heart of the constitutional constraint, that Congress not enclose existing information and facts by granting exclusive rights. Three years later, in another unanimous opinion, Campbell v. Acuff-Rose Music, Inc., the Court reiterated its position in Feist, and implied that not only originality, but some form of “fair use” constraint on the scope of intellectual property is implicit in the constitutional constraint itself.32
Although Feist is couched in terms of “copyright,” it would be implausible to claim that the Court could have based its decision so heavily on the constitutionally protected status of public access to facts and information, but expected that the exact same exclusive rights that it invalidated could be enacted simply by inserting in the preamble words to the effect that the Commerce Clause, not the Intellectual Property Clause, is the source of power upon which Congress relies. The Intellectual Property Clause speaks of “exclusive rights,” not of the technical terms—well known to its drafters—of “patents” or “copyrights.” It grants and restricts Congress’s power to grant such exclusive rights. And the Court has repeatedly, and unanimously, stated that this constraint requires originality as a precondition to the availability of an exclusive right, and that it prevents Congress from creating exclusive rights in information that is already in the public domain—like facts, whether stated or not. Limitations are rare in Article I, Section 8 grants of congressional power—bankruptcy laws must be nationally uniform,33 an army can only be funded for two years.34 As others, like Jane Ginsburg,35 Malla Pollack,36 Theodore Davis,37 and William Patry,38 have exhaustively explained, Congress may no more enact an exclusive right in information that conflicts with the limitations imposed by the Intellectual Property Clause than it may fund an army for three or four years using the general spending power, or aid the employees of a single railroad by enacting special provisions for the distribution of the assets of their employer.39 To paraphrase Chief Justice Rehnquist on Congress’s attempt to circumvent the uniformity requirement of the Bankruptcy Clause, “to hold otherwise would allow Congress to repeal the [originality] requirement from Art. I, § 8, cl. [8], of the Constitution.”40
United States v. Moghadam41 is the only case directly to address an Intellectual Property Clause challenge to congressional legislation sought to be justified under the Commerce Clause. In Moghadam, the United States Court of Appeals for the Eleventh Circuit upheld the anti-bootlegging statute, which prohibits unauthorized recordings of live musical performances.42 The defendant challenged the constitutionality of the statute by arguing that Congress may only give authors and inventors rights to their “writings,”43 and since a live music performance is not fixed in a tangible medium, it is not “a writing.”44 The court held that whether or not legislation under the Intellectual Property Clause could only apply to works fixed in a tangible medium, the anti-bootlegging statute was valid under the Commerce Clause.45 At first glance, the case seems to fly in the face of the analysis in the preceding paragraph. On closer inspection, however, it does not.
The court in Moghadam clearly accepted the general relationship between the Intellectual Property Clause and the Commerce Clause, interpreting Railway Labor Executives to mean that there are “circumstances . . . in which the Commerce Clause cannot be used by Congress to eradicate a limitation placed upon Congress in another grant of power.”46 The court’s holding suggests that it interposed a quite reasonable gloss on this relationship—that only important or central constraints created by the Intellectual Property Clause, and not merely technical constraints, displace Congress’s power under other clauses of Article I, Section 8 of the Constitution. The anti-bootlegging statute deviated, according to the court, merely from the fixation requirement implied by the word “writings” in the Intellectual Property Clause, and that requirement is not so central to the Intellectual Property Clause as to negate a commerce clause-based law that gives copyright-like rights in live performances.47 Significantly, the court partly justified its holding by stating that an exclusive right to record live musical performances does cohere with the requirement of originality,48 which was presumably more important in the court’s eyes. It also expressly stated that its holding “is limited to the fixation requirement, and should not be taken as authority that the other various limitations in the Copyright Clause can be avoided by reference to the Commerce Clause.”49
Whether the case was correctly decided50 is less important than that the court recognized Congress’s power to create intellectual property-like rights under the Commerce Clause only if they do not negate the substantive constraints imposed by the Intellectual Property Clause on the enactment of such rights. Feist excludes the possibility that an exclusive right to information in a database could be treated as conflicting with the Intellectual Property Clause in merely a technical sense. Graham, Feist, and Acuff-Rose similarly foreclose an interpretation that would permit exclusive rights that more generally circumvent the originality requirement or enclose materials in the public domain to be passed under the Commerce Clause.
What power does Congress nonetheless have to regulate information markets after the possibility of using exclusive rights is removed? We know from the Trade-Mark Cases that the commerce power permits Congress to regulate certain uses of information beyond the scope of the Intellectual Property Clause—the Lanham Act is a case in point. What are the boundaries of this power, and how does it relate to the constraints imposed by the Intellectual Property Clause? These questions are answered, albeit indirectly, in a second cluster of opinions concerning the constitutional bounds imposed by the Intellectual Property Clause on state laws that are inconsistent with federal intellectual property laws. These opinions are one step removed from the question of direct constraints the Intellectual Property Clause places on Congress, because they involve the respective powers of the states and the federal Constitution. Nonetheless, in the course of working out to what extent the states are bound by the constraints imposed by the Intellectual Property Clause, the Court further articulated and strengthened the basic propositions set out from the Trade-Mark Cases to Graham and Feist. Namely, the Intellectual Property Clause imposes a constraint on what exclusive rights in information Congress can legislate. And the Court also outlined the types of information exchange regulations that would be justified under the Commerce Clause without running afoul of the substantive constraints created by the Intellectual Property Clause.
In 1964 the Court considered whether state unfair competition law could be used to create a general, property-like entitlement for unpatented or unpatentable designs. In Sears, Roebuck & Co. v. Stiffel Co.51 and Compco Corp. v. Day-Brite Lighting, Inc.,52 the Court interpreted the holdings below to have been “based wholly on the fact that selling an article which is an exact copy of another unpatented article is likely to produce and did in this case produce confusion as to the source of the article.”53 Because the Court saw these cases as equating confusion with copying, it interpreted them as creating, de facto, an alternative property-like right against copying of unpatented designs, rather than a contextually applied prohibition on causing market confusion. Such a right, in turn, “would interfere with the federal policy, found in Art. I, § 8, cl. 8, of the Constitution and in the implementing federal statutes of allowing free access to copy whatever the federal patent and copyright laws leave in the public domain.”54 The Court specifically limited its holding, stating that it did not exclude the possibility of state regulation of unfair trade practices, like preventing palming off by regulating labeling.55 But these laws could not be used in such a manner as effectively to negate the basic constitutional mandate that a design that is “in the public domain . . . can be copied in every detail by whoever pleases.”56 The emphasis on this limitation of the constitutional constraint on unfair competition law—that states retain the power to regulate “specifically oriented predatory business practices”—was the basis of Justice Harlan’s concurrence.57
The Sears and Compco cases came under pressure for, and eventually were abandoned with respect to, their preemptive effect on states. In Bonito Boats, Inc. v. Thunder Craft Boats, Inc.,58 the Court expressly abandoned the proposition that the Intellectual Property Clause directly prevented the states from “adopt[ing] rules for the promotion of intellectual creation within their own jurisdictions.”59 But in that same case the Court reiterated and revitalized the core holding of Sears and Compco, namely that states could not provide rights that offered patent-like protection, but undermined “the ultimate goal of public disclosure and use which is the centerpiece of federal patent policy.”60 In Bonito Boats the Court considered a Florida statute, purportedly concerned with unfair competition, that prohibited use of the most efficient method of copying the design of a boat hull—using an existing boat to create a mold, and then using the mold to create boats that copy the original hull design.61 The statute protected hull designs whether they were patentable or not.62 In this case, the manufacturer had not applied for a patent prior to marketing the boat and the hull design could therefore no longer be patented.63 The Court suggested that nothing prevented a state from passing a real unfair competition law, one concerned with “the protection of consumers, not the protection of producers as an incentive to product innovation.”64 The Florida statute was not that type of law. Rather,
[i]n contrast to the operation of unfair competition law, the Florida statute is aimed directly at preventing the exploitation of the design and utilitarian conceptions embodied in the product itself. . . . To accomplish this goal, the Florida statute endows the original boat hull manufacturer with rights against the world, similar in scope and operation to the rights accorded a federal patentee.65
Given the functional similarity between the protection offered by the Florida statute to unpatented, and unpatentable designs and patent protection, the Court held that the statute was preempted.66
In conjunction, the Graham-Feist and Sears-Compco-Bonito Boats lines of cases suggest that the Intellectual Property Clause imposes a threshold constraint on congressional legislation that regulates information production and exchange. Congress may not create rights that are functionally the type of exclusive right that it is empowered to create under the Intellectual Property Clause, except in compliance with that clause’s limitations. Under the clause, Congress may create such rights if and when they are likely to encourage information production more than they inhibit it, if it makes them available only to those who make original contributions to the wealth of our collective knowledge, and if the exclusive right enacted does not remove, or burden free access to, information already in the public domain. While Congress may regulate information markets under the Commerce Clause as well, it may do so only to the extent that it does not thereby circumvent the limitations placed by the Intellectual Property Clause on its power to create exclusive rights. As Ginsburg explained in her criticism of Feist, this limitation requires that regulations enacted under the commerce power be different in kind, not only in subject matter and degree, from the exclusive property-like rights that are the subject of the Intellectual Property Clause.67 This difference in kind, which is nowhere more clearly articulated than in the Court’s explanation in Bonito Boats of the differences between acceptable consumer protection-like regulation and unacceptable quasi-property rights.68 But when the substantive content and function of the right created by Congress crosses over the line of property-like protection (as some might argue it does in the case of the Federal Trademark Dilution Act of 1995),69 Congress’s actions stand on constitutional quicksand.
Even though a right in information can be constitutionally passed under the Intellectual Property Clause or the Commerce Clause, it may nonetheless involve a restriction on speech that renders it suspect under the First Amendment. This restriction arises because the enforcement of rights in information and cultural products always involves the state in refraining one person or another from using information or communicating in a manner that that person wishes to and could, but for the intervention of a law enforcement officer. The more context-sensitive judicial review that follows from this simple observation is discussed in the following section.
In 1970, Melville Nimmer wrote about the difference between recognizing exclusive rights in the proverbial Blackacre and recognizing similar rights in a novel like Black Beauty. The difference, he wrote, lay in the First Amendment.70 A good number of other commentators have since written about the inherent conflict between exclusive rights in information and the promise that “Congress shall make no law abridging the freedom of speech.”71
The basic conflict is analytically simple. A property right in an information good is created only if Congress prohibits most people from using certain information in certain ways. For example, in order for former President Ford to have a property right in the contents of his memoirs, the law must, at the end of the day, empower the sheriff to come to the door of The Nation magazine and confiscate its copies, or collect a fine, for writing an article that includes in it too many words that are copied from the Ford memoirs.72 Whenever the law permits the sheriff to walk into people’s offices and confiscate their publications, or levy against their belongings because of something they said or how they said it, the First Amendment is deeply implicated.73
The Nation’s quotation from the Ford memoirs was presented to the Supreme Court in Harper & Row. In that case, the Court held that in the context of copyright, the conflict between property rights and the First Amendment was false, because “the Framers intended copyright itself to be the engine of free expression. By establishing a marketable right to the use of one’s expression, copyright supplies the economic incentive to create and disseminate ideas.”74 I have elsewhere expressed the view that this statement should not be viewed as excluding First Amendment review of the actual contours of copyright law. The Court only resolved the conflict by reference to internal copyright doctrines that are increasingly under pressure—like fair use or the idea/fact-expression dichotomy implicated by the database laws.75 Here, I only briefly restate that argument.
First, standard economic understandings of the incentives created by intellectual property suggest that some degree of copyright protection will increase incentives of information producers over time. However, if producers have too complete a set of rights in their information products overall production over time will decline, because the increased costs that future producers must pay to present producers for using their products as inputs will exceed the value of expected returns to present producers. Therefore, Harper & Row should at the very least be read to mean that a properly-tailored copyright scheme is an engine of free expression, leaving room for at least some First Amendment review of the contours of copyright law, such as whether it includes a sufficiently robust fair use provision.76
Moreover, I have elsewhere explained why intellectual property rights prefer large-scale, commercial producers that integrate information production with management of owned information inputs over professional and amateur noncommercial producers, and to a lesser extent over small-scale commercial authors and inventors who do not own an inventory of information goods.77 This is so because the preferred organizations gain from increases in rights in a way that organizations that do not sell access to their products do not, and because they can better absorb the downstream input-cost-increase effects of intellectual property by reutilizing their own inventory.78 If this is in fact the case, then copyright is “the engine of free expression” for some, but not others. It is the engine for Disney, but a drag on the Electronic Freedom Foundation or on the Free Republic Forum, for instance.79 Since copyright provides incentives only to certain kinds of producers who have a limited set of reasons to put pen to paper, but imposes costs on many other types of writers, it is only in a very circumscribed form that it can in fact be “the engine of free expression;” it must leave enough room for political and cultural creation that is not driven by direct sales.80
But one need not revisit and revise, or even clarify Harper & Row to see that the First Amendment imposes distinctive constraints on the regulation of information production and exchange, both when the regulation falls within the Intellectual Property Clause, and most certainly when it falls outside of it.
[C]opyright’s idea/expression dichotomy [strikes] a definitional balance between the First Amendment and the Copyright Act by permitting free communication of facts while still protecting an author’s expression. . . . No author may copyright his ideas or the facts he narrates. . . . As this Court long ago observed: “[The] news element—the information respecting current events contained in the literary production—is not the creation of the writer, but is a report of matters that ordinarily are publici juris; it is the history of the day.”81
It is this fact about copyright law that causes the Court to refuse to recognize The Nation’s claim in Harper & Row. The Court’s refusal comes “[i]n view of the First Amendment protections already embodied in the Copyright Act’s distinction between copyrightable expression and uncopyrightable facts and ideas, and the latitude for scholarship and comment traditionally afforded by fair use. . . .”82 Copyright scholars83 have often and forcefully reiterated that copyright can live in harmony with the First Amendment if copyright law provides enough breathing space in the public domain for writers and speakers to rely on the work of their predecessors, in particular for their ideas, the facts they report, and for transformative uses at the very heart of the fair use doctrine.84
The analysis that applies to the Copyright Act is suggestive of the relationship between the First Amendment and the Intellectual Property Clause. The constraints on laws imposed by the Intellectual Property Clause—the originality requirement, the exclusion of materials already in the public domain, and the express time limitation—are inherent constraints on the tension between property rights in information enacted within the confines of that clause, and the values of free speech. It is not, however, impossible that laws will meet the threshold constraints of the Intellectual Property Clause, but will in context fail the more context-sensitive First Amendment review that must be applied to all content-neutral85 direct regulations of speech. Any rule that says to one person: “you cannot say xyz without the permission of a, and if you say xyz without such permission the sheriff will come knocking at your door” must, at the very least, be shown actually to serve a very important state interest, and must not burden more speech than necessary.86 This is so even if Congress is acting out of the best of intentions to improve the flow of information in society, as has been seen in at least some media regulation cases. One might easily imagine a court finding that applying an extension of the term of copyright by twenty years to new works does not run afoul of the “limited times” constraint in the Intellectual Property Clause,87 but nonetheless imposes a very real restraint on too much valuable speech in return for “too speculative a gain.”88
This care, due in evaluating regulations of information production and exchange enacted under the Intellectual Property Clause, is doubly necessary when considering such regulations that cannot be enacted constitutionally within the bounds imposed by that clause. In the absence of the constraints that require originality, solicitude to the public domain, consistent availability of information and ideas for use by the public, and time limitation, the First Amendment becomes a more important backstop against regulation that restricts information flow too much for too speculative a gain.
The upshot of the analysis is that a law that assigns to some people rights to prevent others from accessing certain information, or communicating in certain ways, must comply with two constitutional constraints.
First, if the nature of the right is an exclusive right intended to create market incentives for its owners, or protect those owners’ investments, by permitting them to exclude others from making valuable uses of the information, then Congress may act only within the confines of the Intellectual Property Clause. It may only give such rights in original works. It cannot create such rights as would enclose or burden access to information or knowledge already available to the public, and it cannot give exclusive rights to control ideas or facts. Furthermore, Congress may only enact rights under other clauses of Article I, Section 8, in particular the Commerce Clause, if these rights are different in kind from the rights that it is empowered, within constitutional bounds, to create under the Intellectual Property Clause. Creating a general right, good against the world, in “the news of the day,” for example, is beyond the power of Congress.
Second, private rights to control the use of information, whether created within the confines of the Intellectual Property Clause or properly created outside of that framework, are regulations on speech. As such, they are subject to independent and cumulative review under the First Amendment. Because of the clause’s solicitude towards the freedom of the public to use the same information in which it grants limited exclusive rights, this review will usually—though not always—result in permitting regulations that pass muster under a robust application of the clause. However, when Congress creates a right outside the bounds of the Intellectual Property Clause, free of the limitations inherent in that clause, First Amendment review must become a more robust constraint. Before applying these constraints to the pending database protection bills, however, one must first consider what might justify such an elaborate two-tier constitutional limitation on legislation of private rights in information.
Why would anyone want to erect such cumbersome constitutional hurdles to enacting private rights in information products? The answers fall into two broad categories. First, free information flows are central to freedom—both in the sense of personal autonomy, and in the sense of political or democratic self-governance. Policies that block off various flows of information, or maldistribute access to and control over our information environment, affect all of our abilities—both as individuals and as members of communities—to be self-governing. Constitutionally based judicial review is most valuable to prevent the implementation of policies that exert these kinds of costs on individual autonomy and the proper functioning of democratic processes.
Second, the production of information, knowledge, and culture, is a conversation over time and space among creative human beings, one whose impediment will fundamentally challenge our welfare as a society and a culture. This centrality is expressed in the constitutional concern with “Progress,” and requires robust institutional constraints to prevent present generations from appropriating too much of this grand conversation at the expense of participants from future generations.
Grafted over this somewhat romantic concept of “Progress” is the hard-nosed realization that the politics of private rights in information involve a systematic imbalance between the private stakes for those who benefit from them and for those who suffer their costs. The social gains of private rights are perceivable as private gains by their beneficiaries. These beneficiaries therefore fully represent in the political process the value of the benefit. But the social losses caused by such rights are diffused among all potential users of the information and therefore no one knows the full scope of these costs. In particular, these losses are borne by future generations that cannot be represented in present political processes, and thus, the costs of expanded property rights in information are systematically under-represented. The result is that legislatures suffer from structural rosy vision when they consider creating private rights in information, and judges must step in to re-evaluate laws passed using this skewed perspective when the costs become apparent enough to a party to a suit.
Why should laws regulating information and communications markets be subject to more extensive judicial review than laws regulating grain or securities markets? After all, one might say, it is one thing to have very close scrutiny of government when it attempts to censor individual expression or public discourse, but what has that got to do with review of normal market-creating and market-regulating functions that the state has always filled?89
The answer that has co-evolved with First Amendment law in the twentieth century90 has been that markets in information affect democracy and autonomy in ways that suggest an important role for courts, whereas other markets affect welfare in ways that do not similarly call for close judicial scrutiny, at least not under the federal Constitution. One can look anywhere from Madison91 to Brandeis92 for inspiration and will meet a similar set of conceptual commitments. For a community to be democratically self-governing its members must have access to information, this information must not be too tightly controlled by one group or another, and constituents must be able to express their views as well as receive information.93 These commitments are most urgently engaged when the government purposefully tries to regulate debate in order to assure that the message it prefers will prevail in the public arena. But the extremity of the concern in those situations need not obscure the fact that the same set of commitments is engaged when the government regulates information exchange among members of society for reasons that are understood by lawmakers to be information flow neutral, or even to support the free flow of information.94 If a law “gets it wrong” in the area of information flow regulation, it has negative effects on democracy that in the United States we have chosen to treat as particularly amenable to judicial oversight.95
I have elsewhere retold stories that illustrate this effect of seemingly neutral intellectual property rules on democratic discourse—for example, how the scientologists used copyright to quash a critic,96 or how the Washington Post and the Los Angeles Times sought to prevent readers from sharing copies of articles online in order to stimulate political discourse.97 Another little appreciated example of this effect, particularly pertinent to the object of our analysis here—exclusive rights in facts, history, the news of the day—is the Supreme Court’s decision in International News Service v. Associated Press.98 The case seemed to revolve around a question of whether the Associated Press (“AP”) had a limited property-like right in the news its correspondents collected, at least vis-à-vis its competitor wire service of the time, the International New Service (“INS”). As the facts were presented in the case, INS reporters would take the facts published on the East Coast by AP-associated papers, and wire them to the West Coast, where its affiliated papers published simultaneously with, or even ahead of, AP affiliated papers.99 Over the objections of Justices Holmes and Brandeis, the Court declared that the AP did in fact have a quasi-property right in the information100—a holding whose remnants still occupy the very bills that are the subject of analysis in this article.101
My point here is not, however, to revisit the wisdom or necessity of the “hot news” exception to the general wisdom that facts are not properly the subject of intellectual property. My point, rather, is to illustrate how an intellectual property rule can influence democratic discourse. The dispute revolved around news reports from World War I.102 AP served Pulitzer papers, and its bylaws allowed Pulitzer to exclude Hearst papers from receiving the AP wires.103 The INS was formed as a news service to serve those newspapers excluded from AP—namely, the Hearst papers. During the war, the European Allies excluded INS correspondents from receiving news or using transatlantic cables because the Hearst papers generally espoused an anti-English or anti-Allied position. The effect of the Court’s market-regulatory decision to exclude INS papers from using AP’s hot news was to burden newspapers that espoused a certain position in a hotly contested political issue in the United States. There is no evidence to suggest that the decision was based on a substantive preference for one set of political views or another rather than on the considerations of fairness and efficiency upon which the Court relied. But the point of this story is not to suggest that the Court improperly tweaked intellectual property law to support its members’ views. The point is that, like many other courts when they don their “intellectual property” hats, the Court was blind to the effects of its decision about property rights on democratic discourse. Recognizing a property right in the news of the day made it possible for those who espoused one view of a central political question of the time to control public reporting and debate of that question. It is a particularly clear instance of the general effect of property rights in information; they affect not only, or even primarily, how much information is exchanged in our public conversations, but more importantly who gets to say what to whom, and who decides these questions.
A less commonly recognized effect of property rights in information is that they create a similar set of concerns for our normative commitment to personal autonomy. The point has to do with the degree of control that individuals have over the information environment from within which they make their life choices, and the degree of control others have to shape the information environment within which an individual chooses. To be an autonomous individual is to be, to a significant extent, the author of one’s life through a series of successive choices. To choose our path in the world, we must have a sense of the range of options open to us and an evaluation of the relative worth of these options. These two elements of information—knowledge of options for action and an evaluative framework to compare them—are germane to our capacity to be the authors of our lives. The pattern of distribution in a society of control over the information that defines the range of options open to individuals in that society is central to the extent and distribution of autonomy in that society.104 Property in information is one dimension of information law that can give some people control over the information environment of others, which they can use to alter the likelihood that those others will behave in a manner conducive to the plans of those who own the information. When this happens, property rights increase the ability of their owners to author the lives of others, and decrease the autonomy of those others.
Imagine that Jane wants to buy a home, having settled into a new job. Her job leaves her very little time to shop for a mortgage during times when the two banks in her town are open. Given her personal condition, her search costs are high so she only knows about mortgage rates available from these two banks.105 She soothes her concern over not having time to search for a better mortgage by assuming that the market is efficient, and that all banks offer roughly equal terms. It turns out, however, that the mortgage rates of the two banks are higher than those available in larger urban areas in the state, reflecting the fact that there are many people in town who share Jane’s predicament. At the mortgage rate available from either of these two banks Jane cannot afford a house—she must live in an apartment. At the mortgage rate available from banks at larger urban centers, she could afford a house. But she does not know of this choice. Given her state of knowledge, the only choice she sees as open to her is what apartment to live in, not whether to live in an apartment or a house. Assume that is a decision important enough to one’s well-being to count as a choice, the presence or absence of which can plausibly be seen as important to one’s personal autonomy.106
Now bring in property rights. Along comes a database company that plans to collect data from the lists of mortgages compiled and maintained by all the banks in the state, and compiles them into a single, easily accessible database.107 Assume that, were such a database available to consumers, Jane would have the time and money to buy it and use it to expand her knowledge of available mortgages, and with it her known set of options. The company’s business plan forecasts acceptable returns, assuming the present state of the law on ownership of facts—they are in the public domain, free as the air to common use. As it turns out, Congress passes a law that gives the banks a property right in the facts of their compilation, so the database producer cannot copy them into a larger, value-added database without permission.108 It turns out that there are many banks in this state like the two banks in our town. These banks are quite happy to deny access to their data, or to charge a price that would compensate them for the premiums (over competitive rates) that they would lose from consumers like Jane. Faced with these new, higher costs of creating its database (paying for each bank’s list rather than simply searching for and copying the lists), the database company abandons its project.
One common response to such a story would be to view it as a simple example of the well-known static efficiency losses caused by intellectual property rights, which, if justified, are usually justified by their beneficial incentive effects on dynamic efficiency. An information product that would have been produced is not produced, even though its value will exceed the marginal cost of its production.109 What I would like to focus on, however, is the effect of this decision on Jane’s choice set and on the distribution of power to control the definition of her choice set.
In the absence of the property right, the banks had no formal right to limit Jane’s perception of the life choices open to her. Jane’s perception was limited by her life conditions and the choices of others around her, but not by a legal right that prevented a change in that perception. When the banks’ property rights were recognized, they received a legal right to prevent such a change, to prevent use of their information to change her knowledge of her options. By denying access to their information, a denial similarly exercised by other banks operating under similar circumstances, the banks imposed their preferred picture of the world on Jane, limiting her perception of her life choices so as to increase the likelihood that she will behave in a manner conducive to their interests. The banks did not care whether Jane lived in an apartment or a house. But they did want Jane to buy her mortgage from them, at a high price, and they are now entitled to prevent access to their information so as to affect Jane’s choice. As a by-product of increasing the likelihood that Jane will act in a manner consistent with their plans, they restricted Jane’s life options so that she now may only choose from among apartments, but not between being an apartment-dweller and a homeowner.
Imagine another alternative: that the database compiler did invest in collecting the information, relying precisely on the new property right to justify the investment. This is, after all, the beneficial incentive effect sought for the law. Assume further, however, that the costs of collection are such that only one or two companies can undertake them, given the expected market for the information itself. One of two things will happen. Either the first database compiler will sell the aggregate information to consumers like Jane, or it will sell to small town banks its agreement to abstain from selling the information.110 The banks have better information about the value of the information in the database than do those consumers who should value it most highly—namely, those who, in the absence of the database, do not know about lower mortgages. But these are precisely the consumers least able to determine the value to them of access to this database. So, the company will or will not release the information, based on whether these consumers as an aggregate under- or overestimate the premium they are paying for not knowing what other mortgages are available. Competing database compilers would be warded off from entering the market to serve consumers, however, even if the first database compiler did not sell to consumers. They would expect that once they enter the market, small town banks will no longer value (and pay for) withholding the information, and the first database compiler would have to turn to consumers to appropriate the value of its investment. That provider will then be able to price below the new entrant, because it has already recouped some or all of its compilation costs from the banks, whereas the new entrant has to recoup these costs entirely from consumers.
Recognizing property rights in information created a locus of control that allowed various players in a social environment to try to manipulate the information environment of others in that society so as to affect their behavior. When property rights are designed in ways that have a more-or-less predictable effect of centralizing control over information, or systematically tipping the scales in favor of one group or another in controlling the information environment of some other group, our normative concerns with autonomy ought to be engaged.
This effect of property in information is under-recognized. Indeed, the concern with autonomy is usually cited in opposition to policy recommendations to restrict property in information, which themselves are usually defended in the name of democracy or robust political discourse. The most explicit instance of this perception of the conflict is the debate between Owen Fiss and Robert Post.111 The theory underlying both these commentators’ acceptance of the conflict between democracy and autonomy as ways of approaching information law is their joint acceptance of autonomy as a sphere of noninterference by the state,112 not as a personal condition admitting of degree, whose extent and pattern ought to be the concern of information law and policy.
A review of some central liberal accounts of autonomy,113 however, suggests that personal autonomy is best understood both as a threshold, on/off category—which would justify treating autonomy as inconsistent with constraints on property rights—and as a matter of degree—which would justify paying close attention to the actual effects of property rights in information on personal autonomy. Both aspects of the concept are normatively significant.114 Thinking of autonomy as a threshold capacity to be self-governing, which almost all sane adults have, animates the liberal attention to treating all individuals as self-governing and deserving as such of equal regard. On this conception alone, autonomy calls for abstaining more than for designing an environment within which people can be more autonomous. Thinking of the degree to which individuals can exercise this capacity as normatively relevant alters one’s policy focus. For, if the capacity of individuals to be autonomous while living in a world largely made by others and the amount of autonomy they actually possess, are a matter of degree, and if the degree of autonomy available to individuals should be seen as a normative concern of that society, then increasing people’s autonomy becomes an important concern for policy making.
The argument about information policy and autonomy goes more or less like this. Individuals can have more or less capacity to lead autonomous lives, and their lives can be, as a factual matter, more or less autonomously lived, based on the extent to which they can play a role in authoring their own lives. The quantum of autonomy people can have and do actually enjoy is partly a function of the information they have about the world as it is, and the options open to them to live their lives. This information, central to individual autonomy, can in turn be more or less controlled by the individual. Information policy that moves the center of control over the information environment from within which each of us must choose and live our lives, in turn, affects the extent to which we can be, and are, autonomous. As law permits individuals to play a more significant part in defining the information environment from within which they choose their lives, it enhances the relative role those individuals play in authoring their own lives. As it does so, it can be seen as a more autonomy-loving law. As law shifts control over the information environment from within which its subjects choose away from most individuals, it undermines the personal autonomy of most of its subjects. Courts, acting as institutional defenders of individual rights, are the appropriate locus to protect personal autonomy from information laws that shift control over individuals’ information environment away from them—whatever the supposed beneficial effects of the shift in terms of the general welfare.115
The hypothetical example of Jane’s mortgage focused on direct control over access to information relevant to defining the options open to one person. One finds more murky, but still instructive, examples in actual cases.
Consider the question of how intellectual property interacts with perceptions of matter so central to self-definition as sex and sexuality. At the most basic level we see the case in which the United States Olympic Commission used its statutory exclusive right over the use of the term “Olympic” to obtain a court order preventing organizers of the “Gay Olympic Games” from using this term to name their event.116 The Court understood the case before it as concerned with a right to commercial exploitation. But, as Boyle has explained,117 this commercial right was used in that case for private censorship—to prevent the expression of a cultural message about the acceptability of homosexuality as a way of living. While no one could plausibly claim that this decision eliminated homosexuality from the range of options known to be open to individuals defining their sexuality, it is a rather crisp instance of exclusive property rights intervening in determining who has the power to control public statements about the relative worth of this life choice as opposed to alternative life choices.118
At a more “technologically sophisticated” level, we might see two suits brought by Playboy Enterprises to try to control the social meaning of the alphanumeric string, “playboy.” In one case a former playmate, Terri Welles, used the term in the metatags describing her site to search engines.119 Playboy claimed that since her site was not authorized by the company, Welles could not use the alphanumeric string to identify her site as relevant to those interested in finding documents relevant to the term “playboy.”120 In the other case, Playboy challenged the practice of a search engine, Excite, of treating the term “playboy” as a general request for soft porn.121 Excite returned to users who searched for the term—in addition to sites authorized by the plaintiff—results that were not related to Playboy Enterprises, including advertisements by competitors together with the search results.122 But it is important to understand that what was at stake in these cases was whether a trademark right will give its owner the exclusive right to determine the range of options open to users who rely on its mark to search for sexual materials. Most importantly, they involve the question of whether the initial role played by the mark in signifying interest in sex permits the organization, Playboy Enterprises, to define for those who seek sexual materials what counts as “sex” or “sexually gratifying expressions.”
As information about the world around us becomes increasingly diverse and dispersed, the right to control search engines, filters, and the very definition of relevance becomes the right to control the perceptions of individuals about what options are open to them. The right asserted by Playboy in these cases was a right to determine the range of options presented to users who initially expressed their interests using the term “playboy.” And control over the range of options open to individuals is control over their choice. At stake in the Welles case was the question of how individuals can rely on their own understanding of the cultural signification of “owned” words to present themselves as relevant to others. At stake in Excite was to what extent trademark owners would be able to control the definition of cultural and social signification and whether they would receive an exclusive right to dictate the options presented to those who use their “owned” signifier to represent a choice or “search.” In both cases, the result would be a more or less concentrated structure for the determination of relevant options open to one who uses a given signifier to define the choice one views oneself as making. The more centralized the structure, the more individuals seeing the world through the controlled lens come to be controlled by owners of the rights to define signification and relevance, and the less autonomous these users become.
Information, then, is central both to democratic self-governance and to personal autonomy. Laws regulating control over information flows must therefore be subject to special care in our democracy, a care we have come to institutionalize in judicial review. In the case of exclusive rights to control the use of information, Part II suggested that judicial review should be applied using two vehicles—the threshold constraints imposed by the Intellectual Property Clause and the more context-sensitive constraints imposed by the First Amendment.
In addition to the normative concerns surrounding enclosure of information, enclosure raises immensely important general welfare concerns that are systematically ill-represented in political processes. I will treat these problems using the admittedly quaint heuristic of “Progress,”123 the value to whose service the Intellectual Property Clause is dedicated.124
The structure of the modern idea of Progress from the mid-eighteenth century until quite late in the twentieth century provides a useful guide to understanding the stakes of exclusive rights in information or knowledge.125 In this quintessentially modern of ideas, Progress is the product of human agency, not the Divine. It is a forward movement whereby one generation builds on the attainments of its preceding generation, and in turn becomes the stepping stone of the next generation’s improvements. The engine of Progress is the progress of knowledge. Knowledge itself, like Progress, advances through human agency and improves from one generation to the next in a process of accretion. It feeds technological innovation, which increases the spread of material welfare and the development of better organizational and institutional arrangements, all of which feed back on each other. Over time, these together lead to the intellectual, moral, and aesthetic improvement of the human condition.126 This too was the basic structure of the idea that animated the passage of the Intellectual Property Clause.127
Understanding this structure of the idea of Progress is central to understanding why private rights in information and knowledge were understood by the drafters of the Constitution as “monopolies” to be carefully circumscribed, rather than as property rights to be cherished and protected.128 The object of property is Nature, given to industrious Man to make fruitful and thereby make his own. The object of exclusive rights in one’s contributions to the Progress of knowledge is knowledge itself—Knowledge, not given in some Promethean transfer, but developed progressively over time, as one intellectual laborer adds to the common project, to be followed by others like him. More like Commerce129 than like Nature, Progress is a common enterprise, carried on by all, for the benefit of all. And like exclusive rights to Commerce—monopolies—exclusive rights to segments of the great process of Progress are to be feared and curtailed. Like a monopoly on tea, or a monopoly on printing, they are feared to be the product of faction preference, of a deal between the governor and some of the governed at the expense of the common good.
What is important to understand for contemporary purposes of institutional design is that insofar as the progress of knowledge is concerned, the basic assumption is that the politics of faction will lead to too much recognition of exclusive rights at the expense of the common good, whereas the major concern about the politics of faction with respect to traditional property was that it would lead to too little respect for property, by supporting redistribution at the expense of the common good.130
So, if we try to break down the core components of the concept of “Progress” into contemporary terms, we arrive at the following uncontroversial list. First, both theoretical learning and practical innovation are central to material well-being.131 Second, these advances are by nature incremental, with each new participant and each new generation building on the achievements of its predecessors, and building a platform upon which future participants in this process will themselves build.132 Third, the product of all the individual efforts making up the progress of knowledge leads to large scale, society-wide benefits at the material, organizational, institutional, and cultural levels. Knowledge production is, in other words, a high positive externality activity, whose external benefits accrue to diffuse beneficiaries in diverse ways.133
The first component of the list, the importance of innovation and the progress of knowledge to aggregate welfare, would not, by itself, counsel close judicial review. We do not, after all, think that courts are systematically better than legislatures at determining good policy. But the second and third elements of what might make for a contemporary, non-naïve understanding of “Progress,” do suggest systematic weaknesses in the legislative process, weaknesses that judicial review could remedy.
The second element goes to the systematic intergenerational bias of legislators. If Progress is a joint enterprise of present and future generations, then an optimal policy in the pursuit of Progress must take into consideration the costs and benefits of any exclusive right in information to both present and future generations. But future information producers and future constituencies that benefit from their innovation do not vote for, or contribute to, present legislators. Only present participants do. What we are therefore likely to see is a systematic bias of legislators to provide exclusive rights—which favor present participants over future participants. Courts, not similarly beholden to present participants, are in a better position to take stock of the overall effects of a proposed law on the intergenerational enterprise called Progress. As among potential institutional decision-makers, they are the best residual bearer of the decision-making power, because they are the most insulated from the pressures of today’s participants.
The third element—the recognition of the high positive externalities involved in information production—suggests that laws designed to respond to the interests of present producers by maximizing their private benefits will not systematically coincide with laws best tailored to optimize knowledge production. Say that a new law, for example copyright term extension, slightly increases the ability of producers to appropriate the benefits of their work but significantly limits the ability of society at large to capture the externalities created by a copyrighted work. It will rarely be the case that any individual in society bears the full public cost of these lost positive utilities as private costs. And, whenever no individual or organization sees the full public costs as private costs, the social costs of an increase in property rights will be underrepresented. Moreover, unless there happens to be an individual or organization who sees private costs that are on par with the benefits that the vendors see as private gains from the change in law, the benefits of the legislation will always be presented to legislatures as outweighing the costs. This leads to a systematic information imbalance whenever a legislature considers a measure to increase the scope of intellectual property rights. The public benefits of the measure are more completely reflected as private benefits for participants in the political conversation—the owners of the new rights—while the public costs are usually reflected as much lower private costs, for a much more widely dispersed constituency. Occasionally a set of costs may be reflected as private costs for a given constituency, and in those instances we would tend to see Congress addressing the costs presented to it by that constituency.
What we would expect to see from such a structure of the information flow into a legislative process is an expansion of rights over time, tempered by specific exemptions for those constituencies who happen to experience a particular set of public costs as private. And this is in fact what we see in the area of intellectual property law—systematic expansion of property rights tempered by exceptions and exemptions for groups who presented to Congress their own subset of the total social cost of the new rights. This structure dates back to Noah Webster’s involvement in the passage of copyright acts in the states to protect his dictionary,134 and was nowhere more evident than in the drafting of the Copyright Act of 1976, so comprehensively described by Jessica Litman.135 More recently we saw it in the Digital Millennium Copyright Act,136 which vastly expanded the ability of providers of information in digital media to exclude non-paying users from their products, tempering this expansion with a long list of exemptions for specific constituencies.137 In the database bills considered in Parts IV and V below, there is no clearer example of this than the absolute exemption given in the broader of the two bills to religious use of genealogical information, apparently central to the practices of the Church of Jesus Christ of the Latter-Day Saints.138
Unlike legislatures, which can respond to no wider a range of social benefits and costs than those internalized at the time it is passed, judges encounter the legislation over time, as its social costs come to be borne as private costs by individual parties to the litigation before them. While courts cannot completely compensate for the informational imbalance regarding legislation that encloses information, over time they nonetheless have a potentially more comprehensive view of the social costs of the exclusive rights created by the legislature. Faced with the systematic informational bias in the legislative process, courts should recognize their relative advantage in observing the social costs of these laws and adopt the role of counterweight to balance the bias’s effects.
This section suggested why we might want a robust system of judicial review for what might, at first glance, look like a normal governmental regulatory function. I proposed two answers. First, the way information flows in society, and decisions about who controls information flows under what circumstances, play a central role in defining our democracy, and in defining to what extent, and under what constraints, each of us can be an autonomous human being. Second, the political economy of regulating information flow is subject to systematic imbalance in the representation of present stakeholders relative to future generations, and in the representation of its private benefits relative to its social costs.
Judicial review is an appropriate institutional mechanism to recalibrate the process by which our society creates exclusive rights to control information flows in our information environment. Courts have traditionally played an important role in assuring the conditions of a well-functioning democracy and in protecting the value of personal autonomy against measures intended—at least purportedly—to advance the public good. Courts are also a good backstop against systematically unbalanced legislative and regulatory processes because they march to the beat of a different drum, can take a longer view of the effects of laws, and are presented with arguments from equally engaged parties.
If the House of Representatives had purposefully tried to create a test case of the constitutional bounds imposed on it by the Intellectual Property Clause, its members could not have done better than to propose the two main opposing database protection bills reported to the House on September 30, 1999. House Bill 354, reported from the Committee on the Judiciary by Representative Coble, creates a property right in raw information in all but name. As reported, House Bill 354 prohibits extraction of information from a database, both for reuse and dissemination and simply for use, and gives database owners the right to track and prevent uses of information extracted from their database into downstream products—whether or not they compete with the database. The report suggests that the law is intended to secure incentives for investment in databases and to prohibit most valuable uses of information obtained from the database without the owner’s permission. House Bill 1858, reported from the Committee on Commerce by Representative Bliley, on the other hand, assiduously shies away from property, and attempts to remain within the confines of unfair competition. It addresses competitors only, not users—either consumers or downstream creative users. It prohibits duplication, defined narrowly as slavish copying of the contents of the database without adding value, for sale of this data in competition with the source database. And it vests enforcement in the Federal Trade Commission. As one reviews the background and components of each proposed law, it becomes clear how House Bill 354 fails both the threshold test imposed by the Intellectual Property Clause and the backstop constraint imposed by the First Amendment, while House Bill 1858 survives at least the former, and probably the latter. This part reviews the constitutionality of the two bills under the Intellectual Property Clause, while Part V reviews constitutionality under the First Amendment.
In 1991, in Feist Publications, Inc. v. Rural Telephone Services Co.,139 the Supreme Court held that raw facts in a compilation, or database, were not covered by the Copyright Act, and could not be so protected consistent with the constraints imposed by the Intellectual Property Clause.140 The Court held that the copyright law could protect the creative element of the compilation, its organization or selectivity, for example,141 but could not protect the facts compiled. There are conflicting accounts of whether Feist thereby broke with a widespread practice of protecting “sweat of the brow” works—works created by investment of money and labor, but lacking creativity—or whether the Court merely acknowledged, adopted, and embedded at the constitutional level the general trend among the lower federal courts.142 What is not debated is that the past few years have seen repeated efforts to pass legislation that would remedy the perceived deficiency that the Court’s holding in Feist created—namely, that producers of investment-intensive, uncreative collections of information did not have a property right in data they collected.
That the Court rooted its Feist decision in a robust interpretation of the Intellectual Property Clause created something of a difficulty for those seeking that holding’s legislative nullification—namely, it did not seem possible for them to do so without running afoul of the Intellectual Property Clause.143 Efforts to protect database providers therefore eventually settled not on a sui generis quasi-property right, as that recognized in the European Database Directive,144 but rather on an unfair competition law, which could be based in the Commerce Clause, free and clear of the inconvenient weight of Feist.
This background has created a rather unique natural experiment for testing the bounds imposed by the Intellectual Property Clause on Congress. We know from our general constitutional analysis that intellectual property rights must be passed, if at all, within the confines of the Intellectual Property Clause. Only protections that are different in kind can be passed under the more general commerce power.145 We also know, from Feist, that intellectual property rights cannot, consistent with the requirements of the Intellectual Property Clause, be recognized in the factual contents of a database, but only in its creative elements. What we now face is the task of examining competing efforts in Congress to protect database providers, to see whether they in fact provide a different kind of protection, or whether they are, functionally, an attempt to legislate a quasi-property right.
The Intellectual Property Clause captures not only legislative efforts that are formally aimed at creating intellectual property rights, but also those that create exclusive rights that are functionally equivalent to intellectual property rights.146 I therefore suggest here that the relevant comparison between the database protection bills and intellectual property is one of function, and here I will focus on economic function. The economic function of intellectual property rights is to provide the legal entitlement that makes information goods partly excludable.147 The technological and legal characteristics of every good locate it somewhere along a range from fully a public good to fully a private or economic good, as defined by the degrees to which the good is rival and to which it is excludable.148 Whether a good is rival or nonrival is solely a function of the state of technology. It either can, as a practical matter, be used by many people simultaneously without anyone’s use rivaling that of another, or it cannot. Whether a good is excludable—whether its producer can as a practical matter prevent some or many people from using the good unless they promise to pay—is both a question of technology and of law. Land is excludable partly because fences can be built and partly because trespass is prohibited, so that jumping the fence will trigger a visit from the sheriff. Property rights in any given good define the parameters of its excludability in conjunction with nature.
Intellectual property rights define the excludability of information products. In order to provide producers of information goods with an incentive to produce, intellectual property rights give their owner a degree of exclusivity in controlling the use of “their” information product. They do so by stating an entitlement, and coupling it with a promise by the state to prevent users who seek to make use of the information in ways covered by the law from doing so. Since those uses are valuable to some users of the information, the owners sell permission to use to those users, and thereby appropriate the benefit that their productive enterprise has created for the users.
Intellectual property rights are in this sense market-creating—they are constitutive of the properties of the goods sold in the market. Functionally, this distinguishes them from laws that are market-regulating, or that constrain behavior in a market for goods whose excludability is already defined by other rules—namely, property rights. Now, it should be clear that these definitions are provisional working definitions, and their borders are permeable. But they capture rather well, for example, the distinction between fair use, a limitation on all information product owners’ ability to exclude users from their products, and misuse of copyright, a limitation on some owners’ ability, under certain circumstances, to do things with their products that they are generally entitled to do. Even though it is formally a defense to a copyright claim, the former is a part of the definition of property in information and not part of the definition of the owner’s exclusive rights. The latter, misuse, is a regulation of using already-defined property rights in certain market situations.
In interrogating the two database protection bills, we should seek to identify whether what each one does is market-creating or market-regulating in the sense just described. One inquiry would require us to identify the relationship addressed by the law: is it a relationship between sellers and buyers, producers and consumers, defining what is up for grabs between them, or is it some other relationship in the market, in particular, the relationship among competitors. Another inquiry concerns what actions are made possible or impossible by the law. In particular, we would be interested in whether the actions prohibited include the uses intended by the owner to be the source of the information’s value to users. By prohibiting such actions, the law defines the “goods” that form the basis of a business in selling the information—or, more accurately, permission to make valuable use of the information. In the alternative, we might see that the prohibited actions are not those that make the information valuable to users, but rather those that permit competitors of the producer to provide the same value to users as the owner plans to sell.
House Bill 354 is addressed to anyone who distributes information extracted from a database,149 and to anyone who uses information in a database.150 House Bill 1858 has two parts, the first plainly addressed to competitors,151 the second to anyone who distributes real-time market information.152 As an initial matter, then, House Bill 1858 does not speak to anyone who uses information (including real-time market information) for its value as information, as opposed to its value as goods in trade. House Bill 354, on the other hand, does address anyone who uses the information in its marketable use, and could therefore, at least in principle, speak to a user for whom the value of use is precisely its value as data, not as goods in trade. It speaks to consumers of data, as well as to competitors in the market for serving consumers in data.
To illustrate the difference between the two, let us look at a simple example. Imagine that John purchases a CD-ROM with a list of daily temperatures, barometric pressure, and other such data, in 200 cities and towns throughout the United States between 1900 and 1999. He lends it to Jane. Jane copies the entire factual contents—but not the organization—of the CD-ROM onto her hard drive, where she now accesses the raw information using her own access software, having returned the CD-ROM to John. Clearly what Jane did would violate copyright, if the contents of the CD-ROM had been copyrightable, and no other factors were present. Equally clear, her actions are privileged from copyright law under Feist, because that decision was not based on a notion of fair use or de minimis use, but on the uncopyrightability of facts, which are in the public domain, free for the taking.
House Bill 1858 does not speak to Jane. She is not disseminating the information in commerce. House Bill 354 does speak to her. She has extracted the entire collection of information. Now we must determine whether her extraction harms the producer’s primary market, and if so, whether or not she is privileged under one of the exceptions included in the Act. But we do not need to make that assessment to complete identifying the first central difference between the two bills. House Bill 354 is addressed to users of the information who use it for its value to them as information, as well as to competitors and others who distribute the information extracted from protected databases. House Bill 1858 is addressed only to competitors of the information provider, or with respect to real-time market information, to those who redistribute that information on a real-time basis.
House Bill 1858 prohibits distribution or sale of a duplicate of a database.153 A duplicate must be both “substantially the same” as the original, and must have been created by extraction of the information from the original.154 The Committee Report explaining the definition of duplicate suggests that a database is not “substantially the same” unless the non-identical portion of the duplicate is immaterial to the overall value of the database.155
House Bill 354 prohibits both extracting to make available and making available information from a database in a manner that causes material harm to the primary or a related market of the producer of the source database,156 and prohibits extracting information in a manner that causes harm to the producer’s primary market.157 The Committee Report on Section 1402 suggests that “material harm” is intended to exclude de minimis uses, but not to introduce a more robust limitation,158 expressly rejecting the requirement imposed in National Basketball Ass’n v. Motorola, Inc. that the plaintiff must show that use of the facts in the manner the defendant uses it would threaten its existence.159
One implication of this difference reinforces the point made in the preceding section. The Committee Report on House Bill 354, in explaining the meaning of “material harm,” includes the following statement: “[I]f a person extracts so much of an online database that the person would, in the future, be able to avoid paying a subscription fee for access to the data it contains, that person has harmed the market for the database.”160 Again, House Bill 1858 would not speak to such a person, whose extraction is for personal use and not for competition.
Moreover, this paragraph in the report illustrates that part of the function of House Bill 354 is to create the legal basis upon which the provider of a database can claim an entitlement to payment. In the absence of House Bill 354 the producer can physically block access to the information, and charge for access as a contractual matter. But the producer has no legal right to use the machinery of the state to prevent a user who has gained access to the information from using it, and hence has no legal claim to payment in the absence of a contract. As the quote from the Committee Report in the preceding paragraph makes clear, House Bill 354 creates this right, which is the essence of a property right—the right to exclude another from using one’s property, irrespective of whether one has, or has not, maintained physical control to implement that exclusion.
Another conceptual difference between the two bills, that goes to the “propertyness” of House Bill 354, is how differently they “track” the rights of a producer in something like the “res” of the collection. Imagine that Jane, in our example above, uses the information about the weather in 200 U.S. cities as part of a compilation of similar data for 5000 cities around the world that she is producing. Under House Bill 1858, the value of the 4800 cities that she added in her database (perhaps by similar extraction from other small databases) would not be “immaterial to the overall value” of her database. Accepting the interpretation that the committee placed on what makes one database “substantially the same” as the original of which it is purported to be a duplicate, her new database would not be “a duplicate,” and would therefore not violate the prohibition imposed by House Bill 1858. The report on House Bill 354, on the other hand, expressly contemplates such use of the entire contents of a database, and explicitly treats it as a prohibited act. The example used in the report is of a person who compiles a collection of public domain photographs of famous people around the United States by including, among other things, a collection of public domain photographs of such people from Massachusetts.161 As the Committee Report puts it, “[t]he Act seeks to prevent market harm to the investment in collections of information, and defendants should not be able to escape liability through activity analogous to the use of a stapler.”162 Once the investment in collecting is made, its value to others, including others who use it as input into their own value-added production, is captured by the law for the benefit of the collector. The collection ceases to be a set of facts, and becomes a res that can be tracked back to its original owner through successive compilations. Again, what we see is that the collection of facts becomes an excludable good because of this proposed law, one that can be sold—or not—to future producers who value it, whether or not they compete with the original collector.
Beyond this conceptual level, it is important to understand that House Bill 1858 and House Bill 354 regulate two different kinds of competitive harm—one that has traditionally been the subject of unfair competition, the other of intellectual property. Take the example given in the Report on House Bill 1858.163 One database producer (“A”) collects information about bank loans from in-state individual bank databases, and creates a statewide database more efficient than those each bank produces. A second database producer (“B”) takes A’s collection, combines it with out-of-state individual bank information, and provides a more comprehensive database still. According to the Committee Report on House Bill 1858, the bill privileges the use that A made of the individual bank databases, and that B made of A’s database. In each case, the second database producer in each pair (bank, A), (A, B) added significant information value to the information copied from the first database in each pair. This additional information is material to the value of the second database in each case, the second database is therefore not a “duplicate,” and the use of the existing information collection is privileged. Under House Bill 354 both A’s and B’s actions would be prohibited. House Bill 1858 would, however, cover the behavior of each of A—vis-à-vis the banks—or B—vis-à-vis A, if the second database in each case had done nothing more than copied all of the prior database, and then sold that duplicate without adding much of value.
House Bill 1858 thus provides protection against a competitor who competes by offering the same product, and who can sustain a lower price for that product solely by relying on its competitor’s investments in collecting the information. House Bill 354 protects against this competitor, but in addition protects against a competitor who competes by offering a better and more comprehensive product that is fully or partially substitutable with the original product, which does not even necessarily undercut the original database producer in price, but is cheaper than it would have been had the competitor been prohibited from using existing information collections as raw materials for its better product. Such protection encumbers the subsequent producer with a production price that is higher than the social cost of its production at the time it is undertaken, and hence diminishes the production of these later, better products, to below their socially optimal level at that later time. This intergenerational tradeoff is the standard effect of intellectual property rules, but not of unfair competition law.
In effect, House Bill 354 functions economically like an intellectual property rule. If it is truly about unfair competition through price undercutting made possible by free-riding on the investment of another, it need not cover as many uses of information contained in databases as it does. Protection of the type provided in House Bill 1858 does that job. If it is about requiring the makers of advanced products to pay a fee to the makers of precursors and of information inputs they use in their products, then it is not about unfair competition—it is about dividing the welfare produced by investment in the collection, organization, and dissemination of information, such that subsequent producers pay predecessors who made their new production possible. It is about trading static efficiency of information production for dynamic efficiency. It is about intellectual property.
House Bill 1858 has fewer exemptions and limitations, because the scope of its initial reach is much narrower. House Bill 354 has many more limitations and exceptions. The most important for our purposes is the list of permitted acts under Section 1403—some of which create context-specific exceptions to liability, and others of which in fact limit the extent of the property right created by the bill. The most important limitations on property in facts are that extraction or dissemination of an individual fact contained in a database is permitted, and that collected facts remain in the public domain in the sense that competitors are free to collect them independently. The property right created in the investment in compilation, however, is maintained by prohibiting repeated extraction of individual facts—thus limiting this exemption, as with the definition of “primary market,” to de minimis instances of extraction—and by prohibiting collection of the facts by accessing the collection.
The remaining permitted uses provide exceptions for certain specified uses. These, however, do not change the nature of the right created, but only define its contours. There is a general reasonable use exemption, which requires context-specific weighing of factors, somewhat like the fair use defense in copyright. What is important to note about this exemption is that it does not affect its nature as a property-like right. Just as the scope of “fair use” defines the scope of the copyrights authors have in their writings, but does not change the nature of copyright from an intellectual property right into an unfair competition rule, so too the availability of a reasonable use exemption does not change the nature of the database protection right created in House Bill 354 from an intellectual property-like right into an unfair competition law, but merely defines the scope of the property right.
It is not that a list of permitted uses could not, as a conceptual matter, have, if Congress so chose, fundamentally altered the nature of the right. If the permitted acts had in fact negated the basic reach and effect of the prohibitions stated in Section 1402, they would have, effectively, negated our initial conclusion that House Bill 354 creates a property right. But that is not what they do. A comparison between the exemption for nonprofit educational and research uses and the exemption for genealogical information collected or used for religious purposes is instructive.
Section 1403(c) exempts nonprofit educational and research uses that do not harm the primary market of the information provider. This is an exemption intended to permit education and research uses of investments made in collecting information for various other markets without payment, and is a real limitation on a general property right. Nonetheless, it retains the structure of a property right in the information collected to form the basis for a provider that treats nonprofit educational and research users as a primary market. “Primary market” is defined in the bill to include any market in which the provider offers a service or product incorporating the database, and from which the provider reasonably expects to receive revenue.164 If the collection of information is offered for sale to educators or researchers, even only as a sideline business, then the exemption does not apply, and the provider can charge for the use. The users have no exemption from the prohibition on using the information. This exemption retains, therefore, the market-creating role of House Bill 354, as to the market in information useful in education and research. For example, a company would be protected from this exemption if it invested in collecting information from public records about family histories and offered them in a CD-ROM for use in classrooms in the study of personal family history, or to researchers of migration patterns. The exemption would not apply even if the business rationale that justified the compiler’s investment in compiling the data were based purely on the value of the compilation for family entertainment, so that the existence or non-existence of the nonprofit educational or research market were irrelevant to the investment decision.
The same company that collects family history information cannot, however, decide that there is a market for its efforts among members of the Mormon Church, who might be willing to pay a good deal to find information about their families because this information would help them to perform church ordinances for their deceased ancestors.165 These religious users of genealogical information are entirely exempt from the Act’s coverage, irrespective of the effect on the producer’s intended primary market.166 This is a form of exemption that excludes a property right in certain kinds of information—genealogical information useful for religious observance. It makes religiously useful genealogical information nonexcludable, and hence not the object of market transactions. Were the “permitted acts” enumerated in section 1403 generally of this structure, the property right would, indeed, be negated in many cases.
Another important practical difference between House Bill 1858 and House Bill 354 is in the enforcement mechanism. The general prohibition on duplication in competition under House Bill 1858 is enforceable by the Federal Trade Commission (“FTC”).167 There is no private right of action (although there is such a right for the bill’s second part, dealing with misappropriation of real-time market data). House Bill 354 does provide a private right of action.168 This practical difference is important as a prudential matter. On the one hand, this limitation might render the protection ineffective by subjecting enforcement to the budget and personnel constraints of the FTC. On the other hand, it provides an important backstop against overreaching assertions of rights, if that is what one is concerned with in the database context.
These practical effects may be important, but they do not affect the constitutionality of the database laws under the Intellectual Property Clause. That trademark rights can be asserted by private owners does not make them any more, or any less, distinct from intellectual property rights of the variety that must be passed, if at all, under the Intellectual Property Clause. Similarly, that the securities laws permit private rights of action to enforce their disclosure requirements does not convert the regulatory reporting requirements into property rights of shareholders. They remain nonwaivable background rules intended to assure the efficiency of the market, not rules that define the contours of securities as objects in trade. Nor do private