Is intellectual property trivial?

• Author: Barnett, Jonathan M.

INTRODUCTION I. WHY INTELLECTUAL PROPERTY IS AND IS NOT TRIVIAL A. Why Intellectual Property Is Nontrivial (Always) B. Market Alternatives to Intellectual Property 1. Technology 2. Contract 3. Organization 4. Complementary Assets C. Why Intellectual Property Is Trivial (Sometimes) 1. Simple Scenario I: Certainly Nontrivial, Nonperverse Effect 2. Simple Scenario II: Certainly Trivial, Nonperverse Effect 3. Complex Scenario I: Potentially Nontrivial, Nonperverse Effect 4. Complex Scenario II: Potentially Nontrivial, Perverse Effect II. WHY INTELLECTUAL PROPERTY IS NOT TRIVIAL (SOMETIMES) A. An Unconventional View of Intellectual Property B. Distributive Effects of Intellectual Property 1. Vertical Distributive Effects 2. Horizontal Distributive Effects III. WHY INTELLECTUAL PROPERTY MAY BE INDIRECTLY NONTRIVIAL (SOMETIMES) CONCLUSION INTRODUCTION

Policy, scholarly, and popular discussions of the socially desirable level of protection provided by intellectual property rights typically take for granted that changes in the level of intellectual property protection matter a great deal. It is commonly assumed to make a substantial difference in regulating access to intellectual goods whether patent claims are broadly or narrowly interpreted, the copyright term is longer or shorter, or the fair use exemption is applied more or less generously. This assumption follows what appears to be an uncontroversial proposition commonly set forth in intellectual property jurisprudence and scholarship: patents, copyrights, and other entitlements determine which technologies and creative works fall into the private domain (to which access is constrained) and which remain in the public domain (to which access is unfettered). (1) In this Article, I show that this proposition should be controversial. It is not clear that changes--even substantial changes--in intellectual property protection typically make any meaningful difference in regulating access to the underlying pool of intellectual goods, which in turn means that these changes do not clearly make any meaningful difference in regulating the anticipated profits that drive innovation incentives. Contrary to natural intuitions, the size of the public domain may be substantially invariant to changes in intellectual property coverage. (2) This qualified indifference thesis is founded in a well-established empirical observation: firms generally can--and do--exploit devices other than intellectual property to limit access to, and thereby appropriate returns from, innovation investments. (3) Hence, intellectual goods that are unprotected by intellectual property may still be protected directly or indirectly by other legal or extralegal mechanisms, which broadly include technology, contract, organizational form, and various complementary assets.

If these alternative instruments can substantially replace the appropriation capacities provided by intellectual property rights, then legal changes that constrain those rights and thereby ostensibly expand the public domain have no substantial net effect; conversely, if these alternative instruments can match or exceed the appropriation capacities provided by intellectual property rights, then legal changes that expand these rights and thereby ostensibly narrow the public domain have no substantial net effect. This proposition is self-evidently true in the extreme case where perfect technological locks can be implemented at zero cost: contractions or expansions in intellectual property coverage have no marginal effect on the access costs incurred by third parties and, as a consequence, on the innovation gains anticipated by resource holders. In a broader class of intermediate settings, this proposition retains descriptive force to the extent that firms can exploit alternative instruments substantially to reproduce, or even surpass, the appropriation capacities provided by intellectual property.

If there is reason to doubt that nontrivial changes in intellectual property coverage always yield nontrivial effects on access to intellectual goods, then there must be reason to doubt the incentives/access tradeoff that is the familiar foundation for normative discussions about the desirable scope of intellectual property. (4) This tradeoff assumes that more intellectual property generates social harm by reducing access to intellectual goods, but generates social benefits by enhancing anticipated profits and thereby enhancing innovation incentives. Conversely, less intellectual property generates social benefits by expanding access to intellectual goods but generates social harm by reducing anticipated profits and thereby reducing innovation incentives. Hence, the policy challenge lies in setting intellectual property coverage so as always to yield a net social gain.

But the zero-sum tradeoff that drives this policy calculus does not hold universally, or even typically, as soon as we drop or relax the unstated but critical assumption that firms cannot use substantially cost-equivalent exclusionary devices. Without that assumption, the incentives/access tradeoff is no longer a safe bet. There can be no assurance that (i) nontrivial contractions in entitlement strength will nontrivially reduce the costs of accessing intellectual goods and thereby decrease innovators' anticipated rewards and investment incentives, or (ii) nontrivial expansions in entitlement strength will nontrivially increase the costs of accessing intellectual goods and thereby increase innovators' anticipated rewards and investment incentives. Any reduction in intellectual property coverage will have trivial effects if it simply induces firms to migrate to the next-least-costly alternative instrument by which to maintain reasonably equivalent appropriation capacities; and any expansion of intellectual property coverage will have trivial effects if firms already make use of alternative instruments that deliver equivalent or greater appropriation capacities at a comparable or lower cost.

This line of argument immediately raises a conundrum: if neither more nor less IP exerts a substantial effect on access costs and innovation gains over some meaningful range of circumstances, then why do profit-maximizing firms expend resources on influencing changes in intellectual property coverage? (5) Working out this conundrum yields a nuanced thesis that identifies more precisely the circumstances under which changes in intellectual property coverage do and do not matter. Even in a world of substantially cost-equivalent appropriation instruments, intellectual property coverage still makes a difference so long as we make the reasonable assumption that alternative instruments--or more precisely, the relative costs of using those instruments--are not equally distributed among all existing and potential participants in the relevant market. (6) Where that assumption is satisfied, any change in entitlement strength does have nontrivial effects.

Contrary to conventional assumptions, these are not effects on the total gains available as a result of the appropriation capacities provided by legal instruments, but on the distribution of those gains among firms that exploit the appropriation capacities provided by a portfolio of legal and extralegal instruments. Even if more or less intellectual property makes no difference on the margin so long as the market generally can use substitute instruments to cover shortfalls in intellectual property coverage, it makes considerable difference on the margin if each individual firm incurs nonidentical costs in migrating to those substitute instruments. If alternative instruments are not available at reasonably comparable cost to all actual and potential participants in the relevant market, then relaxations of intellectual property coverage will shift gains to firms that have the lowest-cost access to alternative instruments and away from firms that have the highest-cost access. Hence, even if intellectual property has trivial effects as an incentive instrument with respect to the market as a whole, every individual firm rationally invests resources in influencing intellectual property coverage. Everything else being equal, reducing coverage will shift rents away from firms with higher-cost appropriation technologies (which should lobby for "critical" intellectual property) while increasing coverage will shift economic rents away from firms with lower-cost appropriation technologies (which should lobby against "excessive" intellectual property).

If we recognize the typical availability of substantially cost-equivalent alternative instruments, then intellectual property is trivial with respect to the total rents generated by innovation investment. If we recognize that alternative instruments are typically distributed unequally across firm types, then intellectual property is nontrivial with respect to the distribution of rents in the relevant market. Surprisingly, the typical abundance of alternative instruments among incumbents and the typical paucity of such instruments among entrants imply that the distributive effects of relaxing intellectual property may often be "regressive" and the distributive effects of increasing intellectual property may often be "progressive." (7) Commentators usually assume that distributive effects run in precisely the contrary direction: stronger intellectual property coverage presumably increases the entry costs incurred by small-firm entrants and therefore increases the pricing power exercised by large-firm incumbents, which in turn punishes end-users. (8) But if intellectual property typically has a differential, nontrivial impact on smaller firms that have the highest-cost access to alternative instruments, then the relationship may be reversed. Weaker legal protections exacerbate large firms' inherent appropriation-cost advantage over small firms, which in turn implies that incumbents' market share is more securely...