Darknets, DRM, and Trusted Computing: Economic Incentives for Platform Providers EXTENDED SUMMARY

In recent years, content providers and the computer industry have joined forces to protect intellectual property rights on digital information. Music publishers and movie studios have experimented with various digital rights management (DRM) initiatives in order to control access, use, and dissemination of their products. Software and hardware providers have developed various technological forms of protection of such DRM. Those protections have been routinely broken, and those property rights have been bypassed in peer-to-peer file sharing networks. Some (Biddle, England, Peinado, and Willman [2003]) have coined the term “darknets” to refer to such networks, concluding that there exist no technical impediments to their growing in convenience and efficiency. At the same time, however, new research in computer science (Arbaugh, Farber, and Smith [1997]) has made the prospect of building difficult to break DRM technologies actually plausible. In 1999, Intel, Microsoft, HP, Compaq, and IBM founded the Trusted Computing Platform Alliance (TCPA), with the goal of creating open industry standards for a trusted computing subsystem to be added to personal computers. “Trusted computing” (TC) refers to standards (and the combination of hardware and software built on them) aimed at increasing computer security of personal computers, by letting certain components verify the trustworthiness of others before interacting with them. Although it is not explicitly mentioned in the original TCPA specifications (Trusted Computing Group [2003]), one possible and natural application of trusted computing is the enforcement of DRM (Anderson [2003a,b]). TC can be used as a building block for an architecture of protected computing environments, where every application, file, or communication needs specific approval in order to be allowed to run on a system. In order for DRM to be effectively deployed through trusted computing, however, both software and hardware platform providers need to modify their products. Although the TCPA has grown to include most of the big players in the computer industry, compliance by several players is costly. Recently, Microsoft announced a delay in the deployment of TC initiatives due to the concerns of enterprise and independent application developers associated with those costs (Evers [2004]). Historically, most DRM systems (from key lock-floppies to dongles) have been competed away in the marketplace. Under the opposite forces of network effects (associated with non-DRM systems, which are preferred by users) and rights control (associated with DRM systems and preferred by content providers), most DMR-platforms have failed to gain widespread adoption, their producers eventually releasing their products in non-DRM forms. Will things be different this time? The alliance between key members of the computer industry and the higher costs necessary to by-pass the protection they would ensure to content suggest that trusted computing initiatives may actually be successful. According to its proponents, TC guarantees a future of security and privacy of digital information. According to its detractors, TC can switch almost unrestricted control of a computer away from its user and owner, to the vendors. Artificially created lock-ins, incompatibilities, barriers to entry, and other obstacles are described not only as possible, but as the actual business goals of the TC alliance (Green [2002]). Anderson (2003a,b,c) offers the most comprehensive analysis so far of the economic motivations and implications of TC. Anderson (2003b) concludes that the value to corporate and government users of TC is unclear, since new powers could be granted, but also new risks and liability may be faced. According to Anderson, also the content industry an obvious candidate to benefit from TC-enforced DRM may not really benefit from limiting the diffusion of its products with controls and constraints. Anderson concludes that hardware vendors and software vendors have most to gain. Through an interplay of network externalities and technological lock-ins, Microsoft could“[invest] in equipping the operating system platform (NGSCB and Windows2003+) with TC mechanisms in order to reap a reward through higher fee income from its applications.” The likely consequences are that TC will centralize economic power, favoring large companies over small ones. Bergemann, Feigenbaum, Shenker, and Smith (2004) present a position paper with the first attempt to formally represent TC markets. Based on Caillaud and Jullien (2004) and Rochet and Tirole (2004), they present the interaction between content providers, platform providers, and consumers as a form of bilateral competition. The content provider needs to choose amongst competing platforms for its product, but also satisfy the largest possible network of users. The authors use this setup to raise several interesting questions such as what is the likely effect of varying amounts of competing platform providers and their governance structures; and how do adoption decisions by content providers and consumers depend on the distribution of consumers’ valuations of different products. In our paper we present a model less general than the one by Bergemann et al (2004), in order to discuss a narrower set of questions. We study the incentives for platform providers to produce trusted computing, DRM compliant products in a market in which consumers can choose between trusted platforms and platforms that do not enforce TC or DRM. Under what conditions does it make sense for platform providers to second content providers’ requests to protect the latter’s content? Under what conditions