INTRODUCTION

Outlining the elements of an industrial organization of the Internet requires an opera­tional delineation of its boundaries and reliance on theoretical frameworks appropriate to its rapid pace of change.

Narrowly construed, the Internet can be operationalized as a physical and logical network of networks based on open shared protocols that enable digital information flows among an increasing number of individuals, organizations and devices. From this vantage point attention is directed to issues such as technological change in the Internet; network platforms and their economic characteristics; or the structure and performance of component, access, backbone, and storage markets. The analysis may focus on par­ticular segments of this infrastructure (e.g., the industrial organization of access markets or the economics of cloud infrastructure) or at emergent properties at an aggregated level such as the overall patterns of innovation or the effects of the Internet on productivity and employment (in a particular country, region, or the world).

Concentrating on the network infrastructure sharpens the focus but it also excludes important aspects of the Internet that are critical for its overall organization. Additional theoretical and empirical questions emerge by widening the perspective to broader Internet ecosystems, including applications and the proliferating number of services offered over the Internet.

We combine these two perspectives to examine the economics of the Internet from micro and macro perspectives. At a ‘lower’ micro level we see multiple technologies and market segments (e.g., access, backbones, Internet exchanges, hosting, and specific services) (Comer, 2006) with specific economic characteristics. With the transition from

23

narrowband to broadband Internet the evolution of alternative fixed and mobile access infrastructures and the services they enable becomes particularly relevant. At a ‘higher’ level we see properties that emerge from these component markets, such as investment and innovation patterns at the sectoral level. This multi-pronged way allows systematic economic theorizing on how decisions and behavior at lower levels of the Internet con­tribute to (emergent) properties at higher levels of the system.

A second challenge is the selection of theoretical frameworks appropriate for analyzing this dynamic system of interrelated technologies, markets, and non-market activities. At a first glance many of the issues arising in the Internet economy look similar to those of traditional industrial economics and the economics of networks. Nonetheless, the tight combination of high fixed and low incremental cost, the pervasive presence of increas­ing returns, the rapidity and frequency of entry and exit, high rates of innovation, and economies of scale in consumption (positive network externalities) have created unique economic conditions that have inspired the term ‘new economy’.¹ One approach is to apply concepts of traditional industrial organization to the Internet, building on insights from game theory, the economics of networks, and institutional economics, among others.

However, the methodological conventions of these approaches, such as the analysis of market equilibria and the treatment of technological change as exogenous, while allow­ing analytical rigor, may not reflect the unique dynamics of the Internet with sufficient accuracy. An alternative route is to seek radically new frameworks that allow better capture of the new attributes of the Internet economy. Interesting strands of research have emanated from evolutionary and complexity theory (e.g., Dorogovtsev and Mendes, 2003; see also Schultze and Whitt, Chapter 3 in this volume) and from network theory (e.g., Jackson, 2008; Easley and Kleinberg, 2010; see also Schneider and Bauer, Chapter 4 in this volume). Promising avenues are also opened by computational and agent-based approaches that allow the simulation of highly non-linear systems such as the Internet.

We opt for a middle ground by combining insights from traditional industrial organi­zation with concepts rooted in the works of Schumpeter (1934), who emphasized the dynamic forces of innovation, and an evolutionary view of markets as dynamic mecha­nisms of discovery and coordination (e.g., Hayek [1968] 2002). As several authors have recognized, digital technology has changed and accelerated the process of innovation by allowing continuous experimentation, rapid feedback and evaluation, and imitation and scaling of successful business models (Brynjolfsson, 2011; Brynjolfsson and McAfee, 2014). The Internet is an integral driver of these developments and is affected itself simi­larly. Not surprisingly, digital industries have also inspired new research on the role of competition and antitrust in interrelated markets (e.g., Posner, 2000; Farrell and Weiser, 2003; Evans and Schmalensee, 2007).

A basic message of much of this research is that dynamic markets like the Internet should neither be exempted from general competition policy nor do they justify new and specific regulatory and antitrust policies.

Since legal entry barriers have been lifted in network industries worldwide, end-to-end vertically integrated network systems owned and operated by legal monopolists have been replaced by liberalized markets for network services differentiated from the markets for network infrastructures capacities. Consequently, horizontal and vertical coordination issues arise, including interconnection and mutual compensation, that are addressed in innovative ways and open new opportunities for innovative entrepreneurial bundling and unbundling decisions (Knieps, 1997b). Since the different components of the Internet (e.g., computers, software) are not isolated but connected via networks, questions regard­ing the system character of networks and the boundaries of markets in network indus­tries have to be considered. The study of other large technical systems (e.g., Mayntz and Hughes, 1988) shows that an analytical separation of networks into individual compo­nents and decision-makers risks neglecting important system interdependencies.

This velocity of change is visible in the changing focus of the overarching paradigm of the Internet, which has shifted from characterizing it as a ‘network of networks’ to a ‘network of platforms’ (Greenstein, 2009, 2012b) and an ‘Internet of Internets’ (Noam, Chapter 27 this volume). Several authors have pointed to the increasing role of communi­cation processes and transactions outside of the public Internet (e.g., in private Internets and in darknets; see claffy and Clark, 2014). Apart from illegal activity that seeks to avoid the public Internet, this development seems to be primarily a response to the quest of finding sustainable business models (a walled garden facilitates establishing a payment system), to challenges of information security (keeping sensitive information behind secure walls), and a desire to escape more stringent regulation imposed on the public Internet. A key dynamic is the interdependent innovation process unfolding between the vibrant applications and services sector and the network infrastructure.

To fully understand these economic and technological dynamics a closer look at the interplay of innovations in application services (at the logical edge)³ and innovations for traffic services (at the core) is required. The capabilities supported by the core network enable and constrain the types and variety of innovation possible at the edge. At the same time, the diversity and number of innovative services at the edge influence the value of the network and have implications for innovation and investment incentives at the core layers. Whereas much of the public debate focuses on edge innovations, innovations at the core are also relevant. With heterogeneous applications and user groups it is useful to look at markets for Internet traffic services as a separate market, providing platforms for Internet application service providers and end users. From a technical vantage point, platforms facilitate the building of related products and services; from an economic perspective, they facilitate transactions between related players and the internalization of externalities (Gawer and Cusumano, 2002; Hagiu and Wright, 2012).

In this historical development the search for appropriate governance mechanisms, particularly the ‘proper’ role of markets, the state, and forms of networked governance (e.g., standardization committees) is an ongoing process. Non-government and non­market organizations of networked governance such as the Internet Corporation for Assigned Names and Numbers (ICANN), the Internet Governance Forum (IGF), and other voluntary organizations such as the Internet Society address coordination tasks across the entire Internet ecosystem (Mueller, 2010). An important example is the chang­ing role of the Internet Engineering Task Force (IETF) as a forum to achieve universal­ity and interconnectivity in the best-effort Internet and the development of innovative quality of service (QoS) differentiated traffic architectures (Knieps, 2015b). At the same time, traditional government players, including national regulatory agencies, legislatures, and intergovernmental organizations like the International Telecommunication Union (ITU) have increasingly asserted a role in addition to decentralized markets.

The remainder of the chapter is organized as follows. Section 2.2 briefly discusses the development and unique attributes of the Internet through an economic lens. With this background, section 2.3 focuses on the industrial organization of the dynamic environ­ment of the Internet. Section 2.4 discusses technology and cost conditions followed by an exploration of demand and pricing in section 2.5. Section 2.6 focuses on the interplay of innovation and service quality in the Internet. The chapter returns to one of the opening questions - what can and cannot be coordinated by market forces - in section 2.7. An outlook on possible future developments and emerging research questions concludes.

2.2