Economic Perspectives on the Internet

Alan E. Wiseman

Executive Summary

The information technology sector of the United States economy, spurred on by the expansion of the Internet, has undergone explosive growth. According to one estimate, in 1998 Internet-related industries led to the creation of over 1.2 million jobs and generated more than $300 billion in revenue. At the same time the Internet has expanded immensely. As of December 1999, it was estimated that more than 4.9 million commercial websites had emerged, and in the closing months of the year, that number was increasing at a rate of almost 500,000 per month. The rapid growth of the "Information Economy" has drastically changed the manner in which commercial transactions are conducted, as anything from material goods (such as books) to information goods (such as databases) can be procured with the ease of a mouse click. This report investigates the underlying economics of certain facets of the Internet. By examining recent literature addressing economic aspects of the Internet and electronic commerce, as well as more traditional literature on pricing practices and market structure, this report will serve as a roadmap of the current terrain in Internet-related economic matters, as well as a framework for future analysis.

The report covers five primary areas of inquiry. The first section provides a brief technical overview of the Internet and discusses the transmission technologies employed. What is currently known as the "Internet" is a descendent of federally-funded research that began in the Department of Defense in the early 1960's and was maintained by the National Science Foundation until the mid 1990's. Unlike telephone networks where conducting a call requires a dedicated circuit for the duration of the transmission, information is sent across the Internet via packet-switching technology that breaks the transmission down into data "packets" of approximately 200 bytes and sends them from origin to destination. While traveling from end to end, the packets may traverse several different computer networks, and the Transmission Control Protocol/Internet Protocol (TCP/IP) helps facilitate the arrival and reassembly of all packets at their destination. High levels of user activity can lead to congestion on the network, which might slow down, or prevent, packet transmission. No centralized authority governs the Internet, but several industry consortiums and volunteer nonprofit groups help devise standards for interoperability. For the matter of domain name assignment (assigning the proper names for webpages), Network Solutions Incorporated (NSI) currently maintains control over the domain name registry.

The second section of the report discusses different methods employed and theoretical models proposed for pricing user access to the Internet. The most commonly used method is known as flat pricing, where users pay a flat fee to have unlimited access for a fixed period of time (e.g., per month). While flat pricing is very convenient in that it simplifies accounting issues, certain economic properties may make it less than optimal. First, the flat price does not induce users to take into account the congestion costs (an externality) that they impose on other users during peak periods, which can lead to further network congestion. Second, the flat price cannot discriminate between "high" and "low" priority applications, leading to cases in which those who value services the most are not receiving them. To address these problems, several alternative models of access pricing are being studied. The auction approach would require users to attach bids to each packet they submit to the network, and any bid above the market clearing price (as defined by the network's congestion constraints) would be transmitted. Static and dynamic priority pricing models present users with a menu of applications, which vary in price as a function of the priority level the user attaches to the application in question, which will designate its place in line should network congestion occur. Finally, the Paris Metro Pricing model partitions the network into subsections and charges different prices for access to each subsection; users will sort themselves into divisions depending on their willingness to pay, leading to an efficient allocation of network resources. With the impending growth of Internet usage, in the absence of some sort of technological change and/or pricing policy change, it is possible that the quality of service associated with a flat-pricing regime will deteriorate. Among the technological alternatives being considered is capacity expansion, "caching", and new protocols to facilitate differentiated service levels. The pricing models proposed might succeed in alleviating congestion, but as of yet, none have been widely implemented in a real-world setting. Future research might aim at testing these various models against each other, either through simulations or in experimental settings, to determine which options perform the best in managing congestion.

The third section of the report examines economic issues surrounding pricing of goods and services on the Internet. Theory might imply that a reduction in the costs of transactions will lead to lower prices, less price dispersion, and frequent price adjustments by online firms in comparison to their offline counterparts. Early empirical analysis offers mixed support of these hypotheses; in the markets for books and compact discs, prices seem to follow these trends, but these findings do not hold across all markets. Because the Internet environment facilitates easy (and costless) reproduction of information goods, theory might lead us to believe that one should expect to see many products and services being packaged together and sold as bundles. While there are efficiencies to be gained from such ventures, there are also possibilities that such activities might stifle innovation and deter entry of potential competitors into the market. The utilization of technologies that allow firms to monitor surfers' travels across the Internet present numerous opportunities for price discrimination and product customization. The use of electronic agents known as "shopbots" may lead to lower prices which benefits consumers, but may also have the effects of adding to Internet congestion and facilitating industry collusion. Theory and preliminary empirical analysis suggests that regardless of the reduction in transactions costs, there will still be a need for institutions that can provide product expertise and vouch for firm reputations on the Internet. Future research might study the robustness of price differences between online and offline firms across additional markets, the prospects for collusion in electronic markets, and the manner in which the Internet serves as a clearinghouse for second-hand durable goods.

The fourth section of the report examines the role of network externalities in the development of the Internet, and how such effects might affect firms' activities. A good is commonly referred to as possessing network externalities if the value of the good to an individual user increases as more individuals use it. Given that the Internet, by design, is a "network of networks", one would suspect that conventional economic principles should apply to much of its underlying mechanics. Theory implies that in network industries, a consumer's choice of a given technology depends directly on their perceptions about the likelihood of that product's dominance. Once a technology has attracted enough users, it is said to achieve "critical mass," after which it can survive as a viable product, and possibly be the dominant standard. In trying to achieve critical mass, firms might employ various strategies, such as low introductory prices, aggressive advertising, etc., in the hopes of attracting consumers as quickly as possible. On the Internet, network externalities arise in considering network interconnection agreements between Internet service providers, as well as in considering standards adoption by emerging technologies such as digitized media. The ownership of an (arguably) essential facility, such as an Internet backbone, may allow large Internet service providers to squeeze rents from those firms that require interconnection to conduct business. In establishing new standards, there is the potential for excluding competitors from the standard, who may then be harmed because of their incompatiblity with the existing network. Furthermore, in trying to achieve critical mass, some of the strategies employed by firms may be viewed as anticompetitive. At the same time, standards can obviously benefit consumers in that they expand the size of the network. Future research might aim at uncovering the differences between market power that arises from normal competition versus market power that arises from anticompetitive exploitation of network externalities.

The final section of this report considers the issue of taxation of Internet commerce. There is a debate ensuing between the proponents and opponents of online taxation. Those in favor of effective collection of existing sales taxes on the web cite economic research arguing that such taxes are necessary to ensure that the sales tax system is not distortionary, as well as ensuring that state and local governments will not lose sizable revenues as more consumers migrate to the Internet. Those opposed to Internet taxes claim that mandating such taxes will place a large burden on retailers who would likely be required to collect and remit the taxes, as well as stifling the growth of electronic commerce. The existing empirical research has focused on estimating the likely effects of instituting Internet taxes with respect to consumer activity, state revenues, and the level of compliance costs. With respect to consumer purchasing patterns, early analysis has shown that those who live in areas with high sales taxes purchase more often online (and actually buy more) than those who do not; hence, implementing current sales taxes online would likely decrease online purchases. It does not appear currently that governments are losing a sizable portion of their revenues to the web. Finally, there is empirical evidence to suggest that smaller businesses would bear the burden of collecting and remitting sales taxes more so than larger businesses. While the empirical evidence might offer a modicum of support to those who oppose taxation, the Internet as a transactions medium is immature, and further research, studying sales, rates of Internet diffusion, and revenue streams of local governments and bricks-and-mortar stores, is necessary to determine if these findings are representative of a long-run stable outcome.

The author served as an economic research analyst with the Bureau of Economics, Federal Trade Commission in 1999, and is currently a doctoral candidate in political economics, Stanford University, Graduate School of Business. The views expressed in the report are those of the author and do not necessarily reflect the views of the Commission or any individual Commissioner. Furthermore, the views expressed herein do not necessarily reflect the views of the Commission or its staff regarding any pending matter.