June 15, 2005

How competitive is the Internet?

The Internet's image is that of a decentralized industry with low barriers to entry. But a study of market concentration in the Internet sector and related industries yields some surprising results.


During the Internet’s boom years, when financing was easy and entry barriers were low, the Internet sector had a reputation for being wide open and competitive. But now that the easy financing has dried up, are a few firms with market power coming to dominate the sector? If so, what are the implications of greater market concentration?

In an effort to answer these questions, Professor Eli Noam conducted the first empirical study of market concentration of the Internet sector. His definition of the sector includes eight subindustries: backbones, service providers, broadband providers, portals, search engines, modems, browsers and working equipment. This definition excludes Internet applications, content, computer hardware and the underlying telecom/cable infrastructure.

For each of the eight subindustries, Noam calculated the market share of individual firms. Then, using those market shares and the Herfindahl-Hirschman index (HHI), the major concentration index used by the U.S. Department of Justice, Noam tracked the concentration of each industry from 1984 to 2003 (or, for industries that don’t go back that far, since the industry’s inception).

The study’s findings show that the Internet sector’s overall concentration has never been low. In fact, for most of the past 20 years, the sector has been in the highly concentrated range. Graphically, the trend line forms a U-shaped curve. From its highest point in 1984, concentration declined steadily for more than a decade, dipping into the moderately concentrated range during the mid-1990s. The trend reversed after 1996, and since then the sector has become increasingly concentrated.

Broadening the definition of the Internet sector to include such related industries as PCs, operating systems and software, Noam found a similar U-shaped curve. To put these trends into perspective, Noam also analyzed related media industries and subsectors of the information sector. He found that the newer the medium, the more concentrated it is and, in general, the more its concentration has been rising in recent years. Internet media — particularly the newest medium, broadband — have the highest concentration.

Noam went on to compare the concentration of the Internet sector with that of the telecommunications, mass media and information technology (IT) sectors. He found that, despite justified public concern over mass media consolidation, mass media is actually the least concentrated of the four sectors. Like the Internet, telecom is above the Department of Justice’s threshold for “highly concentrated” industries, although telecom concentration has fallen from its postdivestiture levels. Concentration in telecom, IT and the Internet increased after 1996, when the deregulatory Telecommunications Act opened several sectors to further competition.

To analyze vertical integration in the Internet sector, Noam looked at three indicators: the 10 largest Internet firms’ share of the Internet sector; the 10 largest Internet firms’ share of the entire information sector; and the “Power Index,” which aggregates the HHI scores of top Internet firms. All three indicators confirmed that the Internet sector is becoming more vertically integrated, with the largest firms gaining market share and expanding into multiple industries.

What explains the growing concentration of the Internet sector? The most likely answer, according to Noam, is that since a new media generation requires large infusions of capital, there is a natural tendency toward concentration and oligopoly in newer media industries. Each of the eight Internet subindustries has a different story, but the common elements are economies of scale on the supply side and network effects on the demand side. These characteristics encouraged rapid expansions and created a period of intense competition, driving prices to levels that couldn’t sustain costs. The eventual result was the failure of some market participants and consolidation of the survivors with the aim of creating a market structure that could sustain higher prices.

If the concentration of the Internet sector continues, we can expect to see

  • Higher prices for consumers and higher profitability for major Internet firms
  • A slowing of innovation
  • Increased control by major firms of the sector’s standards and protocols and of access to the market for content and application providers (unless these firms have countervailing market power)
  • Cross-subsidies within major firms from segments with market power to segments that are more competitive, distorting competition
  • The emergence of regulation to mitigate the sector’s anticompetitive features

Thus, the Internet sector may move from its entrepreneurial and libertarian model to one of market power resembling that of other electronic media.


Eli Noam is professor of finance and economics and director of the Columbia Institute of Tele-Information at Columbia Business School.

Have You Read Columbia Business?

Columbia Business School’s alumni magazine connects alumni with each other and the School; celebrates alumni milestones and accomplishments; and chronicles the impact of Columbia Business School alumni, faculty members, and students on the global business landscape.

Read Columbia Business>