The Internet Economy


(B) I had a chance to read some of the research of the 2014 Economy Nobel Prize winner Jean Tirole about the innovation and the adoption of new technologies. Probably his most well-known paper is “Platform Competition in Two-Sided Markets”. Examples of well-known two-sided markets are those for instance of Apple, Google, and Facebook where the platform such as a software platform or an Internet service has both an ecosystem of users and apps developers or online advertisers. The user of the platform or the service is both the “customer” and the “product”, by being the target for the advertising or developer revenues. The provider of the platform of a two-sided market can gain a monopoly position if the platform scale and is successful. The more users use the platform, the more the platform grows in scale and in profitability and the more impossible it becomes for a new player to invest in a new platform that can match that scale.

Following is a summary of Tirole’s work related to the technology industry from the Nobel Prize Web site in particular two-sided markets, technology patents and open source software.

Two-Sided Markets

The current industrial policy has to deal with new forms of competition, often linked to the introduction of new technology. Two examples are network competition (Laffont, Rey, and Tirole) and two-sided (platform) markets (Rochet and Tirole). Tirole is a leader in the study of these new forms of competition. In a two-sided market, the two sides (say, buyers and sellers) interact via a platform. Examples include operating systems, payment cards (credit, debit or charge cards), shopping malls, and TV-channels.

A concrete example of a two-sided market is given by credit-card networks (such as Visa, Mastercard, or American Express). The two sides of the market are the consumers and the retailers. If a certain credit-card company charges retailers a high transactions fee, a retailer might decide to not accept this card. This might, however, lead consumers who prefer this card to shop elsewhere. On the other hand, there is a positive feedback loop Tirole analyzed the equilibrium of this kind of two-sided market and studied its questions addressed in these articles include the equilibrium pricing structure and the extent to which consumers and retailers use more than one network (“multi-homing”).

In platform markets, demands from the two sides can be very different. For example, advertisers might desire that there be many viewers or readers, whereas viewers and readers often prefer that there be few advertisers. As a result, prices that would be clearly anti-competitive in a one-sided market can be highly competitive in a two-sided market. For example, offering newspapers for free would be a sign of predatory pricing if the newspaper’s only source of revenue came from readers, but may be entirely consistent with competitive pricing if advertising revenues are important. Because conventional tests for anti-competitive behavior are not applicable in platform markets, the work by Rochet and Tirole has had an immediate impact on competition policy (see Evans). This work has also influenced the ongoing debate about network neutrality — i.e., a regulation prohibiting broadband-access providers from charging content providers for access to broadband customers (see Musacchio, Schwartz, and Walrand).

Patents From Research & Development

R&D competition among firms is often portrayed as a race to obtain a patent, where a firm that spends more resources increases its chances of winning. The firm that has invested the most in R&D in the past is the leader in that race — i.e., it has the highest probability of winning. Thus, past (irreversible) investments can create a first-mover advantage and a follower may not even want to participate in a race he is unlikely to win. This logic was formalized by Fudenberg, Gilbert, Stiglitz, and Tirole. They showed that if past R&D expenditures are perfectly observed, and if the patent is a one-time discovery (there are no intermediate discoveries), then the follower will indeed be deterred from entering the race. However, if past R&D expenditures are not perfectly observed, the more plausible case, the follower may be able to leapfrog the leader, and the competition to obtain the patent becomes very stiff (unless the leader is very far ahead, in which case the follower gives up). Similarly, competition is stiff if intermediate discoveries can be made — in this case, the follower can leapfrog the leader by making the first intermediate discovery.

The results of Fudenberg et al. suggest that a patent race will be very intense, with high R&D expenditures from both firms, when the race is close. Thus, at least under some conditions, intense competition will stimulate R&D efforts. If the leader is too far ahead, the follower gives up, and the leader reduces his R&D expenditures, as he is no longer challenged.

Open-Source Software (OSS)

The literature on patent races assumes that R&D investments are driven by the traditional profit-motive. However, economists are increasingly realizing that not all new technologies originate in this way. An important example is Open Source Software (OSS), where programmers at many different locations and organizations share code to develop new software. This process was important in the development of the Internet, contributing such software as TCP/IP, BIND, Perl, Sendmail, Linux, Mozilla, and Apache. What initially baffled economists was that OSS developers did not seem to benefit financially from their efforts. OSS is, in effect, a public good, which raises the question why OSS programmers contribute voluntarily, without pay, to the public good.

Lerner and Tirole argued that economic theory may, in fact, be able to answer this question. Their main hypothesis is that software developers have career concerns. Contributing to the OSS may be a credible signal of one’s programming ability, which may lead to job offers, shares in commercial open source-based companies, or access to the venture-capital market. Drawing on a previous literature on career concerns (e.g., Holmstrom), they argue that the signaling motive is stronger the more visible is performance, the higher the impact of effort on performance, and the more informative performance is about talent. They find support for the signaling motive in four case studies (Apache, Linux, Perl, and Sendmail). Subsequent research also found support for Lerner and Tirole’s signaling theory, at least for some OSS projects (e.g., Hann, Roberts, and Slaughter for Apache). However, programmers seem to contribute to OSS for a variety of reasons, perhaps including altruism. A satisfactory explanation of OSS development may require a mixture of signaling theory and insights, like those of 2009 Economics Laureate Elinor Ostrom, into how cooperative behavior can be supported by norms and other social mechanisms (O’Mahony).


Jean-Charles Rochet and Jean Tirole, “Platform Competition in Two-Sided Markets
The Royal Swedish Academy of Sciences, “Jean Tirole: Market Power and Regulation

Note: The picture above is Van Gogh’s painting “La Nuit Etoilee” from the Musee d’Orsay.

Copyright © 2005-2015 by Serge-Paul Carrasco. All rights reserved.
Contact Us: asvinsider at gmail dot com.

Categories: Economy