August 24, 2010 | Event Highlights

Shifting Out of Neutral

Game theory shows how flexibility on net neutrality principles could benefit consumers, ISPs, and content providers.


In 2005 the Federal Communications Commission (FCC) changed the Internet’s designation from telecommunication services to information services, lifting non-discrimination requirements on Internet service providers (ISPs). Since then, a debate has simmered over the Internet’s fundamental neutrality.

The notion of the Internet as a neutral space stems from the fact that no single transmission has priority over any other transmission and that ISPs cannot discriminate against any transmission by slowing or restricting it; traffic from all content providers (CPs) is treated equally. For example, traffic generated by a search in Google takes the same time to travel through the system and has the same priority level as traffic generated by a startup search engine running in a garage.

In the eyes of neutrality advocates, the Internet breeds entrepreneurial equalization and innovation: anyone can put up a website to promote a concept, product or service and easily place it within reach of a worldwide audience. For those who hold this view, the Internet’s neutral default is its essence, and is central to its success.

On the other side of the debate is a group mainly represented by ISPs and the telecom industry. These firms invest significant resources to introduce the physical network: digging trenches, laying optic fiber and cable, and going the costly last mile, where the signal from a large main cable must be split into a number of much smaller lines that fan out to individual users. And, maintaining and upgrading the infrastructure brings commensurate costs. For ISPs in the wireless market, investments are even bigger and recurrent because they need to buy rights to use the spectrum from the government at hefty prices, and airwaves offer limited capacity compared to the wired Internet. But ISPs and telecoms have little flexibility in the contracts they can offer to CPs. In their view they should be permitted to offer a greater variety of contract terms, such as charging higher rates for faster or more secure delivery, to recoup their investment. Others view the current Internet as neutral in a limited sense, because CPs with the deepest pockets can (and do), for example, strategically deploy server farms and contract with content distribution networks to reach consumers more quickly.

Connection quality — a connection’s speed and priority among all users vying to connect through an ISP or telecom platform — is at the heart of the neutrality debate. CPs prefer high-quality platforms, which transmit data faster and are less likely to drop packets containing user data, because they allow CPs to collect more ad revenue than with low-quality ones. Of course, consumers also prefer high-quality platforms because these improve their Internet experience.

While the FCC’s most recent moves suggest it will uphold key principles of neutrality, so far the agency has not imposed formal regulations, and the Internet’s neutral future remains unclear. A recent joint proposal from Google and Verizon, which proposes more flexibility on neutrality principles, especially in the wireless market, has amplified the debate. While lawyers consider the contractual implications of a neutral versus a non-neutral Internet and engineers study technical considerations, others, including Professor Nicolás Stier-Moses and Professor Gabriel Weintraub are beginning to examine the pros and cons from an economic perspective. Internet stakeholders such as ISPs and CPs tend to view their interactions as a zero-sum game, but the researchers’ work suggests otherwise.

In the current pricing scheme, Google generates a tremendous amount of web traffic but does not pay ISPs to deliver that traffic. Put simply, Google pays an ISP to get Internet access once. Because the Internet is a complex structure of interconnected ISPs and networks, another service provider may end up delivering the information to end users; this provider does not charge fees to Google. “A central issue in the net neutrality debate is whether ISPs, on the argument that such a great proportion of traffic is generated by a relatively small number of CPs, should be allowed to charge those CPs for access to consumers,” Stier says.

To find out, Stier and Weintraub used principles of game theory to create a model that mimics a two-sided market for Internet access. In their model, an ISP or telecom platform receives payment streams on both the consumer side and the content side, so that platforms compete for consumer and CP fees.

When the researchers modeled investment in quality in the neutral scenario between two platforms they found that it is optimal for the platforms to maximally differentiate and segment the costumer market — in other words, for one platform to invest in infrastructure, improving quality and attracting customers (but also incurring higher costs related to establishing and maintaining quality) and for the other platform to invest as little as possible, attracting customers who will forsake a high-quality connection for low fees. If the low-quality platform were to increase its investment in quality, it could not attract enough additional customers to offset its spending. Should a high-quality platform invest less, it would lose too many customers to remain profitable.

In the non-neutral scenario of the Stier and Weintraub model, a platform can charge all CPs, not only those who connect directly to it, granting platforms monopoly power over access to their consumer bases. The ISPs become gatekeepers; CPs must navigate through ISPs to get access to their consumers. This presents a tremendous revenue opportunity and gives low-quality platforms a much bigger incentive to invest in quality. The low-quality platforms can charge higher prices to CPs, and this extra revenue offsets the inevitable loss on the consumer side, where the increase in platform quality drives up competition for consumers and pushes prices down.

Somewhat ironically, the model suggests that CPs will end up better off in a non-neutral scenario, because overall quality increases. “A poor connection to YouTube,” Stier says, “is not worth much.” As the overall Internet experience improves, users will consume more content, enabling CPs to benefit from higher ad revenue.

Consumers, too, would be better off because they would pay competitive rates for higher quality service. The only player in the non-neutral model that is worse off, the researchers found, is the high-quality platform, which faces more competition because the other platforms invest more, eroding its consumer base.

No model can fully mimic the complexity of real interactions between all stakeholders, so the researchers don’t offer a strict prescription based on their work. But they believe their findings can help policymakers consider a full range of approaches to debate, particularly because it highlights how the incentives of stakeholders may be more aligned than it first appears. In addition, the prediction of the model that stakeholders would improve their welfare without net neutrality sheds some light on why in today’s Internet we see de facto non-neutrality, since CPs employ costly solutions to improve the quality of their users' experiences.

The Stier and Weintraub model focuses on quality investment by ISPs. “We suspect that in some sense we would see similar results if we had explicitly modeled innovation investments by CPs,” Stier says. “CPs and ISPs offer services that are complements to one another, so if CPs are given incentives to innovate, it’s likely that ISPs would be better off.”

“Ultimately, all stakeholders may have something to gain by conceding some ground,” Weintraub says, and points to the Google-Verizon proposal. “The goal of their agreement is to simultaneously encourage investment in infrastructure with openness and innovation on the content side. Google’s willingness to concede on neutrality principles demonstrates that a compromise may offer gains for both sides. The goal of the FCC should be to encourage investments, innovation, and competition in the CP and ISP markets, so that these gains translate into benefits for consumers as well.”

Nicolás Stier-Moses and Gabriel Weintraub are associate professors of decision, risk, and operations at Columbia Business School.

Read the Research

Paul Njoroge, Asuman Ozdaglar, Stier-Moses, Gabriel Weintraub

"Investment in Two Sided Markets and the Net Neutrality Debate"

View abstract/citation  Download PDF  

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>