Investment in Two Sided Markets and the Net Neutrality Debate Nicolas S;er‐Moses Columbia Business School Joint work with Paul Njoroge, Asuman Ozdaglar Gabriel Weintraub MIT Columbia Business School Modélisa;on en Economie des réseaux et NEUtRalité du Net Kick‐off Mee;ng, May 26‐27 2011, INRIA Paris
The Net Neutrality Debate Net neutrality: Internet Service Providers (ISPs) cannot exercise discrimina;on against content. All bits should be treated equal; no “priority lanes”. ISPs should not ver;cally integrate into content provisioning. Network providers should not be allowed to charge off‐network content providers (CPs) for consumer access. Generated heated public debate in US and in Europe .
Net Neutrality Advocates Traffic discrimina;on by ISPs will reduce content innova;on
Net Neutrality Opponents Lack of flexibility for ISPs reduces their investment incen;ves Former CEO of AT&T, Ed Whitacre: “they would like to use my pipes free, but I ain’t going to let them”
Open Debate, … S;ll Lots of Controversy Chopra (CTO Obama): “Important step in preven;ng abuses and con;nuing to advance the Internet” Al Franken (Democrat Senator): “weak rule”
Research Ques;ons Provide formal economic analysis to shed light on the debate. What is the effect of a non‐neutral regime on: ISPs’ investments? CPs’ innova;on? Consumer surplus and social welfare? Contribu;on to literature Net‐neutrality models: Hermalin and Katz 07, Choi and Kim 08, Economides and Tag 08, Schwartz, Musacchio, and Walrand 09, and many more yesterday and today in the mee7ng . Two‐sided markets: Armstrong 06, Rochet and Tirole 03, Farhi and Hagiu 07.
Our Approach We plan to provide some answers by looking at a model whose players are: A con;nuum of heterogenous users who connect to 1 ISP and many CPs A con;nuum of heterogenous CPs A doupoly of ISPs that invest in QoS
Net Neutrality as a Pricing Rule: “Last Mile” Access Fees Neutral regime: CPs are not charged Non‐Neutral regime: CPs are charged to access offnet consumers. to access offnet consumers. CP j CP j $ $ $ ISP α ISP β ISP α ISP β q α q β q α q β
Preview of Main Results Aggregate ISPs investments are higher in the non‐neutral regime. CPs’ profits and consumer surplus are generally superior in the non‐neutral regime. Social welfare is also generally superior in the non‐ neutral regime.
Extensive Form Game: Players Duopoly of ISPs
Solve for subgame perfect Nash equil. 1. Quality Choice : Plamorms make costly investments in QoS y α and y β 2. CP Prices : Plamorms choose CPs connec;on fees w α and w β 3. CP Connec7ons : CPs decide which ˆ plamorm to join φ ( j ) 4. Consumer Prices : Plamorms choose consumer prices p α and p β 5. Consumer Connec7ons : Consumers decide which plamorm to join φ ( i ) 6. Consump7on Decisions : Consumers decide CPs to patronize
Consumer U;lity Connec;on quality consumer i – CP j : The min captures a “ bo0leneck effect ” u ij = q ij f ( γ j ) U;lity consumer i – CP j : This f ( . ) depends on addi;onal services offered by ISP and on conges;on U;lity consumer i :
CP U;lity g ( γ j , q ij ) Adver;sing rate charged by CP j for consumer i : Neutral model: connect to α or β � π j = g ( γ j , q ij ) di − w ˆ φ ( j ) all consumers Non‐neutral model: directly connect to α and / or β � Connect to α : j = g ( γ j , y α ) di − w α π α consumers in α � π β Connect to β : j = g ( γ j , y β ) di − w β consumers in β j + π β π α Connect to both : j
Main Results Theorem Neutral Model: Under mild technical condi;ons, there exists a unique subgame perfect equilibrium (SPE): one ISP invests in posi;ve quality and the other chooses not to invest. ISPs differen;ate in quality to sooen consumer price compe;;on Theorem Non‐neutral Model: Under mild technical condi;ons, there exists a unique subgame perfect equilibrium (SPE): both firms choose to invest in quality. ISPs trade off differen;a;on on consumer side (that sooens compe;;on) for the opportunity to make profits in CP side.
Welfare Comparisons CPs: prefer the non‐neutral regime because investments by both plamorms increase adver;sing revenue. Non‐neutral regime doesn’t reduce market coverage. Consumers: prefer the non‐neutral regime because increased compe;;on between plamorms lowers consumer prices. Low‐quality plamorm: prefers non‐neutral regime because it exploits its monopoly over access. High‐quality plamorm: prefers the neutral regime because plamorms are maximally differen;ated. Theorem: The non‐neutral regime has a higher social welfare compared to the neutral regime.
Observa;ons / Some Limita;ons of Model Defini7on of Net Neutrality? Is it realis;c that consumers can’t access some CPs? We understand this as basic vs. premium service. It would be nice to let CPs choose investment in quality. Probably can be done but at the expense of ISP investments or not having compe;;on among ISPs. It would be also be nice to add conges7on on the consumer side . May be more realis;c than conges;on on the CP side. The last stages become harder to solve (more like a Wardrop equilibrium) making it much harder to provide a closed‐form equilibrium for first stages.
Conclusions Results suggest that non‐neutral regime increases aggregate social welfare as well as individual welfare of almost all agents. Mechanisms related to ISP investments seem robust. Investment incen;ves play key role in net‐neutrality debate: In the non‐neutral regime it’s easier to extract surplus through appropriate pricing, so aggregate investment is larger. Net neutrality: NOT a zero‐sum game. CP and consumer u;li;es are enhanced by plamorm investments. Inform the ongoing debate with formal economic analysis
Investment in Two Sided Markets and the Net Neutrality Debate Nicolas S;er‐Moses Columbia Business School Joint work with Paul Njoroge, Asuman Ozdaglar Gabriel Weintraub MIT Columbia Business School Modélisa;on en Economie des réseaux et NEUtRalité du Net Kick‐off Mee;ng, May 26‐27 2011, INRIA Paris
Extensive Form Game: Players Duopoly of ISPs
Extensive Form Game: Players Mass of CPs with heterogeneous quality γ j with endogenous par;cipa;on Plamorms ISP α ISP β with y α y β quali;es: Mass of consumers with heterogeneous preferences for quality θ i (full coverage)
Open Debate, … S;ll Lots of Controversy
Recommend
More recommend