제로레이팅(zero-rating)이란 이동통신 이용자로 하여금 특정 온라인 콘텐츠를 이용자의 월 데이터 한도에서 데이터의 차감 없이 업로드/다운로드 할 수 있도록 하는 것을 의미한다. Somogyi(2017)를 비롯한 기존 연구는 독점 ISP를 가정하고, 독점 ISP가 특정CP와 배타적인 계약을 체결하고 제로레이팅을 도입하는 이슈를 분석하였다. 기존 연구와는 달리 본 연구는 경쟁 ISP를 가정하고, 영향력 있는 CP가 특정 ISP와 배타적인 제로레이팅을 도입하는 이슈를 분석하였다. 경쟁 ISP 시장을 분석하기 위하여 Somogyi(2017)의 모형을 차별화된 서비스를 제공하는 경쟁 시장을 분석하는데 광범위하게 사용되는 호텔링 모형으로 확장시켰다.
분석의 결과는 다음과 같다. 첫째, ISP가 이용자에게 제공하는 월 데이터 한도(data cap)가 충분히 커서 이용자가 원하는 최대 데이터 이용량과 차이가 크지 않을 때, 부분게임 완전 내쉬 균형에서 어떤 ISP도 제로레이팅을 제공하지 않는다. 둘째, 월 데이터 한도가 충분히 작은 경우, 균형에서 콘텐츠 공급자(content provider, CP)는 두 ISP 모두와 제로레이팅 계약을 체결한다. CP가 하나의 ISP와 배타적으로 제로레이팅 계약을 체결하는 것은 균형이 아니다. 이러한 분석 결과는 영향력 있는 CP와의 배타적 거래를 통한 경쟁 ISP의 봉쇄의 가능성이 높지 않다는 점을 시사한다.

When a mobile operator offers a zero-rating plan for data of a content provider(CP), an end user can upload/download the data of the CP without counting against her monthly data cap. Recently this business model is very much popular with a lot of mobile operators worldwide.
According to Open Internet Order, FCC(Federal Communications Commission) does not ban zero-rating. FCC rules that zero-rating is subject to general regulations and that lawsuits against zero-rating could be evaluated case by case. On the other hand, EU Guidelines(BEREC, 2016) put more emphasis on the conformity of zero-rating with network neutrality principle than the effects on market competition. EU views that a negative impact on CP competition would impair consumer choice of contents.
Eisenach(2015) classifies zero-rating practices into two categories according to the relationship between Internet service providers(ISPs) and content providers(CPs): carrier initiated zero-rating and sponsored data. Schnurr & Wiewiorra(2018) argue that sponsored data violates network neutrality since a CP entering into an exclusive zero-rating contract with an ISP obtains a competitive advantage over other competing CPs. However, they consider carrier initiated zero-rating non-discriminatory and in line with network neutrality.
Existing studies on zero-rating focus on the exclusivity of the contract and its anticompetitive effects on market competition. They model a monopolistic ISP and examine anticompetitive effects of an exclusive zero-rating between the monopolistic ISP and one of the CPs on the CPs whose data are not zero-rated. Specifically, Somogyi(2017) models a monopolistic ISP, 3 CPs and end users with a Cobb-Douglas utility function. Two of the CPs are perfect substitutes and could be zero-rated by the ISP. The remaining CP represents the rest of CPs existing in the real world and is not subject to zero-rating. It is shown that high or low utility contents are zero-rated depending upon the size of parameters in the model.
This paper extends Somogyi(2017) by assuming duopoly ISPs in a Hotelling model. The results of the analysis are as follows. First, when data cap of an end-user provided by an ISP is sufficiently large so that the difference between the maximum data traffic the end-user wants to consume and the data cap is sufficiently small, no ISP adopts zero-rating in sub-game perfect Nash equilibrium of the game. Meanwhile when the data cap is relatively small, CP enters into a contract with both ISP’s to zero-rate its contents. In equilibrium, CP never enters into an exclusive zero-rating contract with one of the ISPs. This result hints that it is not very likely that an exclusive zero-rating contract between an ISP and a very attractive CP works as a vehicle facilitating foreclosure of competing ISPs.
However, we do not attempt to make a strong stance that zero-rating in a competitive Internet service market does not cause an anticompetitive effects on the ISP out of a zero-rating contract. For a more complete analysis of the issue, we need to examine zero-rating practices when there is a dominant player in Internet service market. It is also necessary to investigate zero-rating in a model with a multiple number of CPs and ISPs since the incentive to engage in an exclusive zero-rating contract between an ISP and a CP could be affected by the existence of competition in both ISP and CP markets. Finally, we need to take into account of competition in pricing for the contents by CPs.