Regulating Paid Peering: A Two-Sided Market Perspective

Discussions over paid peering and usage fees have become a global phenomenon, with ISPs insisting content providers pay for the substantial downstream traffic they generate, while content providers argue for settlement-free agreements, citing that consumers already pay ISPs for content delivery and content providers reduce costs for ISPs by delivering traffic close to customers. This contention is echoed in debates around net neutrality and related policies across the United States, Europe, and South Korea, prompting regulators to consider whether to regulate peering prices and/or impose usage fees. This question gains significance in the United States in light of the recent Federal Communications Commission (FCC) decision to consider the reinstatement of net neutrality rules. Our study examines whether the peering price set by the market aligns with the socially optimal peering price. We first analyze the cost-based peering price, which reflects the ISP’s incremental costs for direct content delivery versus via transit providers. We discover that this price can effectively be zero with sufficient content localization, yet remains above zero without it, influenced by the number of interconnection points. We then evaluate the profit-maximizing peering price using a two-sided market model, finding that increased content localization decreases this price. These prices establish a range if the peering price is unregulated, from the cost-based peering price (at the low end) to the profit-maximizing peering price (at the high end). Regulatory oversight of peering prices may be warranted when there is a substantial difference between cost-based and profit-maximizing prices. Finally, our study informs the debate on potential regulatory interventions about peering prices or usage fees. We compare unregulated market outcomes with those under regulatory measures aimed at maximizing consumer surplus or social welfare. Through our analysis, we identify optimal peering prices within our determined range, considering the influence of content localization and interconnection points on these prices.