Fair cost sharing for infrastructure development: A cooperative game-theoretic approach

In the European Union, a heated debate has emerged about whether the current practice of network neutrality regulation might endanger the development of very high-capacity networks and thus threaten the competitiveness of the European telecommunications sector. Operators insist that large content providers, who are effectively free riding on operator infrastructure, should be required to contribute a "fair share" of the costs of delivering their content; they believe new policies should facilitate this. Yet in most policy discussions, "fairness" appears as a purely rhetorical concept, seldom accompanied by concrete methods of quantification. Many policy recommendations suggest that cost shares should be bargained by operators and content providers, and regulators should step in if negotiations are not successful. Although this paper does not take a position on either side of the debate, by employing a cooperative game-theoretic framework it seeks to clarify what "fair share" could mean in rigorous economic terms. My theoretical model builds on the externality created by content providers originating large volumes of internet traffic, who do not pay for and have no incentive to limit the resulting infrastructure costs. Central to our cooperative framework is the hypothetical "grand coalition," in which all players (operators, content providers, and consumers) cooperate to maximize a joint surplus, thereby fully internalizing this externality. Hypothetical coalitions, including only part of the players, might also partially reduce it. Each coalition's value is defined in its ability to mitigate the externality, and each player's contribution to that mitigation provides a basis for allocating costs. I rely on the Shapley value to examine how operators, content providers, and consumers should share trafficgenerated costs under this widely used conception of fairness. My results suggest that, according to the Shapley value, operators and content providers should split costs equally, leaving no direct burden on consumers. In the case of multiple competing operators, their half of the cost should be further split between them equally. The model aims for ease of practical implementation; hence, the suggested share measures depend only on observable cost parameters. With such a cost-sharing regulation in place, and taken into account by players in their optimization, the market equilibrium shifts closer to the social optimum, infrastructure deployment rises, and at high infrastructure cost levels, both consumer surplus and total welfare improve.