Regarding some regulation fields, such as optimal investments and pricing policies, marginal cost estimations for infrastructure-intensive transport services is always a challenging task. The lack of comparable data among airports is one of the causes which could explain the relative scarcity of this literature in the past. In this paper, the returns to scale and marginal costs are estimated using single-and multi-product translog specifications of a long-run cost function. A pooled database of financial data on 41 airports across Europe, North America, Asia, and Australia for the period 1991-2005 was used. Significant economies of scale and little degree of technological progress are found using Work Load Units (WLU) and Air Traffic Movements (ATM) as output measures. Additionally, individual long-run marginal cost estimates are provided for each output measure, and for every airport under study.