This licentiate thesis examines marginal cost estimation and pricing in transport infrastructure, with empirical applications to Swedish roads and railways. The thesis consists of three papers addressing how infrastructure charges can be aligned with short-run marginal social costs.

The first paper estimates marginal road renewal costs per heavy vehicle kilometre using panel data on approximately 1.3 million 100-meter road segments (1999–2022). A two-part discrete–continuous framework is applied, combining models of renewal timing (probit and Weibull survival specifications) with renewal expenditures (a log-log function). The results show substantial heterogeneity in marginal renewal costs across road types, with heavy vehicles generating higher costs on low-standard roads than on motorways or other high-standard roads.

The second paper estimates short-run marginal road maintenance costs per heavy vehicle kilometre using administrative maintenance data at the Maintenance District Unit (MDU) level (2015–2024). A log-log cost function relates annual maintenance expenditure to heavy-vehicle traffic intensity by road type, controlling for network scale and weather conditions. The results indicate that the marginal maintenance cost per heavy vehicle kilometre is about 50 percent higher on low-standard roads than on high-standard roads. Combined with the renewal estimates from Paper 1, total marginal infrastructure costs range from 0.03 €/heavy-vkm on motorways to 0.51 €/heavy-vkm on low-standard roads, indicating that the marginal infrastructure cost imposed by heavy vehicles varies greatly across the road network.

The third paper analyses optimal rail access charges in a vertically separated railway market, where marginal capacity costs reflect track congestion. A calibrated demand and supply model for the Stockholm–Gothenburg corridor is used to simulate how track charges affect fares, frequencies, welfare, and modal competition. The results show that increases in track charges to reflect the marginal cost of congestion primarily reduce service frequency rather than substantially increasing fares, with welfare-optimal charges depending critically on congestion externalities.

Together, the papers show that marginal infrastructure pricing requires both disaggregated cost estimates and an understanding of how transport operators respond to infrastructure charges.