Govaerts, NielsBruninx, KennethLe Cadre, HeleneMeeus, LeonardoDelarue, Erik2021-11-302021-11-3020210922-680X10.1007/s11149-021-09432-5http://hdl.handle.net/20.500.12127/6993Many regulators are pushing for more cost-reflective distribution network charges to inform end users of the grid infrastructure costs their behavior causes. Since future investment costs can be avoided by reducing simultaneous peak loads, forward-looking, coincident peak charges are often proposed. Under the assumption of convex network costs, it has been shown that optimal charges signal long-run marginal network costs, triggering an optimal trade-off between network expansion and peak load reduction. In practice, however, network investments are lumpy, requiring engineering methods to estimate ill-defined marginal costs based on long-term peak demand forecasts. In this paper, we derive the optimal forward-looking network charge set by a social welfare maximizing regulator, endogenously considering investment lumpiness and uncertain consumer demand. While the optimal tariff still equals marginal network costs in essence, it now depends on a multitude of network- and demand-related parameters. Our results demonstrate that forward-looking network charges require accurate information on willingness to pay for peak demand, which currently is typically unknown to regulators.enRegulatorsJournal of Regulatory Economics1573-0468151626