How Electric Vehicles Affect Power Distribution Companies’ Efficiency: A Spatial Dependency Analysis

The rapid increase in the number of electric vehicles (EVs) in Norway presents both challenges and opportunities for power distribution companies (PDCs). Growing EV adoption increases electricity demand and places pressure on local distribution networks, requiring substantial investments in grid infrastructure. This study examines the impact of EV adoption on distribution costs and technical efficiency using a spatial stochastic frontier analysis applied to a balanced panel data for Norwegian PDCs over the period 2016 to 2021. The results show that higher EV adoption is associated with increased distribution costs, reflecting the need for network reinforcements and grid modernization. At the same time, Norwegian PDCs exhibit generally high levels of technical efficiency, indicating that distribution networks are able to adapt effectively to rising demand. Importantly, we find that spatial dependence in costs operates primarily through correlated random shocks and unobserved factors rather than through inefficiency spillovers. Once spatial dependence is modeled jointly in both inefficiency and noise, we find that spatial correlation is concentrated in the random noise component, suggesting that shared demand pressures, weather conditions, and other common shocks play a central role in shaping observed cost dynamics. These findings highlight the importance of spatial modeling in electricity network regulation, as ignoring spatial dependence may lead to biased efficiency assessments. From a policy perspective, the results underscore the need for coordinated regional planning, smart grid development, and tariff designs that account for spatially correlated demand shocks to ensure a reliable and cost-efficient integration of EVs into electricity distribution systems. JEL Classification : Q41, Energy: Demand and Supply; Prices; Q50, Environmental Economics: General; L94, Electric Utilities; C23, Single Equation Models; Single Variables: Panel Data Models; Spatio-temporal Models; D24, Production; Cost; Capital; Capital, Total Factor, and Multifactor Productivity; Capacity