Regulation Methods for Assessing Cost Efficiency of Electricity Grid Companies in Sweden

Sammanfattning

Sweden’s electricity distribution regulation relies on benchmarking to assess cost efficiency, yet the current methodology, primarily based on DEA, lacks flexibility and fairness. Especially when it comes to rising demands for electrification and reinvestment. This thesis compares two benchmarking methods, Data Envelopment Analysis (DEA) and Stochastic Frontier Analysis (SFA), using eleven model simulations based on data from 148 distribution system operators (DSOs). The study examines how methodological choices, parameter selection, and company characteristics influence efficiency scores. DEA often overestimates efficiency for structurally advantaged firms, while SFA may understate high performers by attributing part of their success to statistical noise. Parameters like number of customers, network stations, transferred energy, and outage costs showed the strongest impact. Firms with aging, depreciated infrastructure were often misclassified as efficient in DEA due to low input levels. Efficiency score differences between DEA and SFA exceeded 20–30% in many cases, especially for smaller or frontier-defining firms. These differences reflect core model assumptions, such as scale returns and noise treatment. No single method proved optimal, but combining DEA and SFA, or applying hybrid models like SDEA or StoNED, could provide more balanced and fair assessments. The study highlights the need for more robust and transparent benchmarking frameworks in Sweden, better aligned with long-term goals for electrification, reinvestment, and grid sustainability.