Market Power and Platform Design in Decentralized Electricity Trading

A new research paper by economists Nicolas Eschenbaum and Nicolas Greber uses a sophisticated multi-agent AI framework to model strategic behavior in decentralized electricity markets. The study examines how 'prosumers'—households with solar panels (PV) and batteries—trade energy on a peer-to-peer platform. The computational model reveals a Cournot-like market power mechanism: because the price producers receive falls as total exports rise, strategic players intentionally under-utilize their storage and withhold supply to keep prices higher. This behavior raises overall grid settlement costs by about 6% compared to a scenario where all participants are simple price-takers.

The research's key finding is that platform design is critical for mitigating these inefficiencies. The AI simulations show that the 6% cost increase from strategic play is not inevitable. Certain pricing rules and information disclosure policies can largely eliminate the incentive to game the system. Most strikingly, increased competition in battery ownership dramatically reduces supply withholding; the distortion nearly disappears once storage capacity is split among more than three owners. Despite these strategic frictions, the decentralized trading platform remains immensely valuable, slashing a typical consumer's annual electricity bill by approximately 40% compared to relying solely on the traditional grid, with strategic behavior clawing back only about 8% of those total potential savings.