時変係数同時方程式モデルによる電力市場売買応札曲線の推定

摘要

Simultaneous-equations model (SEM) is generally used in economics to estimate interdependent endogenous variables such as price and quantity in a competitive, equilibrium market. In this paper, we have attempted to apply SEM to JEPX (Japan Electric Power eXchange) spot market, a single-price auction market, using the publicly available data of selling and buying bid volumes, system price and traded quantity. The aim of this analysis is to understand the magnitude of influences to the auctioned prices and quantity from the selling and buying bids, than to forecast prices and quantity for risk management purposes. In comparison with the Ordinary Least Squares (OLS) estimation where the estimation results represent average values that are independent of time, we employ a time-varying simultaneous-equations model (TV-SEM) to capture structural changes inherent in those influences, using State Space models with Kalman filter stepwise estimation. The results showed that the buying bid volumes has that highest magnitude of influences among the factors considered, exhibiting time-dependent changes, ranging as broad as about 240% of its average. The slope of the supply curve also varies across time, implying the elastic property of the supply commodity, while the demand curve remains comparatively inelastic and stable over time.%現時点でわが国随一の卸電力取引所である日本卸電力取rn引所（Japan Electric Power eXchange: JEPX）における取rn引が開始されて2年あまりが経過し，市場への関心や活用rn度の高まりとともに，価格予測や市場構造分析等に関rnする研究が近年蓄積されつつある。今後とも市場の厚みのrn増大と，参加者から見た利便性の向上などが期待されていrnる。