A Model Research on PJM Day-Ahead Market Locational Marginal Price Spread

In this paper, day-ahead electricity price spread data in the PJM power market are modeled by Markov-switching generalized autoregressive conditional heteroskedasticity (MS-GARCH) model and the standard GARCH model. The MS-GARCH models are respectively estimated with two and three regimes. Research results show that locational marginal price spread volatility is not only highly volatile but also has significantly strong regime dependence, besides, also find that both the model fitting and for one-and multi-step ahead prediction according to such as MSE, RMSE and MAE index calculation, the results of MS-GARCH model are better than the standard GARCH model in fitting and forecasting. This electricity price spread forecast result can provide a reference for power market members to bid financial transmission rights in an annual auction, so as to maximize the benefits of market participants.