Panel cointegration modelling and forecasting of power tariff

Power tariff is an important economic lever of power market. Accurate relationship analysis and forecast of power tariff has played an important role on the rational adjustment of power demand by using the pricing policies. This paper applies recently developed panel unit root and panel cointegration techniques to estimate the long-run and short-run relationship between the power tariff and demand. It takes advantage of an unusual combination of cross-sectional and time-series data sets in order to improve the performance of small sample data and solve the spurious regression problem. With the panel data of power tariff and demand from 1990 to 2005 in United States, the results indicate that the power tariff and demand are non-stationary and have obvious panel cointegration relationship. Then, the fixed effects model based on cross-section weightings is proposed in order to accurately forecast the power tariff.