Forecasting Electricity Consumption for Pakistan

Now-a-days, different sectors of the economy are being significantly affected by the electricity variable. In this research, we analyzed the monthly electricity consumption in Pakistan for the period of January 1990 through December 2011, using linear and non linear modeling techniques. They include ARIMA, Seasonal ARIMA (SARIMA) and ARCH/GARCH models. Electricity consumption model reveals a significant trend due to socioeconomic factors. The monthly behavior of our forecast values depicts that the electricity consumption is more for summer season and this demand is expected to increase in future. Forecast model and the forecast values reveal that the electricity consumption is increasing with the passage of time. We have checked the accuracy of the models by diagnostic tests and then compared the forecast values to select the more appropriate model. In this study, the least out of sample forecast performance i.e. Mean Absolute Percentage Error (MAPE) value and the minimum forecast standard deviation value of ARIMA (3, 1, 2) show that among the four competing time series models, ARIMA(3, 1, 2) model is the most appropriate model to forecast electricity consumption in Pakistan. Keywords— ARCH/GARCH, ARIMA, Electricity Consumption, Forecast accuracy, SARIMA.