A Laspeyres Index Decomposition-based Multivariable Grey Prediction Model for Forecasting Energy Consumption: A Case Study of Ghana

Energy consumption is closely linked to a country's economic activity. For most developing countries making efforts to shift to industry-driven economies, the relationship between energy consumption and economic activity cannot be overemphasized. This study, therefore, employs the Laspeyres Index Decomposition (LID) analysis to decompose the change in energy demand into five driving factors according to three effects. The derived factors are then combined with the first order multivariable grey forecast model to form the hybrid model, LID-GM(1,6). The model is applied to the energy consumption situation of Ghana as a case study. The decomposition analysis gives insight into which economic sectors are accountable for the energy demand changes that occurred during the period 2006-2019, and thus serves as a guide for policymaking. The significance of this paper lies in its contribution to the development of the GM(1,N) prediction models. The grey forecast model, based on factors derived from an index decomposition analysis, is used to predict total energy consumed annually in Ghana from 2020 to 2030. The LID-GM(1,6) is evaluated for forecast accuracy and compared with other models. The LID-GM(1,6) has an out-of-sample MAPE of 3.77%, signifying an accuracy of approximately 96%.