 This paper proposes a novel approach for forecasting monthly electricity consumption based on decomposition techniques followed by various time series models. First, the authors decompose the time series into three components, the long-term trend, seasonality, and stochastic component. Then, they compared the forecast accuracy of various combinations of time series models for each component. Finally, the forecasted values were combined to form the overall forecast. The authors found that the proposed modeling and forecasting framework gave more accurate and efficient results than the benchmarks. Additionally, the final forecasting models produced lower mean errors than those reported in the literature. This article was authored by Hassan Iftiker, Nadeela Bidi, Paolo Conest Rodriguez, and others.