Global Earth Systems Integrated Assessment Models are used by policymakers to understand the complex interactions between anthropogenic activity, including energy use, and its environmental impacts. However, they are computationally expensive, so spatiotemporal resolution is kept low as a trade-off. Generally, the smallest geopolitical component is a large nation. Yet, much of energy transition planning occurs at subnational levels. Downscaling is the process of turning low resolution model outputs into higher resolution ones and has been used extensively in hydrological and weather modelling. Since downscaling in the energy sector is nascent, this presentation reviews the trade-offs between statistical downscaling methods within the energy domain using ten newly developed criteria. Highlighting such trade-offs can better equip policymakers as they craft energy transition policies. These criteria fit within three main categories: replicability, coherence to the parent model (the IAM), and handling of energy-specific insights. The linear downscaling method was highly replicable and, though coupled to the parent model, assumes that all its constituent regions are homogeneous and distributes outputs blindly. It performs well to introduce technologies that do not exist yet, but does not consider geographic to resource use here. The convergence method similarly struggles with geospatial limitations, but obfuscates energy sector nuances described by the original model. It is less replicable than the linear method. Modifying these approaches could resolve some of their issues, but they will likely never be useful for serious policymaking. Two other methods described in the literature are also discussed that could overcome the limitations of these two approaches, though they are not implemented here. Downscaling energy systems still holds some promise, though significant research is required to integrate technological innovation and diffusion considerations to the methods.