In this study, we develop a series of technology diffusion formulations that endogenously
represent empirically observed spatial diffusion patterns. We implement these formulations in
the energy system optimization model MESSAGE to assess their implications for the market
penetration of low-carbon electricity generation technologies. In our formulations, capacity
growth is constrained by a technology's knowledge stock, which is an accumulating and
depreciating account of prior capacity additions. Diffusion from an innovative core to less
technologically adept regions occurs through knowledge spillover effects (international spillover
effect). Within a cluster of closely related technologies, knowledge gained through deployment of
one technology spills over to other technologies in the cluster (technology spillover effect).
Parameters are estimated using historical data on the expansion of extant electricity technologies.
Based on our results, if diffusion in developing regions relies heavily on earlier deployment in
advanced regions, projections for certain technologies (e.g., bioenergy with carbon capture and
storage) should be tempered. Our model illustrates that it can be globally optimal when
innovative economies deploy some low-carbon technologies more than is locally optimal as it
helps to accelerate diffusion (and learning effects) elsewhere. More generally, we demonstrate
that by implementing a more empirically consistent diffusion formulation in an energy system
optimization model, the traditionally crude—or nonexistent—representation of technology
diffusion in energy-climate policy models can be significantly improved. This methodological
improvement has important implications for the market adoption of low-carbon technologies.