Because they lower the cost of using energy in production and consumption activities, energy efficiency improvements may lead to a rebound effect, in which the demand for energy increases to offset partly or completely the initial energy saving. The magnitude of the rebound effect depends on: (1) the extent to which the effective costs of energy services (capital and operating) actually decrease from efficiency improvements, (2) the technical and economic ease with which energy and other inputs to production and consumption (capital, labour, materials) can be substituted when the effective cost of using energy does in fact decrease, (3) the response of intermediate and final demands to changes in the cost of energy services (structural change), and (4) the response of energy service demand to changes in income. Estimates of (1) and (2) together, which is the elasticity of substitution (ESUB) between capital and energy, have been criticized in the past for either not being explicit about what is technically possible in future (critique of top-down) or ignoring the intangible costs facing firms and households when considering and implementing energy efficiency investments (critique of bottom-up). In this study, we attempt to address these two concerns by using a technologically explicit and behaviourally realistic, technology simulation model to estimate long-run, future ESUB values for capital and energy for the Canadian economy. Our simulations suggest that the capital-energy ESUB may be lower than assumed by most studies that ignore behaviour, suggesting only weak substitution between capital and energy. Ironically, this could mean that our energy efficiency efforts will be less effective than sometimes thought but that the rebound effect will also be relatively small. However, estimates of (3) and (4) must be combined with our findings to assess the full magnitude of the rebound effect.