Given recent developments on energy markets and skyrocketing oil prices, we argue for an urgent need to study the potential effects of world oil production reaching a maximum (Peak Oil) in order to facilitate the development of adaptation policies. We consider input-output (IO) modelling as a powerful tool for this purpose. However, the standard Leontief type model implicitly assumes that all necessary inputs to satisfy a given demand can and will be supplied. This is problematic if the availability of certain key inputs becomes restricted and it is therefore only of limited usefulness for the study of the phenomenon of Peak Oil. Hence this paper firstly reviews two alternative modelling tools within the IO framework: supply-driven and mixed models. The former has been severely criticised for its problematic assumption of perfect factor substitution and perfect elasticity of demand as revealed by Oosterhaven [Oosterhaven J. On the plausibility of the supply-driven IO model. J Reg Sci 1988: 28:203-17. [1]]. The supply-constrained model on the other hand proved well suited to analyse the quantity dimension of Peak Oil and is therefore applied empirically in the second part of the paper, using data for the UK, Japanese and Chilean economy. Results show how differences in net-oil exporting and net-oil importing countries are clearly visible in terms of final demand. Industries, most affected in all countries, include transportation, electricity production and financial and trade services. (C) 2008 Elsevier Ltd. All rights reserved.