This paper describes a physically based model of the energy use of the housing stock (called BREHOMES) and explains how it has been used to develop two scenarios for energy use and carbon dioxide emissions. The general structure of the model is outlined together with the data on which it relies, These data have been systematically collected for several years, and this allows the important trends to be identified, These historical trends form the basis of the scenarios that have been developed using the model. One of the scenarios (called the Reference scenario in this paper) represents what is likely to happen if current trends continue. The other (called the Efficiency scenario) represents what could happen if the uptake rates of energy efficiency measures were tell increase to levels that are feasible in so far as they have been seen to occur in the past, Neither scenario attempts to consider the effects of new technologies being introduced, Everything is based on current well-tried technology only, Thus, the savings of the Efficiency scenario relative to the Reference scenario are conservative estimates of what should be possible compared to current trends, The results indicate that savings of about 20 PJ year(-1) (around 21 million tonnes CO2 year(-1)) are possible by 2020. A cost-benefit analysis suggests that these savings are cost-effective for all discount rates below 11%.