Non-residential buildings comprise a quarter of the European Union (EU) building stock and cause approximately one-third of this stock's energy consumption. The imperative to reduce their energy consumption is widely accepted in EU countries. It is now common to upgrade both residential and nonresidential buildings to increase energy efficiency, but studies indicate a significant gap between energy performance and design. In the residential sector, this gap has been extensively investigated and quantified under the heading of the 'rebound effect', but there do not yet appear to be studies on rebound effects in non-residential buildings. This paper develops a methodology to fill this gap. Through technical measurements in four recently retrofitted German public service case study buildings, rebound effects were estimated under two definitions: the 'elasticity' rebound effect and the 'energy performance gap' (EPG). Cross-sectional data of actual and calculated consumption in 176 comparable buildings was used to estimate typical rebound effects which could be expected in such buildings. The case study results indicated comparatively low rebound effects but high EPGs. Semi-structured interviews with building occupants and operators identified credible determinants of energy consumption. The methodology proved effective and robust, and is offered for further use and development. (C) 2015 Elsevier B.V. All rights reserved.