We apply a quadratic expenditure system to estimate price and expenditure elasticities of residential energy demand (electricity and healing) in Germany. Using official expenditure data from 1993 to 2008, we estimate an expenditure elasticity for electricity of 0.3988 and of 0.4055 for space heating. The own price elasticity for electricity is 0.4310 and 0.5008 in the case of space heating. Disaggregation of households by expenditure and socio-economic composition reveals that the behavioural response to energy price changes is weaker (stronger) for low-income (top -income) households. There are considerable economies of scale in residential energy use but scale effects are not well approximated by the new OECD equivalence scale. Real increases in energy prices show a regressive pattern of incidence, implying that the welfare consequences of direct energy taxation are larger for low income households. The application of zero-elasticities in assessments of welfare consequences of energy taxation strongly underestimates potential welfare effects. The increase in inequality is 22% smaller when compared to the application of disaggregated price and income elasticities as estimated in this paper.