There is currently a high level of interest among policy review agencies in using water usage prices to balance the supply and demand for water (scarcity pricing). The potential effectiveness of scarcity pricing depends upon the price elasticity of demand. This study applies panel data analysis to household level data to obtain short- and long-run estimates of the price elasticity of demand in Sydney. The period of the analysis spans June 2004 to June 2009, during which water usage prices increased in real terms by over 45 per cent. The data are disaggregated into residential user groups of owner-occupied and tenanted houses that did and did not participate in a subsidised water appliance efficiency programme, and housing units. Owner-occupied and tenanted houses are further disaggregated through clustering analysis based on property size. The estimated real price elasticities vary both across and within user groups. The weighted average real price elasticities are small, particularly in the short run. As well as overall price elasticities, the grouping and clustering of households provides insights into the way different households respond to changes in water usage prices, weather conditions and participation in water appliance efficiency programmes. The heterogeneity in responses is found to generate statistically different results compared with an aggregate model of all residential households.