During recent decades, water demand estimation has gained considerable attention from scholars. From an econometric perspective, the most used functional forms include log‐log and linear specifications. Despite the advances in this field and the relevance for policymaking, little attention has been paid to the functional forms used in these estimations, and most authors have not provided justifications for their selection of functional forms. A discrete continuous choice model of the residential water demand is estimated using six functional forms (log‐log, full‐log, log‐quadratic, semilog, linear, and Stone‐Geary), and the expected consumption and price elasticity are evaluated. From a policy perspective, our results highlight the relevance of functional form selection for both the expected consumption and price elasticity.