We describe the theory, computation and results of a multiperiod general equilibrium model designed to assist an urban water authority in Its pricing and investment decisions. The model includes gestation periods in the creation of dams, main sewers and treatment plants. It allows for lumpy capital items and recognizes cost differences in the provision of services in peak and non-peak times. Its general equilibrium framework is convenient for handling links between the water authority and the rest of the economy, especially the housing sector. We have used two computational approaches. In the first, we reformulate the model as a single-entity optimization problem and then apply a linear programming package. We have found that a better approach is to apply Newton-Raphson methods to a formulation of the model as a set of equations depicting purely competitive behaviour in all productive activities. A special feature of this paper is an integration of the model's results, obtained under the assumption of certainty, with data on weather-induced variations in streamflow and demand. Using Monte Carlo techniques we assess the risks of water shortages associated with the investment and pricing strategies that our model indicates. Key Words and Phrases : water pricing and investment: uncertain streamflow: water policy in a general equilibrium model: water policy for Melbourne: linear programming: Newton-Raphson: Monte Carlo: peak and non-peak.