Abstract
Allocating limited water resources among competing uses is increasingly challenging as water demand rises and climate variability intensifies. Because water is commonly managed through administrative rules rather than market mechanisms, allocation outcomes reflect implicit policy priorities whose effects on performance and resilience are often difficult to assess. This paper develops a structural model of dynamic water allocation in a multipurpose reservoir system to recover the sectoral priorities embedded in observed management decisions. By employing stochastic dual dynamic programming and the simulated method of moments, we estimate preference weights that account for historical patterns in water storage and release. Applied to the Panama Canal Watershed, a rain-fed man-made reservoir that supports global trade, energy production, and municipal water supply, the model reveals a high implicit priority on drinking-water reliability and navigation, with hydropower acting as a flexible adjustment margin. Counterfactual drought simulations show that adaptive reallocation can preserve storage and reduce navigation losses during extreme climate stress.