Transboundary water resources under uncertainty: Informal agreements, bargaining power and reservoir capacity

Many freshwater resources are located across national boundaries where governments must rely on institutional schemes to manage their shared resources. These schemes are formalized by institutional agreements that usually operate in uncertain hydrologic conditions. Should the rate of change of the conditions governing the natural resource be faster than the rate at which formal institutions adapt, the chance of conflicts among governments may increase. Flexible formal institutions are desirable, but the transaction costs associated with renegotiation expensive. Thus users rely on other institutional schemes, namely informal agreements, which are policy instruments used by governments to overcome crises, uncertain events, or variable conditions that are not always considered in formal agreements. I developed a cooperative model to measure the potential gains of establishing informal agreements and show that the gains are greater than or equal to the gains of formal agreements. I found that after considering all bargaining power parameters, there are maximum gains of establishing informal agreements. I analyzed 39 bilateral basins over time and found that on average these maximum gains are 6.5%. I also analyze the relationship between gains and variability of water flow. Finally, I studied the implications of water storage by trying to understand the tension between how much water to store and to release given that water flow is subject to variation and that the current distribution of flow into reservoirs is variable and uncertain. I develop an analytical method to analyze the effects in optimal reservoir capacity for various climate change conditions that alter the current mean and variance of water flow. I estimate an optimal reservoir capacity by stochastic dynamic programming and parameterize the model with real time data from five major reservoirs in Northern California. I illustrate that for a decrease in annual mean inflow with no corresponding change in standard deviation my results suggests a need to increase reservoir capacity; however if the decrease in the mean is met with a corresponding decrease in the standard deviation no change in reservoir size is required.