A key determinant of high performing water resource systems is an effective institutional arrangement. An essentfal research question is what is an effective institutional arrangement? We argue that there is no single best institutional model, as institutional requirements vary depending on the phase of development of the basin. Based on water accounting studies we observe that water use patterns can be used to identify stages of river basin development. The hypothesis presented in this paper is that depending on the phase of development, institutions wlff be concerned with different tasks. Thus an important feature of a well-functioning set of water management institutions is the ability to adapt to changes. We use a water accounting methodology to illustrate the concept of phases of development in river basins. We argue that to meet increases in demand over time, institutions must change their focus from development of infrastructure, to better utilising and conserving water resources, then to improving allocation and regulation of water resources. Institutions must be dynamic entities that change with changing phases of development of the basin. At their inception, serve a single purpose. Then as the basin develops, they either expand their number of functions, or other institutions evolve to fulfil management requirements. These concepts are illustrated in two cases derived from Nepal and China.