The paper describes how "allocation stress" of water resources is often identified as resulting from four different observations:

• agriculture receives the bulk of all diverted water resources
• agriculture is an activity that incurs by far the largest wastage
• cities are "thirsty"
• water productivity or efficiency in non-agricultural sectors is far higher than in agriculture

The authors draw a number of surprising conclusions from their study of a large number of cities worldwide:

• increased urban supply is often realised through transfers of water away from agriculture, or from ecological reserves and aquifers. Tapping of any new resources frequently happens in a non-sustainable or costly manner
• most cities in developing country cities have poor water supply and sanitation infrastructure. This can be a more limiting supply factor than a lack of access to water resources
• the main issue is where to find the money rather than where to find the water
• urban water supply decisions are a reflection of local politics: popular pressure, gains to elites and decision makers, and privileged relations largely determine the level of water services and how these are spatially and socially differentiated
• there is no strong evidence that indicates that non-agricultural activities are significantly constrained by a lack of water, both in their daily functioning and in their capacity to expand
• sectoral water allocation may be hindered where formal water rights systems exist, but much less so where allocation is centrally administered
• contrary to common knowledge, states have consistently given priority to cities and nonagricultural activities where water has been reallocated, and the large economic gains anticipated for intersectoral transfers are greatly overstated
• gains from better water management are fairly limited over the medium-term
• planners should acknowledge that water systems will have to cope with growing uncertainty and fluctuations in supply, contingency planning should allow for short-term transfers
• better access to hydrological data, the use of multi-stakeholder platforms, and patterns of governance which include empowerment of marginalised social/ethnic groups have the potential for ensuring fairer reallocation of water.