Programme Poster session 3 abstract 645
Sharing water at catchment level: the impact of context on cooperative
game solutions
Author(s): Mathieu Désolé, Stefano Farolfi, Fioravante Patrone, Patrick Rio, Mabel Tidball, Sophie
Thoyer,
Cirad
University of Genoa
Inra-Lameta
Supagro-Montpellier (Lameta)
Keyword(s): negotiation-support tool, water allocation, cooperative games,
experimental economics, South-Africa
Session: Poster session 3
Abstract Thirteen years after the
beginning of the democratization process in South Africa, many radical socio-political and institutional
transformations have taken place in the country. Unlike during the Apartheid era, natural resource management and
governance, particularly in the water sector, is nowadays based on concepts and criteria such as decentralization,
economic efficiency, environmental sustainability and social equity. These criteria, which represent the pillars of the
South African National Water Act (NWA-1998), are universally recognized as the fundaments of sustainable
development and are widely employed in the definition of the environmental policies of the industrialized
countries.
To accompany this socio- political revolution, a process of institutional building is taking place in the
South African water sector. New organizations, namely the Catchment Management Agencies (CMAs) and the
Water Users Associations (WUAs) in charge of local governance of water are being established in the country.
These new organizations urgently need tools, methods, processes that can help them in their difficult task of
implementing locally the NWA by promoting the participation of local stakeholders in the process of water
management and allocation. In particular, there is a need to gain a better understanding of collective decision
mechanisms. In the South African context, participatory approaches – involving local stakeholders, decision-makers
and researchers- have been implemented to accompany the water allocation process at catchment scale : the Kat
River catchment was used as a pilot study to develop a role-playing game (KatAware –based on multi-agent
simulations) accompanying the negotiation between water users on the allocation rules of irrigation water. However,
the lessons of this experience are not easily transferable to other sites or situations. The objectives of our work are
therefore to assess the impact of the context on the outcome of a collective decision: by identifying the contextual
elements influencing most on individual behaviour and on collective action, we could then build negotiation-support
tools which are more easily transferable from one context to another.
This paper presents the results of theis
joint research program called SAFeWATER, undertaken the University of Pretoria , the Centre for Environmental
Economics and Policy in Africa and two French research institutions, CIRAS and INRA. We conduct laboratory
experiments to identify the way subjects play a cooperative game – sharing a common resource leading to higher
joint revenue when all users agree than when users form partial coalitions or try to play single-handedly. The
allocations obtained in the laboratory are compared to well-known cooperative game solutions such as the Shapely
values or the nucleolus solutions. This allows us to build a reference game. The same experiments are then conducted
with a protocol which is progressively enriched with contextual details: ie the resource is named, subjects are
attributed roles, etc. and compared to the reference game results. A final phase consists in conducting the same
experiments in the field with concerned subjects (farmers instead of students). Our results confirm previous field
experiments that context matters. If time allows, test-bed experiments are then used to compare the acceptability of
various policy instruments yielding the same allocation as the copperative solution foud in the previous experiments.
This work should help to measure more precisely to what extent context changes behaviour and allocation
solutions in cooperative game situations. It should then allow us to build accompanying models facilitating
negotiations between water users at catchment levels, which can be adjusted to each catchment, therefore
overcoming the disposable model issue.