Multi-Agent Simulations to Explore Rules for Rural Credit Management in a Highland Farming Community of Northern Thailand

Introduction Thanks to recent advances in the field of distributed artificial intelligence, agent-based models (ABM) can now be used to run simulations of social phenomena based on their computerized representations, and to apply experimental methods in social sciences (Axelrod 1997, Gilbert and Troitzsch 1999, Jager 2000). In the field of renewable resource management and environmental sciences, several ABM simulation platforms offer the possibility to explore interactions between social and ecological dynamics (Costanza and Ruth 1998, Bousquet et al. 1998, Lansing 2002). In these complex systems, the social and economic dynamics can be viewed as a set of interactions among heterogeneous agents, generating aggregate phenomena that are different from the behaviour of groups of average individuals considered in classical economic thinking (Rouchier and Bousquet 1998). Such a view was adopted in the research presented here. The agent-based model presented in this paper was built to explore the interrelated roles of formal and informal credit in a socially heterogeneous community of small farmers exploiting a highland catchment of mountainous upper northern Thailand. Formal credit corresponds to institutionalized credit funds whereas informal credit is seen as loans settled among villagers, either without interests within networks of acquaintances, or with high interest rates when loan sharks are involved. An original characteristic of the companion modelling approach (Bousquet and Trébuil 2005, http://commod.org) and the simulation process adopted in this case study is the co-construction of the model with the farmers and the use of simulations with them in their village. The objective was to facilitate collective decision-making regarding the adaptation of the local rules for the allocation of rural credit to allow a more equitable and extensive process of expansion of perennial crops (Barnaud et al. 2005). Following a description of the methodology and tools used in this experiment, the results of a series of multi-agent system simulations are presented and analyzed. To end with, the specific questions and challenges raised by this type of social modelling and simulation process are discussed, particularly its use and usefulness to local stakeholders.