Towards analysing social norms in microfinance groups Pablo Lucas dos Anjos1, Federico Morales2, Ignacio Garcia3 firstname.lastname@example.org, email@example.com, firstname.lastname@example.org 1 Centre for Policy Modelling, Manchester Metropolitan University 2 PROIMMSE-IIA-National Autonomous University of Mexico 3 Facultad de Filosofía y Letras, Universidad de Buenos Aires Abstract: This research focuses on the social organisation and commitment dynamics among borrower groups at a microfinance institution (MFI) in Mexico. Due to our publishing agreement, their identity is omitted. The MFI manages micro-credit loans given to geographically distributed groups in the southern state of Chiapas, each with 3 to 7 women only. This non-governmental organisation has adapted in 1998 the Grameen Foundation methodology and use guidelines from the Consultive Group to Assist the Poor to implement their own solidarity lending, life insurance in cooperation with Zurich Financial Services, educational and nutritional programs that prioritise the local rural community . Technical advisers are trained to facilitate, using Spanish or one of the 8 regional Mayan languages, the process of managing quota repayments that are periodically expected from individuals in every group. As financial techniques are employed to support a social mission, lending does not rely on traditional assets required by private and public banks in order to consider a credit application. Instead, social collateral is assessed according to socio-economic situation of every applicant and a reference poverty line. Although there is vast collection of published academic and third sector literature on MFI good practices, little is known about social norms that influence social collateral among borrowers at MFIs. Given this clear gap of dedicated studies analysing the internal structure and supporting mechanisms of such micro-credit groups, our 2008 research project is analysing data from 5 MFI financial databases, collected interview data from technical advisers and clients in order to both better understand their social context and guide the development of an agent-based computer simulation. In addition to contribute with a clear sociological and socio-economical analysis, the computational modelling approach is being informed by available data to describe and simulate (1) the evolution –not necessarily optimisation– of commitment to quota repayments, (2) which aspects in a group can contribute, or deteriorate, the individual reliance on the social collateral, and (3) assess if this evidence- driven simulation has potential to relevant stakeholders. The proposed simulation is focused on exploring the dynamics of social collateral among groups of borrowers participating in microfinance. In this sense, it is essential to guide the individual agent development with reliable and statistically significant data extracted from questionnaires about the behaviour of the MFI clientele and numerical evidence from their financial databases. Apart from the initial socio-economical assessment and technical advisors throughout their loan period, there is very scarce understanding of how social networks and trust mechanisms are structured within those microfinance groups. Their financial data have detailed information tracking every individual payment according to the interest rate associated with MFI approved loans, but no register is made on social behaviours thay influence individuals to pay or cover quotas. Existing software such as Microfin  and Symbanc  are only suitable to manage or analyse MFI financial processes, but offer no feature to analyse data regarding the internal mechanisms that can influence the social collateral of groups and cooperative behaviour of its members. Agent behaviour and structure of social networks among microfinance groups are being implemented according to the available MFI evidence. That is, two online questionnaires administered to 35 credit officers, one form to 600 borrowers and a semi- structured interviews carried during a fieldwork visit in May 2008. The model is being tested using these retrospective datasets in comparison to outcomes from what-if simulated scenarios. References  AlSol Chiapas AC, Background, Client profiles and monthly Operation Report: Grameen Foundation, USA, June 2007.  Anthony Sheldon, Chuck Waterfield, Business planning and financial modeling for MFI: a Microfin handbook, CGAP, 1998.  G. Hirsch,J. Rosengard, G. Stuart, D. Johnston, A Simulator for Microfinance Institutions. ADB Finance for the Poor 6.4, Dec 2005.