The purpose of this article is to analyse the efficiency of the repayment incentive mechanisms used for individual microcredit, the guarantor mechanism in particular. Little academic research has been devoted to this instrument whereas it is very frequently used. Churchill (1999) notably underscored the fact that, within the framework of microfinance institutions (MFI), the guarantor acts as a vector of social pressure on the borrower rather than as an alternative source of reimbursement. Little to no econometric studies have been devoted to testing this assumption. This is what we propose to do in this article, based on original data from a Brazilian microcredit programme, VivaCred, operating in the Rio de Janeiro favelas. Using a simple theoretical model, we will first of all describe the relations between an individual microcredit institution and a borrower in order to highlight the role of social sanction. Next, with an ordered multinomial logit model, we will study the probability of a change in behaviour by the borrower in terms of adherence to the repayment schedule. To be more precise, we will analyse the behaviour of borrowers with repayments significantly overdue in order to determine what encourages them to adhere to repayment deadlines in the future (economic conditions, policy of the MFI, etc.). In particular, results show that the number of guarantors has a positive impact on the likelihood of adherence to the loan repayment schedule in the future. However, it is a one-shot mechanism: the MFI cannot increase the number of guarantors required in the event of repeated delays in repayment over a period of time. On the contrary, the guarantor mechanism does not affect the likelihood of reverse transition. This mechanism thus proves efficient ex post, that is to say to encourage borrowers to adhere to the schedule in the event of repayment difficulties but has no preventive properties.
