This project develops a credit scoring system that is based on customer transactions rather than a purely characteristics-based screening approach. In developing countries where credit bureau information is often not available or unreliable, transaction-based lending models allow ‘good’ applicants to demonstrate their quality. The project will identify different customer transaction patterns and account usage behaviors and test their predictive power for repayment behavior of SMEs. The project will randomize access to the different credit scoring models (transaction versus characteristics based) and compare default rates on loans between the groups. By exploring different methods of screening, the project can help financial institutions increase lending to SMEs and thus improve access to finance.