THE DEVELOPMENT OF A DEFAULT PREDICTION MODEL FOR MICRO FINANCE CREDIT ANALYSIS (Implementation at a microfinance organization)

• Main Author: PERMANA (NIM 23407039); Pembimbing: Prof.Dr. Ir. Bermawi P. Iskandar, EKA
• Format: Theses
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/15500
• Institution: Institut Teknologi Bandung

Up to now, the credit analysis process at microfinance organization has been using a financial ratio analysis and scoring method. This method is a univariate method and gives partial description about measures contributing to the organization performance. Meanwhile the scoring method used to manage the credit data has not standardized weigthing system. The scoring system depends on the management jugment in each financial organization.<p>This research introduces an improved approach for the credit analysis process by using a statistical method to develop default prediction model. Model developed incorporated 15 financial and non financial variables that represent 5C (Capital, Capacity, Character, Colleteral, Condition and Constraint) approach. From the literature study, asumption and data type available lead to the logistic regression as an appropriate statistical method to develop a default prediction model.<p>The model developed was implemented at a microfinance organization using historical credit data for 3 years (2004-2007). Result of the model provides a classification debtor candidate as a default or a non default debtor group. Model gives a comprehensive credit analyst with a total hit ratio 78%.<p>For ease and effective implementation, the result model is integrated to a web base DSS (Desicion Support System) application. This application used as a risk assesment tool for credit risk in a risk management system and framework. The use of application makes the model more easy to implement and can be further developed for other purposes such as alarm system, a credit note, crosstab analysis and report. <br />