Credit Risk Modeling of Housing Finance Loans

Abstract

This report consists of project undertaken during my summer internship at National Institute of Bank Management, Pune from 9th April to 8th June 2020. I was assigned to work upon ‘Credit Risk Modeling of Housing Finance Loans in India.’ This project report contains three divisions- A, B and C respectively. Part A comprises of explanation of Training and Education Industry, details about NIBM including its vision, mission, role, services catered, governing board members, etc. Part B consists of thesis that examined the factors considering for sanctioning housing loans and predicting the probability of loan default. It covers methodology to conduct research work through logistic regression. The process contains data handling, calculation of derived variables, binning and conversion of continuous to categorical variables, descriptive statistics, hypothesis formation, variable selection, running regression model through machine learning and final model testing. It has mentioned the model output that rejects null hypothesis and conclusion part has shown the positive/negative and magnitude of relationship between independent and dependent variable. Managerial implication of credit risk modeling and drawbacks & recommendation is also presented in the report. Part C talks about learning from my summer internship project. This opportunity has given me an extensive exposure to learn about credit risk and its modeling in housing finance industry. Application of already known techniques and methods like descriptive statistics, hypothesis, regression analysis, p value, etc. were used during the research and lot of new concepts and processes like machine learning, logistic regression, correlation matrix, WOE and IV test, KS Statistics, confusion matrix and ROC curve I encountered during these two months. It has also mentioned the insights about the managerial role that I’ve learnt during working in this domain.