Context
Stikcredit is a lending institution and focuses on issuing short-term unsecured consumer loans. For any lending institution, it is crucial to accurately estimate the future performance of its loans. One of the most important performance measures is related to the expected cash flow from a given set of loans. In other words, if we provide a 100 EUR loan to a borrower, what amount do we expect to receive back in 1, 2, 3 or more months after the loan has been granted.
This may seem like a simple task, but the repayment behavior of the borrowers is erratic. Some borrowers repay their loans early, others follow the repayment schedule of the loan or maybe don’t repay the loan at all.
Accurately modeling the future cash flows arising from a loan is crucial for financial planning and budgeting. Creating a successful model will help the finance team in any lending organization to accurately estimate the amount of money that the lender will receive in any given month.
This requires forecasting the month-by-month cash flows over the entire loan portfolio, looking a month, a quarter or a year in advance.
Idea
The goal of this project is to answer one question: What do we expect the repayment cash flow to look like for an individual loan in the future?
To achieve this you will have to develop a model based on Markov property which calculates the transition probability for every loan between different periods. Select appropriate machine learning algorithm to develop, train and validate the model. Use different loan contextual features (e.g. amount, term, age of the loan, credit risk, etc.) to predict the probability of the loan to be in any of the given states in time t+1.
Use the forecasted transition matrixes to calculate the future cash flow expected from any given loan next month or any other future period.
Benefit
If successfully implemented, this project will have a great impact on the financial position of any lending institution. The benefits are:
- Enhanced risk management and underwriting decisions.
- Improved forecasting capability, financial planning, and budgeting.
- Reduced liquidity and maturity mismatch risk – the lenders will be able to accurately predict future cash flows and adjust their funding and lending strategies accordingly.
- The lenders will be able to increase their profitability while being able to better understand and manage their risks.
Data and support
The company will provide the participants with a proprietary database to train, validate and test their models. The company will provide further information to the problem and industry knowledge and will suggest possible ways to approach the development of the service.
About the company
Stikcredit is a European FinTech company underwriting short-term consumer loans online. We’ve also built a national network of 50 offices. We operate on the Bulgarian market since 2013 and aim to become the go-to-platform for lending loans online. We lend money where traditional banks don’t and we are on a mission to make consumer credit as easily accessible as 1 click on any device. We underwrite short-term single-payment and installment loans for up to EUR 2,500 and with maturity of up to 24 months.