Credit Scoring: How Banks and Microfinance Institutions Decide Whether to Grant You a Loan

19 April 645

You are planning to take out a large loan and assumed the bank would approve it since you’ve never had missed payments. However, banks evaluate potential clients not only based on payment history and income but also using many other parameters. Sometimes they even analyze social media photos or the model of your phone. This evaluation system is called credit scoring. Let’s explore how it works.

What Is Credit Scoring?

Credit scoring is a system used to assess a borrower’s likelihood of repaying a loan on time. It relies on mathematical calculations and statistical models.

It is assumed that people with similar habits handle finances in similar ways. Therefore, banks and MFIs compare potential borrowers with others who have taken similar loans. Scoring models use algorithms to evaluate a person’s creditworthiness.

Most lenders use several scoring systems simultaneously for different types of borrowers or loans. Each borrower characteristic is given a score. For example, employment length: someone on a probation period receives fewer points than a long-term employee.

After evaluating all factors, the algorithm calculates the total credit score. The most reliable clients get the highest scores and the best chance of receiving loans on favorable terms.

Where Do Lenders Get Data for Scoring Models?

The main sources of information are the borrower’s credit history, application form, and the lender’s internal financial data (e.g., payment history, account top-ups). Additional sources may also be used.

Credit History

This is the most important resource. The more data the bank or MFI has, the more accurate the prediction, and the more likely a responsible borrower will get a favorable loan.

Lenders want to know:

How many loans or credits the borrower currently has.
Past delinquencies — how often and how long payments were late. A few days’ delay differs from several months in scoring.
Whether the borrower has taken microloans — frequent MFI use may suggest hidden financial risks, reducing the score.
The amount of credit repaid on time — banks define relevant periods differently, some check one year, others the full history.
How often the borrower was rejected by other lenders.

Credit bureaus often provide scoring data, which lenders may use directly or combine with their models. Most banks and MFIs prefer to calculate scores themselves.

Borrower Application

When requesting a loan, borrowers fill out an application. Information from the form is also used to calculate the credit score.

Lenders usually collect similar data but interpret it differently. For example, gender is recorded, but banks may have different statistics on whether men or women are more reliable. This data feeds into the scoring model.

Typical information requested includes:

Income — higher earnings increase the score.
Address — people in regions with higher living standards are considered more reliable.
Age — students and retirees are usually higher risk; for others, older borrowers are considered more responsible.
Marital status — married people are generally more disciplined in repayment.
Profession and work experience — some jobs or industries are considered more stable and reliable.

Loan parameters also influence the score. For example, short-term loans of 3–7 days are repaid more reliably than 30-day loans. Selecting the maximum loan period without reason may reduce the score. It’s better to realistically indicate what you can manage.

Lender’s Internal Financial Data

If the borrower uses their payroll bank, the lender knows their income. Account balances, debit/credit cards, or deposits are also factored into the scoring model to assess financial discipline.

Additional Data

Lenders may use any other available data. For example, the borrower’s phone model can influence the score. Banks can know this if the borrower uses their app or mobile site. Payment behavior observed through card transactions is also included.