Identification of Borrower Creditworthiness Using Neural Network Model

In our work on smallholder farmers in Africa and Asia, we have often been confronted with a few common challenges. First was how to identify the right type of farmers having an entrepreneurial mindset and secondly who among these have creditworthiness. Both these issues have significance because a lot of funding goes to those who do not possess any entrepreneurial capabilities and traits. Secondly, many who borrow funds do not repay the loans. Thus, any pilot project with them does not become scalable. As it is Africa is full of such short-lived pilots that never have gone to scale. The smallholder farmers are perennially short of funds. As stated earlier those who get funds from the financial institutions do not repay the borrowed money. The percentage of loan defaults in Africa is staggeringly high to the tune of 55%. Again, part of the reason is that banks and microfinance institutions have a very narrow and straight-jacketed method of determining eligibility for loans. These are mostly centred around collaterals and some perfunctory information about the borrower. The loan officers have no clue about the psychological determinants affecting both entrepreneurship and credit worthiness. Some serious research has been undertaken to develop statistical models to arrive at a reliable credit scoring model. Conventional statistical models used logistic regression analysis. Recent advances in statistical research are using neural network models for decision accuracy. A neural network is a system made of highly interconnected and interacting processing units that are based on neurobiological models mimicking the way our nervous system works.

Over the years based on the many projects we carried out in Sub-Saharan Africa (SSA), we have developed highly accurate and sophisticated psychometric profiling tools under the patented name mPower-U that helps measure entrepreneurship and creditworthiness. In our view measuring only entrepreneurship or only creditworthiness does not yield reliable result. Someone who has a lot of entrepreneurial abilities may not be creditworthy and someone highly creditworthy may lack leadership skills. Thus, we need to measure both these factors. The tool/battery mPower-U has been successfully deployed in Senegal, Tanzania, Kenya, Bangladesh and India over a varied sample consisting of farmers, extension officers, dealers of input companies as well as marketing staff of the input companies.

In doing the analysis and profiling we have adopted the constellation of traits approach as well as the neural network logic. This theory assumes that a major trait is like a major planet or a star around which several minor stars form a galaxy. Each major trait may have a few positive and a few negative attributes. Through psychometric instruments, we figure out the constellation of these traits and sub-traits. Using the neural network example, a major trait or group of traits are like the brain that has linkages to several sub-traits that exchange and process information exactly the way the interactions between the brain and the nervous system take place.

In this short note let me articulate our findings on creditworthiness. Our research has indicated that the creditworthiness of a borrower is reflected through his/her food habits, physical habits, social habits, financial management habits and achievement motivation. The manifestations of these habits determine the person’s creditworthiness. Unless we understand the interplay of this constellation simply looking at the collateral of the borrower and determining his/her creditworthiness is not the right way to go about it.

Given below is the constellation of creditworthiness:

Food habits of creditworthy borrowers:

• Purchasing only those food items that are healthy, necessary
• No wasting of food
• Planning what to buy is done beforehand
• Eating routine well established
• Regular medical check-ups
• No craving for junk food

Food habits of less creditworthy borrowers:

• Impulsive purchase of food items
• Ordering more than what is needed
• Wastage
• Unplanned kitchen
• No routine pattern in eating
• Craving for junk food
• Excessive and unnecessary usage of debit/credit cards for purchasing junk food items

Physical habits of creditworthy borrowers:

• Regular exercise
• Able to control physical desires
• Can sacrifice physical desires for long-term goals
• Believes in dating than seeking physical desire right away
• Tolerance for physical discomfort

Physical habits of less creditworthy borrowers:

•   Unnecessary risk-taking
•   Impulsiveness in thought and action
•   Obsession with selfies
•   Looking for sensation
•   Urge to impress everyone
•   Pretentious

Social habits of creditworthy borrowers:

• Quite but not very sociable
• Seamless networking
• Quiet assertiveness
• Good listening
• Ability to introspect

Social habits of less creditworthy borrowers:

• Withdrawn and loner
• Difficulty in forming relationships
• No listening
• Non-conformism
• Argumentative

The money management of creditworthy borrowers:

• Proper financial planning with adequate diligence
• No impulsive purchase/investment
• Live within means
• No unrealistic risk appetite
• Take loans only when it is a must

The money management of less creditworthy borrowers:

• No financial planning
• Impulsive spending
• Buying things that are really unnecessary
• Getting swayed by temporary advantages/benefits
• Living on credit cards

Achievement motivation of creditworthy borrowers:

• A high degree of conscientiousness
• Internal locus of control
• Passion for work
• Sound planning and forecasting
• Maintaining equilibrium

Achievement motivation of less creditworthy borrowers:

• Poor work ethic
• Tendency to adopt shortcuts
• No due diligence
• Take it easy approach
• Prefer means to end