How you shop and what you buy at the supermarket can predict whether you pay your credit card bills on time based on your current research.

As marketing professors, we wanted to learn more about alternatives to traditional credit scores, so we partnered with a multinational conglomerate that operates, among other things, a major supermarket chain and a credit card issuer.

By analyzing consumer data from these two business units, we were able to see how 30,089 people shop and manage their finances.

We’ve found that people with more regular shopping habits are more likely to pay their credit card bills on time. These are people who typically shop on the same day of the week, spend roughly the same amount each month, buy similar items in multiple purchases, and regularly take advantage of special offers.

We also found that people’s purchasing habits provide insight into how they manage their finances. For example, shoppers who frequently buy cigarettes or energy drinks are more likely to miss credit card payments. Shoppers who frequently buy fresh milk or salad dressing are more likely to pay their bills more conscientiously.

In general, purchasing healthier but less convenient foods was an indicator of responsible payment behavior, even when we held constant consumer characteristics such as income, occupation, credit score, and family size.

Building on these insights, we developed a credit scoring algorithm that evaluates individuals based on their shopping habits and traditional credit risk indicators. When we simulated approval decisions using this algorithm, we found that using grocery data can help lenders more accurately predict defaults while increasing their profit per customer.

Why it is important

According to the World Bank, more than a billion people worldwide do not have access to formal financial systems and therefore do not have a credit score. In the United States alone, approximately 45 million adults have no credit history or insufficient credit history to obtain credit.

This makes it harder for them to access credit, even if they are responsible borrowers. And without credit, it is harder to get a car, a job, or even an apartment. It is a problem that disproportionately affects disadvantaged groups, including people of color and women.

In response, policymakers and researchers are increasingly interested in using alternative data sources to assess creditworthiness. Fannie Mae, for example, now considers mortgage applicants’ rental payment history, allowing people without a traditional credit history to demonstrate their creditworthiness.

Grocery data is particularly promising because there is so much of it. Pretty much everyone buys groceries, and not just once. Information about consumer preferences is constantly being generated in every department of grocery stores around the world.

Our study shows that this data is valuable far beyond the food industry.

What happens next?

We believe our study serves as a proof of concept and provides insights for designing and conducting future research. However, some important questions remain. For example, what if this approach affects different groups unequally? And what about privacy concerns?

Our follow-up research aims to address these issues. We are working with a conglomerate in Peru, a country that relies on cash and where a large portion of the population is unbanked. Building on our current findings, we are working closely with this company to test the impact of our approach on low-income populations. We will help assess loan applicants using retail transaction data, aiming not only to improve profitability but also to promote social inclusion in the region.

The Research Brief is a summary of interesting scientific papers.