BNPL – Borrowing No Different
by Edge Insights | Buy Now, Pay Later
Buy Now, Pay Later (BNPL) plans are a recent innovation that enjoyed steady growth through the Covid-19 pandemic and resulting downturn. Now with BNPL delinquencies rising and consumers squeezed from inflation outpacing income growth, investors and lenders alike are closely watching the coming quarters to see how the industry fares in a potentially meaningful recession.
Until recently, hypotheses about through-cycle performance of BNPL lending were based on analysis by proxy of installment loans. The interrelationship of BNPL with repayment of other credit products has been a blind spot in lending because BNPL liabilities are mostly hidden from other lenders.
With the leading data lake combining bank transactions and loan performance, Edge has unique insight into the impact of BNPL on other forms of credit. Our findings indicate that borrowers exhibit behaviors similar to what lenders have seen across other credit products for decades.
Stacking, Stress & Default
We’ve learned that BNPL stacking is highly correlated with broader credit utilization – more traditional tradelines correspond with more BNPL obligations:
- Someone with only one traditional credit product is most likely to have zero (40%) or one (26%) BNPL loan.
- On the other hand, someone with three or more traditional credit products is most likely to have two, three, or more BNPL loans (55%).
Financial stress is closely linked to unsecured debt levels. Some believe that the quick and easy nature of BNPL credit makes it easy to get overextended. From our unique vantage point, we’re able to demonstrate that borrowers who incur more BNPL obligations show signs of greater financial stress.
Lenders often use payment-to-income ratio (PTI) to ensure a consumer generates enough disposable income (after) to service their debts. For this reason, PTI is also known as debt service ratio or financial obligations ratio.
A typical credit active subprime or near-prime borrower with no BNPL loans currently has a PTI ratio of 20-25%. For the most active BNPL borrowers paying three or more BNPL plans, the average PTI more than doubles to roughly 50% payment-to-income.
Default risk increases with higher leverage, and economists have observed financial stress driving higher delinquencies and charge-offs since antiquity. We don’t have the crystal ball to call a credit cycle, but what we have observed with the rise in BNPL is a higher risk of default for consumers with additional indebtedness from BNPL plans.
Edge data indicates that borrowers with BNPL debt in 2022 have been 10-15% more likely to default on one or more of their obligations – whether it’s their BNPL loans or a traditional credit product.
When BNPL obligations are rightly analyzed as another tradeline, we see consumers behave just as they have with other financial products. History repeats itself, and lenders can only apply empirical risk techniques to properly assess BNPL borrowers when they have visibility across all financial obligations.
Deriving Actionable Insights
Seeing BNPL obligations is an important starting point, but accessing bank transaction data is not enough to make informed risk decisions. You need consistent data to extract meaningful insights and calibrate relevant insights to assess their implications for a consumer’s risk profile.
Bank transaction data is surprisingly disparate as you look across data providers and banks, whether you’re reviewing a PDF or ingesting a JSON payload. Our analysis at Edge starts with a “universal translator” to harmonize and normalize raw data from any provider. This unlocks the ability to perform real-time, automated analysis of data from nearly every bank in the U.S.
The next step in deriving actionable insights is to systematically extract measurable properties, something our data scientists call features (others like credit bureaus would call attributes). Examples of attributes extracted from traditional credit data include tradeline data, inquiries, and revolving versus transacting behavior.
With BNPL loans, the amount and frequency of payments – all relative to a consumer’s income – are part of a core Edge feature set that includes other indebtedness, income, balances, and high-risk behaviors.
Edge was uniquely born out of multiple decades of data-driven digital lending experience and incubated in a consumer lending business where every feature is pressure-tested for predictive value in underwriting. We curate the features delivered to our customers, so you’re not boiling one ocean of data just to end up in another. Remarkably, it’s possible to over-analyze a year of bank transactions and end up with ten times more features than the underlying transaction count!
When the questions you’re trying to answer come down to “Do I want to lend to this consumer?” “How much do I want to lend them?” and “What interest rate should I charge?” then fewer, more powerful insights will ultimately deliver the most value for your business. Features are the building blocks for predictive models, and Edge distills our features into a simple three-digit score that predicts default risk along the lines of credit scores widely used in consumer lending.
Edge is able to rank and score risk because we’ve not only analyzed over 2.5 billion bank transactions but also calibrated our risk score against over 2.5 million consumer credit applications. This unique dataset enables us to deliver a risk score that’s differentiated from any other data and analytics provider – with demonstrable value for customers.
We’d love to engage on how Edge can improve the top and bottom line for your business. Please reach out today if you’d like to better understand how we can demystify hidden BNPL liabilities in your underwriting and enable decisioning on a consumer’s complete financial picture.