Accelerating the Inevitable with Consumer-Permissioned Data

Today, we are formally (and finally!) announcing our biggest product launch since our founding. It’s a broadening of our vision to modernize the credit system. Let me tell you why.

Credit reporting: an imperfect system

Credit reporting has been around for over a century, and its fundamental building blocks have not meaningfully changed. Every few weeks, lenders furnish data to the credit bureaus, who match that data to an individual’s identity and sell it back to lenders (with an increasing number of bells and whistles). 

But this system was pieced together haphazardly. As more and more lenders emerged to sustain the rapid pace of increasing consumer demand for credit, credit bureau data (and, later, credit scores) became the accepted stand-in for trustworthiness in our financial system. It was a game-changing, democratizing moment. In the matter of a decade, the basis for the ability to borrow became observable merit, rather than who you know. The advantages of this new tool were self-evident: lower risk for lenders and a quicker path to mass adoption of credit products. With an entire country of customers to acquire, financial inclusion was not a priority. 

Today, this initial oversight has come back to bite us. Our society has globalized, digitized, and continues to adopt ever more complex financial products. Unsurprisingly, the conventional rubrics of credit scoring, designed around the needs of 1950s lenders, are as disconnected as ever from the needs of the modern consumer. Millions of people remain underserved by the American financial system.

Luckily, technology has brought about a new paradigm: consumer-permissioned data. 

Nova Credit: where we started

Nova Credit was founded on a strong conviction that consumers should be able to wield the power of their own data to unlock access to financial products. Not only does this principle make for a more empowered consumer and a more inclusive system–it breeds efficiency on all sides. Lenders are enabled to conduct a more informed risk assessment and consumers gain an efficient onramp to credit products at fairer rates.

We started with one of the most obvious applications of consumer-permissioned data: financial access for immigrants. There are nearly 300 million immigrants around the world, and before Nova, they arrived in a new country “credit invisible.” Local credit bureaus had no way of knowing who they were, and as a result, these new arrivals, who needed credit to set up practically every aspect of their new lives–to open a bank account, apply for a credit card, lease a car, rent an apartment, sign up for a cell phone plan, etc.–were shut out from the financial system.

Over the past five years, our team has worked tirelessly to help immigrants overcome this challenge with data. Nearly 70% of recent immigrants can now bring their financial history from their home country directly into financial applications. Nova Credit’s global infrastructure can access over 1 billion of these Credit Passports® around the world and we deliver that information to partners like American Express, Westlake, and Verizon through a real-time API endpoint. This capability has globalized the world’s credit bureau system. It no longer matters where you’re from or where you’re going.

A new challenge: credit-excluded Americans

But immigrants are not the only ones who struggle with financial access, and they’re not the only group who could benefit from consumer-permissioned data. In fact, over 60 million Americans struggle to access the U.S. credit system. Among them are:

  • New-to-credit consumers: Anyone who hasn’t had credit before gets stuck in the infamous “Catch-22” of not having a credit history and therefore not being able to access credit. This paradox has existed since the very foundation of the credit reporting industry and remains largely unsolved. We estimate that there are ~15-20 million new-to-credit consumers that primarily consist of immigrants and young adults (Gen-Z). 

  • Infrequent credit users: These consumers can be of any age, but simply don’t use credit as frequently as most credit algorithms would need to score them. Their consumer data tends to be sparse and fairly outdated, resulting in unrepresentative scores. This segment represents ~10% of credit files. 

  • Other thin-files: These are consumers who have a credit file, but the information is insufficient to garner approval from most lenders because of a limited set of tradelines or limited length of history for any individual trade. They represent another ~10% of credit files.

Without the ability to signal creditworthiness, individuals in these groups are left with a dismally limited menu of financial products. Most of these credit-excluded consumers turn to secured cards or high-interest rates to (re)build their credit histories, effectively paying a tax to the U.S. financial system in order to be able to access it fairly. 

But what if these consumers had some other way to show that they are reliable borrowers? Turns out, they do. For many new-to-credit, infrequent, and thin-file consumers, consumer-permissioned bank transaction data could be the answer. 

(Re-)introducing bank transaction data

Before credit scores became the universal access pass for financial services, bank account information was a key component of credit underwriting. When you’d apply for a loan, your loan officer would ask for a bank statement to understand how you manage your finances. While this is still true today for most mortgage applications, bank transaction data has gradually disappeared from the calculus of broader consumer lending because of how cumbersome it is to gather and process. 

Yet bank transaction data continues to represent an incredible wealth of information about a consumer’s borrowing behavior. Deposits capture a consumer’s complete income profile, whether they receive a recurring paycheck every couple of weeks or earn in a less conventional way (e.g., gig workers, small-business owners or rental income). Expense information reveals discretionary and non-discretionary spending, including any recurring loan payments. And cash balance lays bare the overall financial health and trends of a prospective borrower. 

Moreover, in our increasingly cashless world, bank transaction data captures nearly every financial transaction we make. According to the FDIC, approximately 95% of U.S. households have a checking or savings account, with an estimated 600 million checking accounts across the credit eligible US population of 250 million. By any measure, this is the single most expansive financial data source about U.S consumers. The insights that can be gleaned from it are vast and could be highly complementary to traditional credit bureau data. 

Bank transaction data is now readily available through a range of increasingly sophisticated aggregators (Finicity, Plaid, Yodlee, etc). And permissioning access to this data has already become an accepted feature in many aspects of modern banking with over 80% of Americans having linked an account.

So why, then, is this source of insight not being used more widely for underwriting? 

To implement a scalable solution for bank transaction data underwriting, an individual lender would have to move mountains. They’d need to determine which bank aggregator to work with, assemble a training set to model, spend months doing data engineering to identify key attributes and whether they can be used compliantly, debate how to handle consumer disputes and adverse actions, figure out how to modify user flows while minimizing friction, run retroactive studies with their own clients, and more. And all of this would have to be done prior to actually testing a new lending hypothesis with real data. 

This is a behemoth of an undertaking for even the most nimble financial institution, and it’s a lot to ask when the sad reality is that many banks today don’t even have the ability to check the account balance of an existing customer applying for a new credit card. 

Even if financial institutions could learn to use bank transaction data for underwriting on their own, it would be inefficient, the same way it would be inefficient for every lender to invent a proprietary credit score from scratch. It’s clear that, like lending data, bank transaction data is a dish best served standardized.

Introducing Cash Atlas™ - cash flow-underwriting-as-a-service

Today, I’m excited to unveil what we’ve been working on for the past two years: Cash Atlas™. With Cash Atlas™, consumers are empowered to choose to share their own bank transaction data to paint a more complete picture of who they are when applying for credit.

We want to make it possible for lenders to use bank transaction data in a matter of days. We’ve spent years assembling a statistically representative data set, categorizing and standardizing that data in a transparent and compliant manner, optimizing user experiences and opt-in value propositions, and working with several aggregators to understand how to create the best end-to-end offering. 

We’re certainly not the first to bring bank transaction data into underwriting. Many banks and fintechs have already proven its value across economic cycles. Some of the credit bureaus and FICO have even tried to bring this data into the credit system through initiatives like Experian Boost and UltraFICO. But these solutions don’t always meet the needs of those underserved by our financial system. Only credit superusers have the foresight and sophistication to pre-emptively link their bank transaction data to augment their credit scores in advance. 

As we learned from Credit Passport®, truly inclusive systemic solutions need to meet the consumer at their point of greatest need. By embedding our technology in the middle of a credit application, we give consumers the greatest chance to leverage bank transaction data for approval. 

Our approach with  Cash Atlas™ is designed to balance the competing needs of consumers and lenders. For consumers, we’ve focused on creating the ideal user experience, with the educational content to explain the benefits of linking bank transaction data. For lenders, we’ve developed a rigorous set of attributes to make using this data seamless, scalable and compatible with the FCRA. 

It’s hard to imagine a world in a decade’s time where bank transaction data is not a core component of how lenders make credit decisions. The data is readily available, rich in insight, and orthogonal to credit bureau data. Today, we’re taking a big step forward in accelerating the inevitable for this industry.

As a company, this also marks a step in Nova Credit’s evolution from the world’s leading cross-border credit bureau to a consumer credit bureau that uses permissioned data to enable (i) consumers to paint a more complete picture of who they are and (ii) lenders to make more fair and informed lending decisions. With the Credit Passport® and Cash Atlas™, we are taking a significant leap towards our mission of “powering a more fair and inclusive financial system for the world.”

I want to thank our many partners for helping to make this a reality. Most importantly, the potential impact of this product would not be possible without the hard work, grit, and world-class expertise of our incredible team. You have helped broaden our mission to solve yet another industry problem, once and for all.