Playing Both Sides of the Field: Combining Orthogonal Datasets

Brian Reshefsky - CEO

Super Bowl LVIII is on the books. Whether you were among the 123 million like me who watched the game or the other half of the country who didn’t, you’re probably familiar with the adage “defense wins championships.” Seeing the Chiefs and 49ers go back and forth got me thinking about the legacy consumer risk paradigm where you’re only assessed on your equivalent of defense – how successfully you’ve kept creditors at bay, as reported on traditional credit reports based on historical tradelines.

However, as a borrower you spend most of your time on offense, moving the sticks as you earn a living then save and build assets. Creditors don’t ignore these datapoints – after all, metrics like debt to income ratio are critical in assessing a consumer’s ability to repay. Few lenders have automated this important area of credit risk assessment (though lenders who have automated these datapoints in their real-time risk assessment see meaningful uplift in portfolio performance).

Instead, most lenders today either automatically approve applications that pass other real-time tests (like a bureau score and fraud risk screens) or they push the application to manual review and require paystubs as a condition to approval.

More sophisticated digital lenders are already leveraging open banking data to verify income in real-time, streamlining approvals for higher conversions and fewer drop-offs. Fewer are utilizing balance behaviors and consumer spending habits to go beyond ability to pay for risk insights not seen in any other dataset – and arguably the most predictive alternative data available.

There’s tremendous power to combining the defensive insights from credit reports and scores with open banking data that reveals consumers’ ability to “score points” in keeping with the metaphor. Very different risk profiles from a credit report emerge when you systematically separate applicants who are responsibly spending and saving versus those who live paycheck to paycheck.

Smarter people than myself call the two datasets – bureau data and transaction data – orthogonal, simply meaning they’re uncorrelated and originate from different, unrelated activities. Let’s explore why you’ll win more frequently and by wider margins if you can bring a complete game to your credit decisions if you leverage what’s available to you on both sides of the field.

Why: The Case for Completeness

There are three overarching reasons why every risk decision should include transaction data that’s available in the open banking ecosystem:

  1. The positive – Wouldn’t it be nice to never pause an approval for proof of income again? And not have to pay The Work Number® a boatload to do so? Peering inside a consumer’s primary financial account reveals every ACH within seconds of successfully connecting and enables real-time, automated decisioning on their ability to pay. And doing so is far from a Hail Mary pass, with almost 90% of Americans now willing to share banking activity data.

  1. The negative – Score inflation is real. Credit scores are at all-time highs yet delinquencies have been rising. The disconnect is evident: credit reports are a lagging indicator by about a month, they offer no insight into the income or assets that must be considered in assessing ability to pay, and they miss potentially significant obligations like rent, utilities, and Buy Now, Pay Later plans. Open banking analytics consider all this and more, like balance behaviors  

  1. The unknown – Most of us probably know the parable of the blind men and the elephant where everyone has a different perspective and they’re all wrong. Looking at any one dataset, whether it’s a bureau score and report or banking activity, results in an incomplete picture. Whatever you’re not seeing could be the difference between an auto-approval and a charge-off, including meaningful employment paystreams where you don’t need to delay the application for proof of income and hidden liabilities like Buy Now, Pay Later plans not seen on a bureau report that could evidence real financial strain.

Simply put, most risk decisions will be better informed with better outcomes for lenders and consumers alike when you bring to bear the full offensive and defensive tools at your disposal. The orthogonal bureau and banking activity datasets don’t just seem complementary from an anecdotal standpoint – as I’ll share in the next section, the benefit is readily quantifiable.

How: Layering Insights

If you’ve been in the consumer lending business for more than a few years then you’ve weathered more than your share of ups and downs between pandemic, recession, stimulus, inflation, rate uncertainty, and more. Your credit policy has no doubt evolved, and you’re not about to start over with open banking data.

Instead, we typically see lenders have the most success when you layer rules related to banking activity insights on top of your existing credit box. In effect, you’re taking a second look (in real-time) at some portion of the “tails” that appear from a legacy policy standpoint to have the most and least risk.

The overlay of banking activity insights invariably reveals applicants in both tails who present very different risk profiles for all the above reasons (and more). This enhanced good versus bad risk separation then allows lenders to more accurately rank order risk and, in turn, to “swap in” good risk surfaced with the new insights and “swap out” applicants with inflated bureau scores.

Take this example of a highly regarded digital lender whose overlay of open banking insights onto its legacy approach enabled separation with roughly 2x precision:

For the best 20% of applicants, open banking insights reduced defaults by over one-third – from 9% down to 6% – for the best fifth of applicants in this portfolio. Turning to the worst 20% resulted in similar uplift with over one-quarter more defaults – from 22% up to 27% – in a risk band most lenders aren’t likely to underwrite, freeing up capital to instead approve more applicants with favorable risk profiles.

Naturally, the composition and risk guardrails vary across every applicant pool and resulting portfolio. The important first step is to lace up your cleats, hit the field, and log enough gametime with this alternative data to test, learn, and refine where cutoffs make sense to auto-approve and auto-deny.

What: Get in Your Draft Picks

I’ve seen a number of entrants (and some exits) in this space as I approach my fourth year of selling and delivering risk insights from banking activity data. My perspective from the playing field is that there are several important considerations that make for a strong offense and ensure you’re set up for success with open banking insights:

  • Consumer lending know-how:  Operationalizing open banking insights is not only difficult but requires expertise both with the new data and with how to incrementally improve credit policy with new risk insights. You’re essentially calling audibles each time a new vintage of loans is evaluated then matures, deciding whether to tighten or loosen rules for performant attributes or risk scores. And you’re much better positioned to make those calls with a trusted partner who’s built a playbook from numerous successful adoptions of open banking insights. Ask both EDGE and others which of our customers have made more money and to what extent

  • End-to-end capabilities:  Whether you’re accessing bank data through one of the many aggregators (or aggregators of aggregators like EDGE), there’s demonstrable value in working with a provider who’s active in all aspects of the open banking ecosystem. Aside from the convenience and cost benefits of a single contract that gets you from data access to actionable insights, as the space evolves you’re likely to see more innovation and more value from providers who are equally attuned to developments in data acquisition as to consumer credit risk (and everything in between).

  • Regulatory ducks in a row:  Just as in football, penalties can dramatically impact outcomes. In such a highly regulated area as consumer lending, there’s oversight and enforcement from the CFPB, FTC, OCC, and more. This complex regulatory landscape poses an existential risk for any lending business that trips among the alphabet soup of regulations, and the same is true for your data providers and technology partners who could be out of business or put you out of business if they’re not adhering to the highest levels of regulatory compliance. EDGE became a CRA (consumer reporting agency) in mid-2023 to ensure that we and our customers are squarely on the right side of all these regulators, and we’re starting to see others in the open banking ecosystem follow suit.

If any of this resonates with you, drop a line today and let’s see how we can strengthen your offense or at least start fielding teams on both sides of the field so we can set you up for long-term success with open banking data and risk insights!

Let's start a conversation