Dale Robinson is an early adopter of consumer-permissioned demand-deposit account (DDA) data and scoring in retail credit, and an independent financial services consultant with over 20 years of experience as a credit risk executive at Ameriprise Financial Services, Bluestem Brands, HSBC and USBank. Dale is a featured panelist at FICO World 2023.
Q: Dale, to start with, can you provide a little background on your business motivation as a credit risk executive for exploring the value of consumer-permissioned DDA data?
A: I think players in the non-prime and near-prime space are trying to get the most complete picture of credit worthiness for any potential borrower, so it’s becoming a “must have” in a lot of ways to stay competitive against new Fintechs in the space and not run into adverse selection. I’m a big believer in testing different data sources. Given the movement in the space, consumer-permissioned data seems like a no-brainer as the next generation of alternative data to round out lenders’ suite of existing sources.
While there have been new data sources that move the needle slightly, in the last four or five years, there have only been a handful of new elements out there that have made a true stair-step improvement to credit performance. DDA data has the potential to be one of them for many lenders.
Q: Many banks, especially larger lenders, are building out standalone calls to DDA data by partnering with DDA aggregation providers, but you took a different route to additionally test FICO’s DDA-based score, UltraFICO® Score. Can you explain your rationale?
A: There are certainly pros and cons to both approaches. My priorities in testing the potential of UltraFICO Score with DDA attributes were speed to market, simplicity and leveraging an existing scorecard already built on the data – and it didn’t hurt that FICO was the organization that developed and branded it from a marketing, approval and compliance standpoint.
I think organizations must ask themselves several questions about their approach: Do you have the time to test live consumer permissioned data through multiple cycles, both macroeconomic and annual? Do you have the capacity to build out your own custom scorecards on the DDA data or can your organization get up and running faster with something available off-the-shelf? As step two, could your organization test or develop custom scores from the individual attributes later after completing a proof-of-concept? For me, leveraging existing partnerships with FICO and Experian seemed like the best path offering the quickest time to market while allowing us to see if there was any potential lift without investing resources and time on standalone custom development.
Q: You’ve shared with FICO that you found value both in the UltraFICO® Score - which incorporates consumer-permissioned DDA data– and the trended cash-flow attributes that are bundled with UltraFICO Score. Can you provide any highlights regarding features of the score and attributes you found most valuable?
A: Well, I think the UltraFICO® Score itself added more predictive strength compared to our base FICO® Score 8. In our test, we only used the UltraFICO Score in our decisions if it benefitted the customer, that is, when their UltraFICO Score went up compared to their FICO Score. We found that preliminary credit performance on those booked loans exceeded what was expected from the increase in the score itself, albeit on limited data.
Regarding the trended cash flow attributes that are bundled with UltraFICO Score, the most interesting attributes were around recency-how is the customer trending on cash-flow that isn’t quite picked up in their credit performance yet? Recent overdrafts or significant drops in deposits were strong indicators of financial stress. Obviously, this may be different for other institutions and other attributes will certainly reveal different signal strength depending if lenders focus on super-prime, near-prime, sub-prime, etc.
Q: Some lenders only ask applicants for access to DDA data permissioned by them when it’s an absolute requirement (e.g., Verification of Income for a mortgage) OR when the applicant is “near approval” and additional information could help the applicant. However, you conducted a test where you offered ALL applicants for a retail credit product the opportunity to provide DDA data permissioned by them for consideration. What are the pro’s and con’s in your view between opening up the option to all applicants versus limiting the offer to a narrower band of applicants?
A: My favorite part of the test we conducted was positioning the offer up-front during the application process. In addition to gleaning information from the score and the raw attributes, we were seeking to learn if there can be additional signal strength by changing the position of the offer and the language presented. Is there a positive or negative select bias by asking applicants for this information up front with language explaining to the applicant that providing the DDA information could only help them?
Q: What insights have you gained (risk or otherwise) about consumers who are willing to permission access to their DDA data? Any surprises?
A: I’m not the first to use the term “gamification” when describing this type of testing in the application process. Our initial hypothesis was that those providing the DDA link would be worse than the population on average because the most credit challenged customers would be most willing to provide their DDA data. We did indeed see some scores slip vs. the traditional score, however, as mentioned prior, contrary to our intuition, we saw a positive select bias even with lower UltraFICO scores. Was this simply because of more “friction” in the process – even if voluntary? Further testing would be needed to change the position of the offer, language presented to the customer, etc., but we were very surprised by this initial result. Also, our final consideration for asking for the data during the initial application was re-decisioning. For simplicity as part of a test, we preferred to make only one call through the auto-decision process, so getting the information before submission was important.
Q: Do you have any advice for lenders to make applicants more comfortable with permitting access to their DDA data in the context of a loan application?
A: I’m in the camp that the next great frontier of understanding customers derives from socioeconomics and doing a bit of “nudging” to move customers in a direction that creates a mutually beneficial arrangement with their financial institution. Communicating very clearly that this is a “give-to-get” arrangement or, in the case of the test we performed, would only be to the customer’s benefit, should drive adoption by lenders. However, we were challenged with this and additional testing was planned to increase the take-up rate of the offer.
Q: With the CFPB’s recently released Open Banking framework, what do you see on the horizon as far as retail lenders’ top priorities go for leveraging the framework?
A: Some of my banking partners might not agree, but I think, from a consumer perspective anyway, this type of rule is long overdue. Allowing customers to easily move between financial institutions should drive competition and innovation at established organizations while breaking down barriers of entry for new market participants. This in turn should increase financial inclusion and lending opportunities for credit challenged customers.
Interested in learning more about UltraFICO®, The Open Banking Score™? Join the UltraFICO® Simulation Challenge at FICO World by registering here.
