In the post-credit report lending world, the name of the game is maximizing coverage in credit underwriting. Creditors work hard to attract applicants for their products. The shame is when worthy applicants cannot be approved. The inherent bias of lenders is to grow the topline safely. A key factor is being able to get sufficient information about the applicant to make an appropriate decision.
Maximizing coverage means using alternatives to the classic credit report. In considering alternative data for credit underwriting, three primary factors important:
There are other factors, of course. FCRA compliance is a big one. Cost of the data is another issue, as it has to be economical for the creditor. And there’s the intangible smell test for data; for example, how would lenders use this as a basis for decision: “If you fill in your name in all caps, you’re a much higher risk.”
Here’s a brief look at the key factors to consider in maximizing data coverage.
Consumer will provide
For consumers, in any online realm including credit, providing access to personal information presents three issues to be navigated.
Storage and use of my data: Consumers rightly maintain concern about who has their data and what they’re doing with it. Data breaches have become more prevalent. There naturally is increased reticence to providing personal data. In our online world, providing access to personal information is unavoidable. But consumer will need to feel like the benefits outweigh the concerns.
Remembering access credentials: Quick: tell me the login ID and password of your five favorite sites and apps. It’s not so easy. Differing passwords (e.g. “Your password must be 12 characters, and include a capitalized letter, a number, and a special character.”). Varying logins (email, user name, account number). Asking for login credentials is friction in the process.
Stolen login credentials: Related to both issues above, there is the very legitimate concern about loss of login credentials. It’s so bad, Google Chrome has started telling users when their password showed up in a data breach.
Providing access to my financial data – recurring payments to service providers and bank account transaction information – is an active choice. This differs from the credit report, where consumers do not proactively provide access to payment and debt histories.
Non-financial alternative data can be sourced from a wide variety of places. Generally, the less personally sensitive the information, the easier it should be to get consumer buy-in. Things like social networks and cell phone usage likely cross a threshold that consumers are uncomfortable with. We see this discomfort today in concern over how Facebook uses data. But there is an opportunity for sourcing non-financial alternative data that is not as invasive.
Third party will provide
It’s one thing to earn consumer approval to access data. It’s another thing to get that data. Focusing on two sources for alternative data – recurring payments, bank account data – helps to understand the lay of the land.
Recurring payments: Services for which people pay ongoing, monthly payments vary in availability. A big one that people have looked at is rental payments. Substantial payments that are due every month. The challenge with rental payments is the highly fragmented nature of the rental industry. There is a concentration of landlords who can provide regular standardized data. To collect rental information at scale would integration with thousands of individual companies and proprietors.
Utility providers have more market concentration, making it easier to secure data. However, historically they haven’t provided it. Why? It’s a new layer of red tape. There’s a charge to report payment data to credit bureaus. In addition, once that data is reported, utilities can see an increase in disputes that need resolution.
More recently, recurring payments of lighter nature have emerged as possible alternative data. For instance, Experian Boost lets consumers add payments to four streaming services to their credit report:
Note that to get this information, the consumer needs to provide their checking account login. And banks have generally made it easier to do so.
Bank account information: Access to bank account information is, in a way, a throwback to hyper local banking of a bygone era. When your local banker knew you and your family, and could easily make a lending decision.
While some of the more interesting sources of “more financial facts” – rent payments, utilities – are less likely to be available, many third parties can and do provide more financial facts for use in underwriting.
Of course, it would be great to take third parties out of the equation. This would reduce friction, non-standard information, and changes in industry participants.
Data is predictive
The third requirement of alternative data is that it must be predictive. Obviously, this is table stakes for even being considered.
Experian Boost, which includes recurring payments, reports that a consumers on average see an increase of 13 points in their credit scores. It’s safe to assume that calibration of credit scores is based on statistical rigor as being linked to payment default.
Perhaps the biggest issue in alternative data is the ability of a creditor to use the data in underwriting. For instance, Experian Boost will incorporate payments to streaming services. But lenders who are TransUnion or Equifax “shops” cannot make use of that data.
Of course, some interesting potential alternative data doesn’t pan out as predictive. As an executive told the Wall Street Journal in 2014: “There could come a time where certain social media could be predictive and we’re looking at that, but it isn’t yet,” said Anthony Sprauve, senior consumer-credit specialist at FICO.
Creditors have a natural interest in expanding the number of applicants they can approve safely. The story of alternative data is one continued to expand its coverage.