How most lenders evaluate credit applications has not changed in 20+ years. Real time credit scoring, automated instant decisioning were being utilized by some lenders as early as in 1990s. But the situation is changing rapidly as new technology makes inroad into Financial Services. We might be at an inflection point. Newer FinTechs are often going with self-developed custom credit scoring instead of using canned bureau data based credit scores.
Where are the established banks and credit unions?
· Banks and credit unions (Financial Institutions or FIs) lacked and mostly still lack the infrastructure to leverage data across silos, like using debit card or checking account data for lending decisions.
· Many have antiquated databases that cannot handle or scale with massive amount of unstructured data like customer service chat, customer complaints, mobile or online user behavior tags.
· Large FIs have bench strength in analytical talents but they are often hemmed in by tools not capable of crunching massive amount of data needed for techniques like unsupervised machine learning.
What could the future look like?
Future involves using novel data beyond credit bureau information and powerful algorithms that can shift through massive amount of data to find patterns. For example, the modeling arm under a hedge fund’s umbrella is using social media metadata to predict stock prices and then used the same methodology for credit scoring.
Possibilities are endless. Utility bill payment, person to person payment information, prepaid card information, cable or rent payment- all these are important data about a person’s financial behavior but mostly happen outside of traditional credit bureau information.
With the explosive growth of Buy Now Pay Later, where most consumers use debit card to borrow, there is now growing amount of credit usage and payment behavior outside of traditional lender data sources.
This does not mean a human sitting behind a computer will be reading Facebook posts and LinkedIn profile of every credit applicant. That will be expensive, inefficient and most importantly create delay in a process where instant decisioning is prized.
Public information like social media can be scraped off internet by computer programs. Non-public, personal account level information can be retrieved for a consumer with their express consent when they apply for credit. Services like PLAID already exist for connecting to any FI. Similar service can be created for largest utilities, rental company payment portal or mobile phone carriers.
Who will benefit from changes in underwriting methods?
Beneficiaries would be people who often get declined when applying for credit. The segment consists of two distinct and separate groups - people with bad credit and people with no credit history.
Unfortunately, applicants without credit history are pretty much treated as bad credit while in reality many of them may be exactly the opposite. A good chunk of the no credit population is young adults, some are new to the country and remaining are low income and find banking difficult and expensive. Some of the unbanked Americans receive their pay on prepaid cards which can show steady income but not reflected in traditional bank information.
Even for people with bad credit, the assumption that past behavior is the only future predictor is limiting. This misses specific triggers that may have caused certain credit behavior (for example, job loss, short term disability etc.) and fails to anticipate improving credit behavior until after it is observed.
I have seen instances where a lender forced minimum housing expense if an applicant said no rent or mortgage, especially if the person did not show a paid off mortgage in the credit file. I am against such paternalistic treatment of applicants. Some of these could have been easily verified by using a PLAID like service to verify applicant’s deposit balance or direct deposit which will be a good indicator of customer’s ability to pay and find rent or mortgage payment.
Why should the FIs change their approach?
Because it is good for their bottom line. An evolution in credit underwriting will help them to
· Approve more applicants
· Improve customer experience
· Get loyalty by being the first to extend credit to a new segment
· Lower expenses with higher automation
With number of FinTechs providing both deposit and loan service together, they are already on their way to provide credit using deposit data. FIs like credit union which almost always have deposit account information for their loan applicants should start using it. Other option is to make the data available to a third party who can combine various data sources to make loan decision. If open banking makes inroad in US like in European Union, lot of the bank data will be available to their competition anyway.
The new methodologies can not only open credit to traditionally untouched population, it can also help FIs fine tune underwriting process for customers with credit data. Probability of not paying off a debt is often used not only to decide a credit application but also to determine cost of that credit to consumer. Leveraging new data which can make more precise prediction for a large population can also help to lower APR for many loan products. And that is a good thing for American borrowers……
Commenti