In these times where dispensing consumer credit has been akin to doing it through an eyedropper, the use of alternative data sources is one way to possibly increase the flow of money, one proponent said.
Mark Luber, vice president of analytics and data acquisition for LexisNexis Risk Solutions, Alpharetta, Ga., said, “A historically credit-damaged consumer isn't someone who is necessarily not worthy of credit now. And so what we have is the financial services industry looking for new data that can help them find those consumers who are good risks despite some of their past credit history.”
Another segment lenders are seeking out is the thin file or no credit file consumer, those whose histories are so sparse they do not generate a credit score. Banks have approached LexisNexis Risk Solutions regarding attracting these consumers, he added.
Through its work with banks, the company has determined that up to two-thirds of these thin file customers “are really relatively good risks and have low loss rates. They just don't have a significant current bureau performance,” Luber said.
Bringing in data that LexisNexis has and combining it with whatever information these people could have on file with one of the credit bureaus, the user can generate “a meaningful, predictive score,” he continued, explaining “we're looking for all of the other impressions that a consumer leaves as they operate in the world in a financial context.
“We're primarily looking at a consumer's stability, a consumer's ability to repay a debt and a consumer's willingness to repay a debt. Those are the traditional categories and then we try to build out data assets that speak to one of those categories.”
His company compiles “an incredibly deep history of every consumer whose operated in the U.S. We have more than 30 years of history for each of these consumers.”
That includes their address history. So LexisNexis has created the concept of address stability. It analyzes how much each consumer physically moves around.
Luber said the company has determined that address stability relates significantly to creditworthiness. Very stable individuals are half as risky for credit than on average.
Those individuals who move frequently are significantly riskier, he continued. If the consumer has ever had an eviction when they were renting, they are half as likely to pay back the debt.
In the area of ability to repay, LexisNexis looks at assets—what sort of property assets does that consumer have now and in the past. He said the asset information helps to determine that ability to repay.
Another data point the company collects under the ability to repay header is professional licenses. Someone with a professional license is more likely to have the means to earn income and thus a greater ability to repay a loan.
As for willingness to repay a debt, LexisNexis looks at what Luber termed as some of the more negative parts of a consumer's history.
This includes felony convictions and bankruptcies. Consumers with this type of information are obviously more risky, about 30 times more risky, than on average.
Their overall wealth is also an indicator as to willingness to repay a debt.
He added LexisNexis is not just looking at a point of time in the consumer's history. “What we found is that the trajectory of the consumer is at least as important as their current position. So we've developed the concept of economic trajectory.
“So is that consumer moving up or down relative to their financial position. That trajectory is very predictive for risk,” Luber said.
LexisNexis collects data from thousands of sources, many of which are public records. That includes going to hundreds of courthouses to collect information such as lien history or criminal history. But it also gets data from commercial relationships it has established with third-party information aggregators.
It then makes certain the information is linked to the right consumer, he said, and from that it examines predictive attributes from that data. Those can be used by the lender to develop its own internal model.
“We also do a significant amount of custom modeling where we actually use a financial institution's performance data and develop a model on their behalf and implement it our system,” he said. The consumer's information is pulled, run through the model and LexisNexis returns a score.
Many smaller- and medium-sized institutions use what Luber called one of the company's “flagship models.” These are based on a broad view of particular industries.
Much of these models have been used in credit card and auto lending, as well as for use with demand deposit accounts. In the last few months, LexisNexis has seen a lot of interest from the mortgage industry for this service in two contexts: risk rating a portfolio and generating recovery models to determine the likelihood of someone resuming paying when they have already stopped.
“The mortgage industry has really lagged on the use of alternative data in making better decisions but the work we've done most recently shows that there is significant value in the mortgage industry doing this.
“That doesn't surprise me given how effective this has been across other markets,” Luber said.
Now that there has been penetration using this kind of data in the portfolio analysis stage, it leads to interest on the front end of the mortgage process, he said.
“By using alternative data, you're going to make better decisions for the bank and more fair decisions for the consumer. You're going to be able to offer a lower rate for those risk consumers and a more appropriate rate for a higher risk consumer. Having this information is going to create a more fair result,” Luber said.
Back in 2004, FICO got into the alternative credit game with its Expansion Score. The sources of data for the Expansion Score include deposit accounts, pay day loans and product purchase payment plans.
At that time FICO said there were 50 million American consumers who do not have enough data in one of the credit repositories to generate a traditional score.
