Automating Decisions

“If you want to do this type of lending, the model is the best answer to compliance.” —Robert Hughes, Finance and Thrift Co.
“If you want to do this type of lending, the model is the best answer to compliance.”
—Robert Hughes, Finance and Thrift Co.

A California community bank deploys software to guide and streamline its consumer lending

By Beth Mattson-Teig

When Robert Hughes joined the Finance and Thrift Co. as president and CEO at the end of 2007, it didn’t take him long to realize that he had a big task ahead of him.

At the time, the Porterville, Calif., community bank had $100 million in assets and 22 branches in the Central Valley area in towns ranging from Modesto to Bakersfield. There was a loan policy in place, but the interpretation of that policy varied widely among the different branches. The niche for the bank, both then and now, was and is consumer lending.

“What you had was people at each of those branches making their own decisions. It was 22 branches [each] doing their own thing,” Hughes says.

New Software Tools of the Trade

Online Loan Payments

An online and mobile payments services provider, Allied Payment Network in Fort Wayne, Ind., has introduced a new Web portal to allow banks to accept last-minute online loan payments from borrowers. Through the company’s PortalPay interface, borrowers can make loan payments from any checking account, without the need for an existing Internet banking account. To make a payment, borrowers enter their contact information, loan number and the account information from a checking account of their choice.

A mobile digital bill payments option also allows borrowers to take a picture of a check to send a loan payment from a smartphone.


Consumer Loan Stress Testing

FICO in San Jose, Calif., has launched a predictive credit risk management software system that allows banks to stress-test their consumer loan portfolios. The company’s software module, called Score Economic Calibration Service 2.0, forecasts the affect different economic scenarios would have on both a bank’s individual consumer borrower’s repayment as well as an overall consumer loan portfolio. The projections, based in part on the credit scores of borrowers comprising a bank’s loans, would attempt to measure expected capital losses and capital needs under certain situations.
Read more:

Enhanced Data Analytics

Computer Services Inc. in Paducah, Ky., has released a core banking software version that includes a suite of new business intelligence tools to allow banks to capture and analyze customer data. The CSI IQ module, developed with business analytic software developer SAP America Inc., is designed to allow banks to more quickly and easily access summaries of customer information through an executive dashboard.

Various consumer loans were being made that probably shouldn’t have been granted, such as financing a Hummer sport utility truck for somebody working at a $10-an-hour job, he adds. As a result, the bank was seeing big losses when those loans didn’t work out.

Hughes decided the best solution was to introduce a specialized in-house automated loan underwriting system. “I am more of a traditional banker. I never thought that I would be a person who used a decision model,” he says. “But if you want to do this type of lending, the model is the best answer to compliance.”

An added challenge for Finance and Thrift was being sensitive to its customer base of predominantly middle- and low-income customers. The bank serves an area with a large Hispanic immigrant population. Half of the bank’s customers speak only Spanish. Those customers don’t necessarily have poor credit, but rather limited credit, notes Hughes. The bank is a Minority Depository Institution and a Community Development Financial Institution.

Finance and Thrift worked with Atlanta-based Verde International to develop new loan underwriting software. In the end, the bank bought three software systems—a core banking system, a loan origination system and a decision engine. The decision engine, introduced in 2008, gathers borrower information such as application data, a credit report and other third-party data.

The decision engine generates one of three responses for a loan application—an approval, a decline or a counteroffer, and establishes loan pricing. If the system can’t approve a $10,000 loan for 48 months at a rate of 10 percent, then it searches for the best available loan it can approve in a particular situation.

Now about 99.6 percent of all of the bank’s consumer loans are fully underwritten and priced by the model with no human intervention. “It is all based on criteria in the model,” Hughes says. The only loans that a senior bank executive reviews are those that have, for one reason or another, insufficient data for the bank’s software model or if a customer complains about an application that was turned down.

Bringing in the specialized loan-decision software was a game-changer for Finance and Thrift in several ways, Hughes says. Most notably, it brought efficiency, consistency and objectivity to the loan underwriting process. Since the software was implemented, rates of loan delinquencies have improved significantly and are currently running at about 3 to 4 percent. The bank also has grown to $130 million in assets, while reducing its branch network from 22 to five. In addition, employees have more time to spend with customers.

Finance and Thrift continues to specialize in making consumer loans, particularly used car loans. One of the challenges of that specialty is the speed at which the loans turn over. The bank has a $90 million loan portfolio, and just to break even the bank needs to originate $5 million in new loans each month.

“We do a lot of loans,” Hughes explains. “So having an automated system is a way to re-humanize the process so that our people aren’t just crunching numbers and making a decision.” Now the bank’s computers do much of that work, which frees time for bank employees to engage with customers and try to better meet their needs.

“We moved from just a finance bank to a bank that is more focused on community involvement,” he adds.

Beth Mattson-Teig is a freelance writer in Minnesota.