The mortgage market has been showing signs of life, but anyone over the age of five knows that doesn't mean the market's being filled with good loans and healthy borrowers.
Data mashers like Digital Risk and CoreLogic (CLGX) see an opening in the newly risk-averse mortgage industry. They say state-of-the-art data accrual and analytics can help lenders get a heads-up before a mortgage boom turns into a mountain of bad loans.
Digital Risk and CoreLogic recently unveiled dueling mortgage risk products that aim to broaden the data sourcing and analytics used by originators and investors to vet a borrower's overall health, economic conditions and collateral and a lender's underwriting and lending policies. These vendors are joining a crowded field of analytics providers that also includes credit scoring companies like FICO (FICO), business intelligence providers such as SAS and business information software developers such as Wolters Kluwer (WKL). These vendors have upgraded their systems to use broader data sourcing, including unstructured data such as social media, to provide a view of a borrower's financial health that goes beyond mortgage debt ratios and traditional credit scores.
"For sustainable securitization, there are a number of issues to be resolved, things that got us into the crisis," such as questionable products, low loan quality and "fundamental issues such as lack of due diligence," says Craig Foccardi, a senior research director at CEB TowerGroup.
The companies are targeting an expanding mortgage refinancing market. Mortgage rates are near record lows, boosted by the Federal Reserve's decision to buy millions in home loans. The market was already doing well; the Mortgage Bankers Association's refinance index neared a three-year high in August. "We're in the middle of one of the biggest refinance booms we have ever seen," says Jeff Taylor, a co-founder and managing partner at Digital Risk.
Digital Risk, in business seven years, has developed Veritas, a platform that uses behavior analysis to determine risk levels of potential loans and borrowers. "If you have two houses in the same area, and two people under the same duress, from a behavior perspective they aren't going to act the same way," Taylor says. "If it's become socially acceptable in some areas" to walk away from a home, "the market's dynamic has changed."
Digital Risk, which was founded in Maitland, Fla., in 2005, is positioning Veritas as a tool that models both systemic risk (or market risk) and operational risk, the procedures and metrics used by a lender in making loan decisions. To do this the company leverages borrower, property and local real estate data as part of its scoring methodology. The mix goes beyond the credit score and loan-to-value ratios long used to underwrite mortgages, Digital Risk says.
Veritas uses its data to divide borrowers across a range of 32 clusters. It aims to identify risk targets, such as high combined loan-to-value borrowers, who are likely to default. But it also finds those who are likely to stay current with payments because they tend to prioritize mortgage payments.
The software also determines how a combined loan-to-value ratio will affect a borrower's likelihood of defaulting on a loan modification.
Combined loan-to-value refers to a ratio of the first and second mortgage over the property's value — a measurement of eligibility for a refinance, home equity loan or home equity line of credit.
Veritas' analytics are driven by its database, which includes the number and types of credit relationships that a borrower has, a borrower's monthly debt payment obligations, and how the borrower will likely respond if placed under duress, such as the loss of a job. Digital Risk's property database includes age of the structure and the property's value across a multiyear period. The real estate market data includes housing price changes mapped over 6,000 ZIP codes, market conditions, what kinds of properties are selling and average time on the market.
"The local conditions of a real estate market play a large role in determining how an individual will react in terms of" staying current with a loan or defaulting," Taylor says.
A couple of tier-one banks have gone live with Veritas so far. Digital Risk didn't reveal the clients or reveal the specific data sources that fill Veritas' database, though it says it uses a variety of public and private data sources that can tell how and when people pay debts and what debts they pay. Digital Risk hopes Veritas will help it sustain the strong growth the company has enjoyed over the past few years. Earnings are about $79 million, up from $3 million in 2008, and the company has nearly doubled its head count in the past year. It hopes to double it again within the next 12 months. "I wish I could clone our best people," Taylor says.
CoreLogic's new FICO Mortgage Score Powered by CoreLogic evaluates traditional credit data along with supplemental consumer credit data contained in the CoreLogic CoreScore credit report. That data includes payments such as mobile phones, rental contracts, utilities and other basic payments.
The supplemental data in CoreScore is pulled from the CoreLogic database, which contains about 1 billion consumer transaction records covering 99% of the country. These include county, municipal and tax records, landlord/tenant data, payday lending, installment and tend-to-own information, residential properties and liens an consumer-specific bankruptcies, liens and judgments.
A CoreLogic spokesman said the software also looks at alternative credit data, economic conditions and unemployment numbers. "It used to be that the mortgage was the first thing people paid," he says. Not anymore. The accumulation of rental payments for three or four years — information not always included in traditional credit reports — is a good predictor of mortgage payments, he says.