How Fannie Plans to Use 'Trended Data' to Better Assess Risk

mayopoulos-timothy-bl041113-365.jpg

WASHINGTON — Fannie Mae is planning to add a new data element to its automated underwriting process soon that is designed to give lenders a clearer picture of how consumers pay their bills over time.

So-called "trended data" provides a 24-month snapshot of an applicant's credit card payments. It can show if borrowers are "revolvers" who make at least the minimum payment each month and carry a balance, or "transactors" who pay their credit card bill in full each month.

"Prior to trended data, there was no way to distinguish between a transactor and a revolver," according to Atlanta-based credit expert John Ulzheimer, who has written several books on consumer issues. "It is very valuable and powerful information."

For example, transactors who pay their credit card bill in full each month are considered three to five times less risky than revolvers who are current but do not pay in full. Ulzheimer expects transactors will get better pricing on their mortgages.

Trended data can also show if a balance is growing and whether monthly payments are increasing or decreasing. If consumers are choosing to pay less while balances are growing, it may signal an applicant is becoming more risky and could become delinquent before their credit score changes.

"It provides insight into each borrower's use of credit and provides a more complete picture of their eligibility for mortgage loans as well as the risks inherent in their creditworthiness," said Timothy Mayopoulos, Fannie's president and chief executive, in a recent interview.

Fannie is slated to update Desktop Underwriter in the second quarter with trended data supplied by the credit reporting bureaus TransUnion and Equifax. Lenders should be able to access trended data sometime this summer.

The introduction of trended data is not expected, however, to expand Fannie's credit criteria.

"The challenge for us is that people are not lending to the full scope of our credit box," Mayopoulos said. "We are hoping that by making this kind of credit information available to lenders, it will give them greater confidence that they can accurately assess a borrower's ability to repay a loan."

Nor is trended data expected to increase the pool of eligible mortgage borrowers until the government-sponsored enterprises update their FICO 4 credit-scoring system.

Fannie Mae and Freddie Mac are evaluating whether to adopt the latest FICO or VantageScore credit-scoring models. But the implementation of a new model is still a year or two down the road.

Ulzheimer noted that trended data does not exist unless a borrower already has access to credit. "It is not going to give you a credit history that can help you get a loan. It adds context to an existing credit report," he said.

Fannie, which first announced its decision to adopt trended data in October 2014, is not trying to oversell the impact of trended data, either.

"I won't call this a revolutionary change. I would call this an incremental improvement to credit analytics," Mayopoulos said. "We believe this will help improve risk assessments by allowing for a smarter and more thorough risk analysis of a borrower's credit history."

But the credit bureaus expect trended data to increase access to credit. They claim trended data can show that borrowers with low credit scores due to past financial problems have recovered and are capable of taking on debt.

"We have done studies that show there is a significant percentage of consumers who score better with trended data. That has positive implications for access to loans as well as for rates," said Joe Mellman, head of TransUnion's mortgage group. "Additionally, TransUnion trended data can allow millions of consumers that are 'unscoreable' with traditional credit data to be scored. This also has positive implications for access to credit."

Experian, the third credit bureau, is also expected to participate.

"We are working with Fannie to include Experian's trended data also," said Paul DeSaulniers, senior director for risk scoring and trended-data solutions at Experian.

But DeSaulniers is concerned the adoption of trended data might be compromised if Fannie is using FICO 4, which was built off of how consumers performed 10 years ago.

"Since then, there has been a monumental change in consumer behavior. Newer models like VantageScore 3.0 can score 30-35 million more consumers than these legacy products and 10 million can qualify for a mortgage," DeSaulniers said in an interview.

Charlie Wise, vice president in TransUnion's Innovative Solutions Group, said Fannie will likely use a mix of traditional credit data, trended credit data and FICO scores.

"It is also important to keep in mind that like any tool, utilizing trended credit data can follow an evolution and we're potentially just in the first step of that evolution," Wise said.

For reprint and licensing requests for this article, click here.
Mortgage technology Analytics Risk management Secondary markets Underwriting GSEs
MORE FROM NATIONAL MORTGAGE NEWS