Having collectively made billions in consumer loans that have become delinquent at record rates since 2008, lenders now understand that relying on the credit score as a predictor of future consumer payment behavior is a bad business practice.
While the credit score had worked well in ranking consumer’s risk level, what lenders overlooked is that it is a tool that weights past payment behavior, data found in consumer credit reports and other variables, and assigns a score to the analyzed data. The general thought was the higher the score, the more creditworthy the borrower.
That measure of creditworthiness, however, does not provide any indication of future behavior patterns, such as how the borrower will behave during a recession, because many credit scores don’t change until a consumer significantly alters their payment behavior.
A consumer with a 720 credit score, for example, will retain that score until becoming delinquent. By then, it can be too late for lenders to take action that can minimize risk on the account going forward, such as loan restructuring, because the consumer is buried under a mountain of debt.
“There is no stable relationship between the credit score and a consumer’s ability to pay during a recession,” says Michael Zeltkevi, partner, Finance & Risk Practice for Oliver Wyman. “There is some static thinking around credit scores that the finance industry needs to address.”
Desperate to avoid making the same mistakes going forward, lenders are clamoring for new tools that can help them predict consumer payment behavior as the economy begins to show signs of weakness, as well as signs of strength as the economy moves in to recovery after a recession.
One solution is to build predictive models using macro-economic data to anticipate how underlying economic conditions can change consumer payment behavior. A consumer who has built a high credit score during good economic times may be more prone to economic stresses when a recession occurs than a consumer with a moderate credit score who battled economic stresses such as unemployment during good economic times.
With this in mind, FICO has teamed with Moody’s Analytics and Equifax to create its Economic Impact Index, which uses macro-economic data to create potential scenarios that can influence consumer payment behavior during changing economic times. By running portfolios through the scenarios, lenders arrive at conclusions that help them readjust credit scores as economic conditions change.
A consumer with a 720 score, for example, may behave more like a consumer with 680 score during a recession.
“From a credit score perspective, present behavior can often look like past behavior, which is why lenders want tools that can look forward to gauge risk levels and predict how consumer behavior will change with the economy,” says Careen Foster, Director of Scores and Product Management at FICO. “Using macro-economic data to build economic scenarios can provide quantifiable measures that lenders can take action against.”
The Economic Impact Index (EII), which debuted in April, uses credit bureau data from Equifax and analytic models from Moody’s that forecast potential economic scenarios. Lenders can chose up to six economic scenarios based on such current economic data as gross domestic product, unemployment, housing values and consumer delinquency rates. Lenders can supplement the models with their own consumer behavior data.
The models can be run against national or regional portfolios and lenders can even build models using regional data.
“In a recession, as well as a rebounding economy, there are other factors that determine consumer behavior beyond what have been used to formulate credit scores,” says Tony Hughes, senior director, credit analytics group for Moody’s Analytics. “Bringing macro economic drivers into the equation can help lenders determine the implications on consumer behavior and credit worthiness, which is reflected in an adjusted credit score.”
Consumers who look to be at risk of becoming delinquent during an expected downturn can be proactively reached out to for loan restructuring. Having that kind of foresight can be a real plus to lenders.
Many mortgage lenders are anxiously awaiting the next wave of defaults, which risk management experts predict will be good risks that have been out of work for a year or longer and burned through a large portion of their savings to make ends meet.
If a tool such as the EII had been in place even a year ago, mortgage lenders may have been able to identify potential defaults among consumers with good credit records and reached out to them sooner with loan restructuring programs, according to risk management experts.
On the flipside, lenders can use the EII to identify opportunities for growth as the recession eases.
“A consumer who builds a good credit record during a recession has a more meaningful credit score than a consumer that built their score Christine Pratt, senior analyst for Aite Group. “Banks need to do a better job of managing risk and identifying growth opportunities going forward.”
Still, some risk management experts harbor doubts that integrating macro-economic data and projected economic performance into a credit score will prove the antidote for lenders that rely too heavily on the credit score as an indicator of consumer payment behavior.
“Predicting the macro-economy is difficult and something that not many do well,” says Peter Carroll, partner, Retail Banking Practice for Oliver Wyman. “To think macro-economic data can be baked into credit scores to make the score a better predictor of default probability is not necessarily a credible statement.”
Carroll argues that macro-economic scenarios need to be viewed separately from the credit score. “Data used to build macro-economic models can be biased,” he says.
To illustrate his point, Carroll points to the 2006-2007 timeframe when mortgage lenders thought that debt to income ratios, a macro-economic indicator, did not to add any predictive power to the lending decision because home values were rising. When home prices fell, mortgage lenders realized they had made a serious mistake.
“Interpreting macro-economic factors requires some judgment and trying to automate human judgment by integrating it into a credit score, rather than looking at the credit score and economic predictors separately, may be trying to have the credit score do too much.”
Some risk management experts counter that macro-economic data is more detailed and timely than ever thanks to changes in how fast the federal government publishes economic data, which when added to the credit score formula, can make the credit score more relevant.
“Fifteen years ago, lenders did not have the same depth of macro-economic data to bring to bear on managing risk that they do today and they were using relying on fresh reports with data that is two years old,” says Aite’s Pratt.
FICO’s Foster defends the EII by pointing out it was extensively tested through retro analysis using decades of data leading up the recession. Three things learned were:
1. Consumers that built good credit scores during good economic times will experience the greatest change in payment behavior when recession hits.
2. Consumers with lower scores built during good economic times don’t show much change in payment behavior.
3. Consumers with high scores and access to a lot of credit in good economic times, but that don’t readily use that credit, will access it heavily once they or a spouse becomes unemployed.
“Behavior patterns influence credit scores and being able to predict changes in behavior patterns can help lenders better manage current risk and identify growth opportunities without increasing their risk exposure too much,” says Foster.
If nothing else, bringing more macro-economic data to bear into the credit granting decision, whether it be integrated into a credit score or viewed in addition to the credit score, is expected to help lenders better manage their future risk.
“This is the time overhaul analytic risk management techniques so we don’t repeat the same mistakes that lead to the current crisis in the first place,” says Moody’s Hughes.