There is always risk in lending and recently the risks have become harder to predict. The solution to this issue is to become better prepared to adapt credit policy and business strategies quickly and increase the sophistication of processes for determining who is an acceptable lending risk. The lending pendulum has swung from too permissive to very restrictive over the past couple of years. In 2010 loan growth will be necessary to move toward sustained economic recovery. Banks that embrace change and engage in more sophisticated credit risk policy development processes will help swing the pendulum back toward middle ground.
In 2009 the Federal Deposit Insurance Corp. recorded the largest annual lending decline since the 1940s ($587 billion dollars or 7.5%). Tightening credit is an understandable reaction to rising unemployment, loan default rates, and business failures. Banks are increasing reserve ratios to prepare for the unexpected, but placing score cutoffs high enough above what is believed to be a relatively safe margin of error is not going to bring balance to the economy or profits to the banks. All this accomplishes is a reduction in loan volume and probably eliminates as many good credit candidates as bad.
Meeting current lending demands requires rethinking tried and true methodologies and employing more efficient credit risk policy development processes. They need to be faster than the typical 12-18 month cycle most institutions currently undergo and ideally become a continual process rather than a project that is undertaken every few years. This change in process involves employing continuous development cycles so as consumer behavior changes, it can be spotted immediately and policy can be adapted to respond to that change.
While not every attribute or scoring model that is created under this new development concept will be implemented, a more fluid process provides opportunities to fine tune policy and manage change without being overly restrictive with lending parameters. With consistent monitoring, if a particular scoring model produces better results it can be implemented immediately. On the other hand, if the new model is not effective no time is lost as the development of multiple models is done in parallel and successive ideas can then be introduced. Processes where new credit risk models and attributes are constantly developed and tested in a non-production environment—using the bank’s own performance data for evaluation—will help institutions be more competitive without taking unnecessary chances. It isn’t about automating the development cycle just to become faster. It is a transformation of all aspects of the process (including the automation of certain steps) with the end goal of producing more accurate and timely policy. To speed up the development cycle, inefficiencies and translation issues must be eliminated from traditional processes. Eliminating the redundancy of multiple teams and multiple systems allows banks to quickly evaluate consumer behavior against new ideas and determine whether or not proposed policy will be effective. This strategy provides constant feedback of the most up-to-date information making the development process more proactive.
Golden borrowers have tarnish
This change in process is necessary because consumer behavior is now more dynamic and less predictable. Indicators that someone was almost guaranteed to be a golden borrower or a lending disaster have become much harder to recognize. For example, someone with 20 years of excellent credit history may have a foreclosure on their credit report because they lost their job during the most recent recession and fell behind on their mortgage. That person may be employed again, maintaining all of their current payments and a good credit risk even though they still have the foreclosure and damaged credit score hanging over their head.
Keeping up with payment-hierarchy shifts adds to the complexity of deciding who is a good credit risk and who isn’t. A recent study by TransUnion not only confirms the unprecedented trend of consumers staying current with their credit card bills and falling behind on their mortgage payments, but shows that this number is rising. Fair Isaac Corp. statistics establish that while this trend used to affect mostly subprime borrowers, it is beginning to include more consumers with high end FICO scores and second homes, the so-called “golden borrowers.” In addition, there are more strategic defaults on homes whose value has dropped beneath the amount owed on the property.
This dynamic and complex market requires that banks look at data differently and more deeply and derive greater significance from that data review. A variety of data sources (external, internal, traditional and non-traditional) should be utilized for testing new policy and determining who will be a good credit risk. Banks can also extract a great deal of value by taking into consideration all of the credit products owned and phases of the consumer lifecycle to drive decisions about their current customers (and more importantly avoid making costly ones). This applies to a bank’s best customers in addition to those that may already be headed for trouble.
If a gold card carrying customer hasn’t used her card in years, cancelling it could be a bad move especially if she owns several other high value accounts with the institution and is a good credit customer. Her credit score could be negatively impacted by the cancellation, particularly if it is her oldest line of credit. Instead, offer an upgrade, an incentive to use the card, or at a minimum ask before taking action to maintain a positive customer experience. Many institutions have some level of up-sell/down-sell logic like this in place, but it is often linked to individual products instead of a holistic view of the customer.
Needed: real-time score monitoring
On the “headed for trouble” side, the collections sector has become extremely demanding and requires a level of analysis and real-time monitoring that wasn’t needed in the past. The complex analytics used in originations can be applied across the entire credit lifecycle to provide more effective servicing, portfolio management, and collections. Lenders can begin to look at trends across consumers’ accounts and develop a new model that can be consistently applied across lines of business and phases of the credit lifecycle. If a customer has a FICO score of 790 when they were approved for a home equity line, but in portfolio review they are at a 760, that could be significant. The bank not only needs to know why, but they need to look at other accounts held by the customer and proactively consult with them before those accounts go bad. The reverse is also true. If a customer had a score of 650 at origination and is now at a 700, there may be an opportunity to deepen that relationship with additional product offers. Implementing rules that are based on in-house data sources provides current and very relevant information to protect the bank as well as promote growth.
The change in traditional classifications of people’s behavior deems many institutions’ existing predictive capabilities extinct when it comes to propensity to repay loans. No new standard behavior pattern has emerged but we do know that jobs are much more transitory, traditional credit scores are less reliable, and high salaries today don’t necessarily mean lenders will get paid tomorrow. The current lending terrain requires a holistic view of the customer, more detailed analysis, a variety of data sources, and a more intelligent and far-reaching evaluation of consumer behavior and performance than ever before. Credit risk managers know that to keep up in today’s economy they need to be faster at developing risk policy without sacrificing quality. While individual institutions will need to determine their own risk tolerance, getting through the next phase of financial recovery securely will involve a better assessment of the complete relationships consumers have with their lender and the flexibility to adapt promptly to the unexpected and the next era of lending trends.
Topics: Retail Banking,