Building on early identification of borrowers at risk the FICO/TowerGroup Mortgage Credit Risk Management 2009 Survey I mentioned in the last issue revealed that portfolio segmentation by delinquency sta- tus, credit score and loan product type is critical in credit risk management.
Portfolio segmentation accomplishes many things. Most importantly, it identifies which borrowers are most at risk. Servicers can then form segmentation strategies for contact and collections that optimise costs and minimise losses.
For example, borrowers most at risk need earlier or more intervention than those less so. The latter group can be managed less expensively through automated or less frequent contact. More than half of the firms surveyed use periodic credit score updates and internal scoring models to identify changes in borrowers’ credit risk status.
Some servicers are more sophisticated in their strategies than others. For example, 36% prioritise and target customers with loan products that are about to reset whereas 24% focus only on customers who have already missed payments.
Half of all servicers perform only general outreach to encourage customers to contact them to restructure their mortgages. This ’one size fits all’ approach may be easy to implement but it results in wasted resources and inadequate staff- ing to manage unpredictable inbound volumes.
These disparate approaches to customer segmentation can yield widely different collection, reinstatement and re- default results.
As servicers gain control over operational resource issues their next need is to increase resources in predictive analytics – a critical area that is lagging.
Most appear to be doing this. Half the servicers surveyed are implementing a solution and another 20% are evaluating one with an intention to implement in the next 12 months. Only 20% of servicers are not looking to implement new solutions. All these are smaller firms outside the top 25.
More than two-thirds of credit risk managers use portfolio analytics for the early identification of borrowers at risk but far fewer are able or willing to act on this information. For example, only 36% are able to target customers with products that are about to reset.
Furthermore, 23% report that their institutions focus only on customers who are already delinquent. This means the remaining servicers are aware of the size of their portfolio problems but unable to initiate a plan to do something about it.
Many are well prepared for interaction with delinquent borrowers. They use behavioural scoring and automated scripting to determine borrowers’ behaviour, estimate risk and man- age forbearance programmes.
But a minority of firms are using more innovative technology and analytics to implement best practices.
For example, 38% have implemented online collection self-service which enables borrowers to research their collection options, interact with the coll- ections department and monitor payments. These systems can cut staffing requirements at collections call centres.
Some leading servicers augment online self-service with integrated virtual agents – online email chat or live call-back features – that further encourage otherwise reluctant borrowers to contact servicers to discuss alternative payment options.
Other top firms use what is known as ’best time to call’ software to increase borrower contact rates, promote promises to pay and boost forbearance programme participation.
Servicers have long used technology to manage loan forbearance programmes but they are now adding technology for loan modification programmes, modification having occurred only infrequently before 2008.
Two-thirds use net present value software to automate required borrower eligibility guidelines for government loan modification programmes. Only 36% use decisioning technology to compare forbearance programmes. Most servicers still do this manually, which may be acceptable if experienced collectors are making decisions.
We will be reading about the US mortgage credit crisis well beyond 2010 and at least into 2012. The findings in our survey show that although the US mortgage industry has made significant progress in reforming its lending practices the credit risk assessment area still needs improvement. This revamp should start by revisiting assumptions about credit risk, product features, defaults, employee training and analytical tools.
And for these changes to make a positive difference lenders, regulators and politicians need to acknowledge the business and political failures of the past, admit that not every consumer in default can be saved and change loan collection programmes to identify and help those borrowers who can be spared the pain of foreclosure.
Craig Focardi is senior research director with TowerGroup. He can be contacted at cfocardiLS@towergroup.com