Ignore prepayment risk at your peril.

Author:Bykhovsky, Michael
Position:Executive Essay
 
FREE EXCERPT

FOR FINANCIAL INSTITUTIONS, THE average expected cost of customers exercising their refinancing option is about 10 times greater than the average cost of their defaults. Yet most mortgage bankers focus primarily on credit risk and do not fully analyze their refinancing risk exposure, especially when originating loans. These institutions could benefit from using a tool to accurately evaluate prepayment exposure, similar to the way the FICO[R] score is used to evaluate the credit exposure of loans.

In the last 10 years, refinancings and not defaults have been at the core of nearly all published bankruptcies in the mortgage industry. It is no secret that current low interest rates are causing consumers to scramble for better mortgage rates. Are mortgage bankers accurately evaluating the whole risk picture-both default and prepayment- when considering a mortgage application? And are they focused on retaining the right customers through marketing and service? Are they pricing and hedging the loans on the books properly?

Lenders need to better forecast consumer prepayment behavior. However, evaluating prepayment risk is a complex process and many lenders do not have the resources to fully accomplish the task. What is needed is a methodology that uses the latest advances in analyzing prepayment risk without requiring additional resources in order to incorporate such methodology into the process of loan origination and management. One way of doing this is to provide a prepayment score.

Over the last few years, many in the industry have unsuccessfully attempted to provide a simple way of comparing the prepayment risk of similar loans by assigning prepayment scores to them.

Scores generated in these previous attempts would assign a single number to a loan that corresponded to its probability of prepayment. That number reflected many of the factors that drive a borrower to refinance a loan, including the interest rate incentive, the product type, the age of the loan, the burnout status, the current loan-to-value (LTV) ratio, the source of the loan and local economic conditions. While these scores were better than the previous rules-based models used in retention, and may have had a reasonable short-term predictive power to help servicers achieve some improvement in retention, they were not at all useful for hedging or relative-value analysis. Once interest rates move or time passes, these scores become irrelevant. This may not be an issue for...

To continue reading

FREE SIGN UP