# Point of View: The Servicing Impact

Mr. Showalter is senior director of product management at LoanPerformance. This viewpoint is the second of two parts. The first part ran in the August edition of MSN.

The impact of borrower and product trends on mortgage servicing is becoming apparent. As the mix of borrowers and products becomes more diverse, the complexity of the mortgage servicing problem is growing exponentially. Besides loan volume increases, today's loan portfolios have a higher portion of service-intensive loans (e.g., ARM, option ARM, IO). Moreover, servicing a delinquent ARM is not the same kind of problem as servicing a delinquent IO, neg am or fixed-rate loan.

Furthermore, addressing the issues of a subprime borrower are far different from addressing the servicing issues of the alt-A or prime borrower. A subprime borrower is likely to be more economically unstable and, at times, ill-equipped to handle the rigors of monthly debt service, especially when compared to the more economically stable, experienced bill-paying alt-A or prime borrower.

Clearly, the complexity of servicing loans is on the rise. Unfortunately, only a portion of subprime mortgage servicers are successfully coping with this increased level of complexity. There is a growing gap (as measured by nominal loss rates) between "best" and "worst" servicers, when comparing the performance of servicers who are servicing subprime paper.

In this study we compared actual (nominal) loss rates by subprime servicers using a static pool with a two-year look-back period ending in October 2005. The variance in servicer performance exceeded 300 basis points when comparing the typical top quartile player and the typical bottom quartile player. A comparison of the "best" and "worst" servicers revealed a 600-plus basis point variance. These servicer comparisons were not balanced for product type or vintage, so they only serve as a rough estimate of loss rate performance differences across servicers. Nonetheless, the conclusion that there are substantial (and economically significant) loss rate performance differences across subprime servicers is entirely valid.

The risk-adjusted impact of servicer selection on loss rates escalates with increasing portfolio loss rates. When moving from an expected loss rate of 1% to one of 8%, the potential impact of servicer selection on portfolio losses increases from 100 basis points to over 800 basis points on a risk-adjusted basis. This means that when neutralizing the effect of collateral risk upon servicer performance, loss rate performance can vary as much as 800 basis points between "good" and "not so good" servicers if the overall subprime loss rate equals 8%.

Translating risk-adjusted index performance into value added (or subtracted) by a servicer is relatively easy. If a servicer is servicing a $1 billion portfolio that loses 1.5% ($15 million) and this servicer's loss rate performance generates an index score of 75%, then this servicer would have reduced losses by $3.75 million relative to the typical servicer in this market.

The estimate of 8% expected loss rates may seem exaggerated. Loss rates are not typically that high, even in subprime portfolios. However, in a recent analysis using the LoanPerformance risk model, such loss rates were typical of subprime loans, given today's environment and escalate dramatically given potential changes in housing price index and interest rates.

Given today's conditions of modest HPI increase and stagnant interest rates, exceed 400 basis points (4%). However, with modest changes in HPI and interest rates, loss rates on subprime mortgages could easily exceed the 8% expected loss rate. As the analysis revealed, loss rates exceeded 33% in some of these scenarios.

A decrease in HPI of 10% or 20% could drive loss rates on subprime fixed and subprime ARMs to extremely high levels. Expected loss rates for the "down 10%" scenarios are 10.61% and 18.53% for subprime fixed and subprime ARM, respectively. Moreover, the "down 20%" scenarios generate expected loss rates of 14.75% and 33.48% for subprime fixed and subprime ARMs, respectively.

The impact of interest rate increases on expected loss rates, while noticeably less, is no bargain. The interest rate scenarios "up 100 bp" show expected loss rates of 8.33% and 8.21% for subprime fixed and subprime ARM, respectively. Similarly, the interest rate scenarios "up 200 bp" show expected loss rates of 9.22% and 8.66% for subprime fixed and subprime ARM, respectively.

It should be noted that the risk model expected loss projections are servicer-neutral. That means that whatever loss rates are projected are median rates and the impact of servicer would then add or subtract basis points of losses, as appropriate. For example, if the expected loss rate of 9.22% were realized, the net effect of a "good" servicer may reduce that loss rate by 450 basis points to approximately 4.75%. Moreover, the net effect of a "not so good" servicer would increase that expected loss rate by an estimated 450 basis points to over 13%.

While the expected loss rates generated by the LoanPerformance risk model relate a 30-year window, most mortgage portfolios incur over 85% of their losses within their first five years. This means that the impact of declines in HPI or increases in interest rates on the economics of the mortgage industry (and mortgage industry securities) may be far closer than 30 years.

Let's assume that the most likely loss rate scenario lies somewhere between the scenarios based upon today's conditions (modest HPI increase and stagnant interest rates) and the scenarios involving a decline in HPI and/or a further increase in interest rates (tomorrow's conditions). Given these assumptions, a rough estimate of future Subprime loss rates would most likely be in excess of 5%, possibly approaching 10%.

Ineffective (below median) servicing is expensive, with an impact in the billions of dollars. The impact of below median servicing for the Subprime mortgage industry with loss rates of approximately 8% shows an impact of $10 Million per $ Billion (or $10 Billion on a $1 Trillion Subprime portfolio). Lower loss rates, while reducing the dollar impact, still create a multi-billion dollar problem. And, higher loss rates create an impact well in excess of $10 Billion.

Conclusion