Point of View: Apples to Apples: With Billions at Stake
Mr. Showalter is the senior director of product management at LoanPerformance. This is the first of two parts.
In the not too distant past, the selection of a mortgage servicer was a routine, mundane event. Portfolio delinquency rates were low, defaults and losses were even lower, especially relative to outstanding principal. Much of a servicer's job was customer service, not collections. Processing payments, sending out coupon books and answering the occasional question (What is my outstanding balance?) were the servicer's key tasks.
Things are changing. Rapidly. As the mortgage industry ventures beyond the secure confines of prime lending and 30-year fixed-rate loans, the performance of mortgage portfolios is becoming more servicer dependent. Within the subprime segment alone, the projected impact of a "good" vs. "not so good" mortgage servicer has a multibillion-dollar consequence. The "not so good" servicer generates higher loss rates, sometimes considerably higher.
The selection of a mortgage servicer has become a pivotal decision. As one mortgage executive remarked, "While a 'good' servicer cannot bail you out of a 'bad' deal, a 'bad' servicer can kill a 'good' deal." If you are a capital market player who securitizes loans, if you are a servicer, or if you are an investor, this whitepaper is likely to be interesting reading. Each of you is intimately affected by the choice of servicer and its performance, especially if you are invested in nonprime loans or one of the new (not fixed rate) mortgage products that is likely to be quite interest rate sensitive.
As we will discuss in more detail, mortgage servicers have begun to vary greatly in their performance, especially loss-related performance. However, comparing mortgage servicers is quite challenging. In this article, we will introduce a method of comparing servicers that is very quantitative and granular, especially relative to the methods used by the typical rating agency. This approach has several strong points, one of which is its ability to conduct a risk-adjusted comparison of performance across servicers.
What is a risk-adjusted performance comparison?
A risk-adjusted comparison enables an analyst to address the fundamental issue: is this servicer performing credibly, given the riskiness of the loans it is servicing? Risk adjustment eliminates differences in the riskiness of collateral as a driver of the differences in loss rates between one servicer and another. Typically, a servicer with low-risk loans will have "low" loss rates and a servicer with high-risk loans will have "high" loss rates. The "superior" loss rate performance of the servicer who is servicing the low-risk loans is often due to the decreased likelihood of low risk loans to generate a loss and is not necessarily due to the increased skills of this servicer. A risk-adjusted comparison enables an analyst to determine whether or not low loss rates are being driven by the low-risk nature of the loans or the superior skills of the servicer.
Moreover, in a risk-adjusted paradigm it is possible for the servicer with the higher loss rates to be rated as a better servicer, because this servicer is mitigating losses at a far greater rate than is typical of a servicer who is dealing with loans of comparable risk. For example, let us suppose that a servicer who is servicing low-risk loans is generating a 1% loss rate and that another servicer, who is servicing high-risk loans, is generating a 2% loss rate.
Now then, a simple comparison would identify the 1% loss rate servicer as the "better" servicer because its loss rates were much lower (1% vs. 2%). However, in a risk-adjusted world, this servicer may be doing quite poorly. Let's suppose that servicers who are servicing similar risk collateral are actually generating 0.5% loss rates, not 1% rates. Our "good" servicer would actually be doing far more poorly than expected on a risk-adjusted basis.
In turn, the servicer who is generating a 2% loss rate may be doing quite well. Servicers who are servicing similar risk loans may be generating 3% loss rates, on average. Hence, the "less accomplished" servicer may actually be quite accomplished. It is doing 100 basis points better than expected on a risk-adjusted basis.
Performance Differences Across Servicers Are Projected to Increase
In the subprime market, risk-adjusted loss rate performance between "good" and "not so good" servicing is projected to range from 200-800 basis points. If the subprime industry enjoys low loss rates, the loss rate basis point spread between "good" and "not so good" will be lower (e.g., 300-400 basis points). Higher subprime loss rates will foster a much broader difference between "good" and "not so good" (e.g., 500-800 basis points). Depending upon the scenario, "not so good" servicers are projected to generate an additional $10 billion or more in subprime risk-adjusted losses. That means that servicer deficiencies will cause the subprime industry to lose an estimated $10 Billion more than it otherwise would have lost, had "not so good" servicers performed at today's median performance levels.
Given their service/process-intensive structure, performance differences across servicers are driven by infrastructure issues. Those with exceptional risk-adjusted performance have superb infrastructure. Those will poor risk-adjusted performance have a weaker infrastructure. Moreover, infrastructure-driven performance differences are resistant to change. Infrastructure weaknesses are hard to correct. They take substantial time, focus, resources and skill and involve changes to the business's strategy, structure, processes and people.
The pace of improving servicer infrastructure can be accelerated if the servicer evolves from a "collector is the solution" business model to one that is much more technology and system dominant, or more aptly stated, the "system is the solution" business model. However, migrating to the "system is the solution" business model requires far more than buying an upgrade to the existing systems platform. That platform must embody a potent systems-directed process, where the directions are based upon empirically proven best practices, which often involve investment in predictive technologies at key points in the process.