Printer Friendly

The art of auditing ARM portfolios.

There are several ways to find adjustment errors in an ARM servicing portfolio. Here's some pros and cons of different ways to detect the problems.

Many mortgage servicers today are concerned about the accuracy of the interest rates used to adjust their adjustable-rate loans. In fact, statistics gathered by various regulatory bodies, while somewhat inconclusive, suggest that almost one-third of all ARMs have some sort of adjustment problem.

Most mortgage servicers--especially those with portfolios comprised of many different types of ARMs and/or large portfolios of acquired servicing--will readily admit that at least some of their ARMs have not been adjusted properly in the past. With penalties for incorrect adjustments substantial (see "ARMed and Dangerous," Mortgage Banking, June 1991, and "Adjusting to the Perils of ARMs," Mortgage Banking, August 1991), the servicing industry is well-aware that it needs to take

The upshot of this concern is an attempt by mortgage servicers to determine the best way to audit their ARM portfolios so they may uncover these troublesome misadjustments. Many have turned to the Mortgage Bankers Association of America's (MBA) Adjustable Rate Mortgage Task Force report for guidance. The task force report, published in August 1991, lists a number of common ARM errors and identifies the sole method required to fix these errors as the "re-amortization of the mortgage loan." Knowing a re-amortization process of some type is necessary to supply the data to readjust the ARM; many servicers have used a random re-amortization procedure to find problem loans as well. The process of random re-amortization consists of selecting a sample of loans at random and re-amortizing them to determine if the interest rate charged has been correct. Generally a 10 percent sample is used.

However, this "find and fix" approach has, in many cases, proven to be less than ideal. An increasing number of servicers are discovering that there's a faster, more cost-effective way to audit ARM portfolios--by first performing an automated analysis of the entire portfolio to single out troubled loans, then re-amortizing only the loans that need it.

Random re-amortization

There are two major drawbacks to the random re-amortization approach. First of all, it is very time-consuming. Second, random audits are seldom conclusive. For example, if random re-amortization determines that 10 percent of all ARMs in a portfolio have been incorrectly adjusted and that the incorrect adjustments have occurred on a somewhat random basis, it is impossible to find every misadjusted loan without auditing every loan in the portfolio. Even if there is a pattern of error, conclusive identification of every problem occurrence is virtually impossible.

The re-amortization process usually consists of having a trained mortgage professional operate a PC-resident program that calculates the principal and interest spread, as well as the outstanding balance of a loan on a monthly basis. The auditor begins with source documentation (e.g., note, ARM rider, disclosure document) to "set-up" the mortgage loan independent of the initial rates, margins and so forth that were placed on the servicer's system. Every principal and interest payment collected from the borrower is then reapplied

using correct index values plus the correct margin. The auditor determines what the principal balance and interest rate should be, then compares those results to the information provided by the servicer's computer system.

A properly calculated re-amortization provides critical information concerning the following: whether the loan was set up correctly; whether the proper index values, rounding techniques and the like were used to calculate adjustments; whether the computer system functioned correctly; whether the computer system was used correctly by the servicing staff; and whether the disclosures were made properly at the time of origination.

Even more important, re-amortization provides output in a format that can also calculate necessary corrections where refunds and/or new disclosures might be required.

Obviously, re-amortization provides all the necessary data required to fix a misadjusted ARM. But this does not automatically make it the best vehicle to uncover adjustment errors. Yes, sometimes it provides an excellent way to find problem loans. But more often than not, random re-amortization leaves unidentified misadjusted ARMs lurking in a portfolio.

Computer analysis of the entire portfolio: close scrutiny

To eliminate the problem of inconclusive random audits, Hanover Capital Partners Ltd, Chicago, as well as several large mortgage servicers, has developed highly efficient, automated techniques to determine the accuracy of all ARMs in a portfolio. Through a computer analysis, the servicer's database tapes are examined, using original set-up information (i.e., initial interest rate, margin, servicing fee, initial payment date, index to be used, rounding, caps and so forth). The audit program then goes forward over the life of the loan to re-calculate interest rates with an independently derived index value. The results of the computer analysis are then compared to the present interest rate and principal balance found on the servicer's computer tape, at which point any discrepancies are clearly displayed.

The automated analysis identifies every ARM that is incorrectly adjusted today, and it flags loans that have principal balances that suggest that, although correct today, they might have been incorrectly adjusted in the past. All loans categorized as "incorrect" will need to be re-amortized, so that corrective action can be taken. And, most important, all loans identified as "correct"--which is the majority of loans in most shops--need not be the subject of any further analysis by the servicer.

The benefits of the automated anaylsis are illustrated in Table 1. The table shows two hypothetical ARM-loan portfolios, the first of which has a 5 percent rate of error and the second of which has a 33 percent rate of error. The table shows that servicers using automated analyses to find their ARM errors need only re-amortize those errors identified by the computer program, while servicers using random re-amortization techniques must completely re-amortize their entire portfolios to find every ARM adjustment error. The cleaner a servicer's ARM portfolio is at the time of the audit, the more dramatic are the time savings realized by the automated analysis.

However, there is a potential pitfall to the automated analysis, because it relies on the set-up data that is presently on the lender's database. If, for example, the margin is incorrectly stated, the automated analysis has no way of knowing. As a result, the analysis will not point out the fact that an incorrectly set-up loan is in error.

Nevertheless, the time and cost savings from a computerized analysis compared to a random sample are so dramatic that, in many cases, it is to a lender's advantage to actually go in and look at source documents on each and every ARM to reconfirm the set-up information that is on the system. Table 2 shows a comparison of the elapsed time required to resetup an ARM loan portfolio, run an automated analysis and re-amortize the problem loans identified by the automated analysis, compared to a random re-amortization of the same portfolio. Again, the savings in elapsed time, as well as in dollars, are substantial.

Both Table 1 and Table 2 assume a rather conservative resource commitment to setting up and amortizing mortgage loans. In actual practice, when outside auditing firms are used, the elapsed time to accomplish the tasks is substantially shorter.

Too much streamlining can be hazardous

A common hazard in attempting to audit ARMs too quickly is that some servicers, interested in abbreviating the auditing process, merely recalculate the most recent interest rate change. To do this, they simply use the appropriate index value added to the margin listed in the trial balance, and recalculate the interest rate. However, the margin listed in the trial balance may not be correct, the index listed in the trial balance may be incorrect, the preceding year's interest rate might have been incorrect (causing potential cap problems), and factors such as the rounding technique and look-back periods may be incorrect as well.

In short, quick-and-dirty audits don't find set-up problems and don't find problems that have occurred in preceding adjustment periods.

Different portfolios require different approaches

Which type of ARM audit is best for your portfolio? The definitive answer to that question usually requires a portfolio-specific analysis. However, some general parameters apply.

* Most lenders and servicers do not

have reason to believe that there is

anything wrong with their ARM

portfolios and are almost always

best served, regardless of portfolio

size, by an automated analysis of

their ARM portfolio. The relatively

small number of

problems thus

identified in the audit can

be easily and quickly

re-amortized. * If you have a very

small ARM portfolio

(500 loans or less),

that you have reason

to believe contains a

considerable number

of adjustment

problems, you are almost

always best served

by completely re-amortizing

the entire

portfolio going back

to the original source

documents. An

automated analysis would not

substantially shorten the process. * If you have a larger ARM portfolio

with a lot of problems, but that has

correct set-up information for all of

the loans, then an automated

analysis of your loan database is the best

choice, because the analysis can be

performed in a matter of weeks for

100 percent of the loans rather than

in a matter of months, as would be

the case for a random re-amortization

of the sample. * If you have a larger problem

portfolio and you think the set-up

information was incorrect, you are usually

best served by setting up the loans

again and then performing a

computer audit. * If your organization has a strong

commitment to the random re-amortization

technique (if you definitely

do not want to perform an

automated analysis), the best approach is to

initially target loan types with

known or suspected problems.

Then work with the audit team to

try to identify patterns of errors,

and expand the sample along those

patterns. While not a conclusive

way of identifying every problem,

this technique can minimize the

number of correct loans re-amortized

while attempting to find the

incorrect ones.

When should you audit your ARM portfolio?

Certainly you should audit any ARM portfolio that you are planning to acquire. To the extent that the audit can be performed prior to merging the data onto your database, it will help in isolating any misadjusted ARMs, as well as identifying, for the seller of the servicing, the extent of the problems. This identification provides the basis upon which to establish a mutually agreeable strategy for rectifying the problem loans.

Apart from portfolio acquisition, the liability to a servicer for misadjusted ARMs should be incentive enough to warrant a complete audit of the existing ARM portfolio at least once. After that, an ongoing audit process should become part of every servicer's plans. The costs involved depend on the size of servicing portfolio. The cost of performing an automated audit is generally less than one dollar per loan for a medium-size servicing portfolio, and substantially less than one dollar per loan for a larger servicing portfolio. So even though costs are involved, you will want to be sure to audit your ARM portfolio before an examiner knocks on your door asking to do the same.

Who should perform the audit?

A basic tenet of any audit is that the examination should be handled by a detached party--one that was not involved in the process of performing the original calculations. Nevertheless, it is very common in many servicing organizations for the servicer to audit its own portfolio, usually because its staff is the only one in-house capable of performing the required calculations.

The obvious problem in self-audits is that these "auditors" will frequently be directed by the same individuals who oversaw the initial calculation process. In most servicing organizations, a small number of individuals such as servicing managers, assistant servicing managers and certain supervisors, are ultimately responsible for determining how each and every mortgage loan in the organization is serviced. An honest error made by one of these people--acting responsibly and deliber- ately, but not infallibly--can result in hundreds or even thousands of loans being serviced incorrectly. Any ARM audit is ultimately an audit not only of the occasional errors made by an inattentive clerical person, and of computer systems and their usage, but also of instructions provided by the supervisory staff. For that reason, an organization is usually best served by having large audits of its own portfolio conducted by outsiders.

Another good reason to use a third-party auditor is simply the amount of work involved. Many servicers find themselves running with limited staff today. They cannot take the time or attention required to supervise an auditing process, much less to part with a sufficient number of properly trained staff to audit the loan portfolio in a continuous and diligent manner. Fragmented auditing processes (those performed as time permits) frequently cause uneven results and fail to identify underlying systematic reasons for error. The time crunch is especially apparent when a servicer is taking on a large portfolio of acquired servicing. The best time to audit an acquired portfolio--at acquisition--is the same time that all experienced staff are needed at their regular jobs to ensure a smooth transition.

Choosing an external auditor

External auditors should be capable of staffing your audit with experienced people knowledgeable in the fine points of adjusting ARM loans. The firm should have a lengthy experience in performing ARM audits. And it should be able to demonstrate the accuracy of the techniques it employs to adjust ARMs.

Most important, however, are the tools that your ARM auditor possesses. A good ARM auditing firm should be sufficiently automated to allow quick identification of problem loans, and should have the tools required to re-amortize the loans identified as problems.

As awareness increases, misadjusted ARMs are sure to become even more of an issue in the public's eyes, and rightfully so. No one wants to be overcharged on their mortgage loan, (even though they don't mind being undercharged).

Now that automated audit programs can quickly find problem ARM loans, servicers don't need to think of ARM audits as a time-consuming, attention-demanding process that ties up both staff and management. The means exist to find problem ARMs easily before being forced to do so by pressure from regulators or consumer law suits.

There are companies out there that specialize in assisting servicers with problem ARM portfolios. It behooves the servicer to take advantage of them. [Table 1 and 2 Omitted]

George J. Ostendorf is a managing director of Hanover Capital Partners Ltd., a financial services firm specializing in mortgage finance. Hanover offers an ARM audit service and has offices in Chicago, New York and Boston.
COPYRIGHT 1992 Mortgage Bankers Association of America
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 1992 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Title Annotation:adjustable rate mortgages
Author:Ostendorf, George J.
Publication:Mortgage Banking
Date:Feb 1, 1992
Previous Article:Practical REO management.
Next Article:A roadmap for HECMs.

Related Articles
Chalking up some gains.
An exotic species.
ARM funds for safety and yield.
Armed for competition.
Managing the risk of ARM errors.
Government study finds rampant errors in ARMs.
Government study concludes 50%-60% of all adjustable rate mortgages contain errors.
A COFI break.
Don't be alarmed if you have an ARM: homeowners with adjustable rate mortgages should carefully reassess their situation before refinancing.

Terms of use | Copyright © 2016 Farlex, Inc. | Feedback | For webmasters