September 18, 2002

Mr. Jonathan G. Katz
U.S. Securities and Exchange Commission
450 Fifth Street, NW
Washington, DC 20549-0609

File No.: S7-16-02

Dear Mr. Katz:

The Mortgage Bankers Association of America appreciates the opportunity to comment on the Notice of proposed rulemaking, Disclosure in Management's Discussion and Analysis about the Application of Critical Accounting Policies (the NPR), released by the Securities and Exchange Commission (the Commission) earlier this year. The NPR would require additional disclosures in a new section of Management's Discussion & Analysis (MD&A) regarding: (1) accounting estimates a company makes in applying its accounting policies and (2) the initial adoption by a company of an accounting policy with a material impact on its financial presentation. We have studied the NPR and would like to raise some questions with you regarding the implications of the NPR to disclosures of mortgage servicing rights.

The MBA is the national association representing the real estate finance industry. The association works to ensure the continued strength of the nation's residential and commercial real estate markets; to expand homeownership prospects through increased affordability; and to extend access to affordable housing to all Americans. MBA promotes fair and ethical lending practices and fosters excellence and technical know-how among real estate finance professionals through a wide range of educational programs and technical publications. Its membership of approximately 2,800 companies includes all elements of real estate finance: mortgage companies, mortgage brokers, commercial banks, thrifts, life insurance companies and others in the mortgage lending field. MBA members currently originate over 50% of the residential mortgage loans, and over 20% of the commercial and multifamily loans, in this country every year.

MBA Position

MBA wholeheartedly supports the stated objective of the NPR which is to improve the transparency of companies' financial disclosures and, thereby, to improve investor confidence and promote a higher level of efficiency in the capital markets. Furthermore, we agree that greater corporate disclosure of corporate accounting policies would be helpful to investors in understanding the financial statement impact of critical accounting estimates and recently adopted accounting policies. However, we believe the degree to which any new disclosure will be beneficial to investors will depend on the clarity and consistency of disclosure among companies with similar policies.

In this regard, we are concerned there could be wide diversity in mortgage banking companies' disclosures based on the questions our members have raised about the NPR. In studying the NPR, our members focused primarily on how the proposed disclosure of accounting estimates could affect disclosure of their mortgage servicing rights (MSR) because MSR generally constitute one of their most significant and volatile assets and because their recorded values are based on estimated, projected cash flows. We would like to request your consideration of some general questions that our members have raised about the NPR as well as our responses to two questions in the NPR.

Questions Generated by Proposed Disclosures of Accounting Estimates

The NPR would require disclosure of an accounting estimate if it is deemed a "critical" accounting estimate, as determined by a "yes" response to both of the following questions:

  1. Did the accounting estimate require us to make assumptions about matters that were highly uncertain at the time the accounting estimate was made?

  2. Would different estimates that we reasonably could have used in the current period, or changes in the accounting estimate that are reasonably likely to occur from period to period, have a material impact on the presentation of our financial condition, changes in financial condition or results of operations?

If an accounting estimate meets both of these conditions, it is subject to a hierarchy of disclosure requirements, including disclosure of the methodology and assumptions that are about highly uncertain matters and other material assumptions underlying the estimate. Our members have questioned the circumstances in which MSR would meet the above conditions and the types of disclosures that they might be required to make as indicated by the following questions.

In explaining our questions, we indirectly offer our reactions, within the context of making disclosures about MSR, to some questions raised in the NPR.

MBA Question 1: Are MSR estimated values "critical" accounting estimates?

[NPR Question: Would it be difficult for a company to discern which of its accounting estimates require assumptions about highly uncertain matters?]

Under FASB Statement No. 140, mortgage servicing rights are required to be valued for financial reporting purposes at the aggregate lower of cost or market value of the identified "risk strata" within a portfolio. Moreover, at any given time, the market value of a stratum or portfolio of MSR could represent its market value as determined by reference to "quoted market prices" or as determined "based on the best information available in the circumstances" (See par. 69 of Statement No. 140). Mortgage banking companies utilize highly sophisticated cash flow valuation models to estimate the market values of their MSR in all environments although the values are most easily substantiated in environments in which MSR are actively traded.

If it is presumed that MSR whose recorded values are based on valuation models are subject to evaluation as critical accounting estimates, it is unclear whether they would be considered critical under the NPR. According to the NPR, an estimate is a critical estimate if the following question generates a "yes" response from management:

Did the accounting estimate require us to make assumptions about matters that were highly uncertain at the time the accounting estimate was made?

In clarifying this condition, the NPR states: "...a matter involves a high degree of uncertainty if it is dependent on events remote in time that may or may not occur, or it is not capable of being readily calculated from generally accepted methodologies or derived with some degree of precision from available data."

It is widely acknowledged that the values of MSR are very sensitive to changes in mortgage market interest rates. However, the mortgage banking industry has developed means for reliably estimating MSR values at a given time. Our members agree that most MSR valuation models in use today (using software provided by a few third party vendors) rely on "generally accepted methodologies" and that the values derived from these models are "derived with some degree of precision from available data." Consequently, while a mortgage company's estimated MSR values would likely meet condition two for disclosure (because the estimated value could change from period to period with a material impact on the company's financial statements), the company could take the reasonable position that the portfolio does not meet the first condition. Moreover, since both conditions must be met in order for an estimate to be deemed a critical estimate and subject to reporting, the company could assume that its estimated MSR values are not subject to disclosure as critical accounting estimates.

We believe a defensible argument could be made that MSR portfolios are not subject to disclosure as "critical" accounting estimates under the NPR. However, if it is assumed they are subject to disclosure, the NPR raises the following additional questions.

MBA Question 2: How Should Distinctions Be Drawn Between Estimates and Assumptions?

[NPR Questions: How many accounting estimates would a company typically identify as critical accounting estimates under the proposed definition?

Should we expand the definition to include MD&A disclosure of volatile accounting estimates that use complex methodologies but do not involve significant management judgment? Should we do so only when the underlying assumptions or methodologies of those estimates are not commonly used and therefore not understood by investors?]

In reading the NPR, our members wondered how they should distinguish between estimates and assumptions within the context of disclosing estimated MSR values. For example, should the estimated fair values that are incorporated into a company's recorded MSR value be subject to evaluation as critical accounting estimates, or should the estimates underlying the estimated MSR values1 themselves be subject to evaluation as critical estimates?

If a company's estimated MSR values are subject to disclosure as critical estimates, the company presumably would disclose the methodology used to determine the MSR values (e.g. type of valuation model used) as well as the material assumptions that were incorporated into the values (e.g. estimated prepayment speeds, discount rates, cost to service, etc.). In this case, the methodology and assumptions underlying the "assumptions" (e.g. the prepayment model used and the assumptions embedded in the model) may or may not be within the scope of the proposed disclosure requirements.

On the other hand, others have questioned whether the estimates underlying a company's recorded MSR value should be evaluated as critical estimates. Once the critical estimates are identified (e.g. estimated prepayment speeds), the company would disclose the methodology used to determine the estimates (e.g. type of prepayment model used) as well as the material assumptions that were incorporated into the model (e.g. current and projected interest rates and other factors). This interpretation would involve more in-depth disclosure than, perhaps, an approach whereby the recorded MSR values are regarded as the critical estimates.

We believe mortgage companies' disclosures could be significantly different depending upon whether they believe the estimated values of the MSR, or the estimates underlying those estimated values, are critical accounting estimates under the NPR.

We also are concerned about how to define "volatile accounting estimates that use complex methodologies but do not involve significant management judgment" and "not commonly used and therefore not understood by investors." What is considered to be significant management judgment? Once a methodology is defined as not commonly used, what level of disclosure should there be to provide greater understanding by investors? Generally, we are concerned that disclosures that are too in-depth could prove to be difficult for mortgage companies to provide in straightforward "layman's language" and quite difficult for the majority of readers to understand.

MBA Question 3: What Quantitative Information Should Mortgage Bankers Disclose?

[NPR Questions: Are there some types of critical accounting estimates or some circumstances where the proposed disclosure relating to sensitivity would not be meaningful or otherwise helpful to investors? If so, which estimates or what circumstances?

Is a three-year period the most appropriate period of time over which investors should consider changes? If not, why would a shorter or longer period be more appropriate?

Is the additional information elicited by the proposals useful to investors, other users of company disclosure and readers of a company's financial statements? If not, how can it be improved to achieve that goal?

Is disclosure necessary concerning the procedures that management follows in selecting its critical accounting estimates? If so, what additional disclosure should be provided?]

The NPR would require two specific quantitative disclosures relating to critical accounting estimates. In the first instance, a company would be required to present quantitative information about changes in its overall financial performance and, to the extent material, the line items in the financial statements that would result if certain changes relating to a critical accounting estimate were assumed to occur. In the second instance, a company would be required to disclose any material changes in its critical accounting estimates during the past three fiscal years.2 Our members have raised a number of questions about the first proposed disclosure and concerns about the second proposed disclosure.

As explained earlier, mortgage banking companies rely on sophisticated models to generate values for their MSR using methodologies and assumptions that are generally accepted and derived with some degree of precision from available data.3 Furthermore, mortgage companies generally substantiate the values produced by these models by reference to market data. These data can include: formal annual market appraisals performed by servicing brokers; periodic quotes from servicing brokers based on the main characteristics of a portfolio; and recent market transactions of which a mortgage company has direct knowledge.

With regard to the first proposed quantitative disclosure, the NPR would give a mortgage company the option of assuming that: (1) it changed the most material assumption or assumptions underlying its estimated MSR value and discuss the results, or (2) that the estimated value itself changed. If a company chooses to assume that the critical accounting estimate changed, the company would use the upper and the lower ends of the range of reasonably possible estimates which it likely determined in formulating its recorded critical accounting estimate. Our members have raised questions about the first alternative.

In considering the first alternative, our members have questioned whether they would be required to disclose the effects on their companies' overall financial performance due to a change in their estimated MSR values (due to a change in an underlying assumption) in isolation or in combination with changes in mortgage production volume and other factors that would also be affected by a change in the assumption. Because MSR values are affected first and foremost by changes in mortgage interest rates, a change in assumption would undoubtedly require a change in interest rate/s underlying the MSR value. However, because changes in interest rates also affect loan production volume and hedging gains and losses, we believe it would be misleading for mortgage companies to disclose the effects on their financial performance of a change in their estimated MSR values due to a change in interest rates without revealing the extent to which other areas of the business would change also.

For example, while MSR impairments increase during periods of declining interest rates, the impairment losses are offset by increases in mortgage production income and, for companies using derivatives as MSR hedges, MSR hedging gains. Conversely, while MSR values improve during periods of rising interest rates, those MSR gains are offset largely by decreases in mortgage production income and, for companies using derivatives as MSR hedges, MSR hedging losses.

Consequently, we believe disclosure of the counterbalancing effects of changes in the production side of the business is critical to an accurate understanding of how a change in the interest rate/s underlying an estimated MSR value would affect the financial performance of the company overall. Consider, for example, a recent press release on the results of a 2001 industry survey which announced, "In terms of total mortgage industry profitability, for most participants the production profits more than offset servicing losses for the year 2001."4 This is noteworthy given that MSR values suffered unprecedented impairments last year.

On the other hand, we are concerned that disclosure of all this information might overwhelm investors at significant cost to mortgage companies. We also question how beneficial the disclosures would be to investors as they would primarily reveal the extent to which changes in estimated MSR values are offset by changes in production income. Furthermore, this information is already reflected in a company's GAAP income.

We are especially concerned, however, about the cost associated with the second proposed quantitative disclosure which would require mortgage companies to disclose information relating to changes in estimated MSR values in the past three years. This disclosure, in particular, would require mortgage companies to produce large quantities of information given the frequency with which MSR values change in response to changes in interest rates. In our estimation, disclosure of past information on changes in MSR values would reveal little more than how often interest rates have changed during that time. Also, FAS 140 already requires fair value disclosures of MSR in the footnotes to the financial statements.

As such, we question whether market driven changes should be subject to the same level of disclosure as changes that are driven primarily by management judgment. In our estimation, a primary purpose of the proposed historical quantitative disclosures is to provide investors with an indication of how reliable management judgment has been in the past. However, because changes in MSR values are affected primarily by interest rates, we believe disclosure of changes in their values would reveal little about management's historical judgment.

Question 4: Would items that meet the definition of critical accounting estimates but that are hedged and discussed in detail in the Risk Management section of the 10-K still be subject to the NPR?

In the first illustrative example in the NPR, Alphabetical Company's disclosure provides information about management's estimate of its warranty obligations (which extend over a period of 6 to 10 years) to repair electrical equipment manufactured and sold by the company. In its disclosure, Alphabetical Company states: "Our hedging programs provide adequate protection against short-term volatility in copper prices, as described in "Risk Management," but our hedging does not extend beyond 5 years. Accordingly, our management must make assumptions about the cost of that raw material in periods 6 to 10 years in the future."

In considering this example, some of our members questioned whether the NPR would require additional disclosures of MSR that are hedged, as described in the "Risk Management" section of a company's MD&A. Or, they wonder, would the NPR require additional disclosures only of MSR that are not hedged? If disclosures of both hedged and unhedged MSR would be required, would the required level of disclosure be the same for both?

In addition to the above questions, we would like to respond directly to the following two questions which are raised in the NPR.

Question 5: Should we require that the critical accounting estimates disclosure in the MD&A undergo an auditor examination comparable to that enumerated in AT §701?

Our members believe any additional assurances that investors might receive as a result of a CPA's performance of an attest engagement with respect to the MD&A (pursuant to AT §701) would not be sufficient to justify the additional cost to companies of hiring CPAs to perform the separate engagements. Our members indicate that the majority of their CPAs already review their MD&A disclosures as part of their annual financial statement audits to ensure, among other things, that the MD&A disclosures and the company's financial statement disclosures are consistent. Consequently, because the financial statement disclosures are already the subject of an audit, we believe that the CPA's review of the MD&A disclosures provides reasonable assurance to investors and that an additional engagement designed to attest to the MD&A disclosures separately would be unnecessary.

Question 6: Should we require in MD&A a discussion of the impact that alternative accounting policies acceptable under GAAP would have had on a company's financial statements even when a company did not choose to apply the alternatives?

We believe that with respect to certain (not all) alternatives available in accounting standards, a requirement to discuss the impact that alternative accounting policies acceptable under GAAP would have had on a company's financial statements could require that companies maintain two sets of books. The effort required to do this would be enormous and costly for certain alternatives. If the SEC decides to expand the scope of the NPR for this purpose, we recommend that it be done on a targeted basis for high risk issues only.


The MBA supports the objective of the NPR to increase investor confidence in corporate financial data by increasing transparency of corporate reporting. However, we believe the extent to which the NPR achieves this objective will depend on the degree accounting policies are consistently disclosed by companies with similar policies.

As explained herein, we are concerned that mortgage companies' disclosures could be dramatically different depending upon how they interpret the NPR, as reflected by the questions raised herein. We are concerned also that the amount of information that a mortgage company might be required to disclose about its estimated MSR values (with respect to the potential disclosure of estimates underlying estimated MSR values as well as the quantitative disclosures) could overwhelm investors at significant expense to mortgage banking companies. We believe that unless our members' questions are addressed, mortgage company disclosures may be inconsistent with the objective of the NPR, despite management's best efforts to comply with it.

For these reasons, we would like to recommend that consideration be given to appointing a joint Commission/industry working group to develop responses to our members' questions which will result in reasonably consistent and helpful information to investors. The MBA would be pleased to recommend members of the industry who would be able to contribute to this effort.

Again, the MBA greatly appreciates the opportunity to comment on the NPR. If you have any questions about our comments, please contact Alison Utermohlen, Staff Representative to MBA's Financial Management Committee, at 202/557-2864.


              Kurt Pfotenhauer
              Senior Vice President
              Government Affairs

1 It should be kept in mind that the recorded value of a company's MSR portfolio is an aggregate number comprised of the "lower of cost or market value" of its different risk strata. As such, the aggregate value could include strata that are recorded at amortized cost and others at estimated market value.
2 Different rules would apply to small businesses.
3 For example, prepayment speed assumptions incorporated in MSR valuation models are derived largely from historical prepayment patterns.
4See June 24, 2002 MBA/The STRATMOR Group Press Release on results of their 2001 Peer Group Roundtables.