Letter of Comment No: 33 July 3,2002

Transcription

1 i!i ERNST & YOUNG 5 Times Square New York, New York Phone: (212) Letter of Comment No: 33 July 3,2002 File Reference: 1100:1~3 Date Received: 7/:YO;;l- Ms. Suzanne Bielstein Director of Major Projects and Technical Activities Financial Accounting Standards Board 401 Merritt 7 P.O. Box 5116 Norwalk, CT Proposed Statement of Financial Accounting Standards, Amendment of Statement 133 on Derivative Instruments and Hedging Activities (File Reference No ) and related Statement 133 Implementation Issues We appreciate the opportunity to provide you with our comments on the Exposure Draft of the Proposed Statement of Financial Accounting Standards, Amendment of Statement 133 on Derivative Instruments and Hedging Activities (the Exposure Draft) and related Implementation Issues B12 and D2. We believe the proposal will unnecessarily add a significant level of additional complexity to an already overly complex standard. The lack of clarity of the drafted model will lead to additional diversity in practice and an increase in the occurrence of errors in financial statements. We therefore urge the Board to abandon this proposal and start over with a simpler model for the accounting for beneficial interests. Specifically, we suggest that: The notion that all beneficial interests issued by a qualifying special purpose entity (QSPE) represent a debt host be abandoned and replaced with an approach consistent with paragraph 60 of Statement 133, the definitions of debt or equity securities in Statement 115 Accounting for Certain Investments in Debt and Equity Securities, and the provisions of Statement of Financial Accounting Concepts No.6, Elements of Financial Statements, a replacement of FASB Concepts Statement No. 3 (incorporating an amendment of FASB Concepts Statement No.2), [Concepts Statement 6], all of which provide criteria for distinguishing debt and equity instruments. Although it usually true that beneficial interests are debt instruments, it is not necessarily always the case. The Board adopt a principles-based approach to the bifurcation issue that focuses on "readily identifiable" embedded derivatives. In our view, embedded derivatives are "readily identifiable" only by examining the contractual cash flows required by the beneficial interest contract or certificate itself. Only when the A Member Practice of Ernst & Young Global

2 EJ ERNST & YOUNG Ms. Suzanne Bielstien Page 2 contractual terms of the certificate are unstated or reference instruments within the SPE should the holdings of the SPE be considered. The impairment model of EITF be expanded to encompass all beneficial interests, including those which do not have "readily identifiable" embedded derivatives. Because debt beneficial interests supported in part by equity assets in an SPE would not be bifurcated, we propose the expansion of the EITF impairment model to address income statement recognition concerns. Paragraph 16 of Statement 133 be amended to allow any hybrid instrument created in a financial asset securitization to be accounted for like a trading security at fair value through earnings, thereby rendering the required bifurcation of embedded derivatives elective. The redrafted paragraph 6(b), regarding initial net investment, be abandoned and replaced with a principles-based approach that requires identification of a debt host instrument only when an instrument is predominantly a financing arrangement. The proposed expansion of the bifurcation of instruments into artificial components that do not reflect the manner in which they will or can be settled does not improve financial reporting. The investor's accounting for its beneficial interest should not cause a QSPE to lose its qualifying status under Statement of Financial Accounting Standards No. 140, Accounting for Transfers and Servicing of Financial Assets and Extinguishments of Liabilities (Statement 140). The consideration of derivatives in assessing the qualification of an SPE should be limited to freestanding derivatives as the Board intended in promulgating Statement 140. If the Board chooses to reject any or all of the above recommendations, then we strongly recommend that the following suggestions be considered to ease the adoption of the complex new model that has been proposed: The transition and grandfathering provisions for beneficial interests issued by QSPEs need to be more generous and clear. All beneficial interests purchased prior to the issuance of the Exposure Draft, whether issued by QSPEs or regular SPEs, should be grandfathered from applying its provisions. We also believe that the Exposure Draft should provide an entity an additional opportunity to transfer any beneficial interest accounted for as an available-for-sale security into the trading category consistent with the provisions of paragraph 55 of Statement 133, upon adoption of the final Statement.

3 ej ERNST & YOUNG Ernst & Young UP Ms. Suzanne Bielstien Page 3 Guidance regarding the application of the Board's model for embedded derivative bifurcation and the related accounting for the bifurcated pieces is ambiguous, complex, and impracticable and must be clarified prior to finalization. We have significant concerns regarding the potential for significant diversity in practice. Examples using journal entries and explanations for calculations are imperative because it requires the identification of components with hypothetical terms. The concerns cited above, as well as our other comments, are elaborated further in the next sections.

4 !J ERNST & YOUNG Ms. Suzanne Bielstien Page 4 Requirement That All Beneficial Interests Issued by a QSPE Contain a Debt Host We concede that most senior level beneficial interests consist of a debt host contract. However, we believe that the nature of the "host" of a residual beneficial interest should be dependent on the existing definitions of liabilities and equity. We do not agree that residual interests issued by QSPEs must always be considered to consist of a debt host. This conclusion is not consistent with paragraph 60 of Statement 133, which states, "if the host contract encompasses a residual interest in an entity, then its economic characteristics and risks should be considered that of an equity instrument and an embedded derivative would need to possess principally equity characteristics to be considered clearly and closely related to the host contract." It is also not consistent with the provisions of Statement of Financial Accounting Concepts No.6, Elements of Financial Statements, a replacement of FASB Concepts Statement No.3 (incorporating an amendment of FASB Concepts Statement No.2), [Concepts Statement 6]. Although voting control is one of the characteristics of an equity owner and equity issued by a QSPE does not have this characteristic, a residual interest may have the residual economic risks and rewards of an entity. We also note that there are many classes of equity that do not carry voting rights such as non-voting common stock and preferred stock. Distinguishing between a debt host and equity host requires judgment and the existing definitions of debt and equity hosts in Statement 133 as well as the definitions in Statement 115 and other existing and upcoming literature, including the FASB's project on Liability and Equity Instruments, should be used to determine the nature of the host contract. As a result, we strongly object to this proposed change and suggest that paragraph 61(m) and Issue B 12 be eliminated in their entirety (the examples from Issue B 12 could be incorporated into Issue D2). Bifurcation of Embedded Derivatives We support the broadening of the paragraph 14 scope exception to include all beneficial interests that arise in a securitization. However, we were disappointed with models proposed in Issues B 12 and D2 and believe that as currently drafted, this guidance will be difficult to implement. The new guidance seems to have moved away from applying the provisions of Statement 133 to the contractual terms of each individual instrument in that it forces entities to create derivatives for accounting purposes that may not actually exist. We believe that under the proposed model, in many cases, an entity will be creating derivatives and/or host instruments by making up hypothetical contractual terms that do not resemble the stated or implied terms of the instrument. We believe that in most

5 EI ERNST & YOUNG Ernst & Young llp Ms. Suzanne Bielstien Page 5 instances a focus on explicit terms and contractual cash flows will prove more operational. This concept is best illustrated in Examples 3 and 4 ofissue B12. In both examples, the transferor retains a residual interest that the exposed guidance indicates must be accounted for as a debt host. But the instrument is a residual interest and the QSPE does not hold debt related assets. Although the guidance does not state how an entity would determine the terms of the synthetically created debt host instrument, we assume the host debt instrument would be a zero coupon debt instrument. It seems inappropriate to permit the accrual of interest income, which will only be realized through the value of equity securities. In addition, it is somewhat artificial to recharacterize the investment as a debt host for accounting pul]joses when that characterization is inconsistent with the economic substance of the residual interest. The following example is another that we have recently encountered in practice that illustrates our concerns: Company transfers lo-year Treasury bonds and equity instruments ($20 million and $80 million of fair value at the time of transfer, respectively) to a qualifying SPE in a transfer that meets the requirements for sale accounting under Statement 140. The senior interest (A-Class-equal to $40 million) pays principal and interest at maturity at a variable rate indexed to LIBOR with a maturity of 10 years. A B-Class interest of $20 million that pays principal and interest at maturity at a variable rate indexed to LIBOR with a maturity of 10 years, that is subordinate to the A-Class interest, is retained by the transferor. The transferor also retains a C-Class interest in the legal form of cumulative preferred stock issued by the qualifying SPE in the amount of $40 million that accumulates dividends at a stated rate of 25 percent over its stated 10-year life and an R-Class, with a nominal initial fair value, which is entitled to any remaining cash flows. The C- and R-Classes are subordinate to both the A-Class and B-Class beneficial interests, but are never required to reimburse them (i.e., the A-Class and B-Class can never receive more than 100 percent of the cash flows from the equity instruments transferred into the qualifying SPE). At the tenth anniversary of the QSPE, enough assets are to be sold to satisfy the A-, B-, and C-Class certificates and the remainder invested in pel]jetuity for the R -Class interest and the proceeds distributed to its certificate holders. The proposed guidance in Issue C 17 suggests that none of the beneficial interests qualify for the paragraph 14 exception because the nature of the cash flows arising from the equity instruments has been changed to incol]jorate a LIBO R interest rate

6 S!J ERNST & YOUNG Ernst & Young LlP Ms. Suzanne Bielstien Page 6 risk and the aggregate sources of cash flows available to service the interest are not consistent with the stated tenns of the interest in the beneficial interests. The second step of the evaluation indicated by the proposed guidance of Issue D2 is an analysis of whether the instrument as a whole meets the definition of a derivative in paragraph 6. In the case under consideration, none of the instruments in their entirety meet the definition of a derivative because they would all be detennined to be option-based contracts and require an initial net investment that is not equal to the fair value of the option component. Although it would not change the conclusion reached, we do not believe that it is intuitively obvious that the A- and B-Class certificates are option-based contracts. We believe that the assessment as to whether a contract is an option-based or nonoption based contract should be based on its tenns and not based on the cash flows of the underlying securitized assets. As a result, we find the conclusion that they are option-based contracts to be confusing and impracticable. The next step in the analysis required by the Exposure Draft is the search for embedded derivatives required by paragraph 12 (the third step). We also find the application of paragraph 12 and the application of Issues B 12 and D2 confusing particularly when analyzing debt instruments with contractual cash flows for embedded derivatives. In these situations, Issue D2 would require the investor to look beyond the tenns of the instrument and to consider the aggregate sources of cash flows that are available to service the interest creating ambiguity as to when an embedded derivative exists and if it does, what it is comprised of. We do not object to the notion that an investor should have to look beyond the tenns of the certificate issued in a securitization to detennine if an embedded derivative exists, but believe such a requirement should be limited to situations in which an instrument does not have contractual tenns (e.g., is a residual interest) and is clearly a debt host contract. For example, the residual tranche of a financial asset securitization (that has a limited life) would most likely be comprised of an embedded derivative if it must absorb the risk of any of the more senior beneficial interests that have interest rate characteristics that are different than those of the assets held by the securitization vehicle. Such an approach could result in a much simpler model that, together with an expansion of the EITF impainnent framework, would lead to an appropriate financial reporting result.

7 i!j ERNST & YOUNG Ms. Suzanne Bielstien Page 7 Ernst & Young's Suggested Model Beneficial interests should be subject to bifurcation only if embedded derivatives are readily identifiable. In our view, embedded derivatives are readily identifiable only by examining the contractual cash flows indicated by the beneficial interest contract or certificate itself and considering whether those contractual terms are clearly and closely related to the host contract. If the beneficial interests are comprised of contractual cash flows and the terms do not indicate that there are any embedded derivatives, then the instrument would be accounted for according to its terms and no bifurcation should be required. In the event that a host debt instrument does not have contractual cash flows and/or the contractual terms of the certificate are unclear or reference instruments within the structure, then the instrument would have to be evaluated for embedded derivatives by considering the holdings of the SPE. We believe that the only way a possible embedded derivative can be reliably identified and accounted for with a high degree of uniformity is if the bifurcation methodology follows the concept of always looking to the contractual cash flows of the instrument. Concepts such as estimates of credit and prepayment risk, or models of projected equity security appreciation or depreciation, should have no role in "imputing" terms to hybrid instruments that do not contractually exist. We acknowledge that this approach may not respond to the Board's concern about the adequacy of the original BI2 model. However, this concern can be more appropriately addressed by an expansion of the impairment model of EITF to encompass all beneficial interests, including those that a readily identifiable bifurcation model does not indicate contain embedded derivatives. In addition, the accounting complexities that arise from bifurcation could be simplified if paragraph 16 of Statement l33 were amended to allow any hybrid instrument created in a securitization, even those that are not technically securities, to be accounted for like a trading security at fair value through earnings. Such an approach would essentially render the required bifurcation of embedded derivatives created in a securitization to be elective (i.e., avoid bifurcation by accounting for the entire hybrid instrument at fair value) thereby relieving the administrative burden of identifying and accounting for embedded derivatives. This modification would also be consistent with the Board's direction to account for all financial instruments at fair value and consistent with the treatment of financial instruments that can contractually be prepaid or otherwise settled in such a way that the holder would not recover substantially all of its recorded investment under paragraph 14 of Statement 140. Because this election would be limited to hybrid instruments created in a securitization, it would be limited enough to keep entities from engineering de minimis embedded derivative features into its liabilities in order to obtain fair value accounting.

8 aj ERNST & YOUNG Ms. Suzanne Bieislien Page 8 Comparison of the Exposure Draft's Model to Our Model The application of the Exposure Draft to our previous example indicates that, either one, or perhaps both, of the A- and B-Class certificate holders may have an embedded derivative because the securitized treasury securities and equities may not provide sufficient contractual cash flows to satisfy the stated terms of the interest even though the terms of the certificates do not include embedded derivatives themselves. The Exposure Draft would also indicate that either one or both of the A- and B-Class interests would be deemed to contain an embedded feature that is a written cap on equity prices equal to the investor's stated fixed return plus a purchased floor on equity prices that effectively limits the potential loss that would result from a decline in value of the equities to the amount beyond the loss absorbed by the more junior interests. Therefore, one or both of the A- and B-Class certificate holders has an embedded equity derivative that is not clearly or closely related to the host contract requiring bifurcation. However, it not clear when the securitized assets may not provide sufficient cash flows to satisfy the stated terms of interest. For example, based on a current forward LIBOR interest rate curve, it could be expected that the treasury securities will provide enough cash flow to satisfy the A-Class certificate. As a result, it is not clear whether both tranches would have embedded derivatives or whether only the B-Class certificate holder would have an embedded derivative. We believe the model in the Exposure Draft is ambiguous. Under our approach, the A-Class and B-Class certificate holders would first look to the terms of their contract to determine if embedded derivatives exist. As both certificates have clear terms that reflect contractual cash flows that are clearly and closely related to their host debt instrument, they would not be considered to contain embedded derivatives. Both certificate holders would account for their investments according the terms of the instrument and no bifurcation would be required. In addition, each certificate would be accounted for in accordance with paragraph 14 of Statement 140. To address income statement recognition concerns as it relates to investments accounted for similar to available-for-sale securities under Statement 115, we would require that these beneficial interests be subject to the EITF impairment model. Following the guidance provided in the Exposure Draft, specifically paragraph 61(m) and Issue B12, also would require the C-Class certificate holder to create a debt host instrument. We agree that the C-Class certificate represents a debt host because of its stated I O-year life. The guidance in the Exposure Draft would also indicate that the C-Class certificate holder would be required to consider the holdings of the SPE because the terms of the

9 $!I ERNST & YOUNG Ernst & Young LlP Ms. Suzanne Bielstien Page 9 C-Class certificate are not comprised of contractual cash flows, but are of a residual nature (we assume that a cumulative dividend rate would not be considered to represent a contractual interest obligation) that are not clearly and closely related to the debt host instrument. Based on that guidance, the C-Class certificate holder would be entitled to the residual cash flow comprised of contractual cash flows arising from an inverse floater (i.e., the difference between the contractual interest rate of the Treasury bonds and LmOR with a notional amount of $20 million) and non-contractual cash flows arising from the equity securities. As neither the inverse floater nor the cash flow arising from the equity securities (i.e., an implicit call spread on the equities) are clearly and closely related to the debt host instrument, we would presume that a compound derivative representing an inverse floater and a call spread on the equities would be required to be bifurcated and recorded based on the fair value of the option component. We would also assume that the debt host instrument would be accounted for similar to a zero coupon bond and initially recorded at amount equal to $40 million less the fair value of the option component of the compound derivative. While we do not technically disagree with the bifurcation that results from applying the guidance in the Exposure Draft to the C-Class certificate, we believe that this model is too complex for the vast majority of financial statement users and preparers to understand and apply, primarily because the embedded derivative consists of a mix of contractual and non-contractual cash flows. As a result, we believe the application of this model will result in potentially significant diversity and lack of comparability. To simplify the application of bifurcation for beneficial interests created in financial asset securitizations that are determined to be debt host instruments, our approach would only require derivatives that are readily identifiable based on contractual cash flows to be bifurcated. As a result, no derivatives would be bifurcated and the C Class certificate would be accounted for in accordance with paragraph 14 of Statement 140. To address income statement recognition concerns as it relates to investments accounted for similar to available-for-sale securities under Statement lis, we would require that these beneficial interests be subject to the EITF impairment model consistent with our approach for the A- and B-Class certificates. Finally, we believe that the R-Class certificate should be analyzed as an equity host instrument, since it represents a residual interest in the SPE and does not have a limited life. Since the instrument would be comprised of an equity host instrument, the R-Class certificate holder would be required to consider whether its cash flows provide a return that is clearly and closely related to a host equity instrument. Since it is only entitled to receive the residual cash flows of the structure, we do not believe it contains an embedded derivative. However we would suggest that the instrument be accounted for in accordance with paragraph 14 of Statement 140 as an available-forsale or trading security so that the fair value of instrument would always be

10 $!I ERNST & YOUNG Ernst & Young LlP Ms. Suzanne Bielstien Page 10 prominently displayed. Furthennore, we would suggest that all beneficial interests (including equity host instruments) that are accounted for similar to available-for-sale securities under Statement 115, be subject to the impainnent model ofeitf We believe that the approach we propose is less complex and more reliable than the bifurcation model contained in the Exposure Draft. It eliminates the creation of artificial host instruments and embedded derivatives when tenns are not explicitly stated or not readily identifiable. It also is more reflective of the actual cash flows and economics of the beneficial interests. Initial Net Investment The amendment of paragraph 6(b) has created more complexities than previous guidance, and we are not clear how the revisions would improve financial reporting. This revision will require an increased amount of bifurcation when derivatives contain an off-market element and potentially require significantly more effort by constituents to identify "hybrid" instruments, bifurcate them into a separate at-market derivative instrument and a separate debt host instrument, and account for the same instrument using both an accrual accounting model and a fair value accounting model on an on-going basis. In certain transactions, we believe that 5 percent may not provide significant leeway for off-market transactions that have a large bid-ask spread and are concerned that a "bright line" introduces a way for entities to structure around the "rules." For example, a 10-year interest rate swap can be structured to have a fair value of $0 at inception (thus, meeting the amended definition of a derivative), but the fixed rate on the swap can be engineered to reset at a later date such that a significant amount of the cash outflows occur in the latter years. While the economic substance of this arrangement is the granting ofloans in the early years, the proposed rules-based model would not require the identification of a debt host. We suggest that contracts should be accounted for based on their substance requiring an element of judgment consistent with a more principles-based approach. A final statement could acknowledge that some derivatives may include financing elements that cause the derivative to contain a financing arrangement without imposing a bright line test. For derivative contracts where the financing element causes the contract to predominately be a financing arrangement, as opposed to a freestanding derivative, and when the tenns of the host instrument are readily detenninable, we agree that bifurcation should be required. However, bifurcation should be mandated only when the derivative contract is predominately a financing arrangement (a substantive judgment). In other cases, bifurcation should be pennitted but not required. Such an approach would provide users of derivatives the flexibility to bifurcate the off-market component of a derivative into a

11 !J ERNST & YOUNG Ms. Suzanne Bielstien Page II debt host because bifurcation may have certain benefits such as being able to designate the bifurcated, at market derivative in a more highly effective hedging relationship. If the F ASB is concerned that users of the financial statements are not giving appropriate consideration to derivative liabilities that are not classified as debt, the Board should consider expanding the disclosure requirements of the expected cash outflow of obligations as of a reporting date to long-term derivative liabilities not accounted for as hedges. The application of paragraph 6(b), as amended, would artificially create non-substantive debt host instruments from compound derivatives with zero net investments. Example 3 of Issue A20 illustrates this scenario. Currently, these instruments are being accounted for in their entirety as derivatives and presented at fair value on the balance sheet. Applying the amended definition of initial net investment and the guidance in Issue A20 does not reflect the economic substance of the contract. This anomaly is especially true when considering an option-based contract. The appropriate initial premium for an option is highly dependent on the terms of the contract and other factors. As a result, and considering the nature of option contracts, we recommend that an option-based contract be considered a derivative in its entirety when the initial investment is between zero and an amount equal to the fair value of the option component (time value of an at-the-money option). To further illustrate our point, consider a loan commitment that relates to the origination or acquisition of a mortgage loan that will be held for sale. We would presume that such a loan commitment would be determined to be an option-based contract. However, we assume that it is possible that a loan commitment would not have an initial net investment equal to the fair value of the option component because an amount equal to the fair value of the option component may not be paid under a particular arrangement. As a result, applying the amended definition of initial net investment in 6(b) could lead an entity to conclude that the loan commitment is not a derivative in its entirety because the initial net investment would be less than the fair value of the option component. We believe the determination of initial net investment could be simplified by leaving the old concept of "smaller than would be required for other contracts that would be expected to have a similar response to change in market factors" and by illustrating the types of arrangements that would be predominately assessed as a financing arrangements requiring bifurcation. There may be instances, based on the substance of the contract, where it would not be appropriate to bifurcate a derivative and the amended guidance should recognize those situations. Consider the common scenario of effectively terminating one derivative by entering into an offsetting derivative. For example, a company enters into an off-market

12 f!i ERNST & YOUNG Ms. Suzanne Bielstien Page 12 non-option-based swap and pays up-front amount equal to the fair value of the swap to economically cancel a swap that was entered into at-market in a previous period without incurring an current tax liability. Currently, both instruments would be accounted for at fair value on the balance sheet and the changes in their fair values would completely offset in the income statement. Assuming the up-front amount that was paid for the second swap was greater than 5 percent of the fully prepaid amount, applying the provisions of the Exposure Draft and Issue A20 would require that the second swap be bifurcated into a debt host instrument and an at-market swap. As a result, the future income statement effect will not be neutral, creating inappropriate income statement volatility. If the Board wishes to pursue this approach, we would be happy to provide sample language to assist in this effort. We also believe that the application of Issue E21 and the amended definition of initial net investment may cause confusion as to when an off-market, non-option-based contract must be viewed as a debt host with an embedded at-market derivative. Specifically, is bifurcation only required at inception? Would it be required after a significant modification to a derivative? If bifurcation is appropriate under these scenarios, would it be permissible to bifurcate an off-market derivative into a debt-host instrument each time a derivative is re-documentedlre-designated in hedging relationship? By introducing the requirement that an off-market derivative be bifurcated into a debt host and an at-market derivative, we believe that it is important that the Exposure Draft clarity what events drive bifurcation. Implications o/embedded Derivatives to QSPEs We do not believe it is logical that an investor's accounting for its beneficial interest would cause a QSPE to lose its qualifying status under Statement 140. We believe that an investor's bifurcated derivative, which is an accounting convention and not the legal form of the instrument, should not be viewed as a derivative for purposes of assessing the provisions of paragraph 35(c)(2) of Statement 140. A QSPE is already restricted as to the types of derivatives and the types of financial assets that it can hold. Broadening paragraph 35(c)(2) of Statement 140 to include embedded derivatives seems egregious and overly complex. As a result, we recommend that the draft guidance provided in the Question and Answers Related to Derivative Financial Instruments Held or Entered into by a QSPE and the guidance in the Exposure Draft be revised to clarity that an embedded derivative within a beneficial interest would be viewed as a beneficial interest and not be considered a derivative as it is addressed in paragraph 35(c)(2) of Statement 140. *********

13 go ERNST & YOUNG Ms. Suzanne Bielstien Page \3 In the event that the Board decides to proceed with the bifurcation model described in the Exposure Draft, we urge the Board to consider the following regarding the transition and grandfathering provisions. Transition and Grandfathering Provisions-Beneficial Interests The transition and grandfathering provisions of the Exposure Draft are neither comprehensive nor clear. We believe that the reason for the lack of clarity is that paragraph 42 tries to sweep the grandfathering of an SPE's qualified status (a Statement 140 concept) into the same guidance as the grandfathering of beneficial interests from the proposed new bifurcation guidance (a Statement 133 concept). We believe it would be clearer and more practical for the Board to separate the grandfathering provisions for these very different concepts. Paragraph 42 currently follows the same general approach to grandfathering that Statement 140 did, with respect to the more restrictive guidance on the powers of QSPEs relative to superseded Statement 125. Accordingly, we do not object to paragraph 42's current approach with respect to the conditions for permitting a formerly qualifying SPE to continue to be accounted for as a QSPE. However, we note that the effective date and transition guidance in Issues B12, C17, and D2 are not consistent with one another and with the Exposure Draft. Issue D2 includes a discussion regarding the application of its guidance to QSPEs and the related grandfathering issues while the others do not. Furthermore, the effective date and transition discussions should be consistent with the final language in paragraphs 40 through 43 of the Exposure Draft. From the discussion in Issue D2, it would seem that a QSPE that meets the applicable criteria is only grandfathered from paragraph 35(c)(2) of Statement 140, but not from the bifurcation provisions of the Exposure Draft, which is not consistent with paragraph 42 of the Exposure Draft as discussed previously above. A second grandfathering paragraph applicable only to the issue of bifurcation of beneficial interests could be styled after the provisions of paragraph 50 of Statement 133. Paragraph 50 originally permitted an entity to elect to grandfather a hybrid instrument that would otherwise be required to be bifurcated under Statement 133 if that instrument were issued, acquired, or substantively modified by the entity before January I, That date was chosen because the F ASB finalized its draft grandfathering guidance at a December 17-18, 1997 public meeting. Thus constituents who desired to avoid the complexities of bifurcation were put on two week's notice as to which hybrid instruments they should avoid issuing, acquiring, or modifying in the near future. This Exposure Draft does not afford constituents the same "fair warning," even though this Exposure Draft would apply to a much smaller universe than the original Statement 133. Current guidance already results in the bifurcation of a number of beneficial

14 i!i ERNST & YOUNG Ms. Suzanne Bie1stien Page 14 interests issued by SPEs and QSPEs. This amendment will encompass a small subset of additional hybrids, so we believe no measurable harm will result by requiring the new bifurcation guidance to apply only to prospectively issued beneficial interests. The Board should simultaneously accelerate the grandfathering date (say to September I, 2002) to prevent abuse of a more generous grandfathering. Providing separate grandfathering for the bifurcation of beneficial interests would also cure an unworkable element of the current provisions. Whether or not a third party beneficial interest holder is exempted from grandfathering should not be determinative as to the continuing qualifying status of an SPE. While theoretically attractive, such a linkage is unworkable in the real world where a third party investor is unlikely to be able to determine the intricacies of a QSPE's accounting issues or whether or not such QSPE is or is not able to meet the three conditions in paragraph 42. Another benefit of creating separate grandfathering guidance is that constituents investing in beneficial interests of QSPEs whose qualifying status is not changed would clearly be eligible for grandfathering. (For example, QSPEs that hold equity securities but no derivatives will not have their qualifying status endangered by the new guidance, but their beneficial interests, the host contracts of which the Exposure Draft requires to be considered "debt," would not be required to be bifurcated.) Paragraph 42 as currently drafted does not address this subset of constituents. We understand from discussions with the FASB staff that such constituents should assume that they too are eligible for grandfathering. Their concerns should be addressed more directly in the final Statement. Our suggestion to allow an election to grandfather all beneficial interests resulting from a securitization, not already required to be bifurcated by existing guidance, and issued, acquired, or substantively modified prior to a selected 2002 date, would solve this inequity. For those constituents who elect to bifurcate their hybrid instruments, the Exposure Draft should clarify how the journal entries should be calculated. A final Statement should explicitly indicate that the beneficial interest holder would have to determine the adoption date carrying value of the debt host instrument by going back to the inception of the instrument and applying the guidance in Issues B20 and B22. The necessity of performing this type of "accounting archeology" to determine the carrying value of the debt host is another reason why we believe the grandfathering provisions related to bifurcation need to be more generous. On a related note, the Exposure Draft should clarify that investors in beneficial interests who do not elect (or who are not permitted) to be grandfathered from the bifurcation provisions should be able to transfer their hybrid instrument into the trading category under Statement 115. Such action would allow such investors to avoid bifurcation accounting by recognizing the entire change in fair value of their hybrid instrument in earnings. Paragraph 15 of Statement 115 states that transfers into or from the trading

15 ill ERNST & YOUNG Ms. Suzanne Bielslien Page 15 category should be "rare," so it would be helpful if a final Statement would clarify that its effective date would constitute one of those "rare" occasions. Transition and Grandfathering Provisions-Initial Net Investment Based on our review of the transition provisions of paragraph 40 of the Exposure Draft, we presume a contract that had been accounted for as a derivative in its entirety (even those that would have been deemed to have "embedded debt hosts" under the provisions of the Exposure Draft) prior to the transition date would not have to apply the provisions of paragraph 6(b) of the Exposure Draft and that the revised guidance should be applied prospectively for future transactions. This presumption is also consistent with the provisions regarding the effective date and transition of Issue A20. We also presume when such contracts were designated in hedging relationships and determined to qualify for the shortcut method, that the application of the shortcut method would be grandfathered. The transition provisions of a final Statement should be clarified in these regards so that companies do not inappropriately account for these contracts at transition under the new bifurcation requirements. In addition, the transition provisions should be clarified to address what events would cause these pre-existing contracts to no longer be "grandfathered" from the bifurcation provisions of the Exposure Draft. For example, would a business combination cause all pre-existing contracts to have to begin complying with the bifurcation provisions of the Exposure Draft? Other examples that the transition provisions should consider are the implication of a substantial modification to a derivative after the effective date of the Exposure Draft and/or the re-documentationlre-designation of an existing derivative. Other Comments on the Exposure Draft ********* We note that the guidance from Issue C13 which was incorporated into the Exposure Draft provides that, "Loan commitments that related to the origination or acquisition of mortgage loans that will be held for sale, as discussed in paragraph 4 of Statement 65, shall be accounted for as derivative instruments by the issuer of the loan commitment... " However, the provision does not address other loan commitments related to non-mortgage loans (e.g., commercial loans) that are accounted for by analogy to paragraph 4 of Statement 65. We assume that such loan commitments would also be accounted for as derivative instruments by the issuer and suggest that the F ASB clarify this point. We also note that there is a lack of valuation guidance related to loan commitments that meet the definition of a derivative and are concerned that different issuers would value the same loan commitment differently. Specifically, the Board should consider stating that there

16 $!I ERNST & YOUNG Ernst & Young LlP Ms. Suzanne Bielstien Page 16 is a rebuttable presumption that the initial fair value of a loan commitment derivative is zero, since no cash has been exchanged. The amendment to paragraph 1 O( d) regarding certain financial guarantee contracts should be revised to incorporate the concept of subrogation as it is discussed in paragraph A30 (Basis of Conclusions) of the Exposure Draft. As written, this concept is missing from the standard. Deletion of paragraph 19---we suggest that the FASB explain why it was eliminated. We have found the paragraph helpful in certain instances in properly assessing the effectiveness of fair value hedges which follow the "long-haul" method. We noted that the amended provisions of paragraph 15 that provide that the evaluation as to whether a contract with an embedded foreign currency derivative qualifies for the bifurcation exception is to be performed only at the inception of the contract is not consistent with the guidance in Issue A18. Issue AI8 states that if events occur subsequent to the inception or acquisition of a contract that cause the contract to meet the definition of derivative, then that contract must be accounted for at that later date as a derivative under Statement 133. It is not clear to us why embedded foreign currency derivatives would not be treated similarly. There were several issues that had previously been brought to the Derivative Implementation Group that have yet to be addressed by the FASB staff. We are concerned that some of these issues have been lost in the shuffle and would like them to be clarified as part of this Amendment. For example, the issue regarding the interpretation of paragraphs 34 and 35 of Statement 133 with regard to how to apply the provision that states, "If... an impairment loss is recognized on an asset or an additional obligation is recognized on a liability to which a hedged forecasted transaction relates, any offsetting net gain related to that transaction in accumulated other comprehensive income shall be reclassified immediately into earnings," was never addressed. There also seems to be a conflict between this provision and the guidance in Statement 144, Accounting for the Impairment or Disposal of Long-Lived Assets. Statement 144 provides implementation examples that seem to allow an entity to include the undiscounted cash flows of derivatives in the recoverability test of a long-lived asset. However, paragraph 34 states that, "The fair value or expected cash flows of a hedging instrument shall not be considered in applying [existing impairment) requirements."

17 $!I ERNST & YOUNG Ernst & Young UP Ms. Suzanne Bielstien Page 17 * * * * * * * We would be pleased to discuss our comments with you or your staff at your convenience. Very truly yours, ~ofhll'p