IRS Outlines Differences Between Experts in Derivative Valuation Dispute

DR. PARSONS' REPORT COMPARED TO PETITIONER'S REPORT

1. DETERMINING MID-MARKET SWAP VALUES

[2] There is no disagreement as to whether the net present value of the projected cash flows for petitioner's swaps, as computed by petitioner's Devon Derivatives System ("Devon") for December 20, 1993, ¹ produce mid market swap values. Petitioner and all experts arrived at, essentially, the same values when projecting swap cash flows using December 20, 1993 public data and then discounting the cash flows as if all counterparties had a credit rating equivalent to a Standards's & Poor's ("S&P") rating of AA. ² This is an extremely significant agreement.

[3] Despite the enormous complexity of some of the calculations and despite the need to refer to market data of various sorts, such as LIBOR deposit rates, quoted swap spreads, Treasury rates, and Eurodollar futures prices, all experts,are able to arrive at roughly the same number. Dr. Parsons' test of the Devon values, using 57 swaps with approximately $2 billion in notional amount and including interest rate swaps of various designs and with various custom tailored payment schedules and other characteristics, found the overall discrepancies in total value are as small as 0.9%. ³ Looking at petitioner's experts' test, ⁴ using 25 swaps with a total notional amount of $978.8 million, ⁵ they found a total discrepancy of 0.3%. ⁶ Thus, despite the differences in the swap structures and different swaps tested, all parties could agree to the same mid-market value. The only real challenge in the process is mastering the mathematics of interest rate calculations, with the complications of interest rate mathematics not at the center of these cases.

[4] All experts agree that petitioner's 1993 mid-market values come from the primary market. Although it is true in some important formal respects that there is no significant secondary market in swaps, in that most swaps are held to maturity or terminated through buy-outs between the two counterparties to the bilateral contract, the fact that the parties can agree on the mid-market value is a reflection of the fact that there is a primary market ⁷ for swaps and a market for related financial instruments which pegs the mid- market value for swaps of all kinds at any date.

[5] All experts further agree that the market values for outstanding swaps can be constructed, with this construction based on observed market values of the component pieces. The prices for benchmark swaps are publicly known. The prices for other swaps are derived from these benchmarks. A custom tailored swap is really just an assemblage of cash flows, and the pieces that are assembled have been priced, or the pieces are closely related to other pieces that have been priced. This is no different in principle from the valuation of real estate from comparables.

[6] Petitioner's experts downplay this consensus as to the mid-market values and focus on the fact swaps are privately negotiated and trade over-the-counter, not on an exchange. Dr. Parsons' view is that this is not an essential distinction and ignores the fact that both experts were able to get data, e.g., from Bloomberg and Reuters, to test the December 20, 1993 Devon values for the actual swaps at issue. Dr. Parsons' opinion is that, just because an instrument trades over-the-counter, does not mean prices are not publicly available. Many stocks trade on the NASDAQ, which is not an exchange. Dealers on the NASDAQ post bids and offers just as do dealers in swaps. It is widely accepted that the 'price' on the NASDAQ is publicly available, at least to a degree of accuracy sufficient for most purposes. United States Treasury bonds trade over-the-counter in a dealer market, yet the prices are considered to be public and are widely used as benchmarks for valuing other instruments.

[7] The consensus on mid-market values cannot be ignored any more than the existence of public data to benchmark swap values can be ignored. Trades in the swaps market occur over-the-counter, just as do trades in stock on the NASDAQ and United States Treasury bonds. Swap dealers post bid and ask prices for benchmark instruments, and these bid and ask prices are broadly disseminated to the public. Market prices for related instruments, like Eurodollar futures contracts, are also broadly disseminated to the public. An information vendor such as Bloomberg provides users with the quoted bid and ask prices and with the swap prices implied from these other market prices.

[8] Thus, despite all the complications involved, the mid- market value of a swap can be readily agreed upon in most cases. This is true for the many kinds of custom tailored swaps. This is true despite the wild swings in interest rates and the correspondingly large swings in the mid-market value of each swap. This fact is brought home in these cases by the ability of petitioner's experts and respondent's expert, Dr. Parsons, to agree on the December 20, 1993 mid-market valuation using two different samples of swaps.

[9] Dr. Parsons does not dispute that one feature of many markets, which makes the prices reliable, is the posting of actual transaction prices (as opposed to bid and ask quotes of the dealers). In the swaps market there are no transaction prices, only quoted bids and asks. Transaction data are useful, and could be used to test derived values, but Dr. Parsons does not view it as a condition precedent to ascertaining the mid-market swap values. The reliability of transaction data is itself a question that needs to be answered on a case specific basis, e.g., buyout and assignment data, if available, must be evaluated with respect to the circumstances of the buy-out or assignment such as a forced liquidation of a position. The lack of transaction prices does not render mid-market values difficult to ascertain, or agree to, relative to the interest rate swaps at issue in these cases. To the extent petitioner's experts do not concur with this view, Dr. Parsons and petitioner's experts hold divergent views on the difficulty of determining mid-market values for the swaps at issue.

2. PETITIONER'S CARVE-OUT ADJUSTMENTS DO NOT REFLECT MARKET

[11] Dr. Parsons and petitioners' experts agree that there is no consensus in the industry on how, if at all, mid-market values are to be adjusted to reach fair market value. Petitioner's Report highlights the differences among dealers within the industry. See Petitioner's Report, at 30-31. ⁸ The differences can be summarized fairly succinctly: some swap dealers use mid-market values without adjustments, some use mid-market values making adjustments only for credit risk, ⁹ some use mid-market making adjustments only for credit risk and swap servicing costs and some use mid-market making adjustments that take into account multiple considerations. Depending on the dealer, the adjustment for the factor referred to as "administrative expenses" may include only swap servicing costs or swap servicing costs with a variety of other considerations such as the cost to hedge the swaps in the future and the allocable costs of their entire business.

[12] Despite this wide divergence in practices among dealers in the industry, petitioner's experts and Dr. Parsons agree that, in theory, adjustments to mid-market values for credit risk and for administrative costs may be appropriate. ¹⁰ Petitioner's experts and Dr. Parsons further agree that no publicly available swap market benchmarks were available in 1993 for making these adjustments; /11 but, there were a few studies related to credit risk such as the 1992 survey relative to swap default rates. See Dr. Parsons' Report, paragraph 70 at 38. Where the experts disagree is in their analysis of the actual credit risk and administrative expense carve-outs claimed in these cases absent such benchmarks.

[13] Petitioner's experts focus on whether the First National Bank of Chicago (hereinafter petitioner or "FNBC") made its adjustments in roughly the same way as other companies may have adjusted their values. ¹² Petitioners' experts present no evidence to establish petitioner's accounting procedures arrive at fair market value and no evidence to establish exactly how other companies adjusted their mid-market values. Petitioner's experts then conclude that, for 1993, the mid-market values with FNBC's adjustments are a better estimate of the fair market value than are the unadjusted mid- market value. Yet, petitioner's experts did not use any benchmarks to measure the appropriateness of the adjustments and did not in any way independently check or validate any of the adjustments against any valuation standards. No analysis of the petitioner's carve-out adjustments in terms of the willing buyer/willing seller was presented by petitioner's experts and no empirical facts were presented in support of their conclusions.

[14] Dr. Parsons focuses on whether the adjustments claimed by petitioner, called credit-risk carve-outs and administrative expense carve-outs by petitioner, adjust the values calculated by petitioner's Devon Derivatives software ("Devon") in such a manner as to arrive at values closer to fair market value than the unadjusted mid-market values. Dr. Parsons concludes, based on an analysis of the facts of this case, that petitioner's carve-outs do not reflect the market value of the relevant factors. Dr. Parsons further concludes that petitioner's methodologies have systematic biases that lead the adjusted mid-market value away from the fair market value. ¹³

[15] Thus, the pivotal difference is the experts' respective opinions on whether petitioner's carve-outs properly measure the theoretically sanctioned adjustments. Petitioner's experts say yes based only on industry accounting practices, without any analysis of empirical facts. Dr. Parsons says no based on empirical facts.

3. DR. PARSONS' DISPUTES PETITIONER'S EXPERTS' THEORETICAL

[F]rom an economic standpoint, it could reasonably be

[16] First, Dr. Parsons disagrees with the assertion that, from an economic standpoint, a swap would have zero value on the day it is negotiated. Many end-users are willing to pay-up for the risk management benefits of swaps. Therefore, the end-user will agree to a swap which has a negative value to the end-user and a positive value to the dealer. The spread charged by the dealer to the counterparty captures this positive market value for the dealer. Valuing the swap using a mid-market method and then adjusting the value by the entire value in the spread would incorrectly assume a zero value to the dealer.

[17] Second, any support for the assertion that swaps may have zero value on the day they are negotiated is limited to the inter- dealer market. Dealers generally charge smaller spreads to other dealers than they do to end-users, much as wholesalers charge a different price from retailers. Just as a mortgage originator can turn around and resell the loan at a profit in the secondary market, so too can a dealer often negotiate a swap with an end-user at one rate, and then turn around and lock in a profit by negotiating a swap with another dealer at a different bid or ask rate. Thus, petitioner's experts' arguments assume away the distinction between segments of the market and act as if there is only one bid-ask spread, one price for swaps to all parties. In addition, petitioner's experts' arguments ignore the reality that FNBC may enter into a swap with another dealer merely to risk manage its swap book, in effect paying a premium to balance its swap book, "losing" money on the transaction in terms of its spread.

[18] Third, to the extent that petitioner's experts equate the value of the bid-ask spread with the required adjustment to the mid- market value, Dr. Parsons disagrees. Many factors may enter into the determination of a dealer's bid-ask spread besides the ongoing administrative costs and remaining anticipated credit risk or any other remaining anticipated cost. For example, the dealer's spread may include compensation for the marketing and other activities related to origination of swaps. Dr. Parsons believes that the costs of the activities related to origination are not appropriate deductions from the mid-market value.

4. DEVON VALUES DO NOT ALWAYS OVERSTATE FAIR MARKET VALUES

[19] Petitioner's Report asserts that petitioner's unadjusted mid-market values always overstates the fair market value of a swap; consequently, petitioner's adjustments are always necessary. Dr. Parsons disputes this assertion; the mid-market value will not always overestimate the fair market value. Unadjusted mid-market values may OVER-ESTIMATE or UNDER-ESTIMATE the fair market value since credit risk is bilateral.

[20] According to Dr. Parsons' Report at paragraphs 63 through 65, pages 34 and 35, the fair market value may be more than the mid- market value of the swap when FNBC enters into a swap contract with a counterparty that has a higher credit rating than FNBC. Moreover, the fair market value may be more than the mid-market value in another situation. If interest rates move so that a swap has become significantly off-market to the disadvantage of FNBC, there may be a basis to consider an upward credit risk adjustment. ¹⁴ Dr. Parsons provides factual examples of under-valuation using swaps entered into by FNBC in the years before the Court at paragraphs 82 through 85, pages 46 through 48, of his report. Petitioner's Report provides no such empirical support for its assertions.

[21] The divergence of the views of Dr. Parsons and Dr. Smithson and Mr. Sullivan on the relationship between unadjusted mid- market values and fair market value is due to the disagreement between them with respect to the significance of credit risk in the pricing of swaps. Dr. Smithson and Mr. Sullivan assert that only the credit rating of the non-FNBC counterparty is relevant and, in so doing, ignore the bilateral nature of swaps. Petitioner's experts make no reference to the practitioner and scientific literature while Dr. Parsons cites both. See Dr. Parsons' Report, paragraphs 63-66, at 34-36, est. fn. 28. Moreover, in the book Dr. Smithson co-authored he asserts that swap "default risk is two-sided." Smith, Jr., Clifford W. and Smithson, Charles W., The Handbook of Financial Engineering, New Financial Products, Innovations, Applications and Analyses (1990), at 219 and 618.

[22] The reason that Dr. Smithson ignores the bilateral nature of swaps in opining on petitioner's swaps is not set forth in Petitioner's Report. Rather, Petitioner's Report appears to make the factual assumption FNBC will always be the transferor, i.e., "seller," of the swaps being valued., so petitioner's credit rating does not matter. See Petitioner's Report, at 18-50.

[23] Dr. Parsons' position is that such a factual assumption is not warranted. The non-FNBC counterparty, e.g., another dealer, could be transferring its position in the swap to a third dealer so FNBC would stay in the swap with a new counterparty. In that case, FNBC is not the seller and does not terminate its status as a counterparty. Also, in buyout situations, sometimes FNBC initiates the buyout and sometimes the counterparty initiates the buyout. In neither case is a third party involved. If credit risk affects the price that one counterparty would demand for a buyout, then credit risk also affects the price that the other counterparty would demand. In addition, if FNBC were to sell its leg of the swap to another, there is no basis to assume the buyer would have the same, or higher, credit rating that FNBC had at the time of the sale. In a sale to a lower credit rated entity, there is no reason FNBC would forgo the economic benefit/value in its leg of the swap if a lower rated dealer acquired the leg. FNBC would be compensated for the value it brought to the swap. Dr. Parsons can provide more details in his re-direct or rebuttal testimony with regard to his positions.

[24] Moreover, Dr. Smithson, in assuming away the bilateral nature of swaps by asserting FNBC is always the seller, misses the most important point. FNBC's credit status affects what FNBC will receive, or will pay, for a swap. Thus, FNBC's credit status impacts the value of the swaps held by FNBC.

[25] Dr. Parsons finds the Sun, Sundaresan and Wang study ¹⁵ cited by petitioner's experts germane to his assertion that the fair market value of any of swap will have to take into account the credit ratings of both the dealer and counterparty, since it shows that a swap dealer with a lower credit rating raises the fixed rate it offers to pay a counterparty on a swap and lowers the fixed rate it offers to receive. See Dr. Parsons' Report, paragraphs 63-66, at 34- 66; paragraphs 81-85, at 45-48; Petitioner's Report, Appendix 5, at 83-84.

[26] Thus, a specific point of disagreement is whether the credit rating of both counterparties affects the fair market value of a swap. This disagreement extends to whether petitioner's Devon values, ipso facto, overstate fair market value as concluded by petitioner's.experts without supporting empirical facts; or, as contended by Dr. Parsons, may OVER-estimate or UNDER-estimate the fair market value, depending on the credit rating of both counterparties to the swap.

5. CREDIT RISK CARVE-OUT ADJUSTMENTS ARE NOT A MEASURE OF FAIR

a. METHOD OF ANALYSIS.

[28] Dr. Parsons disagrees with petitioner's experts' approach. Petitioner's experts, at a very general level, compare petitioner's general approach to taking carve-outs against the claimed practices of other industry participants. Dr. Parsons' view is that the evaluation of the credit risk carve-out amounts should be based on an analysis of the actual methodologies petitioner used and a comparison against the actual market value of credit risk. ¹⁶ This is the first area of disagreement with respect to the credit risk carve-outs amounts, themselves, and permeates all other areas of disagreement on the question of whether petitioner's credit risk carve-outs arrive at the fair market values of its swaps.

b. SWAPS WITH S&P AA RATED COUNTERPARTIES

[29] The second point of disagreement relates to the reduction of Devon values by credit risk carve-outs when the non-FNBC counterparty has a credit rating the equivalent of an S&P rating of AA or higher. Dr. Parsons, opinion is that the public data relied upon to arrive at mid-market values and the discounts rates used by petitioner are valid for counterparties rated AA. Dr. Parsons opines that to reduce the Devon values of swaps with a counterparty having the equivalent of an AA rating double counts the credit adjustment. See Dr. Parsons' Report, paragraph 74, at 41; paragraph 80, at 45.

[30] Petitioner's experts attempt to downplay the importance of the assumptions built into the Devon methodology that already adjust the resulting Devon values for the risk of non-payment by AA rated counterparties by asserting: (1) petitioner stopped taking such adjustments after March of 1993; ¹⁷ and, (2) the reductions claimed are not large. See Petitioner's Report, at 48. Such an approach ignores the facts which Dr. Parsons views as important. Petitioner claimed credit risk carve-outs for swaps with AA rated counterparties from 1990 through the first quarter of 1993. Petitioner claimed credit risk carve-outs for swaps initiated in the first quarter of 1993 for year-end 1993, well after petitioner knew credit risk carve-outs should not be claimed for AA rated counterparties. This double counting demonstrates a key objection Dr. Parsons has to petitioner's methodology. It is arbitrary in that it bears no relationship to the actual discount, if any, for credit risk relative to its Devon values in the marketplace. Petitioner's Report, in focusing on the dollar amount, misses the point that credit risk for AA rated counterparties has already been taken into account, and therefore, petitioner's methodology does not arrive at a value closer to fair market value than petitioner's unadjusted Devon values. See Dr. Parsons' Report, at 45.

[31] Dr. Parsons disagrees with petitioner's experts' theoretical analysis of the AA discount rate at pages 49 through 50 of their report. Petitioner's experts' analysis that the discount rate applied to swap payments reflects the risk that the LIBOR reference banks will not pay back deposits when due 3 or 6 months later is incorrect. Petitioner's experts' assertion that loans contain protection against credit downgrades which are absent for swaps ignores the fact that swaps do have credit risk protection triggers and generally have more protection than do loans (see Dr. Parsons' Report, paragraph 69, at 37-38; see also references in footnotes 30 and 31), and also ignores the fact that swaps have a lower default rate than loans (see Dr. Parsons' Report, at 33). Petitioner's experts' assertion regarding FNBC's decision not to adjust for the cost of economic capital is beside the point. Moreover, the result of adjusting mid-market swap values for the economic cost of capital, as suggested at page 50 of Petitioner's Report, is that no dealer swap would have any value at origination.

c. ACCOUNTING PROCEDURE: CREDIT RISK CARVE-OUTS AND INCOME

[32] Petitioner's experts assert respondent can object to petitioner's accounting procedures for credit risk carve-out only if there is a systematic bias. See Petitioner's Report, paragraph 4, at 56. Dr. Parsons disagrees: (a) a discount or deferral method can be a bad valuation method even if it is not systematically biased, e.g., it could make a purely random adjustment, thereby regularly increasing both the positive and the negative discrepancies between the adjusted mid-market value and the fair market value without any countervailing benefit; and (b) the amortization procedure is systematically biased. The sources of bias are discussed in Dr. Parsons' Report. See Dr. Parsons' Report paragraphs 96-103 at 55-58.

[33] Petitioner's experts address only one potential source of bias, the effect of changes in credit ratings of the counterparties, and ignore the others. Petitioner's experts refer to the Standard & Poor's Transition Matrix for January 1998 to make the argument that a dynamic adjustment would likely result in additions to credit adjustments, not reductions. Dr. Parsons does not argue that a static adjustment necessarily overstates the credit adjustment and that a dynamic adjustment would yield only subtractions from the credit adjustments, but rather that the static adjustment is not a mark-to- market adjustment and will not approximate the fair market value. In this respect, then, petitioner's experts and Dr. Parsons agree.

[34] However, Dr. Parsons asserts petitioner's experts err in arguing from the Standard & Poor's Transition Matrix which is based on historical averages over a long window of time. In any given period the actual transition rates will differ markedly and often systematically from the historical averages. In these periods, FNBC's static adjustment will be biased, although the bias will sometimes be in one direction and at other historical periods in the other direction. There will be certain historical periods, e.g., recessions, when many companies move into lower rated categories. An amortized adjustment will capture this effect with a delay. For a while, the adjustment will be too small, i.e., biased. Conversely, during a long period of economic expansion, when many companies might be improving their balance sheets and moving into higher rated categories, the adjustment will be too large, i.e., biased. See Dr. Parsons' Report at paragraphs 99-103, at 56-58. Such a procedure does not reflect fair market value at each year-end. See Dr. O'Brien's Report, at 22-26. Petitioner's experts also assert a second argument for FNBC's static procedure, relating to the relative size of the loss factors for the higher and lower risks. Dr. Parsons disagrees with this claim.

d. VALUATION PRINCIPLES

[36] With respect to the credit exposure model ("CEM") aspect of petitioner's credit risk carve-out methodology, Dr. Parsons disagrees that it produces a loan equivalent amount; rather, it produces a measure of maximum potential exposure which should be distinguished from a measure of expected exposure. See Dr. Parsons' Report, paragraph 77, at 41-43. Dr. Smithson, one of petitioner's experts, has emphasized the difference between calculating the "maximum exposure", which is what petitioner's model does, and calculating "expected exposure", as would be appropriate for valuing credit risk and pricing a swap. See Smith, Clifford W. and Smithson, Charles W., Managing Financial Risk: A Guide to Derivative Products, Financial Engineering and Value (3rd ed. 1998), at 383-385.

[37] With respect to the risk loss factor, Dr. Parsons disagrees that it corresponds to the market value of potential losses. Dr. Parsons can testify to this disagreement in re-direct or rebuttal testimony.

[38] Finally, Dr. Parsons points out that the components of FNBC's credit adjustment were designed for credit risk management purposes and that the objective of maintaining a conservative approach to allowable credit risk is at odds with the objective of accurately pricing the credit risk. See Dr. Parsons' Report paragraphs 77-79, at 41-44.

e. SWAP PRICING

[40] Petitioner's experts disagree and argue that making an adjustment for credit risk is normal practice, despite the G30 Survey results showing that many dealers took no adjustments. Petitioner's experts do not quantify the appropriate adjustment or provide empirical evidence as to its significance in the swap market. See Petitioner's Report, at 30-31; 45-46. Petitioner's experts fail to acknowledge the small role credit risk plays in swaps as contrasted to other instruments. In the book he coauthored, Dr. Smithson asserts that "the pricing implications of default [on a swap] are significantly smaller for a swap than for a loan." Smith, Jr., Clifford W., Smithson, Charles W. and Wakeman, Lee MacDonald, The Market for Interest Rate Swaps (1988), at 37-38.

[41] Petitioner's experts assert that FNBC's adjustments yield a more accurate estimate of fair market value (see Petitioner's Report, at 5-6), although petitioner's experts do not actually compare FNBC's actual credit adjustment against any standard; not even the study cited by them. Dr. Parsons addresses the views of the industry as to the small role credit plays in actual swap pricing. See Dr. Parsons' Report, at 36-38.

f. SWAP SPREADS AS A REFLECTION OF DEALER'S COSTS

[43] Except at this most general level where there is no disagreement, the literature cited by petitioner's experts is inapplicable to the case at hand. One reflection of this fact is that petitioner's experts do not use the voluminous results cited in order to provide an estimate of the proper adjustments to the mid-market value against which FNBC's adjustments can be compared. In fact, where there is disagreement among the experts, e.g., with respect to petitioner's experts' claim that the entire bid-ask spread represents costs which must be deducted from the mid-market value, this literature does not support petitioner's experts' position. For example, the body of literature, from which petitioner's experts rely, recognizes that "the bid-ask spread is a misleading measure of the true costs of trading" because, among other reasons, trades typically take place within the bid-ask quotes. ¹⁸

[44] Dr. Parsons notes that the literature cited by petitioner's experts is almost exclusively related to the bid-ask spread in stock markets or currency markets. The one study petitioner's cite which relates to the swap market ¹⁹ was also cited by Dr. Parsons. Petitioner's experts treat this study as dispositive on the significance of credit risk in swap valuations, while Dr. Parsons treats it (together with the other study he cited) as informative but not dispositive; it is the beginning of the research in this area.

g. THE EFFECT OF CLOSE-OUT NETTING AGREEMENTS

[46] Petitioner's experts and Dr. Parsons agree that a failure to take netting into account results in an overestimate of the credit adjustment and therefore an underestimate of the fair market value. Petitioner's experts assert that the system necessary for an unbiased estimate of the credit adjustment taking into account netting were not commonly used or were too costly. Lacking an unbiased estimate of the credit adjustment, petitioner's experts do not explain why it is better to use the overestimated credit adjustment as opposed to making no adjustment at all. Petitioner's experts concern with costs does not apply to the case of mirror swaps raised by Dr. Parsons, since a correct adjustment for these does not require sophisticated calculations. Petitioner's experts' only claim that they do not know if FNBC had any mirror swaps, while Dr. Parsons provides an example.

h. PROBABILITY OF DEFAULT

[48] Petitioner's experts also raise questions about how other market participants would estimate the probability of default on swaps. Dr. Parsons presented information that market participants had concluded the probability of default was smaller than for loans and had adjusted the market price accordingly. See Dr. Parsons' Report, paragraphs 67-70, at 36-38. See Smith, Jr., Clifford W., Smithson Charles W. and Wakeman, Lee MacDonald, "Analyzing the Credit Risk of Swaps" Simon Management Review, Winter 1988, at 3.

i. CREDIT CARVE-OUTS SHOULD NOT EXCEED DEVON VALUES

6. ADMINISTRATIVE EXPENSE CARVE-OUT ADJUSTMENTS ARE NOT A

[51] Petitioner's experts argue that, nevertheless, a significant portion of the spread should be treated as compensation for future administrative costs and excluded from income. Dr. Parsons does not agree. Much, if not all, of the spread may be compensation for origination. The fee for origination should be reflected in the fair market value of the swap. Dr. Parsons cannot endorse a procedure for carving-out part of this income that bears no demonstrable relationship to any market benchmarks. Dr. Parsons is troubled by the fact there is no data or evidence that the industry, in fact, considers administrative expenses for outstanding swaps significant enough to adjust the price at which they would be bought-out or assigned.

[52] Petitioner's experts assert there is a significant difference between Devon values before and after adjustment for administrative expenses which is recognized in the market place, but do not provide supporting data, do not identify what portion of the bid-ask spread is attributable to the administrative adjustment, and do not provide any empirical data to support their endorsement of the amount of petitioner's adjustment for administrative expense carve- outs. Petitioner's Report presents no evidence supporting their contention that the difference is significant.

[53] Petitioner's experts assert major dealers have invested millions of dollars in automated valuation systems, so there must be an administrative expense adjustment. Dr. Parsons does not disagree that major dealers in general or FNBC in particular may have spent a good deal of money on systems. That is not the question. The real question is how much, if any, of the spread is related to future administrative costs.

[54] Nowhere do petitioner's experts ever attempt to calculate the size of a spread that would be properly related to ongoing administrative costs. The cost of systems may be recouped through revenues related to origination, and not from revenues related to the ongoing servicing of swaps. Thus, these system costs would not affect the fair market value of an outstanding swap, although they would be related to the fees or the bid-ask spread which a dealer would charge its end-user clients at origination. Although petitioner's experts do not appear to recognize the importance of identifying the portion of costs allocable to these two different activities, FNBC itself did recognize this distinction when it tried to allocate its own expenses into the two categories (i.e., (1) costs for new swaps and (2) costs to manage outstanding swaps to maturity), but used only one to determine the administrative cost adjustment.

[55] The correct allocation can only be answered in the marketplace, and no market data is currently available on this, which petitioner's experts agree. Dr. Parsons does not concur that the marketplace would recognize petitioner's carve-outs based on its own internal policies for allocating costs and its own expenses.

7. PETITIONER'S ANALYSIS OF VALUATION TECHNIQUES

[57] The bid-ask method "effectively adjusts" all swap values so that all costs, including marketing and origination costs, are deducted from income in the first year of the swap. Dr. Parsons' opinion is that this does not reflect the fair market value of the swap at inception, since origination costs should not be deducted in estimating the fair market value of an outstanding swap. Over the life of the swap, both the-bid-ask method and the mid-market method, with or without adjustments, report the same total income over the life of the swaps (considering realized and unrealized income); just the timing of reporting the income is different. Dr. O'Brien's Report also discusses how the income is reported over time with a proper application of mark-to-market accounting. Dr. Parsons and Dr. O'Brien can both testify to this on their re-direct or rebuttal testimony, as necessary.

[58] Petitioner's experts confuse the problem of identifying independent and reliable bid-ask quotes with the problem of using the bid-ask methodology. It is Dr. Parsons' opinion that bid-ask methodology should only be done using market bid and ask quotes, not the trader's own bid and ask quotes, and not the bid or ask rates charged on a particular swap. Independent and reliable bid and ask quotes are available for benchmark securities and can be used to value the portfolio just as can the mid-market rates. It is all a question of timing and fair market value, as discussed above.

DR. O'BRIEN'S REPORT COMPARED TO PETITIONER'S REPORT

[59] Dr. O'Brien relied on petitioner's representations regarding FNBC's accounting procedures. In some instances, Dr. O'Brien could not verify the accounting entries were made according to FNBC's stated policy because of a lack of information. In those instances where Dr. O'Brien lacked information, she provided a description of the information needed to verify FNBC's statements. See Dr. O'Brien's Report, paragraphs 23-25, at 10-12. Petitioner's experts did not attempt to validate the petitioner's actual computations or the underlying documentation. See Petitioner's Report, at 5, 46.

1993

[60] Dr. O'Brien concludes that FNBC's methodology cannot be described as a mark-to-market accounting method if the mid-market values and adjustments do not reflect market values for 1993. Dr. O'Brien then evaluates the other aspects of petitioner's accounting assuming, arguendo, the mid-market values and adjustments petitioner uses for tax purposes reflect fair market value. See Dr. O'Brien's Report, paragraphs 23-24, at 10-12. Petitioner's experts undertake no such analysis.

[61] Petitioner's experts fail to address the propriety of removing "Nonperforming VEP transactions" from the trading portfolio and valuing these swaps at "modified" lower of cost or market value. Dr. O'Brien addresses this issue. See Dr. O'Brien's Report paragraph 36, at 19.

[62] Petitioner's experts fail to address the fact that petitioner makes no adjustment to the credit deferral amortization schedule for subsequent charge-offs of swaps or early terminations of swaps. Dr. O'Brien addresses this issue. See Dr. O'Brien's Report, at 25.

[63] Petitioner's experts fail to address the one-month lag incorporated in the deferral of credit risk adjustments in light of the end-of year requirements of I.R.C. section 475. Dr. O'Brien addresses this issue. See Dr. O'Brien's Report, paragraph 42, at 21-22.

[64] With respect to using year-end Devon values, petitioner's experts state that valuation at a date prior to the last business date of the year is common practice because of the desire to close their books within "two or three days after the end of the year" and the complexity of the calculations involved. See Petitioner's Report, footnote 72, at 44-45. Dr. O'Brien disagrees that the desire to close the books quickly is a relevant concern. Both the Securities Exchange Commission and the IRS allow a 3 1/2 or 2 1/2 month filing period after the end of the fiscal year so that entities have time to gather the data and perform the calculations necessary to report as of the end of the reporting year. In general, numbers reported on the balance sheet at fair value, like FNBC's trading assets, are understood to be valued at the balance sheet date, in this case December 31, 1993. In Dr. O'Brien's opinion, a justification under generally accepted accounting principles for reporting as fair values non-year-end values on the year-end balance sheet may be that the difference between the non-year-end and year-end values are immaterial. FNBC has not asserted that an immaterial difference exists.

[65] Petitioner's experts fail to address, or attempt to reconcile, the problems with the petitioner's amortization schedules. Dr. O'Brien addresses this deficiency. See Dr. O'Brien's Report, at 23-24. Furthermore, Dr. O'Brien cites several specific reasons why the petitioner's fixed amortization schedules may diverge from market values. See Dr. O'Brien's Report, paragraph 50, at 24-25.

1990-1992

[66] Petitioner's experts failures to address any of the accounting practices of petitioner for the years 1990 through 1992. Dr. O'Brien's addresses these years. See Dr. O'Brien's Report, 25, at 12-14.

CONCLUSION

[67] For all the reasons stated above, respondent submits that his experts independently evaluated and correctly opined on petitioner's valuation methodologies and accounting methods. In contrast, petitioner's experts' reports are a restatement of petitioner's theories without independent analysis of empirical facts relative to the claimed credit risk and administrative expense carve- outs.

STUART L. BROWN

[68] This is to certify that a copy of the foregoing RESPONDENT'S MEMORANDUM OF EXPERT REPORT DIFFERENCES was served on Counsel for Petitioner by mailing the same on ___________, in a postage-paid wrapper addressed to Jay Zimbler, SIDLEY & AUSTIN, Bank One Plaza, 10 S. Dearborn St., Suite 4200, Chicago, Illinois 60603.

Date:

² There are different methods for arriving at mid-market values. Petitioner's experts refer to the method that discounts projected cash flows using the same family of curves used to project the cash flows. See Petitioner's Report, Appendix 4, at 79. However, petitioner did not follow that convention; rather, it selected a separate curve to determine discount rates. See Dr. Parsons' Report at Appendix B, paragraph B20, at 10.

³ See Exhibit 9 in Dr. Parsons' Report.

⁴ See table in Appendix 4, Petitioner's Report.

⁵ The swaps petitioner's experts tested are listed in Appendix A, attached to this memorandum.

⁶ Although petitioner's experts' table in Appendix 4 of their report sets forth a notional amount of $1 billion dollars for the swap with transaction identification number 172240, the information in the Devon listings as of December 20, 1993 (Bates stamped pages 57492-57603) reports a notional amount of $100 million for this swap.

⁷ Dr. Parsons' Report points out that, while there is this primary marketplace in which swaps are priced, this marketplace is not differentiated by the credit rating of the counterparties in the same manner as the bond market is differentiated by credit rating. See Dr. Parsons' Report, Appendix A, paragraph A47, at 29.

⁸ The report called "Derivatives: Practices and Principles" by the Group of Thirty included a recommendation section and Appendix I (working papers), Appendix II (Legal Enforcability, Survey of Nine Jurisdictions dated July 1993) and Appendix III (Survey of Industry Practices dated March 1994). There was also a follow-up Survey of Industry Practice dated December 1994. Petitioner's experts attached only the recommendation section to their expert report, but refer to Appendix III and the 1994 follow-up surveys in the body of their report.

⁹ Petitioner refers to its credit risk adjustments as credit risk carve-outs.

¹⁰ Petitioner's Report, at 3-4; Dr. Parsons' Report, paragraphs 36, 42, at 20, 22.

¹¹ Dr. Parsons' Report, paragraphs 36-42, at 20-22.

¹² Petitioner's Report at pp. 3, 6, 30, 45, 60.

¹³ See Dr. Parsons' Report at paragraph 1, at 1; Section III, Summary of Findings, paragraphs 36-43, at 20-22.

¹⁴ See Dr. Parsons' Report, paragraph 84, at 47.

¹⁵ Sun, Tong-Sheng; Suresh Sundaresan, and Ching Wang. 1993. "Interest Rate Swaps: An empirical Investigation," Journal of Finance, 34, at 77-99.

¹⁶ Dr. Parsons does not opine on the years 1990 through 1992 due to lack of data on the actual methodologies for said years.

¹⁷ Respondent asserts this is tantamount to an admission that credit risk carve-outs should not be claimed on swaps with a counterparty with an AA or higher S&P rating. Moreover, given petitioner's method of amortizing the carve-outs back into income, respondent further asserts (since in 1993 petitioner knew that no carve-outs should be claimed on AA rated counterparties), petitioner should have immediately taken the unamortized balance of carve-outs for AA counterparties into income in 1993.

¹⁸ Madhaven, A., M. Richardson and M. Roomans, "Why Do Security Prices Change? A Transaction Level Analysis of NYSE Stocks," Review of Financial Studies 10 (1997), at 1047.

¹⁹ See footnote 15.